diff options
Diffstat (limited to 'test/CodeGen/Mips/msa')
62 files changed, 3528 insertions, 2800 deletions
diff --git a/test/CodeGen/Mips/msa/2r.ll b/test/CodeGen/Mips/msa/2r.ll index da35ad82cad1b..501936c76e73c 100644 --- a/test/CodeGen/Mips/msa/2r.ll +++ b/test/CodeGen/Mips/msa/2r.ll @@ -8,7 +8,7 @@ define void @llvm_mips_nloc_b_test() nounwind { entry: - %0 = load <16 x i8>* @llvm_mips_nloc_b_ARG1 + %0 = load <16 x i8>, <16 x i8>* @llvm_mips_nloc_b_ARG1 %1 = tail call <16 x i8> @llvm.mips.nloc.b(<16 x i8> %0) store <16 x i8> %1, <16 x i8>* @llvm_mips_nloc_b_RES ret void @@ -29,7 +29,7 @@ declare <16 x i8> @llvm.mips.nloc.b(<16 x i8>) nounwind define void @llvm_mips_nloc_h_test() nounwind { entry: - %0 = load <8 x i16>* @llvm_mips_nloc_h_ARG1 + %0 = load <8 x i16>, <8 x i16>* @llvm_mips_nloc_h_ARG1 %1 = tail call <8 x i16> @llvm.mips.nloc.h(<8 x i16> %0) store <8 x i16> %1, <8 x i16>* @llvm_mips_nloc_h_RES ret void @@ -50,7 +50,7 @@ declare <8 x i16> @llvm.mips.nloc.h(<8 x i16>) nounwind define void @llvm_mips_nloc_w_test() nounwind { entry: - %0 = load <4 x i32>* @llvm_mips_nloc_w_ARG1 + %0 = load <4 x i32>, <4 x i32>* @llvm_mips_nloc_w_ARG1 %1 = tail call <4 x i32> @llvm.mips.nloc.w(<4 x i32> %0) store <4 x i32> %1, <4 x i32>* @llvm_mips_nloc_w_RES ret void @@ -71,7 +71,7 @@ declare <4 x i32> @llvm.mips.nloc.w(<4 x i32>) nounwind define void @llvm_mips_nloc_d_test() nounwind { entry: - %0 = load <2 x i64>* @llvm_mips_nloc_d_ARG1 + %0 = load <2 x i64>, <2 x i64>* @llvm_mips_nloc_d_ARG1 %1 = tail call <2 x i64> @llvm.mips.nloc.d(<2 x i64> %0) store <2 x i64> %1, <2 x i64>* @llvm_mips_nloc_d_RES ret void @@ -92,7 +92,7 @@ declare <2 x i64> @llvm.mips.nloc.d(<2 x i64>) nounwind define void @llvm_mips_nlzc_b_test() nounwind { entry: - %0 = load <16 x i8>* @llvm_mips_nlzc_b_ARG1 + %0 = load <16 x i8>, <16 x i8>* @llvm_mips_nlzc_b_ARG1 %1 = tail call <16 x i8> @llvm.mips.nlzc.b(<16 x i8> %0) store <16 x i8> %1, <16 x i8>* @llvm_mips_nlzc_b_RES ret void @@ -113,7 +113,7 @@ declare <16 x i8> @llvm.mips.nlzc.b(<16 x i8>) nounwind define void @llvm_mips_nlzc_h_test() nounwind { entry: - %0 = load <8 x i16>* @llvm_mips_nlzc_h_ARG1 + %0 = load <8 x i16>, <8 x i16>* @llvm_mips_nlzc_h_ARG1 %1 = tail call <8 x i16> @llvm.mips.nlzc.h(<8 x i16> %0) store <8 x i16> %1, <8 x i16>* @llvm_mips_nlzc_h_RES ret void @@ -134,7 +134,7 @@ declare <8 x i16> @llvm.mips.nlzc.h(<8 x i16>) nounwind define void @llvm_mips_nlzc_w_test() nounwind { entry: - %0 = load <4 x i32>* @llvm_mips_nlzc_w_ARG1 + %0 = load <4 x i32>, <4 x i32>* @llvm_mips_nlzc_w_ARG1 %1 = tail call <4 x i32> @llvm.mips.nlzc.w(<4 x i32> %0) store <4 x i32> %1, <4 x i32>* @llvm_mips_nlzc_w_RES ret void @@ -155,7 +155,7 @@ declare <4 x i32> @llvm.mips.nlzc.w(<4 x i32>) nounwind define void @llvm_mips_nlzc_d_test() nounwind { entry: - %0 = load <2 x i64>* @llvm_mips_nlzc_d_ARG1 + %0 = load <2 x i64>, <2 x i64>* @llvm_mips_nlzc_d_ARG1 %1 = tail call <2 x i64> @llvm.mips.nlzc.d(<2 x i64> %0) store <2 x i64> %1, <2 x i64>* @llvm_mips_nlzc_d_RES ret void @@ -176,7 +176,7 @@ declare <2 x i64> @llvm.mips.nlzc.d(<2 x i64>) nounwind define void @llvm_mips_pcnt_b_test() nounwind { entry: - %0 = load <16 x i8>* @llvm_mips_pcnt_b_ARG1 + %0 = load <16 x i8>, <16 x i8>* @llvm_mips_pcnt_b_ARG1 %1 = tail call <16 x i8> @llvm.mips.pcnt.b(<16 x i8> %0) store <16 x i8> %1, <16 x i8>* @llvm_mips_pcnt_b_RES ret void @@ -197,7 +197,7 @@ declare <16 x i8> @llvm.mips.pcnt.b(<16 x i8>) nounwind define void @llvm_mips_pcnt_h_test() nounwind { entry: - %0 = load <8 x i16>* @llvm_mips_pcnt_h_ARG1 + %0 = load <8 x i16>, <8 x i16>* @llvm_mips_pcnt_h_ARG1 %1 = tail call <8 x i16> @llvm.mips.pcnt.h(<8 x i16> %0) store <8 x i16> %1, <8 x i16>* @llvm_mips_pcnt_h_RES ret void @@ -218,7 +218,7 @@ declare <8 x i16> @llvm.mips.pcnt.h(<8 x i16>) nounwind define void @llvm_mips_pcnt_w_test() nounwind { entry: - %0 = load <4 x i32>* @llvm_mips_pcnt_w_ARG1 + %0 = load <4 x i32>, <4 x i32>* @llvm_mips_pcnt_w_ARG1 %1 = tail call <4 x i32> @llvm.mips.pcnt.w(<4 x i32> %0) store <4 x i32> %1, <4 x i32>* @llvm_mips_pcnt_w_RES ret void @@ -239,7 +239,7 @@ declare <4 x i32> @llvm.mips.pcnt.w(<4 x i32>) nounwind define void @llvm_mips_pcnt_d_test() nounwind { entry: - %0 = load <2 x i64>* @llvm_mips_pcnt_d_ARG1 + %0 = load <2 x i64>, <2 x i64>* @llvm_mips_pcnt_d_ARG1 %1 = tail call <2 x i64> @llvm.mips.pcnt.d(<2 x i64> %0) store <2 x i64> %1, <2 x i64>* @llvm_mips_pcnt_d_RES ret void diff --git a/test/CodeGen/Mips/msa/2r_vector_scalar.ll b/test/CodeGen/Mips/msa/2r_vector_scalar.ll index 64e459e4d9a93..ddcd3cf757d9c 100644 --- a/test/CodeGen/Mips/msa/2r_vector_scalar.ll +++ b/test/CodeGen/Mips/msa/2r_vector_scalar.ll @@ -15,7 +15,7 @@ define void @llvm_mips_fill_b_test() nounwind { entry: - %0 = load i32* @llvm_mips_fill_b_ARG1 + %0 = load i32, i32* @llvm_mips_fill_b_ARG1 %1 = tail call <16 x i8> @llvm.mips.fill.b(i32 %0) store <16 x i8> %1, <16 x i8>* @llvm_mips_fill_b_RES ret void @@ -35,7 +35,7 @@ declare <16 x i8> @llvm.mips.fill.b(i32) nounwind define void @llvm_mips_fill_h_test() nounwind { entry: - %0 = load i32* @llvm_mips_fill_h_ARG1 + %0 = load i32, i32* @llvm_mips_fill_h_ARG1 %1 = tail call <8 x i16> @llvm.mips.fill.h(i32 %0) store <8 x i16> %1, <8 x i16>* @llvm_mips_fill_h_RES ret void @@ -55,7 +55,7 @@ declare <8 x i16> @llvm.mips.fill.h(i32) nounwind define void @llvm_mips_fill_w_test() nounwind { entry: - %0 = load i32* @llvm_mips_fill_w_ARG1 + %0 = load i32, i32* @llvm_mips_fill_w_ARG1 %1 = tail call <4 x i32> @llvm.mips.fill.w(i32 %0) store <4 x i32> %1, <4 x i32>* @llvm_mips_fill_w_RES ret void @@ -75,7 +75,7 @@ declare <4 x i32> @llvm.mips.fill.w(i32) nounwind define void @llvm_mips_fill_d_test() nounwind { entry: - %0 = load i64* @llvm_mips_fill_d_ARG1 + %0 = load i64, i64* @llvm_mips_fill_d_ARG1 %1 = tail call <2 x i64> @llvm.mips.fill.d(i64 %0) store <2 x i64> %1, <2 x i64>* @llvm_mips_fill_d_RES ret void diff --git a/test/CodeGen/Mips/msa/2rf.ll b/test/CodeGen/Mips/msa/2rf.ll index b361ef5eae219..1dbfbda1b612a 100644 --- a/test/CodeGen/Mips/msa/2rf.ll +++ b/test/CodeGen/Mips/msa/2rf.ll @@ -8,7 +8,7 @@ define void @llvm_mips_flog2_w_test() nounwind { entry: - %0 = load <4 x float>* @llvm_mips_flog2_w_ARG1 + %0 = load <4 x float>, <4 x float>* @llvm_mips_flog2_w_ARG1 %1 = tail call <4 x float> @llvm.mips.flog2.w(<4 x float> %0) store <4 x float> %1, <4 x float>* @llvm_mips_flog2_w_RES ret void @@ -29,7 +29,7 @@ declare <4 x float> @llvm.mips.flog2.w(<4 x float>) nounwind define void @llvm_mips_flog2_d_test() nounwind { entry: - %0 = load <2 x double>* @llvm_mips_flog2_d_ARG1 + %0 = load <2 x double>, <2 x double>* @llvm_mips_flog2_d_ARG1 %1 = tail call <2 x double> @llvm.mips.flog2.d(<2 x double> %0) store <2 x double> %1, <2 x double>* @llvm_mips_flog2_d_RES ret void @@ -47,7 +47,7 @@ declare <2 x double> @llvm.mips.flog2.d(<2 x double>) nounwind define void @flog2_w_test() nounwind { entry: - %0 = load <4 x float>* @llvm_mips_flog2_w_ARG1 + %0 = load <4 x float>, <4 x float>* @llvm_mips_flog2_w_ARG1 %1 = tail call <4 x float> @llvm.log2.v4f32(<4 x float> %0) store <4 x float> %1, <4 x float>* @llvm_mips_flog2_w_RES ret void @@ -65,7 +65,7 @@ declare <4 x float> @llvm.log2.v4f32(<4 x float> %val) define void @flog2_d_test() nounwind { entry: - %0 = load <2 x double>* @llvm_mips_flog2_d_ARG1 + %0 = load <2 x double>, <2 x double>* @llvm_mips_flog2_d_ARG1 %1 = tail call <2 x double> @llvm.log2.v2f64(<2 x double> %0) store <2 x double> %1, <2 x double>* @llvm_mips_flog2_d_RES ret void @@ -86,7 +86,7 @@ declare <2 x double> @llvm.log2.v2f64(<2 x double> %val) define void @llvm_mips_frint_w_test() nounwind { entry: - %0 = load <4 x float>* @llvm_mips_frint_w_ARG1 + %0 = load <4 x float>, <4 x float>* @llvm_mips_frint_w_ARG1 %1 = tail call <4 x float> @llvm.mips.frint.w(<4 x float> %0) store <4 x float> %1, <4 x float>* @llvm_mips_frint_w_RES ret void @@ -107,7 +107,7 @@ declare <4 x float> @llvm.mips.frint.w(<4 x float>) nounwind define void @llvm_mips_frint_d_test() nounwind { entry: - %0 = load <2 x double>* @llvm_mips_frint_d_ARG1 + %0 = load <2 x double>, <2 x double>* @llvm_mips_frint_d_ARG1 %1 = tail call <2 x double> @llvm.mips.frint.d(<2 x double> %0) store <2 x double> %1, <2 x double>* @llvm_mips_frint_d_RES ret void @@ -125,7 +125,7 @@ declare <2 x double> @llvm.mips.frint.d(<2 x double>) nounwind define void @frint_w_test() nounwind { entry: - %0 = load <4 x float>* @llvm_mips_frint_w_ARG1 + %0 = load <4 x float>, <4 x float>* @llvm_mips_frint_w_ARG1 %1 = tail call <4 x float> @llvm.rint.v4f32(<4 x float> %0) store <4 x float> %1, <4 x float>* @llvm_mips_frint_w_RES ret void @@ -143,7 +143,7 @@ declare <4 x float> @llvm.rint.v4f32(<4 x float>) nounwind define void @frint_d_test() nounwind { entry: - %0 = load <2 x double>* @llvm_mips_frint_d_ARG1 + %0 = load <2 x double>, <2 x double>* @llvm_mips_frint_d_ARG1 %1 = tail call <2 x double> @llvm.rint.v2f64(<2 x double> %0) store <2 x double> %1, <2 x double>* @llvm_mips_frint_d_RES ret void @@ -164,7 +164,7 @@ declare <2 x double> @llvm.rint.v2f64(<2 x double>) nounwind define void @llvm_mips_frcp_w_test() nounwind { entry: - %0 = load <4 x float>* @llvm_mips_frcp_w_ARG1 + %0 = load <4 x float>, <4 x float>* @llvm_mips_frcp_w_ARG1 %1 = tail call <4 x float> @llvm.mips.frcp.w(<4 x float> %0) store <4 x float> %1, <4 x float>* @llvm_mips_frcp_w_RES ret void @@ -185,7 +185,7 @@ declare <4 x float> @llvm.mips.frcp.w(<4 x float>) nounwind define void @llvm_mips_frcp_d_test() nounwind { entry: - %0 = load <2 x double>* @llvm_mips_frcp_d_ARG1 + %0 = load <2 x double>, <2 x double>* @llvm_mips_frcp_d_ARG1 %1 = tail call <2 x double> @llvm.mips.frcp.d(<2 x double> %0) store <2 x double> %1, <2 x double>* @llvm_mips_frcp_d_RES ret void @@ -206,7 +206,7 @@ declare <2 x double> @llvm.mips.frcp.d(<2 x double>) nounwind define void @llvm_mips_frsqrt_w_test() nounwind { entry: - %0 = load <4 x float>* @llvm_mips_frsqrt_w_ARG1 + %0 = load <4 x float>, <4 x float>* @llvm_mips_frsqrt_w_ARG1 %1 = tail call <4 x float> @llvm.mips.frsqrt.w(<4 x float> %0) store <4 x float> %1, <4 x float>* @llvm_mips_frsqrt_w_RES ret void @@ -227,7 +227,7 @@ declare <4 x float> @llvm.mips.frsqrt.w(<4 x float>) nounwind define void @llvm_mips_frsqrt_d_test() nounwind { entry: - %0 = load <2 x double>* @llvm_mips_frsqrt_d_ARG1 + %0 = load <2 x double>, <2 x double>* @llvm_mips_frsqrt_d_ARG1 %1 = tail call <2 x double> @llvm.mips.frsqrt.d(<2 x double> %0) store <2 x double> %1, <2 x double>* @llvm_mips_frsqrt_d_RES ret void @@ -248,7 +248,7 @@ declare <2 x double> @llvm.mips.frsqrt.d(<2 x double>) nounwind define void @llvm_mips_fsqrt_w_test() nounwind { entry: - %0 = load <4 x float>* @llvm_mips_fsqrt_w_ARG1 + %0 = load <4 x float>, <4 x float>* @llvm_mips_fsqrt_w_ARG1 %1 = tail call <4 x float> @llvm.mips.fsqrt.w(<4 x float> %0) store <4 x float> %1, <4 x float>* @llvm_mips_fsqrt_w_RES ret void @@ -269,7 +269,7 @@ declare <4 x float> @llvm.mips.fsqrt.w(<4 x float>) nounwind define void @llvm_mips_fsqrt_d_test() nounwind { entry: - %0 = load <2 x double>* @llvm_mips_fsqrt_d_ARG1 + %0 = load <2 x double>, <2 x double>* @llvm_mips_fsqrt_d_ARG1 %1 = tail call <2 x double> @llvm.mips.fsqrt.d(<2 x double> %0) store <2 x double> %1, <2 x double>* @llvm_mips_fsqrt_d_RES ret void @@ -287,7 +287,7 @@ declare <2 x double> @llvm.mips.fsqrt.d(<2 x double>) nounwind define void @fsqrt_w_test() nounwind { entry: - %0 = load <4 x float>* @llvm_mips_fsqrt_w_ARG1 + %0 = load <4 x float>, <4 x float>* @llvm_mips_fsqrt_w_ARG1 %1 = tail call <4 x float> @llvm.sqrt.v4f32(<4 x float> %0) store <4 x float> %1, <4 x float>* @llvm_mips_fsqrt_w_RES ret void @@ -305,7 +305,7 @@ declare <4 x float> @llvm.sqrt.v4f32(<4 x float>) nounwind define void @fsqrt_d_test() nounwind { entry: - %0 = load <2 x double>* @llvm_mips_fsqrt_d_ARG1 + %0 = load <2 x double>, <2 x double>* @llvm_mips_fsqrt_d_ARG1 %1 = tail call <2 x double> @llvm.sqrt.v2f64(<2 x double> %0) store <2 x double> %1, <2 x double>* @llvm_mips_fsqrt_d_RES ret void diff --git a/test/CodeGen/Mips/msa/2rf_exup.ll b/test/CodeGen/Mips/msa/2rf_exup.ll index 8d7cc367040a3..fd81ff6d112b0 100644 --- a/test/CodeGen/Mips/msa/2rf_exup.ll +++ b/test/CodeGen/Mips/msa/2rf_exup.ll @@ -9,7 +9,7 @@ define void @llvm_mips_fexupl_w_test() nounwind { entry: - %0 = load <8 x half>* @llvm_mips_fexupl_w_ARG1 + %0 = load <8 x half>, <8 x half>* @llvm_mips_fexupl_w_ARG1 %1 = tail call <4 x float> @llvm.mips.fexupl.w(<8 x half> %0) store <4 x float> %1, <4 x float>* @llvm_mips_fexupl_w_RES ret void @@ -28,7 +28,7 @@ declare <4 x float> @llvm.mips.fexupl.w(<8 x half>) nounwind define void @llvm_mips_fexupl_d_test() nounwind { entry: - %0 = load <4 x float>* @llvm_mips_fexupl_d_ARG1 + %0 = load <4 x float>, <4 x float>* @llvm_mips_fexupl_d_ARG1 %1 = tail call <2 x double> @llvm.mips.fexupl.d(<4 x float> %0) store <2 x double> %1, <2 x double>* @llvm_mips_fexupl_d_RES ret void @@ -47,7 +47,7 @@ declare <2 x double> @llvm.mips.fexupl.d(<4 x float>) nounwind define void @llvm_mips_fexupr_w_test() nounwind { entry: - %0 = load <8 x half>* @llvm_mips_fexupr_w_ARG1 + %0 = load <8 x half>, <8 x half>* @llvm_mips_fexupr_w_ARG1 %1 = tail call <4 x float> @llvm.mips.fexupr.w(<8 x half> %0) store <4 x float> %1, <4 x float>* @llvm_mips_fexupr_w_RES ret void @@ -66,7 +66,7 @@ declare <4 x float> @llvm.mips.fexupr.w(<8 x half>) nounwind define void @llvm_mips_fexupr_d_test() nounwind { entry: - %0 = load <4 x float>* @llvm_mips_fexupr_d_ARG1 + %0 = load <4 x float>, <4 x float>* @llvm_mips_fexupr_d_ARG1 %1 = tail call <2 x double> @llvm.mips.fexupr.d(<4 x float> %0) store <2 x double> %1, <2 x double>* @llvm_mips_fexupr_d_RES ret void diff --git a/test/CodeGen/Mips/msa/2rf_float_int.ll b/test/CodeGen/Mips/msa/2rf_float_int.ll index 3b5dfda2d1e60..369015814b0e5 100644 --- a/test/CodeGen/Mips/msa/2rf_float_int.ll +++ b/test/CodeGen/Mips/msa/2rf_float_int.ll @@ -9,7 +9,7 @@ define void @llvm_mips_ffint_s_w_test() nounwind { entry: - %0 = load <4 x i32>* @llvm_mips_ffint_s_w_ARG1 + %0 = load <4 x i32>, <4 x i32>* @llvm_mips_ffint_s_w_ARG1 %1 = tail call <4 x float> @llvm.mips.ffint.s.w(<4 x i32> %0) store <4 x float> %1, <4 x float>* @llvm_mips_ffint_s_w_RES ret void @@ -30,7 +30,7 @@ declare <4 x float> @llvm.mips.ffint.s.w(<4 x i32>) nounwind define void @llvm_mips_ffint_s_d_test() nounwind { entry: - %0 = load <2 x i64>* @llvm_mips_ffint_s_d_ARG1 + %0 = load <2 x i64>, <2 x i64>* @llvm_mips_ffint_s_d_ARG1 %1 = tail call <2 x double> @llvm.mips.ffint.s.d(<2 x i64> %0) store <2 x double> %1, <2 x double>* @llvm_mips_ffint_s_d_RES ret void @@ -51,7 +51,7 @@ declare <2 x double> @llvm.mips.ffint.s.d(<2 x i64>) nounwind define void @llvm_mips_ffint_u_w_test() nounwind { entry: - %0 = load <4 x i32>* @llvm_mips_ffint_u_w_ARG1 + %0 = load <4 x i32>, <4 x i32>* @llvm_mips_ffint_u_w_ARG1 %1 = tail call <4 x float> @llvm.mips.ffint.u.w(<4 x i32> %0) store <4 x float> %1, <4 x float>* @llvm_mips_ffint_u_w_RES ret void @@ -72,7 +72,7 @@ declare <4 x float> @llvm.mips.ffint.u.w(<4 x i32>) nounwind define void @llvm_mips_ffint_u_d_test() nounwind { entry: - %0 = load <2 x i64>* @llvm_mips_ffint_u_d_ARG1 + %0 = load <2 x i64>, <2 x i64>* @llvm_mips_ffint_u_d_ARG1 %1 = tail call <2 x double> @llvm.mips.ffint.u.d(<2 x i64> %0) store <2 x double> %1, <2 x double>* @llvm_mips_ffint_u_d_RES ret void diff --git a/test/CodeGen/Mips/msa/2rf_fq.ll b/test/CodeGen/Mips/msa/2rf_fq.ll index 021dd937fad3f..05c649ee918ad 100644 --- a/test/CodeGen/Mips/msa/2rf_fq.ll +++ b/test/CodeGen/Mips/msa/2rf_fq.ll @@ -9,7 +9,7 @@ define void @llvm_mips_ffql_w_test() nounwind { entry: - %0 = load <8 x i16>* @llvm_mips_ffql_w_ARG1 + %0 = load <8 x i16>, <8 x i16>* @llvm_mips_ffql_w_ARG1 %1 = tail call <4 x float> @llvm.mips.ffql.w(<8 x i16> %0) store <4 x float> %1, <4 x float>* @llvm_mips_ffql_w_RES ret void @@ -28,7 +28,7 @@ declare <4 x float> @llvm.mips.ffql.w(<8 x i16>) nounwind define void @llvm_mips_ffql_d_test() nounwind { entry: - %0 = load <4 x i32>* @llvm_mips_ffql_d_ARG1 + %0 = load <4 x i32>, <4 x i32>* @llvm_mips_ffql_d_ARG1 %1 = tail call <2 x double> @llvm.mips.ffql.d(<4 x i32> %0) store <2 x double> %1, <2 x double>* @llvm_mips_ffql_d_RES ret void @@ -47,7 +47,7 @@ declare <2 x double> @llvm.mips.ffql.d(<4 x i32>) nounwind define void @llvm_mips_ffqr_w_test() nounwind { entry: - %0 = load <8 x i16>* @llvm_mips_ffqr_w_ARG1 + %0 = load <8 x i16>, <8 x i16>* @llvm_mips_ffqr_w_ARG1 %1 = tail call <4 x float> @llvm.mips.ffqr.w(<8 x i16> %0) store <4 x float> %1, <4 x float>* @llvm_mips_ffqr_w_RES ret void @@ -66,7 +66,7 @@ declare <4 x float> @llvm.mips.ffqr.w(<8 x i16>) nounwind define void @llvm_mips_ffqr_d_test() nounwind { entry: - %0 = load <4 x i32>* @llvm_mips_ffqr_d_ARG1 + %0 = load <4 x i32>, <4 x i32>* @llvm_mips_ffqr_d_ARG1 %1 = tail call <2 x double> @llvm.mips.ffqr.d(<4 x i32> %0) store <2 x double> %1, <2 x double>* @llvm_mips_ffqr_d_RES ret void diff --git a/test/CodeGen/Mips/msa/2rf_int_float.ll b/test/CodeGen/Mips/msa/2rf_int_float.ll index 4665ae066a4f7..77d1404f9cfa7 100644 --- a/test/CodeGen/Mips/msa/2rf_int_float.ll +++ b/test/CodeGen/Mips/msa/2rf_int_float.ll @@ -10,7 +10,7 @@ define void @llvm_mips_fclass_w_test() nounwind { entry: - %0 = load <4 x float>* @llvm_mips_fclass_w_ARG1 + %0 = load <4 x float>, <4 x float>* @llvm_mips_fclass_w_ARG1 %1 = tail call <4 x i32> @llvm.mips.fclass.w(<4 x float> %0) store <4 x i32> %1, <4 x i32>* @llvm_mips_fclass_w_RES ret void @@ -31,7 +31,7 @@ declare <4 x i32> @llvm.mips.fclass.w(<4 x float>) nounwind define void @llvm_mips_fclass_d_test() nounwind { entry: - %0 = load <2 x double>* @llvm_mips_fclass_d_ARG1 + %0 = load <2 x double>, <2 x double>* @llvm_mips_fclass_d_ARG1 %1 = tail call <2 x i64> @llvm.mips.fclass.d(<2 x double> %0) store <2 x i64> %1, <2 x i64>* @llvm_mips_fclass_d_RES ret void @@ -52,7 +52,7 @@ declare <2 x i64> @llvm.mips.fclass.d(<2 x double>) nounwind define void @llvm_mips_ftrunc_s_w_test() nounwind { entry: - %0 = load <4 x float>* @llvm_mips_ftrunc_s_w_ARG1 + %0 = load <4 x float>, <4 x float>* @llvm_mips_ftrunc_s_w_ARG1 %1 = tail call <4 x i32> @llvm.mips.ftrunc.s.w(<4 x float> %0) store <4 x i32> %1, <4 x i32>* @llvm_mips_ftrunc_s_w_RES ret void @@ -73,7 +73,7 @@ declare <4 x i32> @llvm.mips.ftrunc.s.w(<4 x float>) nounwind define void @llvm_mips_ftrunc_s_d_test() nounwind { entry: - %0 = load <2 x double>* @llvm_mips_ftrunc_s_d_ARG1 + %0 = load <2 x double>, <2 x double>* @llvm_mips_ftrunc_s_d_ARG1 %1 = tail call <2 x i64> @llvm.mips.ftrunc.s.d(<2 x double> %0) store <2 x i64> %1, <2 x i64>* @llvm_mips_ftrunc_s_d_RES ret void @@ -94,7 +94,7 @@ declare <2 x i64> @llvm.mips.ftrunc.s.d(<2 x double>) nounwind define void @llvm_mips_ftrunc_u_w_test() nounwind { entry: - %0 = load <4 x float>* @llvm_mips_ftrunc_u_w_ARG1 + %0 = load <4 x float>, <4 x float>* @llvm_mips_ftrunc_u_w_ARG1 %1 = tail call <4 x i32> @llvm.mips.ftrunc.u.w(<4 x float> %0) store <4 x i32> %1, <4 x i32>* @llvm_mips_ftrunc_u_w_RES ret void @@ -115,7 +115,7 @@ declare <4 x i32> @llvm.mips.ftrunc.u.w(<4 x float>) nounwind define void @llvm_mips_ftrunc_u_d_test() nounwind { entry: - %0 = load <2 x double>* @llvm_mips_ftrunc_u_d_ARG1 + %0 = load <2 x double>, <2 x double>* @llvm_mips_ftrunc_u_d_ARG1 %1 = tail call <2 x i64> @llvm.mips.ftrunc.u.d(<2 x double> %0) store <2 x i64> %1, <2 x i64>* @llvm_mips_ftrunc_u_d_RES ret void @@ -136,7 +136,7 @@ declare <2 x i64> @llvm.mips.ftrunc.u.d(<2 x double>) nounwind define void @llvm_mips_ftint_s_w_test() nounwind { entry: - %0 = load <4 x float>* @llvm_mips_ftint_s_w_ARG1 + %0 = load <4 x float>, <4 x float>* @llvm_mips_ftint_s_w_ARG1 %1 = tail call <4 x i32> @llvm.mips.ftint.s.w(<4 x float> %0) store <4 x i32> %1, <4 x i32>* @llvm_mips_ftint_s_w_RES ret void @@ -157,7 +157,7 @@ declare <4 x i32> @llvm.mips.ftint.s.w(<4 x float>) nounwind define void @llvm_mips_ftint_s_d_test() nounwind { entry: - %0 = load <2 x double>* @llvm_mips_ftint_s_d_ARG1 + %0 = load <2 x double>, <2 x double>* @llvm_mips_ftint_s_d_ARG1 %1 = tail call <2 x i64> @llvm.mips.ftint.s.d(<2 x double> %0) store <2 x i64> %1, <2 x i64>* @llvm_mips_ftint_s_d_RES ret void @@ -178,7 +178,7 @@ declare <2 x i64> @llvm.mips.ftint.s.d(<2 x double>) nounwind define void @llvm_mips_ftint_u_w_test() nounwind { entry: - %0 = load <4 x float>* @llvm_mips_ftint_u_w_ARG1 + %0 = load <4 x float>, <4 x float>* @llvm_mips_ftint_u_w_ARG1 %1 = tail call <4 x i32> @llvm.mips.ftint.u.w(<4 x float> %0) store <4 x i32> %1, <4 x i32>* @llvm_mips_ftint_u_w_RES ret void @@ -199,7 +199,7 @@ declare <4 x i32> @llvm.mips.ftint.u.w(<4 x float>) nounwind define void @llvm_mips_ftint_u_d_test() nounwind { entry: - %0 = load <2 x double>* @llvm_mips_ftint_u_d_ARG1 + %0 = load <2 x double>, <2 x double>* @llvm_mips_ftint_u_d_ARG1 %1 = tail call <2 x i64> @llvm.mips.ftint.u.d(<2 x double> %0) store <2 x i64> %1, <2 x i64>* @llvm_mips_ftint_u_d_RES ret void diff --git a/test/CodeGen/Mips/msa/2rf_tq.ll b/test/CodeGen/Mips/msa/2rf_tq.ll index 6f3c508f5b8c0..9b7f02a5ef33c 100644 --- a/test/CodeGen/Mips/msa/2rf_tq.ll +++ b/test/CodeGen/Mips/msa/2rf_tq.ll @@ -10,8 +10,8 @@ define void @llvm_mips_ftq_h_test() nounwind { entry: - %0 = load <4 x float>* @llvm_mips_ftq_h_ARG1 - %1 = load <4 x float>* @llvm_mips_ftq_h_ARG2 + %0 = load <4 x float>, <4 x float>* @llvm_mips_ftq_h_ARG1 + %1 = load <4 x float>, <4 x float>* @llvm_mips_ftq_h_ARG2 %2 = tail call <8 x i16> @llvm.mips.ftq.h(<4 x float> %0, <4 x float> %1) store <8 x i16> %2, <8 x i16>* @llvm_mips_ftq_h_RES ret void @@ -32,8 +32,8 @@ declare <8 x i16> @llvm.mips.ftq.h(<4 x float>, <4 x float>) nounwind define void @llvm_mips_ftq_w_test() nounwind { entry: - %0 = load <2 x double>* @llvm_mips_ftq_w_ARG1 - %1 = load <2 x double>* @llvm_mips_ftq_w_ARG2 + %0 = load <2 x double>, <2 x double>* @llvm_mips_ftq_w_ARG1 + %1 = load <2 x double>, <2 x double>* @llvm_mips_ftq_w_ARG2 %2 = tail call <4 x i32> @llvm.mips.ftq.w(<2 x double> %0, <2 x double> %1) store <4 x i32> %2, <4 x i32>* @llvm_mips_ftq_w_RES ret void diff --git a/test/CodeGen/Mips/msa/3r-a.ll b/test/CodeGen/Mips/msa/3r-a.ll index dab15b66b7cec..db772f918614c 100644 --- a/test/CodeGen/Mips/msa/3r-a.ll +++ b/test/CodeGen/Mips/msa/3r-a.ll @@ -15,8 +15,8 @@ define void @llvm_mips_add_a_b_test() nounwind { entry: - %0 = load <16 x i8>* @llvm_mips_add_a_b_ARG1 - %1 = load <16 x i8>* @llvm_mips_add_a_b_ARG2 + %0 = load <16 x i8>, <16 x i8>* @llvm_mips_add_a_b_ARG1 + %1 = load <16 x i8>, <16 x i8>* @llvm_mips_add_a_b_ARG2 %2 = tail call <16 x i8> @llvm.mips.add.a.b(<16 x i8> %0, <16 x i8> %1) store <16 x i8> %2, <16 x i8>* @llvm_mips_add_a_b_RES ret void @@ -40,8 +40,8 @@ declare <16 x i8> @llvm.mips.add.a.b(<16 x i8>, <16 x i8>) nounwind define void @llvm_mips_add_a_h_test() nounwind { entry: - %0 = load <8 x i16>* @llvm_mips_add_a_h_ARG1 - %1 = load <8 x i16>* @llvm_mips_add_a_h_ARG2 + %0 = load <8 x i16>, <8 x i16>* @llvm_mips_add_a_h_ARG1 + %1 = load <8 x i16>, <8 x i16>* @llvm_mips_add_a_h_ARG2 %2 = tail call <8 x i16> @llvm.mips.add.a.h(<8 x i16> %0, <8 x i16> %1) store <8 x i16> %2, <8 x i16>* @llvm_mips_add_a_h_RES ret void @@ -65,8 +65,8 @@ declare <8 x i16> @llvm.mips.add.a.h(<8 x i16>, <8 x i16>) nounwind define void @llvm_mips_add_a_w_test() nounwind { entry: - %0 = load <4 x i32>* @llvm_mips_add_a_w_ARG1 - %1 = load <4 x i32>* @llvm_mips_add_a_w_ARG2 + %0 = load <4 x i32>, <4 x i32>* @llvm_mips_add_a_w_ARG1 + %1 = load <4 x i32>, <4 x i32>* @llvm_mips_add_a_w_ARG2 %2 = tail call <4 x i32> @llvm.mips.add.a.w(<4 x i32> %0, <4 x i32> %1) store <4 x i32> %2, <4 x i32>* @llvm_mips_add_a_w_RES ret void @@ -90,8 +90,8 @@ declare <4 x i32> @llvm.mips.add.a.w(<4 x i32>, <4 x i32>) nounwind define void @llvm_mips_add_a_d_test() nounwind { entry: - %0 = load <2 x i64>* @llvm_mips_add_a_d_ARG1 - %1 = load <2 x i64>* @llvm_mips_add_a_d_ARG2 + %0 = load <2 x i64>, <2 x i64>* @llvm_mips_add_a_d_ARG1 + %1 = load <2 x i64>, <2 x i64>* @llvm_mips_add_a_d_ARG2 %2 = tail call <2 x i64> @llvm.mips.add.a.d(<2 x i64> %0, <2 x i64> %1) store <2 x i64> %2, <2 x i64>* @llvm_mips_add_a_d_RES ret void @@ -115,8 +115,8 @@ declare <2 x i64> @llvm.mips.add.a.d(<2 x i64>, <2 x i64>) nounwind define void @llvm_mips_adds_a_b_test() nounwind { entry: - %0 = load <16 x i8>* @llvm_mips_adds_a_b_ARG1 - %1 = load <16 x i8>* @llvm_mips_adds_a_b_ARG2 + %0 = load <16 x i8>, <16 x i8>* @llvm_mips_adds_a_b_ARG1 + %1 = load <16 x i8>, <16 x i8>* @llvm_mips_adds_a_b_ARG2 %2 = tail call <16 x i8> @llvm.mips.adds.a.b(<16 x i8> %0, <16 x i8> %1) store <16 x i8> %2, <16 x i8>* @llvm_mips_adds_a_b_RES ret void @@ -140,8 +140,8 @@ declare <16 x i8> @llvm.mips.adds.a.b(<16 x i8>, <16 x i8>) nounwind define void @llvm_mips_adds_a_h_test() nounwind { entry: - %0 = load <8 x i16>* @llvm_mips_adds_a_h_ARG1 - %1 = load <8 x i16>* @llvm_mips_adds_a_h_ARG2 + %0 = load <8 x i16>, <8 x i16>* @llvm_mips_adds_a_h_ARG1 + %1 = load <8 x i16>, <8 x i16>* @llvm_mips_adds_a_h_ARG2 %2 = tail call <8 x i16> @llvm.mips.adds.a.h(<8 x i16> %0, <8 x i16> %1) store <8 x i16> %2, <8 x i16>* @llvm_mips_adds_a_h_RES ret void @@ -165,8 +165,8 @@ declare <8 x i16> @llvm.mips.adds.a.h(<8 x i16>, <8 x i16>) nounwind define void @llvm_mips_adds_a_w_test() nounwind { entry: - %0 = load <4 x i32>* @llvm_mips_adds_a_w_ARG1 - %1 = load <4 x i32>* @llvm_mips_adds_a_w_ARG2 + %0 = load <4 x i32>, <4 x i32>* @llvm_mips_adds_a_w_ARG1 + %1 = load <4 x i32>, <4 x i32>* @llvm_mips_adds_a_w_ARG2 %2 = tail call <4 x i32> @llvm.mips.adds.a.w(<4 x i32> %0, <4 x i32> %1) store <4 x i32> %2, <4 x i32>* @llvm_mips_adds_a_w_RES ret void @@ -190,8 +190,8 @@ declare <4 x i32> @llvm.mips.adds.a.w(<4 x i32>, <4 x i32>) nounwind define void @llvm_mips_adds_a_d_test() nounwind { entry: - %0 = load <2 x i64>* @llvm_mips_adds_a_d_ARG1 - %1 = load <2 x i64>* @llvm_mips_adds_a_d_ARG2 + %0 = load <2 x i64>, <2 x i64>* @llvm_mips_adds_a_d_ARG1 + %1 = load <2 x i64>, <2 x i64>* @llvm_mips_adds_a_d_ARG2 %2 = tail call <2 x i64> @llvm.mips.adds.a.d(<2 x i64> %0, <2 x i64> %1) store <2 x i64> %2, <2 x i64>* @llvm_mips_adds_a_d_RES ret void @@ -215,8 +215,8 @@ declare <2 x i64> @llvm.mips.adds.a.d(<2 x i64>, <2 x i64>) nounwind define void @llvm_mips_adds_s_b_test() nounwind { entry: - %0 = load <16 x i8>* @llvm_mips_adds_s_b_ARG1 - %1 = load <16 x i8>* @llvm_mips_adds_s_b_ARG2 + %0 = load <16 x i8>, <16 x i8>* @llvm_mips_adds_s_b_ARG1 + %1 = load <16 x i8>, <16 x i8>* @llvm_mips_adds_s_b_ARG2 %2 = tail call <16 x i8> @llvm.mips.adds.s.b(<16 x i8> %0, <16 x i8> %1) store <16 x i8> %2, <16 x i8>* @llvm_mips_adds_s_b_RES ret void @@ -240,8 +240,8 @@ declare <16 x i8> @llvm.mips.adds.s.b(<16 x i8>, <16 x i8>) nounwind define void @llvm_mips_adds_s_h_test() nounwind { entry: - %0 = load <8 x i16>* @llvm_mips_adds_s_h_ARG1 - %1 = load <8 x i16>* @llvm_mips_adds_s_h_ARG2 + %0 = load <8 x i16>, <8 x i16>* @llvm_mips_adds_s_h_ARG1 + %1 = load <8 x i16>, <8 x i16>* @llvm_mips_adds_s_h_ARG2 %2 = tail call <8 x i16> @llvm.mips.adds.s.h(<8 x i16> %0, <8 x i16> %1) store <8 x i16> %2, <8 x i16>* @llvm_mips_adds_s_h_RES ret void @@ -265,8 +265,8 @@ declare <8 x i16> @llvm.mips.adds.s.h(<8 x i16>, <8 x i16>) nounwind define void @llvm_mips_adds_s_w_test() nounwind { entry: - %0 = load <4 x i32>* @llvm_mips_adds_s_w_ARG1 - %1 = load <4 x i32>* @llvm_mips_adds_s_w_ARG2 + %0 = load <4 x i32>, <4 x i32>* @llvm_mips_adds_s_w_ARG1 + %1 = load <4 x i32>, <4 x i32>* @llvm_mips_adds_s_w_ARG2 %2 = tail call <4 x i32> @llvm.mips.adds.s.w(<4 x i32> %0, <4 x i32> %1) store <4 x i32> %2, <4 x i32>* @llvm_mips_adds_s_w_RES ret void @@ -290,8 +290,8 @@ declare <4 x i32> @llvm.mips.adds.s.w(<4 x i32>, <4 x i32>) nounwind define void @llvm_mips_adds_s_d_test() nounwind { entry: - %0 = load <2 x i64>* @llvm_mips_adds_s_d_ARG1 - %1 = load <2 x i64>* @llvm_mips_adds_s_d_ARG2 + %0 = load <2 x i64>, <2 x i64>* @llvm_mips_adds_s_d_ARG1 + %1 = load <2 x i64>, <2 x i64>* @llvm_mips_adds_s_d_ARG2 %2 = tail call <2 x i64> @llvm.mips.adds.s.d(<2 x i64> %0, <2 x i64> %1) store <2 x i64> %2, <2 x i64>* @llvm_mips_adds_s_d_RES ret void @@ -315,8 +315,8 @@ declare <2 x i64> @llvm.mips.adds.s.d(<2 x i64>, <2 x i64>) nounwind define void @llvm_mips_adds_u_b_test() nounwind { entry: - %0 = load <16 x i8>* @llvm_mips_adds_u_b_ARG1 - %1 = load <16 x i8>* @llvm_mips_adds_u_b_ARG2 + %0 = load <16 x i8>, <16 x i8>* @llvm_mips_adds_u_b_ARG1 + %1 = load <16 x i8>, <16 x i8>* @llvm_mips_adds_u_b_ARG2 %2 = tail call <16 x i8> @llvm.mips.adds.u.b(<16 x i8> %0, <16 x i8> %1) store <16 x i8> %2, <16 x i8>* @llvm_mips_adds_u_b_RES ret void @@ -340,8 +340,8 @@ declare <16 x i8> @llvm.mips.adds.u.b(<16 x i8>, <16 x i8>) nounwind define void @llvm_mips_adds_u_h_test() nounwind { entry: - %0 = load <8 x i16>* @llvm_mips_adds_u_h_ARG1 - %1 = load <8 x i16>* @llvm_mips_adds_u_h_ARG2 + %0 = load <8 x i16>, <8 x i16>* @llvm_mips_adds_u_h_ARG1 + %1 = load <8 x i16>, <8 x i16>* @llvm_mips_adds_u_h_ARG2 %2 = tail call <8 x i16> @llvm.mips.adds.u.h(<8 x i16> %0, <8 x i16> %1) store <8 x i16> %2, <8 x i16>* @llvm_mips_adds_u_h_RES ret void @@ -365,8 +365,8 @@ declare <8 x i16> @llvm.mips.adds.u.h(<8 x i16>, <8 x i16>) nounwind define void @llvm_mips_adds_u_w_test() nounwind { entry: - %0 = load <4 x i32>* @llvm_mips_adds_u_w_ARG1 - %1 = load <4 x i32>* @llvm_mips_adds_u_w_ARG2 + %0 = load <4 x i32>, <4 x i32>* @llvm_mips_adds_u_w_ARG1 + %1 = load <4 x i32>, <4 x i32>* @llvm_mips_adds_u_w_ARG2 %2 = tail call <4 x i32> @llvm.mips.adds.u.w(<4 x i32> %0, <4 x i32> %1) store <4 x i32> %2, <4 x i32>* @llvm_mips_adds_u_w_RES ret void @@ -390,8 +390,8 @@ declare <4 x i32> @llvm.mips.adds.u.w(<4 x i32>, <4 x i32>) nounwind define void @llvm_mips_adds_u_d_test() nounwind { entry: - %0 = load <2 x i64>* @llvm_mips_adds_u_d_ARG1 - %1 = load <2 x i64>* @llvm_mips_adds_u_d_ARG2 + %0 = load <2 x i64>, <2 x i64>* @llvm_mips_adds_u_d_ARG1 + %1 = load <2 x i64>, <2 x i64>* @llvm_mips_adds_u_d_ARG2 %2 = tail call <2 x i64> @llvm.mips.adds.u.d(<2 x i64> %0, <2 x i64> %1) store <2 x i64> %2, <2 x i64>* @llvm_mips_adds_u_d_RES ret void @@ -415,8 +415,8 @@ declare <2 x i64> @llvm.mips.adds.u.d(<2 x i64>, <2 x i64>) nounwind define void @llvm_mips_addv_b_test() nounwind { entry: - %0 = load <16 x i8>* @llvm_mips_addv_b_ARG1 - %1 = load <16 x i8>* @llvm_mips_addv_b_ARG2 + %0 = load <16 x i8>, <16 x i8>* @llvm_mips_addv_b_ARG1 + %1 = load <16 x i8>, <16 x i8>* @llvm_mips_addv_b_ARG2 %2 = tail call <16 x i8> @llvm.mips.addv.b(<16 x i8> %0, <16 x i8> %1) store <16 x i8> %2, <16 x i8>* @llvm_mips_addv_b_RES ret void @@ -440,8 +440,8 @@ declare <16 x i8> @llvm.mips.addv.b(<16 x i8>, <16 x i8>) nounwind define void @llvm_mips_addv_h_test() nounwind { entry: - %0 = load <8 x i16>* @llvm_mips_addv_h_ARG1 - %1 = load <8 x i16>* @llvm_mips_addv_h_ARG2 + %0 = load <8 x i16>, <8 x i16>* @llvm_mips_addv_h_ARG1 + %1 = load <8 x i16>, <8 x i16>* @llvm_mips_addv_h_ARG2 %2 = tail call <8 x i16> @llvm.mips.addv.h(<8 x i16> %0, <8 x i16> %1) store <8 x i16> %2, <8 x i16>* @llvm_mips_addv_h_RES ret void @@ -465,8 +465,8 @@ declare <8 x i16> @llvm.mips.addv.h(<8 x i16>, <8 x i16>) nounwind define void @llvm_mips_addv_w_test() nounwind { entry: - %0 = load <4 x i32>* @llvm_mips_addv_w_ARG1 - %1 = load <4 x i32>* @llvm_mips_addv_w_ARG2 + %0 = load <4 x i32>, <4 x i32>* @llvm_mips_addv_w_ARG1 + %1 = load <4 x i32>, <4 x i32>* @llvm_mips_addv_w_ARG2 %2 = tail call <4 x i32> @llvm.mips.addv.w(<4 x i32> %0, <4 x i32> %1) store <4 x i32> %2, <4 x i32>* @llvm_mips_addv_w_RES ret void @@ -490,8 +490,8 @@ declare <4 x i32> @llvm.mips.addv.w(<4 x i32>, <4 x i32>) nounwind define void @llvm_mips_addv_d_test() nounwind { entry: - %0 = load <2 x i64>* @llvm_mips_addv_d_ARG1 - %1 = load <2 x i64>* @llvm_mips_addv_d_ARG2 + %0 = load <2 x i64>, <2 x i64>* @llvm_mips_addv_d_ARG1 + %1 = load <2 x i64>, <2 x i64>* @llvm_mips_addv_d_ARG2 %2 = tail call <2 x i64> @llvm.mips.addv.d(<2 x i64> %0, <2 x i64> %1) store <2 x i64> %2, <2 x i64>* @llvm_mips_addv_d_RES ret void @@ -512,8 +512,8 @@ declare <2 x i64> @llvm.mips.addv.d(<2 x i64>, <2 x i64>) nounwind define void @addv_b_test() nounwind { entry: - %0 = load <16 x i8>* @llvm_mips_addv_b_ARG1 - %1 = load <16 x i8>* @llvm_mips_addv_b_ARG2 + %0 = load <16 x i8>, <16 x i8>* @llvm_mips_addv_b_ARG1 + %1 = load <16 x i8>, <16 x i8>* @llvm_mips_addv_b_ARG2 %2 = add <16 x i8> %0, %1 store <16 x i8> %2, <16 x i8>* @llvm_mips_addv_b_RES ret void @@ -532,8 +532,8 @@ entry: define void @addv_h_test() nounwind { entry: - %0 = load <8 x i16>* @llvm_mips_addv_h_ARG1 - %1 = load <8 x i16>* @llvm_mips_addv_h_ARG2 + %0 = load <8 x i16>, <8 x i16>* @llvm_mips_addv_h_ARG1 + %1 = load <8 x i16>, <8 x i16>* @llvm_mips_addv_h_ARG2 %2 = add <8 x i16> %0, %1 store <8 x i16> %2, <8 x i16>* @llvm_mips_addv_h_RES ret void @@ -552,8 +552,8 @@ entry: define void @addv_w_test() nounwind { entry: - %0 = load <4 x i32>* @llvm_mips_addv_w_ARG1 - %1 = load <4 x i32>* @llvm_mips_addv_w_ARG2 + %0 = load <4 x i32>, <4 x i32>* @llvm_mips_addv_w_ARG1 + %1 = load <4 x i32>, <4 x i32>* @llvm_mips_addv_w_ARG2 %2 = add <4 x i32> %0, %1 store <4 x i32> %2, <4 x i32>* @llvm_mips_addv_w_RES ret void @@ -572,8 +572,8 @@ entry: define void @addv_d_test() nounwind { entry: - %0 = load <2 x i64>* @llvm_mips_addv_d_ARG1 - %1 = load <2 x i64>* @llvm_mips_addv_d_ARG2 + %0 = load <2 x i64>, <2 x i64>* @llvm_mips_addv_d_ARG1 + %1 = load <2 x i64>, <2 x i64>* @llvm_mips_addv_d_ARG2 %2 = add <2 x i64> %0, %1 store <2 x i64> %2, <2 x i64>* @llvm_mips_addv_d_RES ret void @@ -595,8 +595,8 @@ entry: define void @llvm_mips_asub_s_b_test() nounwind { entry: - %0 = load <16 x i8>* @llvm_mips_asub_s_b_ARG1 - %1 = load <16 x i8>* @llvm_mips_asub_s_b_ARG2 + %0 = load <16 x i8>, <16 x i8>* @llvm_mips_asub_s_b_ARG1 + %1 = load <16 x i8>, <16 x i8>* @llvm_mips_asub_s_b_ARG2 %2 = tail call <16 x i8> @llvm.mips.asub.s.b(<16 x i8> %0, <16 x i8> %1) store <16 x i8> %2, <16 x i8>* @llvm_mips_asub_s_b_RES ret void @@ -620,8 +620,8 @@ declare <16 x i8> @llvm.mips.asub.s.b(<16 x i8>, <16 x i8>) nounwind define void @llvm_mips_asub_s_h_test() nounwind { entry: - %0 = load <8 x i16>* @llvm_mips_asub_s_h_ARG1 - %1 = load <8 x i16>* @llvm_mips_asub_s_h_ARG2 + %0 = load <8 x i16>, <8 x i16>* @llvm_mips_asub_s_h_ARG1 + %1 = load <8 x i16>, <8 x i16>* @llvm_mips_asub_s_h_ARG2 %2 = tail call <8 x i16> @llvm.mips.asub.s.h(<8 x i16> %0, <8 x i16> %1) store <8 x i16> %2, <8 x i16>* @llvm_mips_asub_s_h_RES ret void @@ -645,8 +645,8 @@ declare <8 x i16> @llvm.mips.asub.s.h(<8 x i16>, <8 x i16>) nounwind define void @llvm_mips_asub_s_w_test() nounwind { entry: - %0 = load <4 x i32>* @llvm_mips_asub_s_w_ARG1 - %1 = load <4 x i32>* @llvm_mips_asub_s_w_ARG2 + %0 = load <4 x i32>, <4 x i32>* @llvm_mips_asub_s_w_ARG1 + %1 = load <4 x i32>, <4 x i32>* @llvm_mips_asub_s_w_ARG2 %2 = tail call <4 x i32> @llvm.mips.asub.s.w(<4 x i32> %0, <4 x i32> %1) store <4 x i32> %2, <4 x i32>* @llvm_mips_asub_s_w_RES ret void @@ -670,8 +670,8 @@ declare <4 x i32> @llvm.mips.asub.s.w(<4 x i32>, <4 x i32>) nounwind define void @llvm_mips_asub_s_d_test() nounwind { entry: - %0 = load <2 x i64>* @llvm_mips_asub_s_d_ARG1 - %1 = load <2 x i64>* @llvm_mips_asub_s_d_ARG2 + %0 = load <2 x i64>, <2 x i64>* @llvm_mips_asub_s_d_ARG1 + %1 = load <2 x i64>, <2 x i64>* @llvm_mips_asub_s_d_ARG2 %2 = tail call <2 x i64> @llvm.mips.asub.s.d(<2 x i64> %0, <2 x i64> %1) store <2 x i64> %2, <2 x i64>* @llvm_mips_asub_s_d_RES ret void @@ -695,8 +695,8 @@ declare <2 x i64> @llvm.mips.asub.s.d(<2 x i64>, <2 x i64>) nounwind define void @llvm_mips_asub_u_b_test() nounwind { entry: - %0 = load <16 x i8>* @llvm_mips_asub_u_b_ARG1 - %1 = load <16 x i8>* @llvm_mips_asub_u_b_ARG2 + %0 = load <16 x i8>, <16 x i8>* @llvm_mips_asub_u_b_ARG1 + %1 = load <16 x i8>, <16 x i8>* @llvm_mips_asub_u_b_ARG2 %2 = tail call <16 x i8> @llvm.mips.asub.u.b(<16 x i8> %0, <16 x i8> %1) store <16 x i8> %2, <16 x i8>* @llvm_mips_asub_u_b_RES ret void @@ -720,8 +720,8 @@ declare <16 x i8> @llvm.mips.asub.u.b(<16 x i8>, <16 x i8>) nounwind define void @llvm_mips_asub_u_h_test() nounwind { entry: - %0 = load <8 x i16>* @llvm_mips_asub_u_h_ARG1 - %1 = load <8 x i16>* @llvm_mips_asub_u_h_ARG2 + %0 = load <8 x i16>, <8 x i16>* @llvm_mips_asub_u_h_ARG1 + %1 = load <8 x i16>, <8 x i16>* @llvm_mips_asub_u_h_ARG2 %2 = tail call <8 x i16> @llvm.mips.asub.u.h(<8 x i16> %0, <8 x i16> %1) store <8 x i16> %2, <8 x i16>* @llvm_mips_asub_u_h_RES ret void @@ -745,8 +745,8 @@ declare <8 x i16> @llvm.mips.asub.u.h(<8 x i16>, <8 x i16>) nounwind define void @llvm_mips_asub_u_w_test() nounwind { entry: - %0 = load <4 x i32>* @llvm_mips_asub_u_w_ARG1 - %1 = load <4 x i32>* @llvm_mips_asub_u_w_ARG2 + %0 = load <4 x i32>, <4 x i32>* @llvm_mips_asub_u_w_ARG1 + %1 = load <4 x i32>, <4 x i32>* @llvm_mips_asub_u_w_ARG2 %2 = tail call <4 x i32> @llvm.mips.asub.u.w(<4 x i32> %0, <4 x i32> %1) store <4 x i32> %2, <4 x i32>* @llvm_mips_asub_u_w_RES ret void @@ -770,8 +770,8 @@ declare <4 x i32> @llvm.mips.asub.u.w(<4 x i32>, <4 x i32>) nounwind define void @llvm_mips_asub_u_d_test() nounwind { entry: - %0 = load <2 x i64>* @llvm_mips_asub_u_d_ARG1 - %1 = load <2 x i64>* @llvm_mips_asub_u_d_ARG2 + %0 = load <2 x i64>, <2 x i64>* @llvm_mips_asub_u_d_ARG1 + %1 = load <2 x i64>, <2 x i64>* @llvm_mips_asub_u_d_ARG2 %2 = tail call <2 x i64> @llvm.mips.asub.u.d(<2 x i64> %0, <2 x i64> %1) store <2 x i64> %2, <2 x i64>* @llvm_mips_asub_u_d_RES ret void @@ -795,8 +795,8 @@ declare <2 x i64> @llvm.mips.asub.u.d(<2 x i64>, <2 x i64>) nounwind define void @llvm_mips_ave_s_b_test() nounwind { entry: - %0 = load <16 x i8>* @llvm_mips_ave_s_b_ARG1 - %1 = load <16 x i8>* @llvm_mips_ave_s_b_ARG2 + %0 = load <16 x i8>, <16 x i8>* @llvm_mips_ave_s_b_ARG1 + %1 = load <16 x i8>, <16 x i8>* @llvm_mips_ave_s_b_ARG2 %2 = tail call <16 x i8> @llvm.mips.ave.s.b(<16 x i8> %0, <16 x i8> %1) store <16 x i8> %2, <16 x i8>* @llvm_mips_ave_s_b_RES ret void @@ -820,8 +820,8 @@ declare <16 x i8> @llvm.mips.ave.s.b(<16 x i8>, <16 x i8>) nounwind define void @llvm_mips_ave_s_h_test() nounwind { entry: - %0 = load <8 x i16>* @llvm_mips_ave_s_h_ARG1 - %1 = load <8 x i16>* @llvm_mips_ave_s_h_ARG2 + %0 = load <8 x i16>, <8 x i16>* @llvm_mips_ave_s_h_ARG1 + %1 = load <8 x i16>, <8 x i16>* @llvm_mips_ave_s_h_ARG2 %2 = tail call <8 x i16> @llvm.mips.ave.s.h(<8 x i16> %0, <8 x i16> %1) store <8 x i16> %2, <8 x i16>* @llvm_mips_ave_s_h_RES ret void @@ -845,8 +845,8 @@ declare <8 x i16> @llvm.mips.ave.s.h(<8 x i16>, <8 x i16>) nounwind define void @llvm_mips_ave_s_w_test() nounwind { entry: - %0 = load <4 x i32>* @llvm_mips_ave_s_w_ARG1 - %1 = load <4 x i32>* @llvm_mips_ave_s_w_ARG2 + %0 = load <4 x i32>, <4 x i32>* @llvm_mips_ave_s_w_ARG1 + %1 = load <4 x i32>, <4 x i32>* @llvm_mips_ave_s_w_ARG2 %2 = tail call <4 x i32> @llvm.mips.ave.s.w(<4 x i32> %0, <4 x i32> %1) store <4 x i32> %2, <4 x i32>* @llvm_mips_ave_s_w_RES ret void @@ -870,8 +870,8 @@ declare <4 x i32> @llvm.mips.ave.s.w(<4 x i32>, <4 x i32>) nounwind define void @llvm_mips_ave_s_d_test() nounwind { entry: - %0 = load <2 x i64>* @llvm_mips_ave_s_d_ARG1 - %1 = load <2 x i64>* @llvm_mips_ave_s_d_ARG2 + %0 = load <2 x i64>, <2 x i64>* @llvm_mips_ave_s_d_ARG1 + %1 = load <2 x i64>, <2 x i64>* @llvm_mips_ave_s_d_ARG2 %2 = tail call <2 x i64> @llvm.mips.ave.s.d(<2 x i64> %0, <2 x i64> %1) store <2 x i64> %2, <2 x i64>* @llvm_mips_ave_s_d_RES ret void @@ -895,8 +895,8 @@ declare <2 x i64> @llvm.mips.ave.s.d(<2 x i64>, <2 x i64>) nounwind define void @llvm_mips_ave_u_b_test() nounwind { entry: - %0 = load <16 x i8>* @llvm_mips_ave_u_b_ARG1 - %1 = load <16 x i8>* @llvm_mips_ave_u_b_ARG2 + %0 = load <16 x i8>, <16 x i8>* @llvm_mips_ave_u_b_ARG1 + %1 = load <16 x i8>, <16 x i8>* @llvm_mips_ave_u_b_ARG2 %2 = tail call <16 x i8> @llvm.mips.ave.u.b(<16 x i8> %0, <16 x i8> %1) store <16 x i8> %2, <16 x i8>* @llvm_mips_ave_u_b_RES ret void @@ -920,8 +920,8 @@ declare <16 x i8> @llvm.mips.ave.u.b(<16 x i8>, <16 x i8>) nounwind define void @llvm_mips_ave_u_h_test() nounwind { entry: - %0 = load <8 x i16>* @llvm_mips_ave_u_h_ARG1 - %1 = load <8 x i16>* @llvm_mips_ave_u_h_ARG2 + %0 = load <8 x i16>, <8 x i16>* @llvm_mips_ave_u_h_ARG1 + %1 = load <8 x i16>, <8 x i16>* @llvm_mips_ave_u_h_ARG2 %2 = tail call <8 x i16> @llvm.mips.ave.u.h(<8 x i16> %0, <8 x i16> %1) store <8 x i16> %2, <8 x i16>* @llvm_mips_ave_u_h_RES ret void @@ -945,8 +945,8 @@ declare <8 x i16> @llvm.mips.ave.u.h(<8 x i16>, <8 x i16>) nounwind define void @llvm_mips_ave_u_w_test() nounwind { entry: - %0 = load <4 x i32>* @llvm_mips_ave_u_w_ARG1 - %1 = load <4 x i32>* @llvm_mips_ave_u_w_ARG2 + %0 = load <4 x i32>, <4 x i32>* @llvm_mips_ave_u_w_ARG1 + %1 = load <4 x i32>, <4 x i32>* @llvm_mips_ave_u_w_ARG2 %2 = tail call <4 x i32> @llvm.mips.ave.u.w(<4 x i32> %0, <4 x i32> %1) store <4 x i32> %2, <4 x i32>* @llvm_mips_ave_u_w_RES ret void @@ -970,8 +970,8 @@ declare <4 x i32> @llvm.mips.ave.u.w(<4 x i32>, <4 x i32>) nounwind define void @llvm_mips_ave_u_d_test() nounwind { entry: - %0 = load <2 x i64>* @llvm_mips_ave_u_d_ARG1 - %1 = load <2 x i64>* @llvm_mips_ave_u_d_ARG2 + %0 = load <2 x i64>, <2 x i64>* @llvm_mips_ave_u_d_ARG1 + %1 = load <2 x i64>, <2 x i64>* @llvm_mips_ave_u_d_ARG2 %2 = tail call <2 x i64> @llvm.mips.ave.u.d(<2 x i64> %0, <2 x i64> %1) store <2 x i64> %2, <2 x i64>* @llvm_mips_ave_u_d_RES ret void @@ -995,8 +995,8 @@ declare <2 x i64> @llvm.mips.ave.u.d(<2 x i64>, <2 x i64>) nounwind define void @llvm_mips_aver_s_b_test() nounwind { entry: - %0 = load <16 x i8>* @llvm_mips_aver_s_b_ARG1 - %1 = load <16 x i8>* @llvm_mips_aver_s_b_ARG2 + %0 = load <16 x i8>, <16 x i8>* @llvm_mips_aver_s_b_ARG1 + %1 = load <16 x i8>, <16 x i8>* @llvm_mips_aver_s_b_ARG2 %2 = tail call <16 x i8> @llvm.mips.aver.s.b(<16 x i8> %0, <16 x i8> %1) store <16 x i8> %2, <16 x i8>* @llvm_mips_aver_s_b_RES ret void @@ -1020,8 +1020,8 @@ declare <16 x i8> @llvm.mips.aver.s.b(<16 x i8>, <16 x i8>) nounwind define void @llvm_mips_aver_s_h_test() nounwind { entry: - %0 = load <8 x i16>* @llvm_mips_aver_s_h_ARG1 - %1 = load <8 x i16>* @llvm_mips_aver_s_h_ARG2 + %0 = load <8 x i16>, <8 x i16>* @llvm_mips_aver_s_h_ARG1 + %1 = load <8 x i16>, <8 x i16>* @llvm_mips_aver_s_h_ARG2 %2 = tail call <8 x i16> @llvm.mips.aver.s.h(<8 x i16> %0, <8 x i16> %1) store <8 x i16> %2, <8 x i16>* @llvm_mips_aver_s_h_RES ret void @@ -1045,8 +1045,8 @@ declare <8 x i16> @llvm.mips.aver.s.h(<8 x i16>, <8 x i16>) nounwind define void @llvm_mips_aver_s_w_test() nounwind { entry: - %0 = load <4 x i32>* @llvm_mips_aver_s_w_ARG1 - %1 = load <4 x i32>* @llvm_mips_aver_s_w_ARG2 + %0 = load <4 x i32>, <4 x i32>* @llvm_mips_aver_s_w_ARG1 + %1 = load <4 x i32>, <4 x i32>* @llvm_mips_aver_s_w_ARG2 %2 = tail call <4 x i32> @llvm.mips.aver.s.w(<4 x i32> %0, <4 x i32> %1) store <4 x i32> %2, <4 x i32>* @llvm_mips_aver_s_w_RES ret void @@ -1070,8 +1070,8 @@ declare <4 x i32> @llvm.mips.aver.s.w(<4 x i32>, <4 x i32>) nounwind define void @llvm_mips_aver_s_d_test() nounwind { entry: - %0 = load <2 x i64>* @llvm_mips_aver_s_d_ARG1 - %1 = load <2 x i64>* @llvm_mips_aver_s_d_ARG2 + %0 = load <2 x i64>, <2 x i64>* @llvm_mips_aver_s_d_ARG1 + %1 = load <2 x i64>, <2 x i64>* @llvm_mips_aver_s_d_ARG2 %2 = tail call <2 x i64> @llvm.mips.aver.s.d(<2 x i64> %0, <2 x i64> %1) store <2 x i64> %2, <2 x i64>* @llvm_mips_aver_s_d_RES ret void @@ -1095,8 +1095,8 @@ declare <2 x i64> @llvm.mips.aver.s.d(<2 x i64>, <2 x i64>) nounwind define void @llvm_mips_aver_u_b_test() nounwind { entry: - %0 = load <16 x i8>* @llvm_mips_aver_u_b_ARG1 - %1 = load <16 x i8>* @llvm_mips_aver_u_b_ARG2 + %0 = load <16 x i8>, <16 x i8>* @llvm_mips_aver_u_b_ARG1 + %1 = load <16 x i8>, <16 x i8>* @llvm_mips_aver_u_b_ARG2 %2 = tail call <16 x i8> @llvm.mips.aver.u.b(<16 x i8> %0, <16 x i8> %1) store <16 x i8> %2, <16 x i8>* @llvm_mips_aver_u_b_RES ret void @@ -1120,8 +1120,8 @@ declare <16 x i8> @llvm.mips.aver.u.b(<16 x i8>, <16 x i8>) nounwind define void @llvm_mips_aver_u_h_test() nounwind { entry: - %0 = load <8 x i16>* @llvm_mips_aver_u_h_ARG1 - %1 = load <8 x i16>* @llvm_mips_aver_u_h_ARG2 + %0 = load <8 x i16>, <8 x i16>* @llvm_mips_aver_u_h_ARG1 + %1 = load <8 x i16>, <8 x i16>* @llvm_mips_aver_u_h_ARG2 %2 = tail call <8 x i16> @llvm.mips.aver.u.h(<8 x i16> %0, <8 x i16> %1) store <8 x i16> %2, <8 x i16>* @llvm_mips_aver_u_h_RES ret void @@ -1145,8 +1145,8 @@ declare <8 x i16> @llvm.mips.aver.u.h(<8 x i16>, <8 x i16>) nounwind define void @llvm_mips_aver_u_w_test() nounwind { entry: - %0 = load <4 x i32>* @llvm_mips_aver_u_w_ARG1 - %1 = load <4 x i32>* @llvm_mips_aver_u_w_ARG2 + %0 = load <4 x i32>, <4 x i32>* @llvm_mips_aver_u_w_ARG1 + %1 = load <4 x i32>, <4 x i32>* @llvm_mips_aver_u_w_ARG2 %2 = tail call <4 x i32> @llvm.mips.aver.u.w(<4 x i32> %0, <4 x i32> %1) store <4 x i32> %2, <4 x i32>* @llvm_mips_aver_u_w_RES ret void @@ -1170,8 +1170,8 @@ declare <4 x i32> @llvm.mips.aver.u.w(<4 x i32>, <4 x i32>) nounwind define void @llvm_mips_aver_u_d_test() nounwind { entry: - %0 = load <2 x i64>* @llvm_mips_aver_u_d_ARG1 - %1 = load <2 x i64>* @llvm_mips_aver_u_d_ARG2 + %0 = load <2 x i64>, <2 x i64>* @llvm_mips_aver_u_d_ARG1 + %1 = load <2 x i64>, <2 x i64>* @llvm_mips_aver_u_d_ARG2 %2 = tail call <2 x i64> @llvm.mips.aver.u.d(<2 x i64> %0, <2 x i64> %1) store <2 x i64> %2, <2 x i64>* @llvm_mips_aver_u_d_RES ret void diff --git a/test/CodeGen/Mips/msa/3r-b.ll b/test/CodeGen/Mips/msa/3r-b.ll index a05d19b4d490c..2ecdc42900674 100644 --- a/test/CodeGen/Mips/msa/3r-b.ll +++ b/test/CodeGen/Mips/msa/3r-b.ll @@ -10,8 +10,8 @@ define void @llvm_mips_bclr_b_test() nounwind { entry: - %0 = load <16 x i8>* @llvm_mips_bclr_b_ARG1 - %1 = load <16 x i8>* @llvm_mips_bclr_b_ARG2 + %0 = load <16 x i8>, <16 x i8>* @llvm_mips_bclr_b_ARG1 + %1 = load <16 x i8>, <16 x i8>* @llvm_mips_bclr_b_ARG2 %2 = tail call <16 x i8> @llvm.mips.bclr.b(<16 x i8> %0, <16 x i8> %1) store <16 x i8> %2, <16 x i8>* @llvm_mips_bclr_b_RES ret void @@ -32,8 +32,8 @@ declare <16 x i8> @llvm.mips.bclr.b(<16 x i8>, <16 x i8>) nounwind define void @llvm_mips_bclr_h_test() nounwind { entry: - %0 = load <8 x i16>* @llvm_mips_bclr_h_ARG1 - %1 = load <8 x i16>* @llvm_mips_bclr_h_ARG2 + %0 = load <8 x i16>, <8 x i16>* @llvm_mips_bclr_h_ARG1 + %1 = load <8 x i16>, <8 x i16>* @llvm_mips_bclr_h_ARG2 %2 = tail call <8 x i16> @llvm.mips.bclr.h(<8 x i16> %0, <8 x i16> %1) store <8 x i16> %2, <8 x i16>* @llvm_mips_bclr_h_RES ret void @@ -54,8 +54,8 @@ declare <8 x i16> @llvm.mips.bclr.h(<8 x i16>, <8 x i16>) nounwind define void @llvm_mips_bclr_w_test() nounwind { entry: - %0 = load <4 x i32>* @llvm_mips_bclr_w_ARG1 - %1 = load <4 x i32>* @llvm_mips_bclr_w_ARG2 + %0 = load <4 x i32>, <4 x i32>* @llvm_mips_bclr_w_ARG1 + %1 = load <4 x i32>, <4 x i32>* @llvm_mips_bclr_w_ARG2 %2 = tail call <4 x i32> @llvm.mips.bclr.w(<4 x i32> %0, <4 x i32> %1) store <4 x i32> %2, <4 x i32>* @llvm_mips_bclr_w_RES ret void @@ -76,8 +76,8 @@ declare <4 x i32> @llvm.mips.bclr.w(<4 x i32>, <4 x i32>) nounwind define void @llvm_mips_bclr_d_test() nounwind { entry: - %0 = load <2 x i64>* @llvm_mips_bclr_d_ARG1 - %1 = load <2 x i64>* @llvm_mips_bclr_d_ARG2 + %0 = load <2 x i64>, <2 x i64>* @llvm_mips_bclr_d_ARG1 + %1 = load <2 x i64>, <2 x i64>* @llvm_mips_bclr_d_ARG2 %2 = tail call <2 x i64> @llvm.mips.bclr.d(<2 x i64> %0, <2 x i64> %1) store <2 x i64> %2, <2 x i64>* @llvm_mips_bclr_d_RES ret void @@ -99,9 +99,9 @@ declare <2 x i64> @llvm.mips.bclr.d(<2 x i64>, <2 x i64>) nounwind define void @llvm_mips_binsl_b_test() nounwind { entry: - %0 = load <16 x i8>* @llvm_mips_binsl_b_ARG1 - %1 = load <16 x i8>* @llvm_mips_binsl_b_ARG2 - %2 = load <16 x i8>* @llvm_mips_binsl_b_ARG3 + %0 = load <16 x i8>, <16 x i8>* @llvm_mips_binsl_b_ARG1 + %1 = load <16 x i8>, <16 x i8>* @llvm_mips_binsl_b_ARG2 + %2 = load <16 x i8>, <16 x i8>* @llvm_mips_binsl_b_ARG3 %3 = tail call <16 x i8> @llvm.mips.binsl.b(<16 x i8> %0, <16 x i8> %1, <16 x i8> %2) store <16 x i8> %3, <16 x i8>* @llvm_mips_binsl_b_RES ret void @@ -127,9 +127,9 @@ declare <16 x i8> @llvm.mips.binsl.b(<16 x i8>, <16 x i8>, <16 x i8>) nounwind define void @llvm_mips_binsl_h_test() nounwind { entry: - %0 = load <8 x i16>* @llvm_mips_binsl_h_ARG1 - %1 = load <8 x i16>* @llvm_mips_binsl_h_ARG2 - %2 = load <8 x i16>* @llvm_mips_binsl_h_ARG3 + %0 = load <8 x i16>, <8 x i16>* @llvm_mips_binsl_h_ARG1 + %1 = load <8 x i16>, <8 x i16>* @llvm_mips_binsl_h_ARG2 + %2 = load <8 x i16>, <8 x i16>* @llvm_mips_binsl_h_ARG3 %3 = tail call <8 x i16> @llvm.mips.binsl.h(<8 x i16> %0, <8 x i16> %1, <8 x i16> %2) store <8 x i16> %3, <8 x i16>* @llvm_mips_binsl_h_RES ret void @@ -155,9 +155,9 @@ declare <8 x i16> @llvm.mips.binsl.h(<8 x i16>, <8 x i16>, <8 x i16>) nounwind define void @llvm_mips_binsl_w_test() nounwind { entry: - %0 = load <4 x i32>* @llvm_mips_binsl_w_ARG1 - %1 = load <4 x i32>* @llvm_mips_binsl_w_ARG2 - %2 = load <4 x i32>* @llvm_mips_binsl_w_ARG3 + %0 = load <4 x i32>, <4 x i32>* @llvm_mips_binsl_w_ARG1 + %1 = load <4 x i32>, <4 x i32>* @llvm_mips_binsl_w_ARG2 + %2 = load <4 x i32>, <4 x i32>* @llvm_mips_binsl_w_ARG3 %3 = tail call <4 x i32> @llvm.mips.binsl.w(<4 x i32> %0, <4 x i32> %1, <4 x i32> %2) store <4 x i32> %3, <4 x i32>* @llvm_mips_binsl_w_RES ret void @@ -183,9 +183,9 @@ declare <4 x i32> @llvm.mips.binsl.w(<4 x i32>, <4 x i32>, <4 x i32>) nounwind define void @llvm_mips_binsl_d_test() nounwind { entry: - %0 = load <2 x i64>* @llvm_mips_binsl_d_ARG1 - %1 = load <2 x i64>* @llvm_mips_binsl_d_ARG2 - %2 = load <2 x i64>* @llvm_mips_binsl_d_ARG3 + %0 = load <2 x i64>, <2 x i64>* @llvm_mips_binsl_d_ARG1 + %1 = load <2 x i64>, <2 x i64>* @llvm_mips_binsl_d_ARG2 + %2 = load <2 x i64>, <2 x i64>* @llvm_mips_binsl_d_ARG3 %3 = tail call <2 x i64> @llvm.mips.binsl.d(<2 x i64> %0, <2 x i64> %1, <2 x i64> %2) store <2 x i64> %3, <2 x i64>* @llvm_mips_binsl_d_RES ret void @@ -211,9 +211,9 @@ declare <2 x i64> @llvm.mips.binsl.d(<2 x i64>, <2 x i64>, <2 x i64>) nounwind define void @llvm_mips_binsr_b_test() nounwind { entry: - %0 = load <16 x i8>* @llvm_mips_binsr_b_ARG1 - %1 = load <16 x i8>* @llvm_mips_binsr_b_ARG2 - %2 = load <16 x i8>* @llvm_mips_binsr_b_ARG3 + %0 = load <16 x i8>, <16 x i8>* @llvm_mips_binsr_b_ARG1 + %1 = load <16 x i8>, <16 x i8>* @llvm_mips_binsr_b_ARG2 + %2 = load <16 x i8>, <16 x i8>* @llvm_mips_binsr_b_ARG3 %3 = tail call <16 x i8> @llvm.mips.binsr.b(<16 x i8> %0, <16 x i8> %1, <16 x i8> %2) store <16 x i8> %3, <16 x i8>* @llvm_mips_binsr_b_RES ret void @@ -239,9 +239,9 @@ declare <16 x i8> @llvm.mips.binsr.b(<16 x i8>, <16 x i8>, <16 x i8>) nounwind define void @llvm_mips_binsr_h_test() nounwind { entry: - %0 = load <8 x i16>* @llvm_mips_binsr_h_ARG1 - %1 = load <8 x i16>* @llvm_mips_binsr_h_ARG2 - %2 = load <8 x i16>* @llvm_mips_binsr_h_ARG3 + %0 = load <8 x i16>, <8 x i16>* @llvm_mips_binsr_h_ARG1 + %1 = load <8 x i16>, <8 x i16>* @llvm_mips_binsr_h_ARG2 + %2 = load <8 x i16>, <8 x i16>* @llvm_mips_binsr_h_ARG3 %3 = tail call <8 x i16> @llvm.mips.binsr.h(<8 x i16> %0, <8 x i16> %1, <8 x i16> %2) store <8 x i16> %3, <8 x i16>* @llvm_mips_binsr_h_RES ret void @@ -267,9 +267,9 @@ declare <8 x i16> @llvm.mips.binsr.h(<8 x i16>, <8 x i16>, <8 x i16>) nounwind define void @llvm_mips_binsr_w_test() nounwind { entry: - %0 = load <4 x i32>* @llvm_mips_binsr_w_ARG1 - %1 = load <4 x i32>* @llvm_mips_binsr_w_ARG2 - %2 = load <4 x i32>* @llvm_mips_binsr_w_ARG3 + %0 = load <4 x i32>, <4 x i32>* @llvm_mips_binsr_w_ARG1 + %1 = load <4 x i32>, <4 x i32>* @llvm_mips_binsr_w_ARG2 + %2 = load <4 x i32>, <4 x i32>* @llvm_mips_binsr_w_ARG3 %3 = tail call <4 x i32> @llvm.mips.binsr.w(<4 x i32> %0, <4 x i32> %1, <4 x i32> %2) store <4 x i32> %3, <4 x i32>* @llvm_mips_binsr_w_RES ret void @@ -295,9 +295,9 @@ declare <4 x i32> @llvm.mips.binsr.w(<4 x i32>, <4 x i32>, <4 x i32>) nounwind define void @llvm_mips_binsr_d_test() nounwind { entry: - %0 = load <2 x i64>* @llvm_mips_binsr_d_ARG1 - %1 = load <2 x i64>* @llvm_mips_binsr_d_ARG2 - %2 = load <2 x i64>* @llvm_mips_binsr_d_ARG3 + %0 = load <2 x i64>, <2 x i64>* @llvm_mips_binsr_d_ARG1 + %1 = load <2 x i64>, <2 x i64>* @llvm_mips_binsr_d_ARG2 + %2 = load <2 x i64>, <2 x i64>* @llvm_mips_binsr_d_ARG3 %3 = tail call <2 x i64> @llvm.mips.binsr.d(<2 x i64> %0, <2 x i64> %1, <2 x i64> %2) store <2 x i64> %3, <2 x i64>* @llvm_mips_binsr_d_RES ret void @@ -322,8 +322,8 @@ declare <2 x i64> @llvm.mips.binsr.d(<2 x i64>, <2 x i64>, <2 x i64>) nounwind define void @llvm_mips_bneg_b_test() nounwind { entry: - %0 = load <16 x i8>* @llvm_mips_bneg_b_ARG1 - %1 = load <16 x i8>* @llvm_mips_bneg_b_ARG2 + %0 = load <16 x i8>, <16 x i8>* @llvm_mips_bneg_b_ARG1 + %1 = load <16 x i8>, <16 x i8>* @llvm_mips_bneg_b_ARG2 %2 = tail call <16 x i8> @llvm.mips.bneg.b(<16 x i8> %0, <16 x i8> %1) store <16 x i8> %2, <16 x i8>* @llvm_mips_bneg_b_RES ret void @@ -344,8 +344,8 @@ declare <16 x i8> @llvm.mips.bneg.b(<16 x i8>, <16 x i8>) nounwind define void @llvm_mips_bneg_h_test() nounwind { entry: - %0 = load <8 x i16>* @llvm_mips_bneg_h_ARG1 - %1 = load <8 x i16>* @llvm_mips_bneg_h_ARG2 + %0 = load <8 x i16>, <8 x i16>* @llvm_mips_bneg_h_ARG1 + %1 = load <8 x i16>, <8 x i16>* @llvm_mips_bneg_h_ARG2 %2 = tail call <8 x i16> @llvm.mips.bneg.h(<8 x i16> %0, <8 x i16> %1) store <8 x i16> %2, <8 x i16>* @llvm_mips_bneg_h_RES ret void @@ -366,8 +366,8 @@ declare <8 x i16> @llvm.mips.bneg.h(<8 x i16>, <8 x i16>) nounwind define void @llvm_mips_bneg_w_test() nounwind { entry: - %0 = load <4 x i32>* @llvm_mips_bneg_w_ARG1 - %1 = load <4 x i32>* @llvm_mips_bneg_w_ARG2 + %0 = load <4 x i32>, <4 x i32>* @llvm_mips_bneg_w_ARG1 + %1 = load <4 x i32>, <4 x i32>* @llvm_mips_bneg_w_ARG2 %2 = tail call <4 x i32> @llvm.mips.bneg.w(<4 x i32> %0, <4 x i32> %1) store <4 x i32> %2, <4 x i32>* @llvm_mips_bneg_w_RES ret void @@ -388,8 +388,8 @@ declare <4 x i32> @llvm.mips.bneg.w(<4 x i32>, <4 x i32>) nounwind define void @llvm_mips_bneg_d_test() nounwind { entry: - %0 = load <2 x i64>* @llvm_mips_bneg_d_ARG1 - %1 = load <2 x i64>* @llvm_mips_bneg_d_ARG2 + %0 = load <2 x i64>, <2 x i64>* @llvm_mips_bneg_d_ARG1 + %1 = load <2 x i64>, <2 x i64>* @llvm_mips_bneg_d_ARG2 %2 = tail call <2 x i64> @llvm.mips.bneg.d(<2 x i64> %0, <2 x i64> %1) store <2 x i64> %2, <2 x i64>* @llvm_mips_bneg_d_RES ret void @@ -410,8 +410,8 @@ declare <2 x i64> @llvm.mips.bneg.d(<2 x i64>, <2 x i64>) nounwind define void @llvm_mips_bset_b_test() nounwind { entry: - %0 = load <16 x i8>* @llvm_mips_bset_b_ARG1 - %1 = load <16 x i8>* @llvm_mips_bset_b_ARG2 + %0 = load <16 x i8>, <16 x i8>* @llvm_mips_bset_b_ARG1 + %1 = load <16 x i8>, <16 x i8>* @llvm_mips_bset_b_ARG2 %2 = tail call <16 x i8> @llvm.mips.bset.b(<16 x i8> %0, <16 x i8> %1) store <16 x i8> %2, <16 x i8>* @llvm_mips_bset_b_RES ret void @@ -432,8 +432,8 @@ declare <16 x i8> @llvm.mips.bset.b(<16 x i8>, <16 x i8>) nounwind define void @llvm_mips_bset_h_test() nounwind { entry: - %0 = load <8 x i16>* @llvm_mips_bset_h_ARG1 - %1 = load <8 x i16>* @llvm_mips_bset_h_ARG2 + %0 = load <8 x i16>, <8 x i16>* @llvm_mips_bset_h_ARG1 + %1 = load <8 x i16>, <8 x i16>* @llvm_mips_bset_h_ARG2 %2 = tail call <8 x i16> @llvm.mips.bset.h(<8 x i16> %0, <8 x i16> %1) store <8 x i16> %2, <8 x i16>* @llvm_mips_bset_h_RES ret void @@ -454,8 +454,8 @@ declare <8 x i16> @llvm.mips.bset.h(<8 x i16>, <8 x i16>) nounwind define void @llvm_mips_bset_w_test() nounwind { entry: - %0 = load <4 x i32>* @llvm_mips_bset_w_ARG1 - %1 = load <4 x i32>* @llvm_mips_bset_w_ARG2 + %0 = load <4 x i32>, <4 x i32>* @llvm_mips_bset_w_ARG1 + %1 = load <4 x i32>, <4 x i32>* @llvm_mips_bset_w_ARG2 %2 = tail call <4 x i32> @llvm.mips.bset.w(<4 x i32> %0, <4 x i32> %1) store <4 x i32> %2, <4 x i32>* @llvm_mips_bset_w_RES ret void @@ -476,8 +476,8 @@ declare <4 x i32> @llvm.mips.bset.w(<4 x i32>, <4 x i32>) nounwind define void @llvm_mips_bset_d_test() nounwind { entry: - %0 = load <2 x i64>* @llvm_mips_bset_d_ARG1 - %1 = load <2 x i64>* @llvm_mips_bset_d_ARG2 + %0 = load <2 x i64>, <2 x i64>* @llvm_mips_bset_d_ARG1 + %1 = load <2 x i64>, <2 x i64>* @llvm_mips_bset_d_ARG2 %2 = tail call <2 x i64> @llvm.mips.bset.d(<2 x i64> %0, <2 x i64> %1) store <2 x i64> %2, <2 x i64>* @llvm_mips_bset_d_RES ret void diff --git a/test/CodeGen/Mips/msa/3r-c.ll b/test/CodeGen/Mips/msa/3r-c.ll index 6ec92c284fec3..a3913e0a27fd4 100644 --- a/test/CodeGen/Mips/msa/3r-c.ll +++ b/test/CodeGen/Mips/msa/3r-c.ll @@ -10,8 +10,8 @@ define void @llvm_mips_ceq_b_test() nounwind { entry: - %0 = load <16 x i8>* @llvm_mips_ceq_b_ARG1 - %1 = load <16 x i8>* @llvm_mips_ceq_b_ARG2 + %0 = load <16 x i8>, <16 x i8>* @llvm_mips_ceq_b_ARG1 + %1 = load <16 x i8>, <16 x i8>* @llvm_mips_ceq_b_ARG2 %2 = tail call <16 x i8> @llvm.mips.ceq.b(<16 x i8> %0, <16 x i8> %1) store <16 x i8> %2, <16 x i8>* @llvm_mips_ceq_b_RES ret void @@ -32,8 +32,8 @@ declare <16 x i8> @llvm.mips.ceq.b(<16 x i8>, <16 x i8>) nounwind define void @llvm_mips_ceq_h_test() nounwind { entry: - %0 = load <8 x i16>* @llvm_mips_ceq_h_ARG1 - %1 = load <8 x i16>* @llvm_mips_ceq_h_ARG2 + %0 = load <8 x i16>, <8 x i16>* @llvm_mips_ceq_h_ARG1 + %1 = load <8 x i16>, <8 x i16>* @llvm_mips_ceq_h_ARG2 %2 = tail call <8 x i16> @llvm.mips.ceq.h(<8 x i16> %0, <8 x i16> %1) store <8 x i16> %2, <8 x i16>* @llvm_mips_ceq_h_RES ret void @@ -54,8 +54,8 @@ declare <8 x i16> @llvm.mips.ceq.h(<8 x i16>, <8 x i16>) nounwind define void @llvm_mips_ceq_w_test() nounwind { entry: - %0 = load <4 x i32>* @llvm_mips_ceq_w_ARG1 - %1 = load <4 x i32>* @llvm_mips_ceq_w_ARG2 + %0 = load <4 x i32>, <4 x i32>* @llvm_mips_ceq_w_ARG1 + %1 = load <4 x i32>, <4 x i32>* @llvm_mips_ceq_w_ARG2 %2 = tail call <4 x i32> @llvm.mips.ceq.w(<4 x i32> %0, <4 x i32> %1) store <4 x i32> %2, <4 x i32>* @llvm_mips_ceq_w_RES ret void @@ -76,8 +76,8 @@ declare <4 x i32> @llvm.mips.ceq.w(<4 x i32>, <4 x i32>) nounwind define void @llvm_mips_ceq_d_test() nounwind { entry: - %0 = load <2 x i64>* @llvm_mips_ceq_d_ARG1 - %1 = load <2 x i64>* @llvm_mips_ceq_d_ARG2 + %0 = load <2 x i64>, <2 x i64>* @llvm_mips_ceq_d_ARG1 + %1 = load <2 x i64>, <2 x i64>* @llvm_mips_ceq_d_ARG2 %2 = tail call <2 x i64> @llvm.mips.ceq.d(<2 x i64> %0, <2 x i64> %1) store <2 x i64> %2, <2 x i64>* @llvm_mips_ceq_d_RES ret void @@ -98,8 +98,8 @@ declare <2 x i64> @llvm.mips.ceq.d(<2 x i64>, <2 x i64>) nounwind define void @llvm_mips_cle_s_b_test() nounwind { entry: - %0 = load <16 x i8>* @llvm_mips_cle_s_b_ARG1 - %1 = load <16 x i8>* @llvm_mips_cle_s_b_ARG2 + %0 = load <16 x i8>, <16 x i8>* @llvm_mips_cle_s_b_ARG1 + %1 = load <16 x i8>, <16 x i8>* @llvm_mips_cle_s_b_ARG2 %2 = tail call <16 x i8> @llvm.mips.cle.s.b(<16 x i8> %0, <16 x i8> %1) store <16 x i8> %2, <16 x i8>* @llvm_mips_cle_s_b_RES ret void @@ -120,8 +120,8 @@ declare <16 x i8> @llvm.mips.cle.s.b(<16 x i8>, <16 x i8>) nounwind define void @llvm_mips_cle_s_h_test() nounwind { entry: - %0 = load <8 x i16>* @llvm_mips_cle_s_h_ARG1 - %1 = load <8 x i16>* @llvm_mips_cle_s_h_ARG2 + %0 = load <8 x i16>, <8 x i16>* @llvm_mips_cle_s_h_ARG1 + %1 = load <8 x i16>, <8 x i16>* @llvm_mips_cle_s_h_ARG2 %2 = tail call <8 x i16> @llvm.mips.cle.s.h(<8 x i16> %0, <8 x i16> %1) store <8 x i16> %2, <8 x i16>* @llvm_mips_cle_s_h_RES ret void @@ -142,8 +142,8 @@ declare <8 x i16> @llvm.mips.cle.s.h(<8 x i16>, <8 x i16>) nounwind define void @llvm_mips_cle_s_w_test() nounwind { entry: - %0 = load <4 x i32>* @llvm_mips_cle_s_w_ARG1 - %1 = load <4 x i32>* @llvm_mips_cle_s_w_ARG2 + %0 = load <4 x i32>, <4 x i32>* @llvm_mips_cle_s_w_ARG1 + %1 = load <4 x i32>, <4 x i32>* @llvm_mips_cle_s_w_ARG2 %2 = tail call <4 x i32> @llvm.mips.cle.s.w(<4 x i32> %0, <4 x i32> %1) store <4 x i32> %2, <4 x i32>* @llvm_mips_cle_s_w_RES ret void @@ -164,8 +164,8 @@ declare <4 x i32> @llvm.mips.cle.s.w(<4 x i32>, <4 x i32>) nounwind define void @llvm_mips_cle_s_d_test() nounwind { entry: - %0 = load <2 x i64>* @llvm_mips_cle_s_d_ARG1 - %1 = load <2 x i64>* @llvm_mips_cle_s_d_ARG2 + %0 = load <2 x i64>, <2 x i64>* @llvm_mips_cle_s_d_ARG1 + %1 = load <2 x i64>, <2 x i64>* @llvm_mips_cle_s_d_ARG2 %2 = tail call <2 x i64> @llvm.mips.cle.s.d(<2 x i64> %0, <2 x i64> %1) store <2 x i64> %2, <2 x i64>* @llvm_mips_cle_s_d_RES ret void @@ -186,8 +186,8 @@ declare <2 x i64> @llvm.mips.cle.s.d(<2 x i64>, <2 x i64>) nounwind define void @llvm_mips_cle_u_b_test() nounwind { entry: - %0 = load <16 x i8>* @llvm_mips_cle_u_b_ARG1 - %1 = load <16 x i8>* @llvm_mips_cle_u_b_ARG2 + %0 = load <16 x i8>, <16 x i8>* @llvm_mips_cle_u_b_ARG1 + %1 = load <16 x i8>, <16 x i8>* @llvm_mips_cle_u_b_ARG2 %2 = tail call <16 x i8> @llvm.mips.cle.u.b(<16 x i8> %0, <16 x i8> %1) store <16 x i8> %2, <16 x i8>* @llvm_mips_cle_u_b_RES ret void @@ -208,8 +208,8 @@ declare <16 x i8> @llvm.mips.cle.u.b(<16 x i8>, <16 x i8>) nounwind define void @llvm_mips_cle_u_h_test() nounwind { entry: - %0 = load <8 x i16>* @llvm_mips_cle_u_h_ARG1 - %1 = load <8 x i16>* @llvm_mips_cle_u_h_ARG2 + %0 = load <8 x i16>, <8 x i16>* @llvm_mips_cle_u_h_ARG1 + %1 = load <8 x i16>, <8 x i16>* @llvm_mips_cle_u_h_ARG2 %2 = tail call <8 x i16> @llvm.mips.cle.u.h(<8 x i16> %0, <8 x i16> %1) store <8 x i16> %2, <8 x i16>* @llvm_mips_cle_u_h_RES ret void @@ -230,8 +230,8 @@ declare <8 x i16> @llvm.mips.cle.u.h(<8 x i16>, <8 x i16>) nounwind define void @llvm_mips_cle_u_w_test() nounwind { entry: - %0 = load <4 x i32>* @llvm_mips_cle_u_w_ARG1 - %1 = load <4 x i32>* @llvm_mips_cle_u_w_ARG2 + %0 = load <4 x i32>, <4 x i32>* @llvm_mips_cle_u_w_ARG1 + %1 = load <4 x i32>, <4 x i32>* @llvm_mips_cle_u_w_ARG2 %2 = tail call <4 x i32> @llvm.mips.cle.u.w(<4 x i32> %0, <4 x i32> %1) store <4 x i32> %2, <4 x i32>* @llvm_mips_cle_u_w_RES ret void @@ -252,8 +252,8 @@ declare <4 x i32> @llvm.mips.cle.u.w(<4 x i32>, <4 x i32>) nounwind define void @llvm_mips_cle_u_d_test() nounwind { entry: - %0 = load <2 x i64>* @llvm_mips_cle_u_d_ARG1 - %1 = load <2 x i64>* @llvm_mips_cle_u_d_ARG2 + %0 = load <2 x i64>, <2 x i64>* @llvm_mips_cle_u_d_ARG1 + %1 = load <2 x i64>, <2 x i64>* @llvm_mips_cle_u_d_ARG2 %2 = tail call <2 x i64> @llvm.mips.cle.u.d(<2 x i64> %0, <2 x i64> %1) store <2 x i64> %2, <2 x i64>* @llvm_mips_cle_u_d_RES ret void @@ -274,8 +274,8 @@ declare <2 x i64> @llvm.mips.cle.u.d(<2 x i64>, <2 x i64>) nounwind define void @llvm_mips_clt_s_b_test() nounwind { entry: - %0 = load <16 x i8>* @llvm_mips_clt_s_b_ARG1 - %1 = load <16 x i8>* @llvm_mips_clt_s_b_ARG2 + %0 = load <16 x i8>, <16 x i8>* @llvm_mips_clt_s_b_ARG1 + %1 = load <16 x i8>, <16 x i8>* @llvm_mips_clt_s_b_ARG2 %2 = tail call <16 x i8> @llvm.mips.clt.s.b(<16 x i8> %0, <16 x i8> %1) store <16 x i8> %2, <16 x i8>* @llvm_mips_clt_s_b_RES ret void @@ -296,8 +296,8 @@ declare <16 x i8> @llvm.mips.clt.s.b(<16 x i8>, <16 x i8>) nounwind define void @llvm_mips_clt_s_h_test() nounwind { entry: - %0 = load <8 x i16>* @llvm_mips_clt_s_h_ARG1 - %1 = load <8 x i16>* @llvm_mips_clt_s_h_ARG2 + %0 = load <8 x i16>, <8 x i16>* @llvm_mips_clt_s_h_ARG1 + %1 = load <8 x i16>, <8 x i16>* @llvm_mips_clt_s_h_ARG2 %2 = tail call <8 x i16> @llvm.mips.clt.s.h(<8 x i16> %0, <8 x i16> %1) store <8 x i16> %2, <8 x i16>* @llvm_mips_clt_s_h_RES ret void @@ -318,8 +318,8 @@ declare <8 x i16> @llvm.mips.clt.s.h(<8 x i16>, <8 x i16>) nounwind define void @llvm_mips_clt_s_w_test() nounwind { entry: - %0 = load <4 x i32>* @llvm_mips_clt_s_w_ARG1 - %1 = load <4 x i32>* @llvm_mips_clt_s_w_ARG2 + %0 = load <4 x i32>, <4 x i32>* @llvm_mips_clt_s_w_ARG1 + %1 = load <4 x i32>, <4 x i32>* @llvm_mips_clt_s_w_ARG2 %2 = tail call <4 x i32> @llvm.mips.clt.s.w(<4 x i32> %0, <4 x i32> %1) store <4 x i32> %2, <4 x i32>* @llvm_mips_clt_s_w_RES ret void @@ -340,8 +340,8 @@ declare <4 x i32> @llvm.mips.clt.s.w(<4 x i32>, <4 x i32>) nounwind define void @llvm_mips_clt_s_d_test() nounwind { entry: - %0 = load <2 x i64>* @llvm_mips_clt_s_d_ARG1 - %1 = load <2 x i64>* @llvm_mips_clt_s_d_ARG2 + %0 = load <2 x i64>, <2 x i64>* @llvm_mips_clt_s_d_ARG1 + %1 = load <2 x i64>, <2 x i64>* @llvm_mips_clt_s_d_ARG2 %2 = tail call <2 x i64> @llvm.mips.clt.s.d(<2 x i64> %0, <2 x i64> %1) store <2 x i64> %2, <2 x i64>* @llvm_mips_clt_s_d_RES ret void @@ -362,8 +362,8 @@ declare <2 x i64> @llvm.mips.clt.s.d(<2 x i64>, <2 x i64>) nounwind define void @llvm_mips_clt_u_b_test() nounwind { entry: - %0 = load <16 x i8>* @llvm_mips_clt_u_b_ARG1 - %1 = load <16 x i8>* @llvm_mips_clt_u_b_ARG2 + %0 = load <16 x i8>, <16 x i8>* @llvm_mips_clt_u_b_ARG1 + %1 = load <16 x i8>, <16 x i8>* @llvm_mips_clt_u_b_ARG2 %2 = tail call <16 x i8> @llvm.mips.clt.u.b(<16 x i8> %0, <16 x i8> %1) store <16 x i8> %2, <16 x i8>* @llvm_mips_clt_u_b_RES ret void @@ -384,8 +384,8 @@ declare <16 x i8> @llvm.mips.clt.u.b(<16 x i8>, <16 x i8>) nounwind define void @llvm_mips_clt_u_h_test() nounwind { entry: - %0 = load <8 x i16>* @llvm_mips_clt_u_h_ARG1 - %1 = load <8 x i16>* @llvm_mips_clt_u_h_ARG2 + %0 = load <8 x i16>, <8 x i16>* @llvm_mips_clt_u_h_ARG1 + %1 = load <8 x i16>, <8 x i16>* @llvm_mips_clt_u_h_ARG2 %2 = tail call <8 x i16> @llvm.mips.clt.u.h(<8 x i16> %0, <8 x i16> %1) store <8 x i16> %2, <8 x i16>* @llvm_mips_clt_u_h_RES ret void @@ -406,8 +406,8 @@ declare <8 x i16> @llvm.mips.clt.u.h(<8 x i16>, <8 x i16>) nounwind define void @llvm_mips_clt_u_w_test() nounwind { entry: - %0 = load <4 x i32>* @llvm_mips_clt_u_w_ARG1 - %1 = load <4 x i32>* @llvm_mips_clt_u_w_ARG2 + %0 = load <4 x i32>, <4 x i32>* @llvm_mips_clt_u_w_ARG1 + %1 = load <4 x i32>, <4 x i32>* @llvm_mips_clt_u_w_ARG2 %2 = tail call <4 x i32> @llvm.mips.clt.u.w(<4 x i32> %0, <4 x i32> %1) store <4 x i32> %2, <4 x i32>* @llvm_mips_clt_u_w_RES ret void @@ -428,8 +428,8 @@ declare <4 x i32> @llvm.mips.clt.u.w(<4 x i32>, <4 x i32>) nounwind define void @llvm_mips_clt_u_d_test() nounwind { entry: - %0 = load <2 x i64>* @llvm_mips_clt_u_d_ARG1 - %1 = load <2 x i64>* @llvm_mips_clt_u_d_ARG2 + %0 = load <2 x i64>, <2 x i64>* @llvm_mips_clt_u_d_ARG1 + %1 = load <2 x i64>, <2 x i64>* @llvm_mips_clt_u_d_ARG2 %2 = tail call <2 x i64> @llvm.mips.clt.u.d(<2 x i64> %0, <2 x i64> %1) store <2 x i64> %2, <2 x i64>* @llvm_mips_clt_u_d_RES ret void diff --git a/test/CodeGen/Mips/msa/3r-d.ll b/test/CodeGen/Mips/msa/3r-d.ll index 0099554a8eeab..4fc32b76a7b2b 100644 --- a/test/CodeGen/Mips/msa/3r-d.ll +++ b/test/CodeGen/Mips/msa/3r-d.ll @@ -10,8 +10,8 @@ define void @llvm_mips_div_s_b_test() nounwind { entry: - %0 = load <16 x i8>* @llvm_mips_div_s_b_ARG1 - %1 = load <16 x i8>* @llvm_mips_div_s_b_ARG2 + %0 = load <16 x i8>, <16 x i8>* @llvm_mips_div_s_b_ARG1 + %1 = load <16 x i8>, <16 x i8>* @llvm_mips_div_s_b_ARG2 %2 = tail call <16 x i8> @llvm.mips.div.s.b(<16 x i8> %0, <16 x i8> %1) store <16 x i8> %2, <16 x i8>* @llvm_mips_div_s_b_RES ret void @@ -32,8 +32,8 @@ declare <16 x i8> @llvm.mips.div.s.b(<16 x i8>, <16 x i8>) nounwind define void @llvm_mips_div_s_h_test() nounwind { entry: - %0 = load <8 x i16>* @llvm_mips_div_s_h_ARG1 - %1 = load <8 x i16>* @llvm_mips_div_s_h_ARG2 + %0 = load <8 x i16>, <8 x i16>* @llvm_mips_div_s_h_ARG1 + %1 = load <8 x i16>, <8 x i16>* @llvm_mips_div_s_h_ARG2 %2 = tail call <8 x i16> @llvm.mips.div.s.h(<8 x i16> %0, <8 x i16> %1) store <8 x i16> %2, <8 x i16>* @llvm_mips_div_s_h_RES ret void @@ -54,8 +54,8 @@ declare <8 x i16> @llvm.mips.div.s.h(<8 x i16>, <8 x i16>) nounwind define void @llvm_mips_div_s_w_test() nounwind { entry: - %0 = load <4 x i32>* @llvm_mips_div_s_w_ARG1 - %1 = load <4 x i32>* @llvm_mips_div_s_w_ARG2 + %0 = load <4 x i32>, <4 x i32>* @llvm_mips_div_s_w_ARG1 + %1 = load <4 x i32>, <4 x i32>* @llvm_mips_div_s_w_ARG2 %2 = tail call <4 x i32> @llvm.mips.div.s.w(<4 x i32> %0, <4 x i32> %1) store <4 x i32> %2, <4 x i32>* @llvm_mips_div_s_w_RES ret void @@ -76,8 +76,8 @@ declare <4 x i32> @llvm.mips.div.s.w(<4 x i32>, <4 x i32>) nounwind define void @llvm_mips_div_s_d_test() nounwind { entry: - %0 = load <2 x i64>* @llvm_mips_div_s_d_ARG1 - %1 = load <2 x i64>* @llvm_mips_div_s_d_ARG2 + %0 = load <2 x i64>, <2 x i64>* @llvm_mips_div_s_d_ARG1 + %1 = load <2 x i64>, <2 x i64>* @llvm_mips_div_s_d_ARG2 %2 = tail call <2 x i64> @llvm.mips.div.s.d(<2 x i64> %0, <2 x i64> %1) store <2 x i64> %2, <2 x i64>* @llvm_mips_div_s_d_RES ret void @@ -95,8 +95,8 @@ declare <2 x i64> @llvm.mips.div.s.d(<2 x i64>, <2 x i64>) nounwind define void @div_s_b_test() nounwind { entry: - %0 = load <16 x i8>* @llvm_mips_div_s_b_ARG1 - %1 = load <16 x i8>* @llvm_mips_div_s_b_ARG2 + %0 = load <16 x i8>, <16 x i8>* @llvm_mips_div_s_b_ARG1 + %1 = load <16 x i8>, <16 x i8>* @llvm_mips_div_s_b_ARG2 %2 = sdiv <16 x i8> %0, %1 store <16 x i8> %2, <16 x i8>* @llvm_mips_div_s_b_RES ret void @@ -111,8 +111,8 @@ entry: define void @div_s_h_test() nounwind { entry: - %0 = load <8 x i16>* @llvm_mips_div_s_h_ARG1 - %1 = load <8 x i16>* @llvm_mips_div_s_h_ARG2 + %0 = load <8 x i16>, <8 x i16>* @llvm_mips_div_s_h_ARG1 + %1 = load <8 x i16>, <8 x i16>* @llvm_mips_div_s_h_ARG2 %2 = sdiv <8 x i16> %0, %1 store <8 x i16> %2, <8 x i16>* @llvm_mips_div_s_h_RES ret void @@ -127,8 +127,8 @@ entry: define void @div_s_w_test() nounwind { entry: - %0 = load <4 x i32>* @llvm_mips_div_s_w_ARG1 - %1 = load <4 x i32>* @llvm_mips_div_s_w_ARG2 + %0 = load <4 x i32>, <4 x i32>* @llvm_mips_div_s_w_ARG1 + %1 = load <4 x i32>, <4 x i32>* @llvm_mips_div_s_w_ARG2 %2 = sdiv <4 x i32> %0, %1 store <4 x i32> %2, <4 x i32>* @llvm_mips_div_s_w_RES ret void @@ -143,8 +143,8 @@ entry: define void @div_s_d_test() nounwind { entry: - %0 = load <2 x i64>* @llvm_mips_div_s_d_ARG1 - %1 = load <2 x i64>* @llvm_mips_div_s_d_ARG2 + %0 = load <2 x i64>, <2 x i64>* @llvm_mips_div_s_d_ARG1 + %1 = load <2 x i64>, <2 x i64>* @llvm_mips_div_s_d_ARG2 %2 = sdiv <2 x i64> %0, %1 store <2 x i64> %2, <2 x i64>* @llvm_mips_div_s_d_RES ret void @@ -163,8 +163,8 @@ entry: define void @llvm_mips_div_u_b_test() nounwind { entry: - %0 = load <16 x i8>* @llvm_mips_div_u_b_ARG1 - %1 = load <16 x i8>* @llvm_mips_div_u_b_ARG2 + %0 = load <16 x i8>, <16 x i8>* @llvm_mips_div_u_b_ARG1 + %1 = load <16 x i8>, <16 x i8>* @llvm_mips_div_u_b_ARG2 %2 = tail call <16 x i8> @llvm.mips.div.u.b(<16 x i8> %0, <16 x i8> %1) store <16 x i8> %2, <16 x i8>* @llvm_mips_div_u_b_RES ret void @@ -185,8 +185,8 @@ declare <16 x i8> @llvm.mips.div.u.b(<16 x i8>, <16 x i8>) nounwind define void @llvm_mips_div_u_h_test() nounwind { entry: - %0 = load <8 x i16>* @llvm_mips_div_u_h_ARG1 - %1 = load <8 x i16>* @llvm_mips_div_u_h_ARG2 + %0 = load <8 x i16>, <8 x i16>* @llvm_mips_div_u_h_ARG1 + %1 = load <8 x i16>, <8 x i16>* @llvm_mips_div_u_h_ARG2 %2 = tail call <8 x i16> @llvm.mips.div.u.h(<8 x i16> %0, <8 x i16> %1) store <8 x i16> %2, <8 x i16>* @llvm_mips_div_u_h_RES ret void @@ -207,8 +207,8 @@ declare <8 x i16> @llvm.mips.div.u.h(<8 x i16>, <8 x i16>) nounwind define void @llvm_mips_div_u_w_test() nounwind { entry: - %0 = load <4 x i32>* @llvm_mips_div_u_w_ARG1 - %1 = load <4 x i32>* @llvm_mips_div_u_w_ARG2 + %0 = load <4 x i32>, <4 x i32>* @llvm_mips_div_u_w_ARG1 + %1 = load <4 x i32>, <4 x i32>* @llvm_mips_div_u_w_ARG2 %2 = tail call <4 x i32> @llvm.mips.div.u.w(<4 x i32> %0, <4 x i32> %1) store <4 x i32> %2, <4 x i32>* @llvm_mips_div_u_w_RES ret void @@ -229,8 +229,8 @@ declare <4 x i32> @llvm.mips.div.u.w(<4 x i32>, <4 x i32>) nounwind define void @llvm_mips_div_u_d_test() nounwind { entry: - %0 = load <2 x i64>* @llvm_mips_div_u_d_ARG1 - %1 = load <2 x i64>* @llvm_mips_div_u_d_ARG2 + %0 = load <2 x i64>, <2 x i64>* @llvm_mips_div_u_d_ARG1 + %1 = load <2 x i64>, <2 x i64>* @llvm_mips_div_u_d_ARG2 %2 = tail call <2 x i64> @llvm.mips.div.u.d(<2 x i64> %0, <2 x i64> %1) store <2 x i64> %2, <2 x i64>* @llvm_mips_div_u_d_RES ret void @@ -248,8 +248,8 @@ declare <2 x i64> @llvm.mips.div.u.d(<2 x i64>, <2 x i64>) nounwind define void @div_u_b_test() nounwind { entry: - %0 = load <16 x i8>* @llvm_mips_div_u_b_ARG1 - %1 = load <16 x i8>* @llvm_mips_div_u_b_ARG2 + %0 = load <16 x i8>, <16 x i8>* @llvm_mips_div_u_b_ARG1 + %1 = load <16 x i8>, <16 x i8>* @llvm_mips_div_u_b_ARG2 %2 = udiv <16 x i8> %0, %1 store <16 x i8> %2, <16 x i8>* @llvm_mips_div_u_b_RES ret void @@ -264,8 +264,8 @@ entry: define void @div_u_h_test() nounwind { entry: - %0 = load <8 x i16>* @llvm_mips_div_u_h_ARG1 - %1 = load <8 x i16>* @llvm_mips_div_u_h_ARG2 + %0 = load <8 x i16>, <8 x i16>* @llvm_mips_div_u_h_ARG1 + %1 = load <8 x i16>, <8 x i16>* @llvm_mips_div_u_h_ARG2 %2 = udiv <8 x i16> %0, %1 store <8 x i16> %2, <8 x i16>* @llvm_mips_div_u_h_RES ret void @@ -280,8 +280,8 @@ entry: define void @div_u_w_test() nounwind { entry: - %0 = load <4 x i32>* @llvm_mips_div_u_w_ARG1 - %1 = load <4 x i32>* @llvm_mips_div_u_w_ARG2 + %0 = load <4 x i32>, <4 x i32>* @llvm_mips_div_u_w_ARG1 + %1 = load <4 x i32>, <4 x i32>* @llvm_mips_div_u_w_ARG2 %2 = udiv <4 x i32> %0, %1 store <4 x i32> %2, <4 x i32>* @llvm_mips_div_u_w_RES ret void @@ -296,8 +296,8 @@ entry: define void @div_u_d_test() nounwind { entry: - %0 = load <2 x i64>* @llvm_mips_div_u_d_ARG1 - %1 = load <2 x i64>* @llvm_mips_div_u_d_ARG2 + %0 = load <2 x i64>, <2 x i64>* @llvm_mips_div_u_d_ARG1 + %1 = load <2 x i64>, <2 x i64>* @llvm_mips_div_u_d_ARG2 %2 = udiv <2 x i64> %0, %1 store <2 x i64> %2, <2 x i64>* @llvm_mips_div_u_d_RES ret void @@ -326,8 +326,8 @@ entry: define void @llvm_mips_dotp_s_h_test() nounwind { entry: - %0 = load <16 x i8>* @llvm_mips_dotp_s_h_ARG1 - %1 = load <16 x i8>* @llvm_mips_dotp_s_h_ARG2 + %0 = load <16 x i8>, <16 x i8>* @llvm_mips_dotp_s_h_ARG1 + %1 = load <16 x i8>, <16 x i8>* @llvm_mips_dotp_s_h_ARG2 %2 = tail call <8 x i16> @llvm.mips.dotp.s.h(<16 x i8> %0, <16 x i8> %1) store <8 x i16> %2, <8 x i16>* @llvm_mips_dotp_s_h_RES ret void @@ -353,8 +353,8 @@ declare <8 x i16> @llvm.mips.dotp.s.h(<16 x i8>, <16 x i8>) nounwind define void @llvm_mips_dotp_s_w_test() nounwind { entry: - %0 = load <8 x i16>* @llvm_mips_dotp_s_w_ARG1 - %1 = load <8 x i16>* @llvm_mips_dotp_s_w_ARG2 + %0 = load <8 x i16>, <8 x i16>* @llvm_mips_dotp_s_w_ARG1 + %1 = load <8 x i16>, <8 x i16>* @llvm_mips_dotp_s_w_ARG2 %2 = tail call <4 x i32> @llvm.mips.dotp.s.w(<8 x i16> %0, <8 x i16> %1) store <4 x i32> %2, <4 x i32>* @llvm_mips_dotp_s_w_RES ret void @@ -377,8 +377,8 @@ declare <4 x i32> @llvm.mips.dotp.s.w(<8 x i16>, <8 x i16>) nounwind define void @llvm_mips_dotp_s_d_test() nounwind { entry: - %0 = load <4 x i32>* @llvm_mips_dotp_s_d_ARG1 - %1 = load <4 x i32>* @llvm_mips_dotp_s_d_ARG2 + %0 = load <4 x i32>, <4 x i32>* @llvm_mips_dotp_s_d_ARG1 + %1 = load <4 x i32>, <4 x i32>* @llvm_mips_dotp_s_d_ARG2 %2 = tail call <2 x i64> @llvm.mips.dotp.s.d(<4 x i32> %0, <4 x i32> %1) store <2 x i64> %2, <2 x i64>* @llvm_mips_dotp_s_d_RES ret void @@ -409,8 +409,8 @@ declare <2 x i64> @llvm.mips.dotp.s.d(<4 x i32>, <4 x i32>) nounwind define void @llvm_mips_dotp_u_h_test() nounwind { entry: - %0 = load <16 x i8>* @llvm_mips_dotp_u_h_ARG1 - %1 = load <16 x i8>* @llvm_mips_dotp_u_h_ARG2 + %0 = load <16 x i8>, <16 x i8>* @llvm_mips_dotp_u_h_ARG1 + %1 = load <16 x i8>, <16 x i8>* @llvm_mips_dotp_u_h_ARG2 %2 = tail call <8 x i16> @llvm.mips.dotp.u.h(<16 x i8> %0, <16 x i8> %1) store <8 x i16> %2, <8 x i16>* @llvm_mips_dotp_u_h_RES ret void @@ -436,8 +436,8 @@ declare <8 x i16> @llvm.mips.dotp.u.h(<16 x i8>, <16 x i8>) nounwind define void @llvm_mips_dotp_u_w_test() nounwind { entry: - %0 = load <8 x i16>* @llvm_mips_dotp_u_w_ARG1 - %1 = load <8 x i16>* @llvm_mips_dotp_u_w_ARG2 + %0 = load <8 x i16>, <8 x i16>* @llvm_mips_dotp_u_w_ARG1 + %1 = load <8 x i16>, <8 x i16>* @llvm_mips_dotp_u_w_ARG2 %2 = tail call <4 x i32> @llvm.mips.dotp.u.w(<8 x i16> %0, <8 x i16> %1) store <4 x i32> %2, <4 x i32>* @llvm_mips_dotp_u_w_RES ret void @@ -460,8 +460,8 @@ declare <4 x i32> @llvm.mips.dotp.u.w(<8 x i16>, <8 x i16>) nounwind define void @llvm_mips_dotp_u_d_test() nounwind { entry: - %0 = load <4 x i32>* @llvm_mips_dotp_u_d_ARG1 - %1 = load <4 x i32>* @llvm_mips_dotp_u_d_ARG2 + %0 = load <4 x i32>, <4 x i32>* @llvm_mips_dotp_u_d_ARG1 + %1 = load <4 x i32>, <4 x i32>* @llvm_mips_dotp_u_d_ARG2 %2 = tail call <2 x i64> @llvm.mips.dotp.u.d(<4 x i32> %0, <4 x i32> %1) store <2 x i64> %2, <2 x i64>* @llvm_mips_dotp_u_d_RES ret void diff --git a/test/CodeGen/Mips/msa/3r-i.ll b/test/CodeGen/Mips/msa/3r-i.ll index 2ef30471b026e..7147b756b15fc 100644 --- a/test/CodeGen/Mips/msa/3r-i.ll +++ b/test/CodeGen/Mips/msa/3r-i.ll @@ -10,8 +10,8 @@ define void @llvm_mips_ilvev_b_test() nounwind { entry: - %0 = load <16 x i8>* @llvm_mips_ilvev_b_ARG1 - %1 = load <16 x i8>* @llvm_mips_ilvev_b_ARG2 + %0 = load <16 x i8>, <16 x i8>* @llvm_mips_ilvev_b_ARG1 + %1 = load <16 x i8>, <16 x i8>* @llvm_mips_ilvev_b_ARG2 %2 = tail call <16 x i8> @llvm.mips.ilvev.b(<16 x i8> %0, <16 x i8> %1) store <16 x i8> %2, <16 x i8>* @llvm_mips_ilvev_b_RES ret void @@ -32,8 +32,8 @@ declare <16 x i8> @llvm.mips.ilvev.b(<16 x i8>, <16 x i8>) nounwind define void @llvm_mips_ilvev_h_test() nounwind { entry: - %0 = load <8 x i16>* @llvm_mips_ilvev_h_ARG1 - %1 = load <8 x i16>* @llvm_mips_ilvev_h_ARG2 + %0 = load <8 x i16>, <8 x i16>* @llvm_mips_ilvev_h_ARG1 + %1 = load <8 x i16>, <8 x i16>* @llvm_mips_ilvev_h_ARG2 %2 = tail call <8 x i16> @llvm.mips.ilvev.h(<8 x i16> %0, <8 x i16> %1) store <8 x i16> %2, <8 x i16>* @llvm_mips_ilvev_h_RES ret void @@ -54,8 +54,8 @@ declare <8 x i16> @llvm.mips.ilvev.h(<8 x i16>, <8 x i16>) nounwind define void @llvm_mips_ilvev_w_test() nounwind { entry: - %0 = load <4 x i32>* @llvm_mips_ilvev_w_ARG1 - %1 = load <4 x i32>* @llvm_mips_ilvev_w_ARG2 + %0 = load <4 x i32>, <4 x i32>* @llvm_mips_ilvev_w_ARG1 + %1 = load <4 x i32>, <4 x i32>* @llvm_mips_ilvev_w_ARG2 %2 = tail call <4 x i32> @llvm.mips.ilvev.w(<4 x i32> %0, <4 x i32> %1) store <4 x i32> %2, <4 x i32>* @llvm_mips_ilvev_w_RES ret void @@ -76,8 +76,8 @@ declare <4 x i32> @llvm.mips.ilvev.w(<4 x i32>, <4 x i32>) nounwind define void @llvm_mips_ilvev_d_test() nounwind { entry: - %0 = load <2 x i64>* @llvm_mips_ilvev_d_ARG1 - %1 = load <2 x i64>* @llvm_mips_ilvev_d_ARG2 + %0 = load <2 x i64>, <2 x i64>* @llvm_mips_ilvev_d_ARG1 + %1 = load <2 x i64>, <2 x i64>* @llvm_mips_ilvev_d_ARG2 %2 = tail call <2 x i64> @llvm.mips.ilvev.d(<2 x i64> %0, <2 x i64> %1) store <2 x i64> %2, <2 x i64>* @llvm_mips_ilvev_d_RES ret void @@ -98,8 +98,8 @@ declare <2 x i64> @llvm.mips.ilvev.d(<2 x i64>, <2 x i64>) nounwind define void @llvm_mips_ilvl_b_test() nounwind { entry: - %0 = load <16 x i8>* @llvm_mips_ilvl_b_ARG1 - %1 = load <16 x i8>* @llvm_mips_ilvl_b_ARG2 + %0 = load <16 x i8>, <16 x i8>* @llvm_mips_ilvl_b_ARG1 + %1 = load <16 x i8>, <16 x i8>* @llvm_mips_ilvl_b_ARG2 %2 = tail call <16 x i8> @llvm.mips.ilvl.b(<16 x i8> %0, <16 x i8> %1) store <16 x i8> %2, <16 x i8>* @llvm_mips_ilvl_b_RES ret void @@ -120,8 +120,8 @@ declare <16 x i8> @llvm.mips.ilvl.b(<16 x i8>, <16 x i8>) nounwind define void @llvm_mips_ilvl_h_test() nounwind { entry: - %0 = load <8 x i16>* @llvm_mips_ilvl_h_ARG1 - %1 = load <8 x i16>* @llvm_mips_ilvl_h_ARG2 + %0 = load <8 x i16>, <8 x i16>* @llvm_mips_ilvl_h_ARG1 + %1 = load <8 x i16>, <8 x i16>* @llvm_mips_ilvl_h_ARG2 %2 = tail call <8 x i16> @llvm.mips.ilvl.h(<8 x i16> %0, <8 x i16> %1) store <8 x i16> %2, <8 x i16>* @llvm_mips_ilvl_h_RES ret void @@ -142,8 +142,8 @@ declare <8 x i16> @llvm.mips.ilvl.h(<8 x i16>, <8 x i16>) nounwind define void @llvm_mips_ilvl_w_test() nounwind { entry: - %0 = load <4 x i32>* @llvm_mips_ilvl_w_ARG1 - %1 = load <4 x i32>* @llvm_mips_ilvl_w_ARG2 + %0 = load <4 x i32>, <4 x i32>* @llvm_mips_ilvl_w_ARG1 + %1 = load <4 x i32>, <4 x i32>* @llvm_mips_ilvl_w_ARG2 %2 = tail call <4 x i32> @llvm.mips.ilvl.w(<4 x i32> %0, <4 x i32> %1) store <4 x i32> %2, <4 x i32>* @llvm_mips_ilvl_w_RES ret void @@ -164,8 +164,8 @@ declare <4 x i32> @llvm.mips.ilvl.w(<4 x i32>, <4 x i32>) nounwind define void @llvm_mips_ilvl_d_test() nounwind { entry: - %0 = load <2 x i64>* @llvm_mips_ilvl_d_ARG1 - %1 = load <2 x i64>* @llvm_mips_ilvl_d_ARG2 + %0 = load <2 x i64>, <2 x i64>* @llvm_mips_ilvl_d_ARG1 + %1 = load <2 x i64>, <2 x i64>* @llvm_mips_ilvl_d_ARG2 %2 = tail call <2 x i64> @llvm.mips.ilvl.d(<2 x i64> %0, <2 x i64> %1) store <2 x i64> %2, <2 x i64>* @llvm_mips_ilvl_d_RES ret void @@ -186,8 +186,8 @@ declare <2 x i64> @llvm.mips.ilvl.d(<2 x i64>, <2 x i64>) nounwind define void @llvm_mips_ilvod_b_test() nounwind { entry: - %0 = load <16 x i8>* @llvm_mips_ilvod_b_ARG1 - %1 = load <16 x i8>* @llvm_mips_ilvod_b_ARG2 + %0 = load <16 x i8>, <16 x i8>* @llvm_mips_ilvod_b_ARG1 + %1 = load <16 x i8>, <16 x i8>* @llvm_mips_ilvod_b_ARG2 %2 = tail call <16 x i8> @llvm.mips.ilvod.b(<16 x i8> %0, <16 x i8> %1) store <16 x i8> %2, <16 x i8>* @llvm_mips_ilvod_b_RES ret void @@ -208,8 +208,8 @@ declare <16 x i8> @llvm.mips.ilvod.b(<16 x i8>, <16 x i8>) nounwind define void @llvm_mips_ilvod_h_test() nounwind { entry: - %0 = load <8 x i16>* @llvm_mips_ilvod_h_ARG1 - %1 = load <8 x i16>* @llvm_mips_ilvod_h_ARG2 + %0 = load <8 x i16>, <8 x i16>* @llvm_mips_ilvod_h_ARG1 + %1 = load <8 x i16>, <8 x i16>* @llvm_mips_ilvod_h_ARG2 %2 = tail call <8 x i16> @llvm.mips.ilvod.h(<8 x i16> %0, <8 x i16> %1) store <8 x i16> %2, <8 x i16>* @llvm_mips_ilvod_h_RES ret void @@ -230,8 +230,8 @@ declare <8 x i16> @llvm.mips.ilvod.h(<8 x i16>, <8 x i16>) nounwind define void @llvm_mips_ilvod_w_test() nounwind { entry: - %0 = load <4 x i32>* @llvm_mips_ilvod_w_ARG1 - %1 = load <4 x i32>* @llvm_mips_ilvod_w_ARG2 + %0 = load <4 x i32>, <4 x i32>* @llvm_mips_ilvod_w_ARG1 + %1 = load <4 x i32>, <4 x i32>* @llvm_mips_ilvod_w_ARG2 %2 = tail call <4 x i32> @llvm.mips.ilvod.w(<4 x i32> %0, <4 x i32> %1) store <4 x i32> %2, <4 x i32>* @llvm_mips_ilvod_w_RES ret void @@ -252,8 +252,8 @@ declare <4 x i32> @llvm.mips.ilvod.w(<4 x i32>, <4 x i32>) nounwind define void @llvm_mips_ilvod_d_test() nounwind { entry: - %0 = load <2 x i64>* @llvm_mips_ilvod_d_ARG1 - %1 = load <2 x i64>* @llvm_mips_ilvod_d_ARG2 + %0 = load <2 x i64>, <2 x i64>* @llvm_mips_ilvod_d_ARG1 + %1 = load <2 x i64>, <2 x i64>* @llvm_mips_ilvod_d_ARG2 %2 = tail call <2 x i64> @llvm.mips.ilvod.d(<2 x i64> %0, <2 x i64> %1) store <2 x i64> %2, <2 x i64>* @llvm_mips_ilvod_d_RES ret void @@ -274,8 +274,8 @@ declare <2 x i64> @llvm.mips.ilvod.d(<2 x i64>, <2 x i64>) nounwind define void @llvm_mips_ilvr_b_test() nounwind { entry: - %0 = load <16 x i8>* @llvm_mips_ilvr_b_ARG1 - %1 = load <16 x i8>* @llvm_mips_ilvr_b_ARG2 + %0 = load <16 x i8>, <16 x i8>* @llvm_mips_ilvr_b_ARG1 + %1 = load <16 x i8>, <16 x i8>* @llvm_mips_ilvr_b_ARG2 %2 = tail call <16 x i8> @llvm.mips.ilvr.b(<16 x i8> %0, <16 x i8> %1) store <16 x i8> %2, <16 x i8>* @llvm_mips_ilvr_b_RES ret void @@ -296,8 +296,8 @@ declare <16 x i8> @llvm.mips.ilvr.b(<16 x i8>, <16 x i8>) nounwind define void @llvm_mips_ilvr_h_test() nounwind { entry: - %0 = load <8 x i16>* @llvm_mips_ilvr_h_ARG1 - %1 = load <8 x i16>* @llvm_mips_ilvr_h_ARG2 + %0 = load <8 x i16>, <8 x i16>* @llvm_mips_ilvr_h_ARG1 + %1 = load <8 x i16>, <8 x i16>* @llvm_mips_ilvr_h_ARG2 %2 = tail call <8 x i16> @llvm.mips.ilvr.h(<8 x i16> %0, <8 x i16> %1) store <8 x i16> %2, <8 x i16>* @llvm_mips_ilvr_h_RES ret void @@ -318,8 +318,8 @@ declare <8 x i16> @llvm.mips.ilvr.h(<8 x i16>, <8 x i16>) nounwind define void @llvm_mips_ilvr_w_test() nounwind { entry: - %0 = load <4 x i32>* @llvm_mips_ilvr_w_ARG1 - %1 = load <4 x i32>* @llvm_mips_ilvr_w_ARG2 + %0 = load <4 x i32>, <4 x i32>* @llvm_mips_ilvr_w_ARG1 + %1 = load <4 x i32>, <4 x i32>* @llvm_mips_ilvr_w_ARG2 %2 = tail call <4 x i32> @llvm.mips.ilvr.w(<4 x i32> %0, <4 x i32> %1) store <4 x i32> %2, <4 x i32>* @llvm_mips_ilvr_w_RES ret void @@ -340,8 +340,8 @@ declare <4 x i32> @llvm.mips.ilvr.w(<4 x i32>, <4 x i32>) nounwind define void @llvm_mips_ilvr_d_test() nounwind { entry: - %0 = load <2 x i64>* @llvm_mips_ilvr_d_ARG1 - %1 = load <2 x i64>* @llvm_mips_ilvr_d_ARG2 + %0 = load <2 x i64>, <2 x i64>* @llvm_mips_ilvr_d_ARG1 + %1 = load <2 x i64>, <2 x i64>* @llvm_mips_ilvr_d_ARG2 %2 = tail call <2 x i64> @llvm.mips.ilvr.d(<2 x i64> %0, <2 x i64> %1) store <2 x i64> %2, <2 x i64>* @llvm_mips_ilvr_d_RES ret void diff --git a/test/CodeGen/Mips/msa/3r-m.ll b/test/CodeGen/Mips/msa/3r-m.ll index ddfd720a2f843..39b4f7db3a48b 100644 --- a/test/CodeGen/Mips/msa/3r-m.ll +++ b/test/CodeGen/Mips/msa/3r-m.ll @@ -10,8 +10,8 @@ define void @llvm_mips_max_a_b_test() nounwind { entry: - %0 = load <16 x i8>* @llvm_mips_max_a_b_ARG1 - %1 = load <16 x i8>* @llvm_mips_max_a_b_ARG2 + %0 = load <16 x i8>, <16 x i8>* @llvm_mips_max_a_b_ARG1 + %1 = load <16 x i8>, <16 x i8>* @llvm_mips_max_a_b_ARG2 %2 = tail call <16 x i8> @llvm.mips.max.a.b(<16 x i8> %0, <16 x i8> %1) store <16 x i8> %2, <16 x i8>* @llvm_mips_max_a_b_RES ret void @@ -32,8 +32,8 @@ declare <16 x i8> @llvm.mips.max.a.b(<16 x i8>, <16 x i8>) nounwind define void @llvm_mips_max_a_h_test() nounwind { entry: - %0 = load <8 x i16>* @llvm_mips_max_a_h_ARG1 - %1 = load <8 x i16>* @llvm_mips_max_a_h_ARG2 + %0 = load <8 x i16>, <8 x i16>* @llvm_mips_max_a_h_ARG1 + %1 = load <8 x i16>, <8 x i16>* @llvm_mips_max_a_h_ARG2 %2 = tail call <8 x i16> @llvm.mips.max.a.h(<8 x i16> %0, <8 x i16> %1) store <8 x i16> %2, <8 x i16>* @llvm_mips_max_a_h_RES ret void @@ -54,8 +54,8 @@ declare <8 x i16> @llvm.mips.max.a.h(<8 x i16>, <8 x i16>) nounwind define void @llvm_mips_max_a_w_test() nounwind { entry: - %0 = load <4 x i32>* @llvm_mips_max_a_w_ARG1 - %1 = load <4 x i32>* @llvm_mips_max_a_w_ARG2 + %0 = load <4 x i32>, <4 x i32>* @llvm_mips_max_a_w_ARG1 + %1 = load <4 x i32>, <4 x i32>* @llvm_mips_max_a_w_ARG2 %2 = tail call <4 x i32> @llvm.mips.max.a.w(<4 x i32> %0, <4 x i32> %1) store <4 x i32> %2, <4 x i32>* @llvm_mips_max_a_w_RES ret void @@ -76,8 +76,8 @@ declare <4 x i32> @llvm.mips.max.a.w(<4 x i32>, <4 x i32>) nounwind define void @llvm_mips_max_a_d_test() nounwind { entry: - %0 = load <2 x i64>* @llvm_mips_max_a_d_ARG1 - %1 = load <2 x i64>* @llvm_mips_max_a_d_ARG2 + %0 = load <2 x i64>, <2 x i64>* @llvm_mips_max_a_d_ARG1 + %1 = load <2 x i64>, <2 x i64>* @llvm_mips_max_a_d_ARG2 %2 = tail call <2 x i64> @llvm.mips.max.a.d(<2 x i64> %0, <2 x i64> %1) store <2 x i64> %2, <2 x i64>* @llvm_mips_max_a_d_RES ret void @@ -98,8 +98,8 @@ declare <2 x i64> @llvm.mips.max.a.d(<2 x i64>, <2 x i64>) nounwind define void @llvm_mips_max_s_b_test() nounwind { entry: - %0 = load <16 x i8>* @llvm_mips_max_s_b_ARG1 - %1 = load <16 x i8>* @llvm_mips_max_s_b_ARG2 + %0 = load <16 x i8>, <16 x i8>* @llvm_mips_max_s_b_ARG1 + %1 = load <16 x i8>, <16 x i8>* @llvm_mips_max_s_b_ARG2 %2 = tail call <16 x i8> @llvm.mips.max.s.b(<16 x i8> %0, <16 x i8> %1) store <16 x i8> %2, <16 x i8>* @llvm_mips_max_s_b_RES ret void @@ -120,8 +120,8 @@ declare <16 x i8> @llvm.mips.max.s.b(<16 x i8>, <16 x i8>) nounwind define void @llvm_mips_max_s_h_test() nounwind { entry: - %0 = load <8 x i16>* @llvm_mips_max_s_h_ARG1 - %1 = load <8 x i16>* @llvm_mips_max_s_h_ARG2 + %0 = load <8 x i16>, <8 x i16>* @llvm_mips_max_s_h_ARG1 + %1 = load <8 x i16>, <8 x i16>* @llvm_mips_max_s_h_ARG2 %2 = tail call <8 x i16> @llvm.mips.max.s.h(<8 x i16> %0, <8 x i16> %1) store <8 x i16> %2, <8 x i16>* @llvm_mips_max_s_h_RES ret void @@ -142,8 +142,8 @@ declare <8 x i16> @llvm.mips.max.s.h(<8 x i16>, <8 x i16>) nounwind define void @llvm_mips_max_s_w_test() nounwind { entry: - %0 = load <4 x i32>* @llvm_mips_max_s_w_ARG1 - %1 = load <4 x i32>* @llvm_mips_max_s_w_ARG2 + %0 = load <4 x i32>, <4 x i32>* @llvm_mips_max_s_w_ARG1 + %1 = load <4 x i32>, <4 x i32>* @llvm_mips_max_s_w_ARG2 %2 = tail call <4 x i32> @llvm.mips.max.s.w(<4 x i32> %0, <4 x i32> %1) store <4 x i32> %2, <4 x i32>* @llvm_mips_max_s_w_RES ret void @@ -164,8 +164,8 @@ declare <4 x i32> @llvm.mips.max.s.w(<4 x i32>, <4 x i32>) nounwind define void @llvm_mips_max_s_d_test() nounwind { entry: - %0 = load <2 x i64>* @llvm_mips_max_s_d_ARG1 - %1 = load <2 x i64>* @llvm_mips_max_s_d_ARG2 + %0 = load <2 x i64>, <2 x i64>* @llvm_mips_max_s_d_ARG1 + %1 = load <2 x i64>, <2 x i64>* @llvm_mips_max_s_d_ARG2 %2 = tail call <2 x i64> @llvm.mips.max.s.d(<2 x i64> %0, <2 x i64> %1) store <2 x i64> %2, <2 x i64>* @llvm_mips_max_s_d_RES ret void @@ -186,8 +186,8 @@ declare <2 x i64> @llvm.mips.max.s.d(<2 x i64>, <2 x i64>) nounwind define void @llvm_mips_max_u_b_test() nounwind { entry: - %0 = load <16 x i8>* @llvm_mips_max_u_b_ARG1 - %1 = load <16 x i8>* @llvm_mips_max_u_b_ARG2 + %0 = load <16 x i8>, <16 x i8>* @llvm_mips_max_u_b_ARG1 + %1 = load <16 x i8>, <16 x i8>* @llvm_mips_max_u_b_ARG2 %2 = tail call <16 x i8> @llvm.mips.max.u.b(<16 x i8> %0, <16 x i8> %1) store <16 x i8> %2, <16 x i8>* @llvm_mips_max_u_b_RES ret void @@ -208,8 +208,8 @@ declare <16 x i8> @llvm.mips.max.u.b(<16 x i8>, <16 x i8>) nounwind define void @llvm_mips_max_u_h_test() nounwind { entry: - %0 = load <8 x i16>* @llvm_mips_max_u_h_ARG1 - %1 = load <8 x i16>* @llvm_mips_max_u_h_ARG2 + %0 = load <8 x i16>, <8 x i16>* @llvm_mips_max_u_h_ARG1 + %1 = load <8 x i16>, <8 x i16>* @llvm_mips_max_u_h_ARG2 %2 = tail call <8 x i16> @llvm.mips.max.u.h(<8 x i16> %0, <8 x i16> %1) store <8 x i16> %2, <8 x i16>* @llvm_mips_max_u_h_RES ret void @@ -230,8 +230,8 @@ declare <8 x i16> @llvm.mips.max.u.h(<8 x i16>, <8 x i16>) nounwind define void @llvm_mips_max_u_w_test() nounwind { entry: - %0 = load <4 x i32>* @llvm_mips_max_u_w_ARG1 - %1 = load <4 x i32>* @llvm_mips_max_u_w_ARG2 + %0 = load <4 x i32>, <4 x i32>* @llvm_mips_max_u_w_ARG1 + %1 = load <4 x i32>, <4 x i32>* @llvm_mips_max_u_w_ARG2 %2 = tail call <4 x i32> @llvm.mips.max.u.w(<4 x i32> %0, <4 x i32> %1) store <4 x i32> %2, <4 x i32>* @llvm_mips_max_u_w_RES ret void @@ -252,8 +252,8 @@ declare <4 x i32> @llvm.mips.max.u.w(<4 x i32>, <4 x i32>) nounwind define void @llvm_mips_max_u_d_test() nounwind { entry: - %0 = load <2 x i64>* @llvm_mips_max_u_d_ARG1 - %1 = load <2 x i64>* @llvm_mips_max_u_d_ARG2 + %0 = load <2 x i64>, <2 x i64>* @llvm_mips_max_u_d_ARG1 + %1 = load <2 x i64>, <2 x i64>* @llvm_mips_max_u_d_ARG2 %2 = tail call <2 x i64> @llvm.mips.max.u.d(<2 x i64> %0, <2 x i64> %1) store <2 x i64> %2, <2 x i64>* @llvm_mips_max_u_d_RES ret void @@ -274,8 +274,8 @@ declare <2 x i64> @llvm.mips.max.u.d(<2 x i64>, <2 x i64>) nounwind define void @llvm_mips_min_a_b_test() nounwind { entry: - %0 = load <16 x i8>* @llvm_mips_min_a_b_ARG1 - %1 = load <16 x i8>* @llvm_mips_min_a_b_ARG2 + %0 = load <16 x i8>, <16 x i8>* @llvm_mips_min_a_b_ARG1 + %1 = load <16 x i8>, <16 x i8>* @llvm_mips_min_a_b_ARG2 %2 = tail call <16 x i8> @llvm.mips.min.a.b(<16 x i8> %0, <16 x i8> %1) store <16 x i8> %2, <16 x i8>* @llvm_mips_min_a_b_RES ret void @@ -296,8 +296,8 @@ declare <16 x i8> @llvm.mips.min.a.b(<16 x i8>, <16 x i8>) nounwind define void @llvm_mips_min_a_h_test() nounwind { entry: - %0 = load <8 x i16>* @llvm_mips_min_a_h_ARG1 - %1 = load <8 x i16>* @llvm_mips_min_a_h_ARG2 + %0 = load <8 x i16>, <8 x i16>* @llvm_mips_min_a_h_ARG1 + %1 = load <8 x i16>, <8 x i16>* @llvm_mips_min_a_h_ARG2 %2 = tail call <8 x i16> @llvm.mips.min.a.h(<8 x i16> %0, <8 x i16> %1) store <8 x i16> %2, <8 x i16>* @llvm_mips_min_a_h_RES ret void @@ -318,8 +318,8 @@ declare <8 x i16> @llvm.mips.min.a.h(<8 x i16>, <8 x i16>) nounwind define void @llvm_mips_min_a_w_test() nounwind { entry: - %0 = load <4 x i32>* @llvm_mips_min_a_w_ARG1 - %1 = load <4 x i32>* @llvm_mips_min_a_w_ARG2 + %0 = load <4 x i32>, <4 x i32>* @llvm_mips_min_a_w_ARG1 + %1 = load <4 x i32>, <4 x i32>* @llvm_mips_min_a_w_ARG2 %2 = tail call <4 x i32> @llvm.mips.min.a.w(<4 x i32> %0, <4 x i32> %1) store <4 x i32> %2, <4 x i32>* @llvm_mips_min_a_w_RES ret void @@ -340,8 +340,8 @@ declare <4 x i32> @llvm.mips.min.a.w(<4 x i32>, <4 x i32>) nounwind define void @llvm_mips_min_a_d_test() nounwind { entry: - %0 = load <2 x i64>* @llvm_mips_min_a_d_ARG1 - %1 = load <2 x i64>* @llvm_mips_min_a_d_ARG2 + %0 = load <2 x i64>, <2 x i64>* @llvm_mips_min_a_d_ARG1 + %1 = load <2 x i64>, <2 x i64>* @llvm_mips_min_a_d_ARG2 %2 = tail call <2 x i64> @llvm.mips.min.a.d(<2 x i64> %0, <2 x i64> %1) store <2 x i64> %2, <2 x i64>* @llvm_mips_min_a_d_RES ret void @@ -362,8 +362,8 @@ declare <2 x i64> @llvm.mips.min.a.d(<2 x i64>, <2 x i64>) nounwind define void @llvm_mips_min_s_b_test() nounwind { entry: - %0 = load <16 x i8>* @llvm_mips_min_s_b_ARG1 - %1 = load <16 x i8>* @llvm_mips_min_s_b_ARG2 + %0 = load <16 x i8>, <16 x i8>* @llvm_mips_min_s_b_ARG1 + %1 = load <16 x i8>, <16 x i8>* @llvm_mips_min_s_b_ARG2 %2 = tail call <16 x i8> @llvm.mips.min.s.b(<16 x i8> %0, <16 x i8> %1) store <16 x i8> %2, <16 x i8>* @llvm_mips_min_s_b_RES ret void @@ -384,8 +384,8 @@ declare <16 x i8> @llvm.mips.min.s.b(<16 x i8>, <16 x i8>) nounwind define void @llvm_mips_min_s_h_test() nounwind { entry: - %0 = load <8 x i16>* @llvm_mips_min_s_h_ARG1 - %1 = load <8 x i16>* @llvm_mips_min_s_h_ARG2 + %0 = load <8 x i16>, <8 x i16>* @llvm_mips_min_s_h_ARG1 + %1 = load <8 x i16>, <8 x i16>* @llvm_mips_min_s_h_ARG2 %2 = tail call <8 x i16> @llvm.mips.min.s.h(<8 x i16> %0, <8 x i16> %1) store <8 x i16> %2, <8 x i16>* @llvm_mips_min_s_h_RES ret void @@ -406,8 +406,8 @@ declare <8 x i16> @llvm.mips.min.s.h(<8 x i16>, <8 x i16>) nounwind define void @llvm_mips_min_s_w_test() nounwind { entry: - %0 = load <4 x i32>* @llvm_mips_min_s_w_ARG1 - %1 = load <4 x i32>* @llvm_mips_min_s_w_ARG2 + %0 = load <4 x i32>, <4 x i32>* @llvm_mips_min_s_w_ARG1 + %1 = load <4 x i32>, <4 x i32>* @llvm_mips_min_s_w_ARG2 %2 = tail call <4 x i32> @llvm.mips.min.s.w(<4 x i32> %0, <4 x i32> %1) store <4 x i32> %2, <4 x i32>* @llvm_mips_min_s_w_RES ret void @@ -428,8 +428,8 @@ declare <4 x i32> @llvm.mips.min.s.w(<4 x i32>, <4 x i32>) nounwind define void @llvm_mips_min_s_d_test() nounwind { entry: - %0 = load <2 x i64>* @llvm_mips_min_s_d_ARG1 - %1 = load <2 x i64>* @llvm_mips_min_s_d_ARG2 + %0 = load <2 x i64>, <2 x i64>* @llvm_mips_min_s_d_ARG1 + %1 = load <2 x i64>, <2 x i64>* @llvm_mips_min_s_d_ARG2 %2 = tail call <2 x i64> @llvm.mips.min.s.d(<2 x i64> %0, <2 x i64> %1) store <2 x i64> %2, <2 x i64>* @llvm_mips_min_s_d_RES ret void @@ -450,8 +450,8 @@ declare <2 x i64> @llvm.mips.min.s.d(<2 x i64>, <2 x i64>) nounwind define void @llvm_mips_min_u_b_test() nounwind { entry: - %0 = load <16 x i8>* @llvm_mips_min_u_b_ARG1 - %1 = load <16 x i8>* @llvm_mips_min_u_b_ARG2 + %0 = load <16 x i8>, <16 x i8>* @llvm_mips_min_u_b_ARG1 + %1 = load <16 x i8>, <16 x i8>* @llvm_mips_min_u_b_ARG2 %2 = tail call <16 x i8> @llvm.mips.min.u.b(<16 x i8> %0, <16 x i8> %1) store <16 x i8> %2, <16 x i8>* @llvm_mips_min_u_b_RES ret void @@ -472,8 +472,8 @@ declare <16 x i8> @llvm.mips.min.u.b(<16 x i8>, <16 x i8>) nounwind define void @llvm_mips_min_u_h_test() nounwind { entry: - %0 = load <8 x i16>* @llvm_mips_min_u_h_ARG1 - %1 = load <8 x i16>* @llvm_mips_min_u_h_ARG2 + %0 = load <8 x i16>, <8 x i16>* @llvm_mips_min_u_h_ARG1 + %1 = load <8 x i16>, <8 x i16>* @llvm_mips_min_u_h_ARG2 %2 = tail call <8 x i16> @llvm.mips.min.u.h(<8 x i16> %0, <8 x i16> %1) store <8 x i16> %2, <8 x i16>* @llvm_mips_min_u_h_RES ret void @@ -494,8 +494,8 @@ declare <8 x i16> @llvm.mips.min.u.h(<8 x i16>, <8 x i16>) nounwind define void @llvm_mips_min_u_w_test() nounwind { entry: - %0 = load <4 x i32>* @llvm_mips_min_u_w_ARG1 - %1 = load <4 x i32>* @llvm_mips_min_u_w_ARG2 + %0 = load <4 x i32>, <4 x i32>* @llvm_mips_min_u_w_ARG1 + %1 = load <4 x i32>, <4 x i32>* @llvm_mips_min_u_w_ARG2 %2 = tail call <4 x i32> @llvm.mips.min.u.w(<4 x i32> %0, <4 x i32> %1) store <4 x i32> %2, <4 x i32>* @llvm_mips_min_u_w_RES ret void @@ -516,8 +516,8 @@ declare <4 x i32> @llvm.mips.min.u.w(<4 x i32>, <4 x i32>) nounwind define void @llvm_mips_min_u_d_test() nounwind { entry: - %0 = load <2 x i64>* @llvm_mips_min_u_d_ARG1 - %1 = load <2 x i64>* @llvm_mips_min_u_d_ARG2 + %0 = load <2 x i64>, <2 x i64>* @llvm_mips_min_u_d_ARG1 + %1 = load <2 x i64>, <2 x i64>* @llvm_mips_min_u_d_ARG2 %2 = tail call <2 x i64> @llvm.mips.min.u.d(<2 x i64> %0, <2 x i64> %1) store <2 x i64> %2, <2 x i64>* @llvm_mips_min_u_d_RES ret void @@ -538,8 +538,8 @@ declare <2 x i64> @llvm.mips.min.u.d(<2 x i64>, <2 x i64>) nounwind define void @llvm_mips_mod_s_b_test() nounwind { entry: - %0 = load <16 x i8>* @llvm_mips_mod_s_b_ARG1 - %1 = load <16 x i8>* @llvm_mips_mod_s_b_ARG2 + %0 = load <16 x i8>, <16 x i8>* @llvm_mips_mod_s_b_ARG1 + %1 = load <16 x i8>, <16 x i8>* @llvm_mips_mod_s_b_ARG2 %2 = tail call <16 x i8> @llvm.mips.mod.s.b(<16 x i8> %0, <16 x i8> %1) store <16 x i8> %2, <16 x i8>* @llvm_mips_mod_s_b_RES ret void @@ -560,8 +560,8 @@ declare <16 x i8> @llvm.mips.mod.s.b(<16 x i8>, <16 x i8>) nounwind define void @llvm_mips_mod_s_h_test() nounwind { entry: - %0 = load <8 x i16>* @llvm_mips_mod_s_h_ARG1 - %1 = load <8 x i16>* @llvm_mips_mod_s_h_ARG2 + %0 = load <8 x i16>, <8 x i16>* @llvm_mips_mod_s_h_ARG1 + %1 = load <8 x i16>, <8 x i16>* @llvm_mips_mod_s_h_ARG2 %2 = tail call <8 x i16> @llvm.mips.mod.s.h(<8 x i16> %0, <8 x i16> %1) store <8 x i16> %2, <8 x i16>* @llvm_mips_mod_s_h_RES ret void @@ -582,8 +582,8 @@ declare <8 x i16> @llvm.mips.mod.s.h(<8 x i16>, <8 x i16>) nounwind define void @llvm_mips_mod_s_w_test() nounwind { entry: - %0 = load <4 x i32>* @llvm_mips_mod_s_w_ARG1 - %1 = load <4 x i32>* @llvm_mips_mod_s_w_ARG2 + %0 = load <4 x i32>, <4 x i32>* @llvm_mips_mod_s_w_ARG1 + %1 = load <4 x i32>, <4 x i32>* @llvm_mips_mod_s_w_ARG2 %2 = tail call <4 x i32> @llvm.mips.mod.s.w(<4 x i32> %0, <4 x i32> %1) store <4 x i32> %2, <4 x i32>* @llvm_mips_mod_s_w_RES ret void @@ -604,8 +604,8 @@ declare <4 x i32> @llvm.mips.mod.s.w(<4 x i32>, <4 x i32>) nounwind define void @llvm_mips_mod_s_d_test() nounwind { entry: - %0 = load <2 x i64>* @llvm_mips_mod_s_d_ARG1 - %1 = load <2 x i64>* @llvm_mips_mod_s_d_ARG2 + %0 = load <2 x i64>, <2 x i64>* @llvm_mips_mod_s_d_ARG1 + %1 = load <2 x i64>, <2 x i64>* @llvm_mips_mod_s_d_ARG2 %2 = tail call <2 x i64> @llvm.mips.mod.s.d(<2 x i64> %0, <2 x i64> %1) store <2 x i64> %2, <2 x i64>* @llvm_mips_mod_s_d_RES ret void @@ -626,8 +626,8 @@ declare <2 x i64> @llvm.mips.mod.s.d(<2 x i64>, <2 x i64>) nounwind define void @llvm_mips_mod_u_b_test() nounwind { entry: - %0 = load <16 x i8>* @llvm_mips_mod_u_b_ARG1 - %1 = load <16 x i8>* @llvm_mips_mod_u_b_ARG2 + %0 = load <16 x i8>, <16 x i8>* @llvm_mips_mod_u_b_ARG1 + %1 = load <16 x i8>, <16 x i8>* @llvm_mips_mod_u_b_ARG2 %2 = tail call <16 x i8> @llvm.mips.mod.u.b(<16 x i8> %0, <16 x i8> %1) store <16 x i8> %2, <16 x i8>* @llvm_mips_mod_u_b_RES ret void @@ -648,8 +648,8 @@ declare <16 x i8> @llvm.mips.mod.u.b(<16 x i8>, <16 x i8>) nounwind define void @llvm_mips_mod_u_h_test() nounwind { entry: - %0 = load <8 x i16>* @llvm_mips_mod_u_h_ARG1 - %1 = load <8 x i16>* @llvm_mips_mod_u_h_ARG2 + %0 = load <8 x i16>, <8 x i16>* @llvm_mips_mod_u_h_ARG1 + %1 = load <8 x i16>, <8 x i16>* @llvm_mips_mod_u_h_ARG2 %2 = tail call <8 x i16> @llvm.mips.mod.u.h(<8 x i16> %0, <8 x i16> %1) store <8 x i16> %2, <8 x i16>* @llvm_mips_mod_u_h_RES ret void @@ -670,8 +670,8 @@ declare <8 x i16> @llvm.mips.mod.u.h(<8 x i16>, <8 x i16>) nounwind define void @llvm_mips_mod_u_w_test() nounwind { entry: - %0 = load <4 x i32>* @llvm_mips_mod_u_w_ARG1 - %1 = load <4 x i32>* @llvm_mips_mod_u_w_ARG2 + %0 = load <4 x i32>, <4 x i32>* @llvm_mips_mod_u_w_ARG1 + %1 = load <4 x i32>, <4 x i32>* @llvm_mips_mod_u_w_ARG2 %2 = tail call <4 x i32> @llvm.mips.mod.u.w(<4 x i32> %0, <4 x i32> %1) store <4 x i32> %2, <4 x i32>* @llvm_mips_mod_u_w_RES ret void @@ -692,8 +692,8 @@ declare <4 x i32> @llvm.mips.mod.u.w(<4 x i32>, <4 x i32>) nounwind define void @llvm_mips_mod_u_d_test() nounwind { entry: - %0 = load <2 x i64>* @llvm_mips_mod_u_d_ARG1 - %1 = load <2 x i64>* @llvm_mips_mod_u_d_ARG2 + %0 = load <2 x i64>, <2 x i64>* @llvm_mips_mod_u_d_ARG1 + %1 = load <2 x i64>, <2 x i64>* @llvm_mips_mod_u_d_ARG2 %2 = tail call <2 x i64> @llvm.mips.mod.u.d(<2 x i64> %0, <2 x i64> %1) store <2 x i64> %2, <2 x i64>* @llvm_mips_mod_u_d_RES ret void @@ -714,8 +714,8 @@ declare <2 x i64> @llvm.mips.mod.u.d(<2 x i64>, <2 x i64>) nounwind define void @llvm_mips_mulv_b_test() nounwind { entry: - %0 = load <16 x i8>* @llvm_mips_mulv_b_ARG1 - %1 = load <16 x i8>* @llvm_mips_mulv_b_ARG2 + %0 = load <16 x i8>, <16 x i8>* @llvm_mips_mulv_b_ARG1 + %1 = load <16 x i8>, <16 x i8>* @llvm_mips_mulv_b_ARG2 %2 = tail call <16 x i8> @llvm.mips.mulv.b(<16 x i8> %0, <16 x i8> %1) store <16 x i8> %2, <16 x i8>* @llvm_mips_mulv_b_RES ret void @@ -736,8 +736,8 @@ declare <16 x i8> @llvm.mips.mulv.b(<16 x i8>, <16 x i8>) nounwind define void @llvm_mips_mulv_h_test() nounwind { entry: - %0 = load <8 x i16>* @llvm_mips_mulv_h_ARG1 - %1 = load <8 x i16>* @llvm_mips_mulv_h_ARG2 + %0 = load <8 x i16>, <8 x i16>* @llvm_mips_mulv_h_ARG1 + %1 = load <8 x i16>, <8 x i16>* @llvm_mips_mulv_h_ARG2 %2 = tail call <8 x i16> @llvm.mips.mulv.h(<8 x i16> %0, <8 x i16> %1) store <8 x i16> %2, <8 x i16>* @llvm_mips_mulv_h_RES ret void @@ -758,8 +758,8 @@ declare <8 x i16> @llvm.mips.mulv.h(<8 x i16>, <8 x i16>) nounwind define void @llvm_mips_mulv_w_test() nounwind { entry: - %0 = load <4 x i32>* @llvm_mips_mulv_w_ARG1 - %1 = load <4 x i32>* @llvm_mips_mulv_w_ARG2 + %0 = load <4 x i32>, <4 x i32>* @llvm_mips_mulv_w_ARG1 + %1 = load <4 x i32>, <4 x i32>* @llvm_mips_mulv_w_ARG2 %2 = tail call <4 x i32> @llvm.mips.mulv.w(<4 x i32> %0, <4 x i32> %1) store <4 x i32> %2, <4 x i32>* @llvm_mips_mulv_w_RES ret void @@ -780,8 +780,8 @@ declare <4 x i32> @llvm.mips.mulv.w(<4 x i32>, <4 x i32>) nounwind define void @llvm_mips_mulv_d_test() nounwind { entry: - %0 = load <2 x i64>* @llvm_mips_mulv_d_ARG1 - %1 = load <2 x i64>* @llvm_mips_mulv_d_ARG2 + %0 = load <2 x i64>, <2 x i64>* @llvm_mips_mulv_d_ARG1 + %1 = load <2 x i64>, <2 x i64>* @llvm_mips_mulv_d_ARG2 %2 = tail call <2 x i64> @llvm.mips.mulv.d(<2 x i64> %0, <2 x i64> %1) store <2 x i64> %2, <2 x i64>* @llvm_mips_mulv_d_RES ret void @@ -798,8 +798,8 @@ declare <2 x i64> @llvm.mips.mulv.d(<2 x i64>, <2 x i64>) nounwind define void @mulv_b_test() nounwind { entry: - %0 = load <16 x i8>* @llvm_mips_mulv_b_ARG1 - %1 = load <16 x i8>* @llvm_mips_mulv_b_ARG2 + %0 = load <16 x i8>, <16 x i8>* @llvm_mips_mulv_b_ARG1 + %1 = load <16 x i8>, <16 x i8>* @llvm_mips_mulv_b_ARG2 %2 = mul <16 x i8> %0, %1 store <16 x i8> %2, <16 x i8>* @llvm_mips_mulv_b_RES ret void @@ -814,8 +814,8 @@ entry: define void @mulv_h_test() nounwind { entry: - %0 = load <8 x i16>* @llvm_mips_mulv_h_ARG1 - %1 = load <8 x i16>* @llvm_mips_mulv_h_ARG2 + %0 = load <8 x i16>, <8 x i16>* @llvm_mips_mulv_h_ARG1 + %1 = load <8 x i16>, <8 x i16>* @llvm_mips_mulv_h_ARG2 %2 = mul <8 x i16> %0, %1 store <8 x i16> %2, <8 x i16>* @llvm_mips_mulv_h_RES ret void @@ -830,8 +830,8 @@ entry: define void @mulv_w_test() nounwind { entry: - %0 = load <4 x i32>* @llvm_mips_mulv_w_ARG1 - %1 = load <4 x i32>* @llvm_mips_mulv_w_ARG2 + %0 = load <4 x i32>, <4 x i32>* @llvm_mips_mulv_w_ARG1 + %1 = load <4 x i32>, <4 x i32>* @llvm_mips_mulv_w_ARG2 %2 = mul <4 x i32> %0, %1 store <4 x i32> %2, <4 x i32>* @llvm_mips_mulv_w_RES ret void @@ -846,8 +846,8 @@ entry: define void @mulv_d_test() nounwind { entry: - %0 = load <2 x i64>* @llvm_mips_mulv_d_ARG1 - %1 = load <2 x i64>* @llvm_mips_mulv_d_ARG2 + %0 = load <2 x i64>, <2 x i64>* @llvm_mips_mulv_d_ARG1 + %1 = load <2 x i64>, <2 x i64>* @llvm_mips_mulv_d_ARG2 %2 = mul <2 x i64> %0, %1 store <2 x i64> %2, <2 x i64>* @llvm_mips_mulv_d_RES ret void diff --git a/test/CodeGen/Mips/msa/3r-p.ll b/test/CodeGen/Mips/msa/3r-p.ll index 852023b0824a2..70b98aa8f9a12 100644 --- a/test/CodeGen/Mips/msa/3r-p.ll +++ b/test/CodeGen/Mips/msa/3r-p.ll @@ -10,8 +10,8 @@ define void @llvm_mips_pckev_b_test() nounwind { entry: - %0 = load <16 x i8>* @llvm_mips_pckev_b_ARG1 - %1 = load <16 x i8>* @llvm_mips_pckev_b_ARG2 + %0 = load <16 x i8>, <16 x i8>* @llvm_mips_pckev_b_ARG1 + %1 = load <16 x i8>, <16 x i8>* @llvm_mips_pckev_b_ARG2 %2 = tail call <16 x i8> @llvm.mips.pckev.b(<16 x i8> %0, <16 x i8> %1) store <16 x i8> %2, <16 x i8>* @llvm_mips_pckev_b_RES ret void @@ -32,8 +32,8 @@ declare <16 x i8> @llvm.mips.pckev.b(<16 x i8>, <16 x i8>) nounwind define void @llvm_mips_pckev_h_test() nounwind { entry: - %0 = load <8 x i16>* @llvm_mips_pckev_h_ARG1 - %1 = load <8 x i16>* @llvm_mips_pckev_h_ARG2 + %0 = load <8 x i16>, <8 x i16>* @llvm_mips_pckev_h_ARG1 + %1 = load <8 x i16>, <8 x i16>* @llvm_mips_pckev_h_ARG2 %2 = tail call <8 x i16> @llvm.mips.pckev.h(<8 x i16> %0, <8 x i16> %1) store <8 x i16> %2, <8 x i16>* @llvm_mips_pckev_h_RES ret void @@ -54,8 +54,8 @@ declare <8 x i16> @llvm.mips.pckev.h(<8 x i16>, <8 x i16>) nounwind define void @llvm_mips_pckev_w_test() nounwind { entry: - %0 = load <4 x i32>* @llvm_mips_pckev_w_ARG1 - %1 = load <4 x i32>* @llvm_mips_pckev_w_ARG2 + %0 = load <4 x i32>, <4 x i32>* @llvm_mips_pckev_w_ARG1 + %1 = load <4 x i32>, <4 x i32>* @llvm_mips_pckev_w_ARG2 %2 = tail call <4 x i32> @llvm.mips.pckev.w(<4 x i32> %0, <4 x i32> %1) store <4 x i32> %2, <4 x i32>* @llvm_mips_pckev_w_RES ret void @@ -76,8 +76,8 @@ declare <4 x i32> @llvm.mips.pckev.w(<4 x i32>, <4 x i32>) nounwind define void @llvm_mips_pckev_d_test() nounwind { entry: - %0 = load <2 x i64>* @llvm_mips_pckev_d_ARG1 - %1 = load <2 x i64>* @llvm_mips_pckev_d_ARG2 + %0 = load <2 x i64>, <2 x i64>* @llvm_mips_pckev_d_ARG1 + %1 = load <2 x i64>, <2 x i64>* @llvm_mips_pckev_d_ARG2 %2 = tail call <2 x i64> @llvm.mips.pckev.d(<2 x i64> %0, <2 x i64> %1) store <2 x i64> %2, <2 x i64>* @llvm_mips_pckev_d_RES ret void @@ -98,8 +98,8 @@ declare <2 x i64> @llvm.mips.pckev.d(<2 x i64>, <2 x i64>) nounwind define void @llvm_mips_pckod_b_test() nounwind { entry: - %0 = load <16 x i8>* @llvm_mips_pckod_b_ARG1 - %1 = load <16 x i8>* @llvm_mips_pckod_b_ARG2 + %0 = load <16 x i8>, <16 x i8>* @llvm_mips_pckod_b_ARG1 + %1 = load <16 x i8>, <16 x i8>* @llvm_mips_pckod_b_ARG2 %2 = tail call <16 x i8> @llvm.mips.pckod.b(<16 x i8> %0, <16 x i8> %1) store <16 x i8> %2, <16 x i8>* @llvm_mips_pckod_b_RES ret void @@ -120,8 +120,8 @@ declare <16 x i8> @llvm.mips.pckod.b(<16 x i8>, <16 x i8>) nounwind define void @llvm_mips_pckod_h_test() nounwind { entry: - %0 = load <8 x i16>* @llvm_mips_pckod_h_ARG1 - %1 = load <8 x i16>* @llvm_mips_pckod_h_ARG2 + %0 = load <8 x i16>, <8 x i16>* @llvm_mips_pckod_h_ARG1 + %1 = load <8 x i16>, <8 x i16>* @llvm_mips_pckod_h_ARG2 %2 = tail call <8 x i16> @llvm.mips.pckod.h(<8 x i16> %0, <8 x i16> %1) store <8 x i16> %2, <8 x i16>* @llvm_mips_pckod_h_RES ret void @@ -142,8 +142,8 @@ declare <8 x i16> @llvm.mips.pckod.h(<8 x i16>, <8 x i16>) nounwind define void @llvm_mips_pckod_w_test() nounwind { entry: - %0 = load <4 x i32>* @llvm_mips_pckod_w_ARG1 - %1 = load <4 x i32>* @llvm_mips_pckod_w_ARG2 + %0 = load <4 x i32>, <4 x i32>* @llvm_mips_pckod_w_ARG1 + %1 = load <4 x i32>, <4 x i32>* @llvm_mips_pckod_w_ARG2 %2 = tail call <4 x i32> @llvm.mips.pckod.w(<4 x i32> %0, <4 x i32> %1) store <4 x i32> %2, <4 x i32>* @llvm_mips_pckod_w_RES ret void @@ -164,8 +164,8 @@ declare <4 x i32> @llvm.mips.pckod.w(<4 x i32>, <4 x i32>) nounwind define void @llvm_mips_pckod_d_test() nounwind { entry: - %0 = load <2 x i64>* @llvm_mips_pckod_d_ARG1 - %1 = load <2 x i64>* @llvm_mips_pckod_d_ARG2 + %0 = load <2 x i64>, <2 x i64>* @llvm_mips_pckod_d_ARG1 + %1 = load <2 x i64>, <2 x i64>* @llvm_mips_pckod_d_ARG2 %2 = tail call <2 x i64> @llvm.mips.pckod.d(<2 x i64> %0, <2 x i64> %1) store <2 x i64> %2, <2 x i64>* @llvm_mips_pckod_d_RES ret void diff --git a/test/CodeGen/Mips/msa/3r-s.ll b/test/CodeGen/Mips/msa/3r-s.ll index 581c3bfd78afe..d04c5ff165f2b 100644 --- a/test/CodeGen/Mips/msa/3r-s.ll +++ b/test/CodeGen/Mips/msa/3r-s.ll @@ -11,9 +11,9 @@ define void @llvm_mips_sld_b_test() nounwind { entry: - %0 = load <16 x i8>* @llvm_mips_sld_b_ARG1 - %1 = load <16 x i8>* @llvm_mips_sld_b_ARG2 - %2 = load i32* @llvm_mips_sld_b_ARG3 + %0 = load <16 x i8>, <16 x i8>* @llvm_mips_sld_b_ARG1 + %1 = load <16 x i8>, <16 x i8>* @llvm_mips_sld_b_ARG2 + %2 = load i32, i32* @llvm_mips_sld_b_ARG3 %3 = tail call <16 x i8> @llvm.mips.sld.b(<16 x i8> %0, <16 x i8> %1, i32 %2) store <16 x i8> %3, <16 x i8>* @llvm_mips_sld_b_RES ret void @@ -39,9 +39,9 @@ declare <16 x i8> @llvm.mips.sld.b(<16 x i8>, <16 x i8>, i32) nounwind define void @llvm_mips_sld_h_test() nounwind { entry: - %0 = load <8 x i16>* @llvm_mips_sld_h_ARG1 - %1 = load <8 x i16>* @llvm_mips_sld_h_ARG2 - %2 = load i32* @llvm_mips_sld_h_ARG3 + %0 = load <8 x i16>, <8 x i16>* @llvm_mips_sld_h_ARG1 + %1 = load <8 x i16>, <8 x i16>* @llvm_mips_sld_h_ARG2 + %2 = load i32, i32* @llvm_mips_sld_h_ARG3 %3 = tail call <8 x i16> @llvm.mips.sld.h(<8 x i16> %0, <8 x i16> %1, i32 %2) store <8 x i16> %3, <8 x i16>* @llvm_mips_sld_h_RES ret void @@ -67,9 +67,9 @@ declare <8 x i16> @llvm.mips.sld.h(<8 x i16>, <8 x i16>, i32) nounwind define void @llvm_mips_sld_w_test() nounwind { entry: - %0 = load <4 x i32>* @llvm_mips_sld_w_ARG1 - %1 = load <4 x i32>* @llvm_mips_sld_w_ARG2 - %2 = load i32* @llvm_mips_sld_w_ARG3 + %0 = load <4 x i32>, <4 x i32>* @llvm_mips_sld_w_ARG1 + %1 = load <4 x i32>, <4 x i32>* @llvm_mips_sld_w_ARG2 + %2 = load i32, i32* @llvm_mips_sld_w_ARG3 %3 = tail call <4 x i32> @llvm.mips.sld.w(<4 x i32> %0, <4 x i32> %1, i32 %2) store <4 x i32> %3, <4 x i32>* @llvm_mips_sld_w_RES ret void @@ -95,9 +95,9 @@ declare <4 x i32> @llvm.mips.sld.w(<4 x i32>, <4 x i32>, i32) nounwind define void @llvm_mips_sld_d_test() nounwind { entry: - %0 = load <2 x i64>* @llvm_mips_sld_d_ARG1 - %1 = load <2 x i64>* @llvm_mips_sld_d_ARG2 - %2 = load i32* @llvm_mips_sld_d_ARG3 + %0 = load <2 x i64>, <2 x i64>* @llvm_mips_sld_d_ARG1 + %1 = load <2 x i64>, <2 x i64>* @llvm_mips_sld_d_ARG2 + %2 = load i32, i32* @llvm_mips_sld_d_ARG3 %3 = tail call <2 x i64> @llvm.mips.sld.d(<2 x i64> %0, <2 x i64> %1, i32 %2) store <2 x i64> %3, <2 x i64>* @llvm_mips_sld_d_RES ret void @@ -122,8 +122,8 @@ declare <2 x i64> @llvm.mips.sld.d(<2 x i64>, <2 x i64>, i32) nounwind define void @llvm_mips_sll_b_test() nounwind { entry: - %0 = load <16 x i8>* @llvm_mips_sll_b_ARG1 - %1 = load <16 x i8>* @llvm_mips_sll_b_ARG2 + %0 = load <16 x i8>, <16 x i8>* @llvm_mips_sll_b_ARG1 + %1 = load <16 x i8>, <16 x i8>* @llvm_mips_sll_b_ARG2 %2 = tail call <16 x i8> @llvm.mips.sll.b(<16 x i8> %0, <16 x i8> %1) store <16 x i8> %2, <16 x i8>* @llvm_mips_sll_b_RES ret void @@ -146,8 +146,8 @@ declare <16 x i8> @llvm.mips.sll.b(<16 x i8>, <16 x i8>) nounwind define void @llvm_mips_sll_h_test() nounwind { entry: - %0 = load <8 x i16>* @llvm_mips_sll_h_ARG1 - %1 = load <8 x i16>* @llvm_mips_sll_h_ARG2 + %0 = load <8 x i16>, <8 x i16>* @llvm_mips_sll_h_ARG1 + %1 = load <8 x i16>, <8 x i16>* @llvm_mips_sll_h_ARG2 %2 = tail call <8 x i16> @llvm.mips.sll.h(<8 x i16> %0, <8 x i16> %1) store <8 x i16> %2, <8 x i16>* @llvm_mips_sll_h_RES ret void @@ -170,8 +170,8 @@ declare <8 x i16> @llvm.mips.sll.h(<8 x i16>, <8 x i16>) nounwind define void @llvm_mips_sll_w_test() nounwind { entry: - %0 = load <4 x i32>* @llvm_mips_sll_w_ARG1 - %1 = load <4 x i32>* @llvm_mips_sll_w_ARG2 + %0 = load <4 x i32>, <4 x i32>* @llvm_mips_sll_w_ARG1 + %1 = load <4 x i32>, <4 x i32>* @llvm_mips_sll_w_ARG2 %2 = tail call <4 x i32> @llvm.mips.sll.w(<4 x i32> %0, <4 x i32> %1) store <4 x i32> %2, <4 x i32>* @llvm_mips_sll_w_RES ret void @@ -194,8 +194,8 @@ declare <4 x i32> @llvm.mips.sll.w(<4 x i32>, <4 x i32>) nounwind define void @llvm_mips_sll_d_test() nounwind { entry: - %0 = load <2 x i64>* @llvm_mips_sll_d_ARG1 - %1 = load <2 x i64>* @llvm_mips_sll_d_ARG2 + %0 = load <2 x i64>, <2 x i64>* @llvm_mips_sll_d_ARG1 + %1 = load <2 x i64>, <2 x i64>* @llvm_mips_sll_d_ARG2 %2 = tail call <2 x i64> @llvm.mips.sll.d(<2 x i64> %0, <2 x i64> %1) store <2 x i64> %2, <2 x i64>* @llvm_mips_sll_d_RES ret void @@ -214,8 +214,8 @@ declare <2 x i64> @llvm.mips.sll.d(<2 x i64>, <2 x i64>) nounwind define void @sll_b_test() nounwind { entry: - %0 = load <16 x i8>* @llvm_mips_sll_b_ARG1 - %1 = load <16 x i8>* @llvm_mips_sll_b_ARG2 + %0 = load <16 x i8>, <16 x i8>* @llvm_mips_sll_b_ARG1 + %1 = load <16 x i8>, <16 x i8>* @llvm_mips_sll_b_ARG2 %2 = shl <16 x i8> %0, %1 store <16 x i8> %2, <16 x i8>* @llvm_mips_sll_b_RES ret void @@ -232,8 +232,8 @@ entry: define void @sll_h_test() nounwind { entry: - %0 = load <8 x i16>* @llvm_mips_sll_h_ARG1 - %1 = load <8 x i16>* @llvm_mips_sll_h_ARG2 + %0 = load <8 x i16>, <8 x i16>* @llvm_mips_sll_h_ARG1 + %1 = load <8 x i16>, <8 x i16>* @llvm_mips_sll_h_ARG2 %2 = shl <8 x i16> %0, %1 store <8 x i16> %2, <8 x i16>* @llvm_mips_sll_h_RES ret void @@ -250,8 +250,8 @@ entry: define void @sll_w_test() nounwind { entry: - %0 = load <4 x i32>* @llvm_mips_sll_w_ARG1 - %1 = load <4 x i32>* @llvm_mips_sll_w_ARG2 + %0 = load <4 x i32>, <4 x i32>* @llvm_mips_sll_w_ARG1 + %1 = load <4 x i32>, <4 x i32>* @llvm_mips_sll_w_ARG2 %2 = shl <4 x i32> %0, %1 store <4 x i32> %2, <4 x i32>* @llvm_mips_sll_w_RES ret void @@ -268,8 +268,8 @@ entry: define void @sll_d_test() nounwind { entry: - %0 = load <2 x i64>* @llvm_mips_sll_d_ARG1 - %1 = load <2 x i64>* @llvm_mips_sll_d_ARG2 + %0 = load <2 x i64>, <2 x i64>* @llvm_mips_sll_d_ARG1 + %1 = load <2 x i64>, <2 x i64>* @llvm_mips_sll_d_ARG2 %2 = shl <2 x i64> %0, %1 store <2 x i64> %2, <2 x i64>* @llvm_mips_sll_d_RES ret void @@ -290,8 +290,8 @@ entry: define void @llvm_mips_sra_b_test() nounwind { entry: - %0 = load <16 x i8>* @llvm_mips_sra_b_ARG1 - %1 = load <16 x i8>* @llvm_mips_sra_b_ARG2 + %0 = load <16 x i8>, <16 x i8>* @llvm_mips_sra_b_ARG1 + %1 = load <16 x i8>, <16 x i8>* @llvm_mips_sra_b_ARG2 %2 = tail call <16 x i8> @llvm.mips.sra.b(<16 x i8> %0, <16 x i8> %1) store <16 x i8> %2, <16 x i8>* @llvm_mips_sra_b_RES ret void @@ -314,8 +314,8 @@ declare <16 x i8> @llvm.mips.sra.b(<16 x i8>, <16 x i8>) nounwind define void @llvm_mips_sra_h_test() nounwind { entry: - %0 = load <8 x i16>* @llvm_mips_sra_h_ARG1 - %1 = load <8 x i16>* @llvm_mips_sra_h_ARG2 + %0 = load <8 x i16>, <8 x i16>* @llvm_mips_sra_h_ARG1 + %1 = load <8 x i16>, <8 x i16>* @llvm_mips_sra_h_ARG2 %2 = tail call <8 x i16> @llvm.mips.sra.h(<8 x i16> %0, <8 x i16> %1) store <8 x i16> %2, <8 x i16>* @llvm_mips_sra_h_RES ret void @@ -338,8 +338,8 @@ declare <8 x i16> @llvm.mips.sra.h(<8 x i16>, <8 x i16>) nounwind define void @llvm_mips_sra_w_test() nounwind { entry: - %0 = load <4 x i32>* @llvm_mips_sra_w_ARG1 - %1 = load <4 x i32>* @llvm_mips_sra_w_ARG2 + %0 = load <4 x i32>, <4 x i32>* @llvm_mips_sra_w_ARG1 + %1 = load <4 x i32>, <4 x i32>* @llvm_mips_sra_w_ARG2 %2 = tail call <4 x i32> @llvm.mips.sra.w(<4 x i32> %0, <4 x i32> %1) store <4 x i32> %2, <4 x i32>* @llvm_mips_sra_w_RES ret void @@ -362,8 +362,8 @@ declare <4 x i32> @llvm.mips.sra.w(<4 x i32>, <4 x i32>) nounwind define void @llvm_mips_sra_d_test() nounwind { entry: - %0 = load <2 x i64>* @llvm_mips_sra_d_ARG1 - %1 = load <2 x i64>* @llvm_mips_sra_d_ARG2 + %0 = load <2 x i64>, <2 x i64>* @llvm_mips_sra_d_ARG1 + %1 = load <2 x i64>, <2 x i64>* @llvm_mips_sra_d_ARG2 %2 = tail call <2 x i64> @llvm.mips.sra.d(<2 x i64> %0, <2 x i64> %1) store <2 x i64> %2, <2 x i64>* @llvm_mips_sra_d_RES ret void @@ -383,8 +383,8 @@ declare <2 x i64> @llvm.mips.sra.d(<2 x i64>, <2 x i64>) nounwind define void @sra_b_test() nounwind { entry: - %0 = load <16 x i8>* @llvm_mips_sra_b_ARG1 - %1 = load <16 x i8>* @llvm_mips_sra_b_ARG2 + %0 = load <16 x i8>, <16 x i8>* @llvm_mips_sra_b_ARG1 + %1 = load <16 x i8>, <16 x i8>* @llvm_mips_sra_b_ARG2 %2 = ashr <16 x i8> %0, %1 store <16 x i8> %2, <16 x i8>* @llvm_mips_sra_b_RES ret void @@ -401,8 +401,8 @@ entry: define void @sra_h_test() nounwind { entry: - %0 = load <8 x i16>* @llvm_mips_sra_h_ARG1 - %1 = load <8 x i16>* @llvm_mips_sra_h_ARG2 + %0 = load <8 x i16>, <8 x i16>* @llvm_mips_sra_h_ARG1 + %1 = load <8 x i16>, <8 x i16>* @llvm_mips_sra_h_ARG2 %2 = ashr <8 x i16> %0, %1 store <8 x i16> %2, <8 x i16>* @llvm_mips_sra_h_RES ret void @@ -419,8 +419,8 @@ entry: define void @sra_w_test() nounwind { entry: - %0 = load <4 x i32>* @llvm_mips_sra_w_ARG1 - %1 = load <4 x i32>* @llvm_mips_sra_w_ARG2 + %0 = load <4 x i32>, <4 x i32>* @llvm_mips_sra_w_ARG1 + %1 = load <4 x i32>, <4 x i32>* @llvm_mips_sra_w_ARG2 %2 = ashr <4 x i32> %0, %1 store <4 x i32> %2, <4 x i32>* @llvm_mips_sra_w_RES ret void @@ -437,8 +437,8 @@ entry: define void @sra_d_test() nounwind { entry: - %0 = load <2 x i64>* @llvm_mips_sra_d_ARG1 - %1 = load <2 x i64>* @llvm_mips_sra_d_ARG2 + %0 = load <2 x i64>, <2 x i64>* @llvm_mips_sra_d_ARG1 + %1 = load <2 x i64>, <2 x i64>* @llvm_mips_sra_d_ARG2 %2 = ashr <2 x i64> %0, %1 store <2 x i64> %2, <2 x i64>* @llvm_mips_sra_d_RES ret void @@ -459,8 +459,8 @@ entry: define void @llvm_mips_srar_b_test() nounwind { entry: - %0 = load <16 x i8>* @llvm_mips_srar_b_ARG1 - %1 = load <16 x i8>* @llvm_mips_srar_b_ARG2 + %0 = load <16 x i8>, <16 x i8>* @llvm_mips_srar_b_ARG1 + %1 = load <16 x i8>, <16 x i8>* @llvm_mips_srar_b_ARG2 %2 = tail call <16 x i8> @llvm.mips.srar.b(<16 x i8> %0, <16 x i8> %1) store <16 x i8> %2, <16 x i8>* @llvm_mips_srar_b_RES ret void @@ -483,8 +483,8 @@ declare <16 x i8> @llvm.mips.srar.b(<16 x i8>, <16 x i8>) nounwind define void @llvm_mips_srar_h_test() nounwind { entry: - %0 = load <8 x i16>* @llvm_mips_srar_h_ARG1 - %1 = load <8 x i16>* @llvm_mips_srar_h_ARG2 + %0 = load <8 x i16>, <8 x i16>* @llvm_mips_srar_h_ARG1 + %1 = load <8 x i16>, <8 x i16>* @llvm_mips_srar_h_ARG2 %2 = tail call <8 x i16> @llvm.mips.srar.h(<8 x i16> %0, <8 x i16> %1) store <8 x i16> %2, <8 x i16>* @llvm_mips_srar_h_RES ret void @@ -507,8 +507,8 @@ declare <8 x i16> @llvm.mips.srar.h(<8 x i16>, <8 x i16>) nounwind define void @llvm_mips_srar_w_test() nounwind { entry: - %0 = load <4 x i32>* @llvm_mips_srar_w_ARG1 - %1 = load <4 x i32>* @llvm_mips_srar_w_ARG2 + %0 = load <4 x i32>, <4 x i32>* @llvm_mips_srar_w_ARG1 + %1 = load <4 x i32>, <4 x i32>* @llvm_mips_srar_w_ARG2 %2 = tail call <4 x i32> @llvm.mips.srar.w(<4 x i32> %0, <4 x i32> %1) store <4 x i32> %2, <4 x i32>* @llvm_mips_srar_w_RES ret void @@ -531,8 +531,8 @@ declare <4 x i32> @llvm.mips.srar.w(<4 x i32>, <4 x i32>) nounwind define void @llvm_mips_srar_d_test() nounwind { entry: - %0 = load <2 x i64>* @llvm_mips_srar_d_ARG1 - %1 = load <2 x i64>* @llvm_mips_srar_d_ARG2 + %0 = load <2 x i64>, <2 x i64>* @llvm_mips_srar_d_ARG1 + %1 = load <2 x i64>, <2 x i64>* @llvm_mips_srar_d_ARG2 %2 = tail call <2 x i64> @llvm.mips.srar.d(<2 x i64> %0, <2 x i64> %1) store <2 x i64> %2, <2 x i64>* @llvm_mips_srar_d_RES ret void @@ -555,8 +555,8 @@ declare <2 x i64> @llvm.mips.srar.d(<2 x i64>, <2 x i64>) nounwind define void @llvm_mips_srl_b_test() nounwind { entry: - %0 = load <16 x i8>* @llvm_mips_srl_b_ARG1 - %1 = load <16 x i8>* @llvm_mips_srl_b_ARG2 + %0 = load <16 x i8>, <16 x i8>* @llvm_mips_srl_b_ARG1 + %1 = load <16 x i8>, <16 x i8>* @llvm_mips_srl_b_ARG2 %2 = tail call <16 x i8> @llvm.mips.srl.b(<16 x i8> %0, <16 x i8> %1) store <16 x i8> %2, <16 x i8>* @llvm_mips_srl_b_RES ret void @@ -579,8 +579,8 @@ declare <16 x i8> @llvm.mips.srl.b(<16 x i8>, <16 x i8>) nounwind define void @llvm_mips_srl_h_test() nounwind { entry: - %0 = load <8 x i16>* @llvm_mips_srl_h_ARG1 - %1 = load <8 x i16>* @llvm_mips_srl_h_ARG2 + %0 = load <8 x i16>, <8 x i16>* @llvm_mips_srl_h_ARG1 + %1 = load <8 x i16>, <8 x i16>* @llvm_mips_srl_h_ARG2 %2 = tail call <8 x i16> @llvm.mips.srl.h(<8 x i16> %0, <8 x i16> %1) store <8 x i16> %2, <8 x i16>* @llvm_mips_srl_h_RES ret void @@ -603,8 +603,8 @@ declare <8 x i16> @llvm.mips.srl.h(<8 x i16>, <8 x i16>) nounwind define void @llvm_mips_srl_w_test() nounwind { entry: - %0 = load <4 x i32>* @llvm_mips_srl_w_ARG1 - %1 = load <4 x i32>* @llvm_mips_srl_w_ARG2 + %0 = load <4 x i32>, <4 x i32>* @llvm_mips_srl_w_ARG1 + %1 = load <4 x i32>, <4 x i32>* @llvm_mips_srl_w_ARG2 %2 = tail call <4 x i32> @llvm.mips.srl.w(<4 x i32> %0, <4 x i32> %1) store <4 x i32> %2, <4 x i32>* @llvm_mips_srl_w_RES ret void @@ -627,8 +627,8 @@ declare <4 x i32> @llvm.mips.srl.w(<4 x i32>, <4 x i32>) nounwind define void @llvm_mips_srl_d_test() nounwind { entry: - %0 = load <2 x i64>* @llvm_mips_srl_d_ARG1 - %1 = load <2 x i64>* @llvm_mips_srl_d_ARG2 + %0 = load <2 x i64>, <2 x i64>* @llvm_mips_srl_d_ARG1 + %1 = load <2 x i64>, <2 x i64>* @llvm_mips_srl_d_ARG2 %2 = tail call <2 x i64> @llvm.mips.srl.d(<2 x i64> %0, <2 x i64> %1) store <2 x i64> %2, <2 x i64>* @llvm_mips_srl_d_RES ret void @@ -651,8 +651,8 @@ declare <2 x i64> @llvm.mips.srl.d(<2 x i64>, <2 x i64>) nounwind define void @llvm_mips_srlr_b_test() nounwind { entry: - %0 = load <16 x i8>* @llvm_mips_srlr_b_ARG1 - %1 = load <16 x i8>* @llvm_mips_srlr_b_ARG2 + %0 = load <16 x i8>, <16 x i8>* @llvm_mips_srlr_b_ARG1 + %1 = load <16 x i8>, <16 x i8>* @llvm_mips_srlr_b_ARG2 %2 = tail call <16 x i8> @llvm.mips.srlr.b(<16 x i8> %0, <16 x i8> %1) store <16 x i8> %2, <16 x i8>* @llvm_mips_srlr_b_RES ret void @@ -675,8 +675,8 @@ declare <16 x i8> @llvm.mips.srlr.b(<16 x i8>, <16 x i8>) nounwind define void @llvm_mips_srlr_h_test() nounwind { entry: - %0 = load <8 x i16>* @llvm_mips_srlr_h_ARG1 - %1 = load <8 x i16>* @llvm_mips_srlr_h_ARG2 + %0 = load <8 x i16>, <8 x i16>* @llvm_mips_srlr_h_ARG1 + %1 = load <8 x i16>, <8 x i16>* @llvm_mips_srlr_h_ARG2 %2 = tail call <8 x i16> @llvm.mips.srlr.h(<8 x i16> %0, <8 x i16> %1) store <8 x i16> %2, <8 x i16>* @llvm_mips_srlr_h_RES ret void @@ -699,8 +699,8 @@ declare <8 x i16> @llvm.mips.srlr.h(<8 x i16>, <8 x i16>) nounwind define void @llvm_mips_srlr_w_test() nounwind { entry: - %0 = load <4 x i32>* @llvm_mips_srlr_w_ARG1 - %1 = load <4 x i32>* @llvm_mips_srlr_w_ARG2 + %0 = load <4 x i32>, <4 x i32>* @llvm_mips_srlr_w_ARG1 + %1 = load <4 x i32>, <4 x i32>* @llvm_mips_srlr_w_ARG2 %2 = tail call <4 x i32> @llvm.mips.srlr.w(<4 x i32> %0, <4 x i32> %1) store <4 x i32> %2, <4 x i32>* @llvm_mips_srlr_w_RES ret void @@ -723,8 +723,8 @@ declare <4 x i32> @llvm.mips.srlr.w(<4 x i32>, <4 x i32>) nounwind define void @llvm_mips_srlr_d_test() nounwind { entry: - %0 = load <2 x i64>* @llvm_mips_srlr_d_ARG1 - %1 = load <2 x i64>* @llvm_mips_srlr_d_ARG2 + %0 = load <2 x i64>, <2 x i64>* @llvm_mips_srlr_d_ARG1 + %1 = load <2 x i64>, <2 x i64>* @llvm_mips_srlr_d_ARG2 %2 = tail call <2 x i64> @llvm.mips.srlr.d(<2 x i64> %0, <2 x i64> %1) store <2 x i64> %2, <2 x i64>* @llvm_mips_srlr_d_RES ret void @@ -744,8 +744,8 @@ declare <2 x i64> @llvm.mips.srlr.d(<2 x i64>, <2 x i64>) nounwind define void @srl_b_test() nounwind { entry: - %0 = load <16 x i8>* @llvm_mips_srl_b_ARG1 - %1 = load <16 x i8>* @llvm_mips_srl_b_ARG2 + %0 = load <16 x i8>, <16 x i8>* @llvm_mips_srl_b_ARG1 + %1 = load <16 x i8>, <16 x i8>* @llvm_mips_srl_b_ARG2 %2 = lshr <16 x i8> %0, %1 store <16 x i8> %2, <16 x i8>* @llvm_mips_srl_b_RES ret void @@ -762,8 +762,8 @@ entry: define void @srl_h_test() nounwind { entry: - %0 = load <8 x i16>* @llvm_mips_srl_h_ARG1 - %1 = load <8 x i16>* @llvm_mips_srl_h_ARG2 + %0 = load <8 x i16>, <8 x i16>* @llvm_mips_srl_h_ARG1 + %1 = load <8 x i16>, <8 x i16>* @llvm_mips_srl_h_ARG2 %2 = lshr <8 x i16> %0, %1 store <8 x i16> %2, <8 x i16>* @llvm_mips_srl_h_RES ret void @@ -780,8 +780,8 @@ entry: define void @srl_w_test() nounwind { entry: - %0 = load <4 x i32>* @llvm_mips_srl_w_ARG1 - %1 = load <4 x i32>* @llvm_mips_srl_w_ARG2 + %0 = load <4 x i32>, <4 x i32>* @llvm_mips_srl_w_ARG1 + %1 = load <4 x i32>, <4 x i32>* @llvm_mips_srl_w_ARG2 %2 = lshr <4 x i32> %0, %1 store <4 x i32> %2, <4 x i32>* @llvm_mips_srl_w_RES ret void @@ -798,8 +798,8 @@ entry: define void @srl_d_test() nounwind { entry: - %0 = load <2 x i64>* @llvm_mips_srl_d_ARG1 - %1 = load <2 x i64>* @llvm_mips_srl_d_ARG2 + %0 = load <2 x i64>, <2 x i64>* @llvm_mips_srl_d_ARG1 + %1 = load <2 x i64>, <2 x i64>* @llvm_mips_srl_d_ARG2 %2 = lshr <2 x i64> %0, %1 store <2 x i64> %2, <2 x i64>* @llvm_mips_srl_d_RES ret void @@ -820,8 +820,8 @@ entry: define void @llvm_mips_subs_s_b_test() nounwind { entry: - %0 = load <16 x i8>* @llvm_mips_subs_s_b_ARG1 - %1 = load <16 x i8>* @llvm_mips_subs_s_b_ARG2 + %0 = load <16 x i8>, <16 x i8>* @llvm_mips_subs_s_b_ARG1 + %1 = load <16 x i8>, <16 x i8>* @llvm_mips_subs_s_b_ARG2 %2 = tail call <16 x i8> @llvm.mips.subs.s.b(<16 x i8> %0, <16 x i8> %1) store <16 x i8> %2, <16 x i8>* @llvm_mips_subs_s_b_RES ret void @@ -844,8 +844,8 @@ declare <16 x i8> @llvm.mips.subs.s.b(<16 x i8>, <16 x i8>) nounwind define void @llvm_mips_subs_s_h_test() nounwind { entry: - %0 = load <8 x i16>* @llvm_mips_subs_s_h_ARG1 - %1 = load <8 x i16>* @llvm_mips_subs_s_h_ARG2 + %0 = load <8 x i16>, <8 x i16>* @llvm_mips_subs_s_h_ARG1 + %1 = load <8 x i16>, <8 x i16>* @llvm_mips_subs_s_h_ARG2 %2 = tail call <8 x i16> @llvm.mips.subs.s.h(<8 x i16> %0, <8 x i16> %1) store <8 x i16> %2, <8 x i16>* @llvm_mips_subs_s_h_RES ret void @@ -868,8 +868,8 @@ declare <8 x i16> @llvm.mips.subs.s.h(<8 x i16>, <8 x i16>) nounwind define void @llvm_mips_subs_s_w_test() nounwind { entry: - %0 = load <4 x i32>* @llvm_mips_subs_s_w_ARG1 - %1 = load <4 x i32>* @llvm_mips_subs_s_w_ARG2 + %0 = load <4 x i32>, <4 x i32>* @llvm_mips_subs_s_w_ARG1 + %1 = load <4 x i32>, <4 x i32>* @llvm_mips_subs_s_w_ARG2 %2 = tail call <4 x i32> @llvm.mips.subs.s.w(<4 x i32> %0, <4 x i32> %1) store <4 x i32> %2, <4 x i32>* @llvm_mips_subs_s_w_RES ret void @@ -892,8 +892,8 @@ declare <4 x i32> @llvm.mips.subs.s.w(<4 x i32>, <4 x i32>) nounwind define void @llvm_mips_subs_s_d_test() nounwind { entry: - %0 = load <2 x i64>* @llvm_mips_subs_s_d_ARG1 - %1 = load <2 x i64>* @llvm_mips_subs_s_d_ARG2 + %0 = load <2 x i64>, <2 x i64>* @llvm_mips_subs_s_d_ARG1 + %1 = load <2 x i64>, <2 x i64>* @llvm_mips_subs_s_d_ARG2 %2 = tail call <2 x i64> @llvm.mips.subs.s.d(<2 x i64> %0, <2 x i64> %1) store <2 x i64> %2, <2 x i64>* @llvm_mips_subs_s_d_RES ret void @@ -916,8 +916,8 @@ declare <2 x i64> @llvm.mips.subs.s.d(<2 x i64>, <2 x i64>) nounwind define void @llvm_mips_subs_u_b_test() nounwind { entry: - %0 = load <16 x i8>* @llvm_mips_subs_u_b_ARG1 - %1 = load <16 x i8>* @llvm_mips_subs_u_b_ARG2 + %0 = load <16 x i8>, <16 x i8>* @llvm_mips_subs_u_b_ARG1 + %1 = load <16 x i8>, <16 x i8>* @llvm_mips_subs_u_b_ARG2 %2 = tail call <16 x i8> @llvm.mips.subs.u.b(<16 x i8> %0, <16 x i8> %1) store <16 x i8> %2, <16 x i8>* @llvm_mips_subs_u_b_RES ret void @@ -940,8 +940,8 @@ declare <16 x i8> @llvm.mips.subs.u.b(<16 x i8>, <16 x i8>) nounwind define void @llvm_mips_subs_u_h_test() nounwind { entry: - %0 = load <8 x i16>* @llvm_mips_subs_u_h_ARG1 - %1 = load <8 x i16>* @llvm_mips_subs_u_h_ARG2 + %0 = load <8 x i16>, <8 x i16>* @llvm_mips_subs_u_h_ARG1 + %1 = load <8 x i16>, <8 x i16>* @llvm_mips_subs_u_h_ARG2 %2 = tail call <8 x i16> @llvm.mips.subs.u.h(<8 x i16> %0, <8 x i16> %1) store <8 x i16> %2, <8 x i16>* @llvm_mips_subs_u_h_RES ret void @@ -964,8 +964,8 @@ declare <8 x i16> @llvm.mips.subs.u.h(<8 x i16>, <8 x i16>) nounwind define void @llvm_mips_subs_u_w_test() nounwind { entry: - %0 = load <4 x i32>* @llvm_mips_subs_u_w_ARG1 - %1 = load <4 x i32>* @llvm_mips_subs_u_w_ARG2 + %0 = load <4 x i32>, <4 x i32>* @llvm_mips_subs_u_w_ARG1 + %1 = load <4 x i32>, <4 x i32>* @llvm_mips_subs_u_w_ARG2 %2 = tail call <4 x i32> @llvm.mips.subs.u.w(<4 x i32> %0, <4 x i32> %1) store <4 x i32> %2, <4 x i32>* @llvm_mips_subs_u_w_RES ret void @@ -988,8 +988,8 @@ declare <4 x i32> @llvm.mips.subs.u.w(<4 x i32>, <4 x i32>) nounwind define void @llvm_mips_subs_u_d_test() nounwind { entry: - %0 = load <2 x i64>* @llvm_mips_subs_u_d_ARG1 - %1 = load <2 x i64>* @llvm_mips_subs_u_d_ARG2 + %0 = load <2 x i64>, <2 x i64>* @llvm_mips_subs_u_d_ARG1 + %1 = load <2 x i64>, <2 x i64>* @llvm_mips_subs_u_d_ARG2 %2 = tail call <2 x i64> @llvm.mips.subs.u.d(<2 x i64> %0, <2 x i64> %1) store <2 x i64> %2, <2 x i64>* @llvm_mips_subs_u_d_RES ret void @@ -1012,8 +1012,8 @@ declare <2 x i64> @llvm.mips.subs.u.d(<2 x i64>, <2 x i64>) nounwind define void @llvm_mips_subsus_u_b_test() nounwind { entry: - %0 = load <16 x i8>* @llvm_mips_subsus_u_b_ARG1 - %1 = load <16 x i8>* @llvm_mips_subsus_u_b_ARG2 + %0 = load <16 x i8>, <16 x i8>* @llvm_mips_subsus_u_b_ARG1 + %1 = load <16 x i8>, <16 x i8>* @llvm_mips_subsus_u_b_ARG2 %2 = tail call <16 x i8> @llvm.mips.subsus.u.b(<16 x i8> %0, <16 x i8> %1) store <16 x i8> %2, <16 x i8>* @llvm_mips_subsus_u_b_RES ret void @@ -1036,8 +1036,8 @@ declare <16 x i8> @llvm.mips.subsus.u.b(<16 x i8>, <16 x i8>) nounwind define void @llvm_mips_subsus_u_h_test() nounwind { entry: - %0 = load <8 x i16>* @llvm_mips_subsus_u_h_ARG1 - %1 = load <8 x i16>* @llvm_mips_subsus_u_h_ARG2 + %0 = load <8 x i16>, <8 x i16>* @llvm_mips_subsus_u_h_ARG1 + %1 = load <8 x i16>, <8 x i16>* @llvm_mips_subsus_u_h_ARG2 %2 = tail call <8 x i16> @llvm.mips.subsus.u.h(<8 x i16> %0, <8 x i16> %1) store <8 x i16> %2, <8 x i16>* @llvm_mips_subsus_u_h_RES ret void @@ -1060,8 +1060,8 @@ declare <8 x i16> @llvm.mips.subsus.u.h(<8 x i16>, <8 x i16>) nounwind define void @llvm_mips_subsus_u_w_test() nounwind { entry: - %0 = load <4 x i32>* @llvm_mips_subsus_u_w_ARG1 - %1 = load <4 x i32>* @llvm_mips_subsus_u_w_ARG2 + %0 = load <4 x i32>, <4 x i32>* @llvm_mips_subsus_u_w_ARG1 + %1 = load <4 x i32>, <4 x i32>* @llvm_mips_subsus_u_w_ARG2 %2 = tail call <4 x i32> @llvm.mips.subsus.u.w(<4 x i32> %0, <4 x i32> %1) store <4 x i32> %2, <4 x i32>* @llvm_mips_subsus_u_w_RES ret void @@ -1084,8 +1084,8 @@ declare <4 x i32> @llvm.mips.subsus.u.w(<4 x i32>, <4 x i32>) nounwind define void @llvm_mips_subsus_u_d_test() nounwind { entry: - %0 = load <2 x i64>* @llvm_mips_subsus_u_d_ARG1 - %1 = load <2 x i64>* @llvm_mips_subsus_u_d_ARG2 + %0 = load <2 x i64>, <2 x i64>* @llvm_mips_subsus_u_d_ARG1 + %1 = load <2 x i64>, <2 x i64>* @llvm_mips_subsus_u_d_ARG2 %2 = tail call <2 x i64> @llvm.mips.subsus.u.d(<2 x i64> %0, <2 x i64> %1) store <2 x i64> %2, <2 x i64>* @llvm_mips_subsus_u_d_RES ret void @@ -1108,8 +1108,8 @@ declare <2 x i64> @llvm.mips.subsus.u.d(<2 x i64>, <2 x i64>) nounwind define void @llvm_mips_subsuu_s_b_test() nounwind { entry: - %0 = load <16 x i8>* @llvm_mips_subsuu_s_b_ARG1 - %1 = load <16 x i8>* @llvm_mips_subsuu_s_b_ARG2 + %0 = load <16 x i8>, <16 x i8>* @llvm_mips_subsuu_s_b_ARG1 + %1 = load <16 x i8>, <16 x i8>* @llvm_mips_subsuu_s_b_ARG2 %2 = tail call <16 x i8> @llvm.mips.subsuu.s.b(<16 x i8> %0, <16 x i8> %1) store <16 x i8> %2, <16 x i8>* @llvm_mips_subsuu_s_b_RES ret void @@ -1132,8 +1132,8 @@ declare <16 x i8> @llvm.mips.subsuu.s.b(<16 x i8>, <16 x i8>) nounwind define void @llvm_mips_subsuu_s_h_test() nounwind { entry: - %0 = load <8 x i16>* @llvm_mips_subsuu_s_h_ARG1 - %1 = load <8 x i16>* @llvm_mips_subsuu_s_h_ARG2 + %0 = load <8 x i16>, <8 x i16>* @llvm_mips_subsuu_s_h_ARG1 + %1 = load <8 x i16>, <8 x i16>* @llvm_mips_subsuu_s_h_ARG2 %2 = tail call <8 x i16> @llvm.mips.subsuu.s.h(<8 x i16> %0, <8 x i16> %1) store <8 x i16> %2, <8 x i16>* @llvm_mips_subsuu_s_h_RES ret void @@ -1156,8 +1156,8 @@ declare <8 x i16> @llvm.mips.subsuu.s.h(<8 x i16>, <8 x i16>) nounwind define void @llvm_mips_subsuu_s_w_test() nounwind { entry: - %0 = load <4 x i32>* @llvm_mips_subsuu_s_w_ARG1 - %1 = load <4 x i32>* @llvm_mips_subsuu_s_w_ARG2 + %0 = load <4 x i32>, <4 x i32>* @llvm_mips_subsuu_s_w_ARG1 + %1 = load <4 x i32>, <4 x i32>* @llvm_mips_subsuu_s_w_ARG2 %2 = tail call <4 x i32> @llvm.mips.subsuu.s.w(<4 x i32> %0, <4 x i32> %1) store <4 x i32> %2, <4 x i32>* @llvm_mips_subsuu_s_w_RES ret void @@ -1180,8 +1180,8 @@ declare <4 x i32> @llvm.mips.subsuu.s.w(<4 x i32>, <4 x i32>) nounwind define void @llvm_mips_subsuu_s_d_test() nounwind { entry: - %0 = load <2 x i64>* @llvm_mips_subsuu_s_d_ARG1 - %1 = load <2 x i64>* @llvm_mips_subsuu_s_d_ARG2 + %0 = load <2 x i64>, <2 x i64>* @llvm_mips_subsuu_s_d_ARG1 + %1 = load <2 x i64>, <2 x i64>* @llvm_mips_subsuu_s_d_ARG2 %2 = tail call <2 x i64> @llvm.mips.subsuu.s.d(<2 x i64> %0, <2 x i64> %1) store <2 x i64> %2, <2 x i64>* @llvm_mips_subsuu_s_d_RES ret void @@ -1204,8 +1204,8 @@ declare <2 x i64> @llvm.mips.subsuu.s.d(<2 x i64>, <2 x i64>) nounwind define void @llvm_mips_subv_b_test() nounwind { entry: - %0 = load <16 x i8>* @llvm_mips_subv_b_ARG1 - %1 = load <16 x i8>* @llvm_mips_subv_b_ARG2 + %0 = load <16 x i8>, <16 x i8>* @llvm_mips_subv_b_ARG1 + %1 = load <16 x i8>, <16 x i8>* @llvm_mips_subv_b_ARG2 %2 = tail call <16 x i8> @llvm.mips.subv.b(<16 x i8> %0, <16 x i8> %1) store <16 x i8> %2, <16 x i8>* @llvm_mips_subv_b_RES ret void @@ -1228,8 +1228,8 @@ declare <16 x i8> @llvm.mips.subv.b(<16 x i8>, <16 x i8>) nounwind define void @llvm_mips_subv_h_test() nounwind { entry: - %0 = load <8 x i16>* @llvm_mips_subv_h_ARG1 - %1 = load <8 x i16>* @llvm_mips_subv_h_ARG2 + %0 = load <8 x i16>, <8 x i16>* @llvm_mips_subv_h_ARG1 + %1 = load <8 x i16>, <8 x i16>* @llvm_mips_subv_h_ARG2 %2 = tail call <8 x i16> @llvm.mips.subv.h(<8 x i16> %0, <8 x i16> %1) store <8 x i16> %2, <8 x i16>* @llvm_mips_subv_h_RES ret void @@ -1252,8 +1252,8 @@ declare <8 x i16> @llvm.mips.subv.h(<8 x i16>, <8 x i16>) nounwind define void @llvm_mips_subv_w_test() nounwind { entry: - %0 = load <4 x i32>* @llvm_mips_subv_w_ARG1 - %1 = load <4 x i32>* @llvm_mips_subv_w_ARG2 + %0 = load <4 x i32>, <4 x i32>* @llvm_mips_subv_w_ARG1 + %1 = load <4 x i32>, <4 x i32>* @llvm_mips_subv_w_ARG2 %2 = tail call <4 x i32> @llvm.mips.subv.w(<4 x i32> %0, <4 x i32> %1) store <4 x i32> %2, <4 x i32>* @llvm_mips_subv_w_RES ret void @@ -1276,8 +1276,8 @@ declare <4 x i32> @llvm.mips.subv.w(<4 x i32>, <4 x i32>) nounwind define void @llvm_mips_subv_d_test() nounwind { entry: - %0 = load <2 x i64>* @llvm_mips_subv_d_ARG1 - %1 = load <2 x i64>* @llvm_mips_subv_d_ARG2 + %0 = load <2 x i64>, <2 x i64>* @llvm_mips_subv_d_ARG1 + %1 = load <2 x i64>, <2 x i64>* @llvm_mips_subv_d_ARG2 %2 = tail call <2 x i64> @llvm.mips.subv.d(<2 x i64> %0, <2 x i64> %1) store <2 x i64> %2, <2 x i64>* @llvm_mips_subv_d_RES ret void @@ -1297,8 +1297,8 @@ declare <2 x i64> @llvm.mips.subv.d(<2 x i64>, <2 x i64>) nounwind define void @subv_b_test() nounwind { entry: - %0 = load <16 x i8>* @llvm_mips_subv_b_ARG1 - %1 = load <16 x i8>* @llvm_mips_subv_b_ARG2 + %0 = load <16 x i8>, <16 x i8>* @llvm_mips_subv_b_ARG1 + %1 = load <16 x i8>, <16 x i8>* @llvm_mips_subv_b_ARG2 %2 = sub <16 x i8> %0, %1 store <16 x i8> %2, <16 x i8>* @llvm_mips_subv_b_RES ret void @@ -1315,8 +1315,8 @@ entry: define void @subv_h_test() nounwind { entry: - %0 = load <8 x i16>* @llvm_mips_subv_h_ARG1 - %1 = load <8 x i16>* @llvm_mips_subv_h_ARG2 + %0 = load <8 x i16>, <8 x i16>* @llvm_mips_subv_h_ARG1 + %1 = load <8 x i16>, <8 x i16>* @llvm_mips_subv_h_ARG2 %2 = sub <8 x i16> %0, %1 store <8 x i16> %2, <8 x i16>* @llvm_mips_subv_h_RES ret void @@ -1333,8 +1333,8 @@ entry: define void @subv_w_test() nounwind { entry: - %0 = load <4 x i32>* @llvm_mips_subv_w_ARG1 - %1 = load <4 x i32>* @llvm_mips_subv_w_ARG2 + %0 = load <4 x i32>, <4 x i32>* @llvm_mips_subv_w_ARG1 + %1 = load <4 x i32>, <4 x i32>* @llvm_mips_subv_w_ARG2 %2 = sub <4 x i32> %0, %1 store <4 x i32> %2, <4 x i32>* @llvm_mips_subv_w_RES ret void @@ -1351,8 +1351,8 @@ entry: define void @subv_d_test() nounwind { entry: - %0 = load <2 x i64>* @llvm_mips_subv_d_ARG1 - %1 = load <2 x i64>* @llvm_mips_subv_d_ARG2 + %0 = load <2 x i64>, <2 x i64>* @llvm_mips_subv_d_ARG1 + %1 = load <2 x i64>, <2 x i64>* @llvm_mips_subv_d_ARG2 %2 = sub <2 x i64> %0, %1 store <2 x i64> %2, <2 x i64>* @llvm_mips_subv_d_RES ret void diff --git a/test/CodeGen/Mips/msa/3r-v.ll b/test/CodeGen/Mips/msa/3r-v.ll index c9693f90d556f..2d36da40d2b16 100644 --- a/test/CodeGen/Mips/msa/3r-v.ll +++ b/test/CodeGen/Mips/msa/3r-v.ll @@ -11,9 +11,9 @@ define void @llvm_mips_vshf_b_test() nounwind { entry: - %0 = load <16 x i8>* @llvm_mips_vshf_b_ARG1 - %1 = load <16 x i8>* @llvm_mips_vshf_b_ARG2 - %2 = load <16 x i8>* @llvm_mips_vshf_b_ARG3 + %0 = load <16 x i8>, <16 x i8>* @llvm_mips_vshf_b_ARG1 + %1 = load <16 x i8>, <16 x i8>* @llvm_mips_vshf_b_ARG2 + %2 = load <16 x i8>, <16 x i8>* @llvm_mips_vshf_b_ARG3 %3 = tail call <16 x i8> @llvm.mips.vshf.b(<16 x i8> %0, <16 x i8> %1, <16 x i8> %2) store <16 x i8> %3, <16 x i8>* @llvm_mips_vshf_b_RES ret void @@ -36,9 +36,9 @@ declare <16 x i8> @llvm.mips.vshf.b(<16 x i8>, <16 x i8>, <16 x i8>) nounwind define void @llvm_mips_vshf_h_test() nounwind { entry: - %0 = load <8 x i16>* @llvm_mips_vshf_h_ARG1 - %1 = load <8 x i16>* @llvm_mips_vshf_h_ARG2 - %2 = load <8 x i16>* @llvm_mips_vshf_h_ARG3 + %0 = load <8 x i16>, <8 x i16>* @llvm_mips_vshf_h_ARG1 + %1 = load <8 x i16>, <8 x i16>* @llvm_mips_vshf_h_ARG2 + %2 = load <8 x i16>, <8 x i16>* @llvm_mips_vshf_h_ARG3 %3 = tail call <8 x i16> @llvm.mips.vshf.h(<8 x i16> %0, <8 x i16> %1, <8 x i16> %2) store <8 x i16> %3, <8 x i16>* @llvm_mips_vshf_h_RES ret void @@ -61,9 +61,9 @@ declare <8 x i16> @llvm.mips.vshf.h(<8 x i16>, <8 x i16>, <8 x i16>) nounwind define void @llvm_mips_vshf_w_test() nounwind { entry: - %0 = load <4 x i32>* @llvm_mips_vshf_w_ARG1 - %1 = load <4 x i32>* @llvm_mips_vshf_w_ARG2 - %2 = load <4 x i32>* @llvm_mips_vshf_w_ARG3 + %0 = load <4 x i32>, <4 x i32>* @llvm_mips_vshf_w_ARG1 + %1 = load <4 x i32>, <4 x i32>* @llvm_mips_vshf_w_ARG2 + %2 = load <4 x i32>, <4 x i32>* @llvm_mips_vshf_w_ARG3 %3 = tail call <4 x i32> @llvm.mips.vshf.w(<4 x i32> %0, <4 x i32> %1, <4 x i32> %2) store <4 x i32> %3, <4 x i32>* @llvm_mips_vshf_w_RES ret void @@ -86,9 +86,9 @@ declare <4 x i32> @llvm.mips.vshf.w(<4 x i32>, <4 x i32>, <4 x i32>) nounwind define void @llvm_mips_vshf_d_test() nounwind { entry: - %0 = load <2 x i64>* @llvm_mips_vshf_d_ARG1 - %1 = load <2 x i64>* @llvm_mips_vshf_d_ARG2 - %2 = load <2 x i64>* @llvm_mips_vshf_d_ARG3 + %0 = load <2 x i64>, <2 x i64>* @llvm_mips_vshf_d_ARG1 + %1 = load <2 x i64>, <2 x i64>* @llvm_mips_vshf_d_ARG2 + %2 = load <2 x i64>, <2 x i64>* @llvm_mips_vshf_d_ARG3 %3 = tail call <2 x i64> @llvm.mips.vshf.d(<2 x i64> %0, <2 x i64> %1, <2 x i64> %2) store <2 x i64> %3, <2 x i64>* @llvm_mips_vshf_d_RES ret void diff --git a/test/CodeGen/Mips/msa/3r_4r.ll b/test/CodeGen/Mips/msa/3r_4r.ll index b7fd7283788c6..73d104c68c537 100644 --- a/test/CodeGen/Mips/msa/3r_4r.ll +++ b/test/CodeGen/Mips/msa/3r_4r.ll @@ -11,9 +11,9 @@ define void @llvm_mips_maddv_b_test() nounwind { entry: - %0 = load <16 x i8>* @llvm_mips_maddv_b_ARG1 - %1 = load <16 x i8>* @llvm_mips_maddv_b_ARG2 - %2 = load <16 x i8>* @llvm_mips_maddv_b_ARG3 + %0 = load <16 x i8>, <16 x i8>* @llvm_mips_maddv_b_ARG1 + %1 = load <16 x i8>, <16 x i8>* @llvm_mips_maddv_b_ARG2 + %2 = load <16 x i8>, <16 x i8>* @llvm_mips_maddv_b_ARG3 %3 = tail call <16 x i8> @llvm.mips.maddv.b(<16 x i8> %0, <16 x i8> %1, <16 x i8> %2) store <16 x i8> %3, <16 x i8>* @llvm_mips_maddv_b_RES ret void @@ -36,9 +36,9 @@ declare <16 x i8> @llvm.mips.maddv.b(<16 x i8>, <16 x i8>, <16 x i8>) nounwind define void @llvm_mips_maddv_h_test() nounwind { entry: - %0 = load <8 x i16>* @llvm_mips_maddv_h_ARG1 - %1 = load <8 x i16>* @llvm_mips_maddv_h_ARG2 - %2 = load <8 x i16>* @llvm_mips_maddv_h_ARG3 + %0 = load <8 x i16>, <8 x i16>* @llvm_mips_maddv_h_ARG1 + %1 = load <8 x i16>, <8 x i16>* @llvm_mips_maddv_h_ARG2 + %2 = load <8 x i16>, <8 x i16>* @llvm_mips_maddv_h_ARG3 %3 = tail call <8 x i16> @llvm.mips.maddv.h(<8 x i16> %0, <8 x i16> %1, <8 x i16> %2) store <8 x i16> %3, <8 x i16>* @llvm_mips_maddv_h_RES ret void @@ -61,9 +61,9 @@ declare <8 x i16> @llvm.mips.maddv.h(<8 x i16>, <8 x i16>, <8 x i16>) nounwind define void @llvm_mips_maddv_w_test() nounwind { entry: - %0 = load <4 x i32>* @llvm_mips_maddv_w_ARG1 - %1 = load <4 x i32>* @llvm_mips_maddv_w_ARG2 - %2 = load <4 x i32>* @llvm_mips_maddv_w_ARG3 + %0 = load <4 x i32>, <4 x i32>* @llvm_mips_maddv_w_ARG1 + %1 = load <4 x i32>, <4 x i32>* @llvm_mips_maddv_w_ARG2 + %2 = load <4 x i32>, <4 x i32>* @llvm_mips_maddv_w_ARG3 %3 = tail call <4 x i32> @llvm.mips.maddv.w(<4 x i32> %0, <4 x i32> %1, <4 x i32> %2) store <4 x i32> %3, <4 x i32>* @llvm_mips_maddv_w_RES ret void @@ -86,9 +86,9 @@ declare <4 x i32> @llvm.mips.maddv.w(<4 x i32>, <4 x i32>, <4 x i32>) nounwind define void @llvm_mips_maddv_d_test() nounwind { entry: - %0 = load <2 x i64>* @llvm_mips_maddv_d_ARG1 - %1 = load <2 x i64>* @llvm_mips_maddv_d_ARG2 - %2 = load <2 x i64>* @llvm_mips_maddv_d_ARG3 + %0 = load <2 x i64>, <2 x i64>* @llvm_mips_maddv_d_ARG1 + %1 = load <2 x i64>, <2 x i64>* @llvm_mips_maddv_d_ARG2 + %2 = load <2 x i64>, <2 x i64>* @llvm_mips_maddv_d_ARG3 %3 = tail call <2 x i64> @llvm.mips.maddv.d(<2 x i64> %0, <2 x i64> %1, <2 x i64> %2) store <2 x i64> %3, <2 x i64>* @llvm_mips_maddv_d_RES ret void @@ -111,9 +111,9 @@ declare <2 x i64> @llvm.mips.maddv.d(<2 x i64>, <2 x i64>, <2 x i64>) nounwind define void @llvm_mips_msubv_b_test() nounwind { entry: - %0 = load <16 x i8>* @llvm_mips_msubv_b_ARG1 - %1 = load <16 x i8>* @llvm_mips_msubv_b_ARG2 - %2 = load <16 x i8>* @llvm_mips_msubv_b_ARG3 + %0 = load <16 x i8>, <16 x i8>* @llvm_mips_msubv_b_ARG1 + %1 = load <16 x i8>, <16 x i8>* @llvm_mips_msubv_b_ARG2 + %2 = load <16 x i8>, <16 x i8>* @llvm_mips_msubv_b_ARG3 %3 = tail call <16 x i8> @llvm.mips.msubv.b(<16 x i8> %0, <16 x i8> %1, <16 x i8> %2) store <16 x i8> %3, <16 x i8>* @llvm_mips_msubv_b_RES ret void @@ -136,9 +136,9 @@ declare <16 x i8> @llvm.mips.msubv.b(<16 x i8>, <16 x i8>, <16 x i8>) nounwind define void @llvm_mips_msubv_h_test() nounwind { entry: - %0 = load <8 x i16>* @llvm_mips_msubv_h_ARG1 - %1 = load <8 x i16>* @llvm_mips_msubv_h_ARG2 - %2 = load <8 x i16>* @llvm_mips_msubv_h_ARG3 + %0 = load <8 x i16>, <8 x i16>* @llvm_mips_msubv_h_ARG1 + %1 = load <8 x i16>, <8 x i16>* @llvm_mips_msubv_h_ARG2 + %2 = load <8 x i16>, <8 x i16>* @llvm_mips_msubv_h_ARG3 %3 = tail call <8 x i16> @llvm.mips.msubv.h(<8 x i16> %0, <8 x i16> %1, <8 x i16> %2) store <8 x i16> %3, <8 x i16>* @llvm_mips_msubv_h_RES ret void @@ -161,9 +161,9 @@ declare <8 x i16> @llvm.mips.msubv.h(<8 x i16>, <8 x i16>, <8 x i16>) nounwind define void @llvm_mips_msubv_w_test() nounwind { entry: - %0 = load <4 x i32>* @llvm_mips_msubv_w_ARG1 - %1 = load <4 x i32>* @llvm_mips_msubv_w_ARG2 - %2 = load <4 x i32>* @llvm_mips_msubv_w_ARG3 + %0 = load <4 x i32>, <4 x i32>* @llvm_mips_msubv_w_ARG1 + %1 = load <4 x i32>, <4 x i32>* @llvm_mips_msubv_w_ARG2 + %2 = load <4 x i32>, <4 x i32>* @llvm_mips_msubv_w_ARG3 %3 = tail call <4 x i32> @llvm.mips.msubv.w(<4 x i32> %0, <4 x i32> %1, <4 x i32> %2) store <4 x i32> %3, <4 x i32>* @llvm_mips_msubv_w_RES ret void @@ -186,9 +186,9 @@ declare <4 x i32> @llvm.mips.msubv.w(<4 x i32>, <4 x i32>, <4 x i32>) nounwind define void @llvm_mips_msubv_d_test() nounwind { entry: - %0 = load <2 x i64>* @llvm_mips_msubv_d_ARG1 - %1 = load <2 x i64>* @llvm_mips_msubv_d_ARG2 - %2 = load <2 x i64>* @llvm_mips_msubv_d_ARG3 + %0 = load <2 x i64>, <2 x i64>* @llvm_mips_msubv_d_ARG1 + %1 = load <2 x i64>, <2 x i64>* @llvm_mips_msubv_d_ARG2 + %2 = load <2 x i64>, <2 x i64>* @llvm_mips_msubv_d_ARG3 %3 = tail call <2 x i64> @llvm.mips.msubv.d(<2 x i64> %0, <2 x i64> %1, <2 x i64> %2) store <2 x i64> %3, <2 x i64>* @llvm_mips_msubv_d_RES ret void diff --git a/test/CodeGen/Mips/msa/3r_4r_widen.ll b/test/CodeGen/Mips/msa/3r_4r_widen.ll index 7063e4566a786..fe248eeb566b3 100644 --- a/test/CodeGen/Mips/msa/3r_4r_widen.ll +++ b/test/CodeGen/Mips/msa/3r_4r_widen.ll @@ -12,9 +12,9 @@ define void @llvm_mips_dpadd_s_h_test() nounwind { entry: - %0 = load <8 x i16>* @llvm_mips_dpadd_s_h_ARG1 - %1 = load <16 x i8>* @llvm_mips_dpadd_s_h_ARG2 - %2 = load <16 x i8>* @llvm_mips_dpadd_s_h_ARG3 + %0 = load <8 x i16>, <8 x i16>* @llvm_mips_dpadd_s_h_ARG1 + %1 = load <16 x i8>, <16 x i8>* @llvm_mips_dpadd_s_h_ARG2 + %2 = load <16 x i8>, <16 x i8>* @llvm_mips_dpadd_s_h_ARG3 %3 = tail call <8 x i16> @llvm.mips.dpadd.s.h(<8 x i16> %0, <16 x i8> %1, <16 x i8> %2) store <8 x i16> %3, <8 x i16>* @llvm_mips_dpadd_s_h_RES ret void @@ -37,9 +37,9 @@ declare <8 x i16> @llvm.mips.dpadd.s.h(<8 x i16>, <16 x i8>, <16 x i8>) nounwind define void @llvm_mips_dpadd_s_w_test() nounwind { entry: - %0 = load <4 x i32>* @llvm_mips_dpadd_s_w_ARG1 - %1 = load <8 x i16>* @llvm_mips_dpadd_s_w_ARG2 - %2 = load <8 x i16>* @llvm_mips_dpadd_s_w_ARG3 + %0 = load <4 x i32>, <4 x i32>* @llvm_mips_dpadd_s_w_ARG1 + %1 = load <8 x i16>, <8 x i16>* @llvm_mips_dpadd_s_w_ARG2 + %2 = load <8 x i16>, <8 x i16>* @llvm_mips_dpadd_s_w_ARG3 %3 = tail call <4 x i32> @llvm.mips.dpadd.s.w(<4 x i32> %0, <8 x i16> %1, <8 x i16> %2) store <4 x i32> %3, <4 x i32>* @llvm_mips_dpadd_s_w_RES ret void @@ -62,9 +62,9 @@ declare <4 x i32> @llvm.mips.dpadd.s.w(<4 x i32>, <8 x i16>, <8 x i16>) nounwind define void @llvm_mips_dpadd_s_d_test() nounwind { entry: - %0 = load <2 x i64>* @llvm_mips_dpadd_s_d_ARG1 - %1 = load <4 x i32>* @llvm_mips_dpadd_s_d_ARG2 - %2 = load <4 x i32>* @llvm_mips_dpadd_s_d_ARG3 + %0 = load <2 x i64>, <2 x i64>* @llvm_mips_dpadd_s_d_ARG1 + %1 = load <4 x i32>, <4 x i32>* @llvm_mips_dpadd_s_d_ARG2 + %2 = load <4 x i32>, <4 x i32>* @llvm_mips_dpadd_s_d_ARG3 %3 = tail call <2 x i64> @llvm.mips.dpadd.s.d(<2 x i64> %0, <4 x i32> %1, <4 x i32> %2) store <2 x i64> %3, <2 x i64>* @llvm_mips_dpadd_s_d_RES ret void @@ -87,9 +87,9 @@ declare <2 x i64> @llvm.mips.dpadd.s.d(<2 x i64>, <4 x i32>, <4 x i32>) nounwind define void @llvm_mips_dpadd_u_h_test() nounwind { entry: - %0 = load <8 x i16>* @llvm_mips_dpadd_u_h_ARG1 - %1 = load <16 x i8>* @llvm_mips_dpadd_u_h_ARG2 - %2 = load <16 x i8>* @llvm_mips_dpadd_u_h_ARG3 + %0 = load <8 x i16>, <8 x i16>* @llvm_mips_dpadd_u_h_ARG1 + %1 = load <16 x i8>, <16 x i8>* @llvm_mips_dpadd_u_h_ARG2 + %2 = load <16 x i8>, <16 x i8>* @llvm_mips_dpadd_u_h_ARG3 %3 = tail call <8 x i16> @llvm.mips.dpadd.u.h(<8 x i16> %0, <16 x i8> %1, <16 x i8> %2) store <8 x i16> %3, <8 x i16>* @llvm_mips_dpadd_u_h_RES ret void @@ -112,9 +112,9 @@ declare <8 x i16> @llvm.mips.dpadd.u.h(<8 x i16>, <16 x i8>, <16 x i8>) nounwind define void @llvm_mips_dpadd_u_w_test() nounwind { entry: - %0 = load <4 x i32>* @llvm_mips_dpadd_u_w_ARG1 - %1 = load <8 x i16>* @llvm_mips_dpadd_u_w_ARG2 - %2 = load <8 x i16>* @llvm_mips_dpadd_u_w_ARG3 + %0 = load <4 x i32>, <4 x i32>* @llvm_mips_dpadd_u_w_ARG1 + %1 = load <8 x i16>, <8 x i16>* @llvm_mips_dpadd_u_w_ARG2 + %2 = load <8 x i16>, <8 x i16>* @llvm_mips_dpadd_u_w_ARG3 %3 = tail call <4 x i32> @llvm.mips.dpadd.u.w(<4 x i32> %0, <8 x i16> %1, <8 x i16> %2) store <4 x i32> %3, <4 x i32>* @llvm_mips_dpadd_u_w_RES ret void @@ -137,9 +137,9 @@ declare <4 x i32> @llvm.mips.dpadd.u.w(<4 x i32>, <8 x i16>, <8 x i16>) nounwind define void @llvm_mips_dpadd_u_d_test() nounwind { entry: - %0 = load <2 x i64>* @llvm_mips_dpadd_u_d_ARG1 - %1 = load <4 x i32>* @llvm_mips_dpadd_u_d_ARG2 - %2 = load <4 x i32>* @llvm_mips_dpadd_u_d_ARG3 + %0 = load <2 x i64>, <2 x i64>* @llvm_mips_dpadd_u_d_ARG1 + %1 = load <4 x i32>, <4 x i32>* @llvm_mips_dpadd_u_d_ARG2 + %2 = load <4 x i32>, <4 x i32>* @llvm_mips_dpadd_u_d_ARG3 %3 = tail call <2 x i64> @llvm.mips.dpadd.u.d(<2 x i64> %0, <4 x i32> %1, <4 x i32> %2) store <2 x i64> %3, <2 x i64>* @llvm_mips_dpadd_u_d_RES ret void @@ -162,9 +162,9 @@ declare <2 x i64> @llvm.mips.dpadd.u.d(<2 x i64>, <4 x i32>, <4 x i32>) nounwind define void @llvm_mips_dpsub_s_h_test() nounwind { entry: - %0 = load <8 x i16>* @llvm_mips_dpsub_s_h_ARG1 - %1 = load <16 x i8>* @llvm_mips_dpsub_s_h_ARG2 - %2 = load <16 x i8>* @llvm_mips_dpsub_s_h_ARG3 + %0 = load <8 x i16>, <8 x i16>* @llvm_mips_dpsub_s_h_ARG1 + %1 = load <16 x i8>, <16 x i8>* @llvm_mips_dpsub_s_h_ARG2 + %2 = load <16 x i8>, <16 x i8>* @llvm_mips_dpsub_s_h_ARG3 %3 = tail call <8 x i16> @llvm.mips.dpsub.s.h(<8 x i16> %0, <16 x i8> %1, <16 x i8> %2) store <8 x i16> %3, <8 x i16>* @llvm_mips_dpsub_s_h_RES ret void @@ -187,9 +187,9 @@ declare <8 x i16> @llvm.mips.dpsub.s.h(<8 x i16>, <16 x i8>, <16 x i8>) nounwind define void @llvm_mips_dpsub_s_w_test() nounwind { entry: - %0 = load <4 x i32>* @llvm_mips_dpsub_s_w_ARG1 - %1 = load <8 x i16>* @llvm_mips_dpsub_s_w_ARG2 - %2 = load <8 x i16>* @llvm_mips_dpsub_s_w_ARG3 + %0 = load <4 x i32>, <4 x i32>* @llvm_mips_dpsub_s_w_ARG1 + %1 = load <8 x i16>, <8 x i16>* @llvm_mips_dpsub_s_w_ARG2 + %2 = load <8 x i16>, <8 x i16>* @llvm_mips_dpsub_s_w_ARG3 %3 = tail call <4 x i32> @llvm.mips.dpsub.s.w(<4 x i32> %0, <8 x i16> %1, <8 x i16> %2) store <4 x i32> %3, <4 x i32>* @llvm_mips_dpsub_s_w_RES ret void @@ -212,9 +212,9 @@ declare <4 x i32> @llvm.mips.dpsub.s.w(<4 x i32>, <8 x i16>, <8 x i16>) nounwind define void @llvm_mips_dpsub_s_d_test() nounwind { entry: - %0 = load <2 x i64>* @llvm_mips_dpsub_s_d_ARG1 - %1 = load <4 x i32>* @llvm_mips_dpsub_s_d_ARG2 - %2 = load <4 x i32>* @llvm_mips_dpsub_s_d_ARG3 + %0 = load <2 x i64>, <2 x i64>* @llvm_mips_dpsub_s_d_ARG1 + %1 = load <4 x i32>, <4 x i32>* @llvm_mips_dpsub_s_d_ARG2 + %2 = load <4 x i32>, <4 x i32>* @llvm_mips_dpsub_s_d_ARG3 %3 = tail call <2 x i64> @llvm.mips.dpsub.s.d(<2 x i64> %0, <4 x i32> %1, <4 x i32> %2) store <2 x i64> %3, <2 x i64>* @llvm_mips_dpsub_s_d_RES ret void @@ -237,9 +237,9 @@ declare <2 x i64> @llvm.mips.dpsub.s.d(<2 x i64>, <4 x i32>, <4 x i32>) nounwind define void @llvm_mips_dpsub_u_h_test() nounwind { entry: - %0 = load <8 x i16>* @llvm_mips_dpsub_u_h_ARG1 - %1 = load <16 x i8>* @llvm_mips_dpsub_u_h_ARG2 - %2 = load <16 x i8>* @llvm_mips_dpsub_u_h_ARG3 + %0 = load <8 x i16>, <8 x i16>* @llvm_mips_dpsub_u_h_ARG1 + %1 = load <16 x i8>, <16 x i8>* @llvm_mips_dpsub_u_h_ARG2 + %2 = load <16 x i8>, <16 x i8>* @llvm_mips_dpsub_u_h_ARG3 %3 = tail call <8 x i16> @llvm.mips.dpsub.u.h(<8 x i16> %0, <16 x i8> %1, <16 x i8> %2) store <8 x i16> %3, <8 x i16>* @llvm_mips_dpsub_u_h_RES ret void @@ -262,9 +262,9 @@ declare <8 x i16> @llvm.mips.dpsub.u.h(<8 x i16>, <16 x i8>, <16 x i8>) nounwind define void @llvm_mips_dpsub_u_w_test() nounwind { entry: - %0 = load <4 x i32>* @llvm_mips_dpsub_u_w_ARG1 - %1 = load <8 x i16>* @llvm_mips_dpsub_u_w_ARG2 - %2 = load <8 x i16>* @llvm_mips_dpsub_u_w_ARG3 + %0 = load <4 x i32>, <4 x i32>* @llvm_mips_dpsub_u_w_ARG1 + %1 = load <8 x i16>, <8 x i16>* @llvm_mips_dpsub_u_w_ARG2 + %2 = load <8 x i16>, <8 x i16>* @llvm_mips_dpsub_u_w_ARG3 %3 = tail call <4 x i32> @llvm.mips.dpsub.u.w(<4 x i32> %0, <8 x i16> %1, <8 x i16> %2) store <4 x i32> %3, <4 x i32>* @llvm_mips_dpsub_u_w_RES ret void @@ -287,9 +287,9 @@ declare <4 x i32> @llvm.mips.dpsub.u.w(<4 x i32>, <8 x i16>, <8 x i16>) nounwind define void @llvm_mips_dpsub_u_d_test() nounwind { entry: - %0 = load <2 x i64>* @llvm_mips_dpsub_u_d_ARG1 - %1 = load <4 x i32>* @llvm_mips_dpsub_u_d_ARG2 - %2 = load <4 x i32>* @llvm_mips_dpsub_u_d_ARG3 + %0 = load <2 x i64>, <2 x i64>* @llvm_mips_dpsub_u_d_ARG1 + %1 = load <4 x i32>, <4 x i32>* @llvm_mips_dpsub_u_d_ARG2 + %2 = load <4 x i32>, <4 x i32>* @llvm_mips_dpsub_u_d_ARG3 %3 = tail call <2 x i64> @llvm.mips.dpsub.u.d(<2 x i64> %0, <4 x i32> %1, <4 x i32> %2) store <2 x i64> %3, <2 x i64>* @llvm_mips_dpsub_u_d_RES ret void diff --git a/test/CodeGen/Mips/msa/3r_splat.ll b/test/CodeGen/Mips/msa/3r_splat.ll index 6b0cb26f8c818..56d26b030de9a 100644 --- a/test/CodeGen/Mips/msa/3r_splat.ll +++ b/test/CodeGen/Mips/msa/3r_splat.ll @@ -11,7 +11,7 @@ define void @llvm_mips_splat_b_test(i32 %a) nounwind { entry: - %0 = load <16 x i8>* @llvm_mips_splat_b_ARG1 + %0 = load <16 x i8>, <16 x i8>* @llvm_mips_splat_b_ARG1 %1 = tail call <16 x i8> @llvm.mips.splat.b(<16 x i8> %0, i32 %a) store <16 x i8> %1, <16 x i8>* @llvm_mips_splat_b_RES ret void @@ -32,7 +32,7 @@ declare <16 x i8> @llvm.mips.splat.b(<16 x i8>, i32) nounwind define void @llvm_mips_splat_h_test(i32 %a) nounwind { entry: - %0 = load <8 x i16>* @llvm_mips_splat_h_ARG1 + %0 = load <8 x i16>, <8 x i16>* @llvm_mips_splat_h_ARG1 %1 = tail call <8 x i16> @llvm.mips.splat.h(<8 x i16> %0, i32 %a) store <8 x i16> %1, <8 x i16>* @llvm_mips_splat_h_RES ret void @@ -53,7 +53,7 @@ declare <8 x i16> @llvm.mips.splat.h(<8 x i16>, i32) nounwind define void @llvm_mips_splat_w_test(i32 %a) nounwind { entry: - %0 = load <4 x i32>* @llvm_mips_splat_w_ARG1 + %0 = load <4 x i32>, <4 x i32>* @llvm_mips_splat_w_ARG1 %1 = tail call <4 x i32> @llvm.mips.splat.w(<4 x i32> %0, i32 %a) store <4 x i32> %1, <4 x i32>* @llvm_mips_splat_w_RES ret void @@ -74,7 +74,7 @@ declare <4 x i32> @llvm.mips.splat.w(<4 x i32>, i32) nounwind define void @llvm_mips_splat_d_test(i32 %a) nounwind { entry: - %0 = load <2 x i64>* @llvm_mips_splat_d_ARG1 + %0 = load <2 x i64>, <2 x i64>* @llvm_mips_splat_d_ARG1 %1 = tail call <2 x i64> @llvm.mips.splat.d(<2 x i64> %0, i32 %a) store <2 x i64> %1, <2 x i64>* @llvm_mips_splat_d_RES ret void diff --git a/test/CodeGen/Mips/msa/3rf.ll b/test/CodeGen/Mips/msa/3rf.ll index ae665afcc9509..dce0c275e8da3 100644 --- a/test/CodeGen/Mips/msa/3rf.ll +++ b/test/CodeGen/Mips/msa/3rf.ll @@ -9,8 +9,8 @@ define void @llvm_mips_fadd_w_test() nounwind { entry: - %0 = load <4 x float>* @llvm_mips_fadd_w_ARG1 - %1 = load <4 x float>* @llvm_mips_fadd_w_ARG2 + %0 = load <4 x float>, <4 x float>* @llvm_mips_fadd_w_ARG1 + %1 = load <4 x float>, <4 x float>* @llvm_mips_fadd_w_ARG2 %2 = tail call <4 x float> @llvm.mips.fadd.w(<4 x float> %0, <4 x float> %1) store <4 x float> %2, <4 x float>* @llvm_mips_fadd_w_RES ret void @@ -31,8 +31,8 @@ declare <4 x float> @llvm.mips.fadd.w(<4 x float>, <4 x float>) nounwind define void @llvm_mips_fadd_d_test() nounwind { entry: - %0 = load <2 x double>* @llvm_mips_fadd_d_ARG1 - %1 = load <2 x double>* @llvm_mips_fadd_d_ARG2 + %0 = load <2 x double>, <2 x double>* @llvm_mips_fadd_d_ARG1 + %1 = load <2 x double>, <2 x double>* @llvm_mips_fadd_d_ARG2 %2 = tail call <2 x double> @llvm.mips.fadd.d(<2 x double> %0, <2 x double> %1) store <2 x double> %2, <2 x double>* @llvm_mips_fadd_d_RES ret void @@ -49,8 +49,8 @@ declare <2 x double> @llvm.mips.fadd.d(<2 x double>, <2 x double>) nounwind define void @fadd_w_test() nounwind { entry: - %0 = load <4 x float>* @llvm_mips_fadd_w_ARG1 - %1 = load <4 x float>* @llvm_mips_fadd_w_ARG2 + %0 = load <4 x float>, <4 x float>* @llvm_mips_fadd_w_ARG1 + %1 = load <4 x float>, <4 x float>* @llvm_mips_fadd_w_ARG2 %2 = fadd <4 x float> %0, %1 store <4 x float> %2, <4 x float>* @llvm_mips_fadd_w_RES ret void @@ -65,8 +65,8 @@ entry: define void @fadd_d_test() nounwind { entry: - %0 = load <2 x double>* @llvm_mips_fadd_d_ARG1 - %1 = load <2 x double>* @llvm_mips_fadd_d_ARG2 + %0 = load <2 x double>, <2 x double>* @llvm_mips_fadd_d_ARG1 + %1 = load <2 x double>, <2 x double>* @llvm_mips_fadd_d_ARG2 %2 = fadd <2 x double> %0, %1 store <2 x double> %2, <2 x double>* @llvm_mips_fadd_d_RES ret void @@ -85,8 +85,8 @@ entry: define void @llvm_mips_fdiv_w_test() nounwind { entry: - %0 = load <4 x float>* @llvm_mips_fdiv_w_ARG1 - %1 = load <4 x float>* @llvm_mips_fdiv_w_ARG2 + %0 = load <4 x float>, <4 x float>* @llvm_mips_fdiv_w_ARG1 + %1 = load <4 x float>, <4 x float>* @llvm_mips_fdiv_w_ARG2 %2 = tail call <4 x float> @llvm.mips.fdiv.w(<4 x float> %0, <4 x float> %1) store <4 x float> %2, <4 x float>* @llvm_mips_fdiv_w_RES ret void @@ -107,8 +107,8 @@ declare <4 x float> @llvm.mips.fdiv.w(<4 x float>, <4 x float>) nounwind define void @llvm_mips_fdiv_d_test() nounwind { entry: - %0 = load <2 x double>* @llvm_mips_fdiv_d_ARG1 - %1 = load <2 x double>* @llvm_mips_fdiv_d_ARG2 + %0 = load <2 x double>, <2 x double>* @llvm_mips_fdiv_d_ARG1 + %1 = load <2 x double>, <2 x double>* @llvm_mips_fdiv_d_ARG2 %2 = tail call <2 x double> @llvm.mips.fdiv.d(<2 x double> %0, <2 x double> %1) store <2 x double> %2, <2 x double>* @llvm_mips_fdiv_d_RES ret void @@ -125,8 +125,8 @@ declare <2 x double> @llvm.mips.fdiv.d(<2 x double>, <2 x double>) nounwind define void @fdiv_w_test() nounwind { entry: - %0 = load <4 x float>* @llvm_mips_fdiv_w_ARG1 - %1 = load <4 x float>* @llvm_mips_fdiv_w_ARG2 + %0 = load <4 x float>, <4 x float>* @llvm_mips_fdiv_w_ARG1 + %1 = load <4 x float>, <4 x float>* @llvm_mips_fdiv_w_ARG2 %2 = fdiv <4 x float> %0, %1 store <4 x float> %2, <4 x float>* @llvm_mips_fdiv_w_RES ret void @@ -141,8 +141,8 @@ entry: define void @fdiv_d_test() nounwind { entry: - %0 = load <2 x double>* @llvm_mips_fdiv_d_ARG1 - %1 = load <2 x double>* @llvm_mips_fdiv_d_ARG2 + %0 = load <2 x double>, <2 x double>* @llvm_mips_fdiv_d_ARG1 + %1 = load <2 x double>, <2 x double>* @llvm_mips_fdiv_d_ARG2 %2 = fdiv <2 x double> %0, %1 store <2 x double> %2, <2 x double>* @llvm_mips_fdiv_d_RES ret void @@ -161,8 +161,8 @@ entry: define void @llvm_mips_fmin_w_test() nounwind { entry: - %0 = load <4 x float>* @llvm_mips_fmin_w_ARG1 - %1 = load <4 x float>* @llvm_mips_fmin_w_ARG2 + %0 = load <4 x float>, <4 x float>* @llvm_mips_fmin_w_ARG1 + %1 = load <4 x float>, <4 x float>* @llvm_mips_fmin_w_ARG2 %2 = tail call <4 x float> @llvm.mips.fmin.w(<4 x float> %0, <4 x float> %1) store <4 x float> %2, <4 x float>* @llvm_mips_fmin_w_RES ret void @@ -183,8 +183,8 @@ declare <4 x float> @llvm.mips.fmin.w(<4 x float>, <4 x float>) nounwind define void @llvm_mips_fmin_d_test() nounwind { entry: - %0 = load <2 x double>* @llvm_mips_fmin_d_ARG1 - %1 = load <2 x double>* @llvm_mips_fmin_d_ARG2 + %0 = load <2 x double>, <2 x double>* @llvm_mips_fmin_d_ARG1 + %1 = load <2 x double>, <2 x double>* @llvm_mips_fmin_d_ARG2 %2 = tail call <2 x double> @llvm.mips.fmin.d(<2 x double> %0, <2 x double> %1) store <2 x double> %2, <2 x double>* @llvm_mips_fmin_d_RES ret void @@ -205,8 +205,8 @@ declare <2 x double> @llvm.mips.fmin.d(<2 x double>, <2 x double>) nounwind define void @llvm_mips_fmin_a_w_test() nounwind { entry: - %0 = load <4 x float>* @llvm_mips_fmin_a_w_ARG1 - %1 = load <4 x float>* @llvm_mips_fmin_a_w_ARG2 + %0 = load <4 x float>, <4 x float>* @llvm_mips_fmin_a_w_ARG1 + %1 = load <4 x float>, <4 x float>* @llvm_mips_fmin_a_w_ARG2 %2 = tail call <4 x float> @llvm.mips.fmin.a.w(<4 x float> %0, <4 x float> %1) store <4 x float> %2, <4 x float>* @llvm_mips_fmin_a_w_RES ret void @@ -227,8 +227,8 @@ declare <4 x float> @llvm.mips.fmin.a.w(<4 x float>, <4 x float>) nounwind define void @llvm_mips_fmin_a_d_test() nounwind { entry: - %0 = load <2 x double>* @llvm_mips_fmin_a_d_ARG1 - %1 = load <2 x double>* @llvm_mips_fmin_a_d_ARG2 + %0 = load <2 x double>, <2 x double>* @llvm_mips_fmin_a_d_ARG1 + %1 = load <2 x double>, <2 x double>* @llvm_mips_fmin_a_d_ARG2 %2 = tail call <2 x double> @llvm.mips.fmin.a.d(<2 x double> %0, <2 x double> %1) store <2 x double> %2, <2 x double>* @llvm_mips_fmin_a_d_RES ret void @@ -249,8 +249,8 @@ declare <2 x double> @llvm.mips.fmin.a.d(<2 x double>, <2 x double>) nounwind define void @llvm_mips_fmax_w_test() nounwind { entry: - %0 = load <4 x float>* @llvm_mips_fmax_w_ARG1 - %1 = load <4 x float>* @llvm_mips_fmax_w_ARG2 + %0 = load <4 x float>, <4 x float>* @llvm_mips_fmax_w_ARG1 + %1 = load <4 x float>, <4 x float>* @llvm_mips_fmax_w_ARG2 %2 = tail call <4 x float> @llvm.mips.fmax.w(<4 x float> %0, <4 x float> %1) store <4 x float> %2, <4 x float>* @llvm_mips_fmax_w_RES ret void @@ -271,8 +271,8 @@ declare <4 x float> @llvm.mips.fmax.w(<4 x float>, <4 x float>) nounwind define void @llvm_mips_fmax_d_test() nounwind { entry: - %0 = load <2 x double>* @llvm_mips_fmax_d_ARG1 - %1 = load <2 x double>* @llvm_mips_fmax_d_ARG2 + %0 = load <2 x double>, <2 x double>* @llvm_mips_fmax_d_ARG1 + %1 = load <2 x double>, <2 x double>* @llvm_mips_fmax_d_ARG2 %2 = tail call <2 x double> @llvm.mips.fmax.d(<2 x double> %0, <2 x double> %1) store <2 x double> %2, <2 x double>* @llvm_mips_fmax_d_RES ret void @@ -293,8 +293,8 @@ declare <2 x double> @llvm.mips.fmax.d(<2 x double>, <2 x double>) nounwind define void @llvm_mips_fmax_a_w_test() nounwind { entry: - %0 = load <4 x float>* @llvm_mips_fmax_a_w_ARG1 - %1 = load <4 x float>* @llvm_mips_fmax_a_w_ARG2 + %0 = load <4 x float>, <4 x float>* @llvm_mips_fmax_a_w_ARG1 + %1 = load <4 x float>, <4 x float>* @llvm_mips_fmax_a_w_ARG2 %2 = tail call <4 x float> @llvm.mips.fmax.a.w(<4 x float> %0, <4 x float> %1) store <4 x float> %2, <4 x float>* @llvm_mips_fmax_a_w_RES ret void @@ -315,8 +315,8 @@ declare <4 x float> @llvm.mips.fmax.a.w(<4 x float>, <4 x float>) nounwind define void @llvm_mips_fmax_a_d_test() nounwind { entry: - %0 = load <2 x double>* @llvm_mips_fmax_a_d_ARG1 - %1 = load <2 x double>* @llvm_mips_fmax_a_d_ARG2 + %0 = load <2 x double>, <2 x double>* @llvm_mips_fmax_a_d_ARG1 + %1 = load <2 x double>, <2 x double>* @llvm_mips_fmax_a_d_ARG2 %2 = tail call <2 x double> @llvm.mips.fmax.a.d(<2 x double> %0, <2 x double> %1) store <2 x double> %2, <2 x double>* @llvm_mips_fmax_a_d_RES ret void @@ -337,8 +337,8 @@ declare <2 x double> @llvm.mips.fmax.a.d(<2 x double>, <2 x double>) nounwind define void @llvm_mips_fmul_w_test() nounwind { entry: - %0 = load <4 x float>* @llvm_mips_fmul_w_ARG1 - %1 = load <4 x float>* @llvm_mips_fmul_w_ARG2 + %0 = load <4 x float>, <4 x float>* @llvm_mips_fmul_w_ARG1 + %1 = load <4 x float>, <4 x float>* @llvm_mips_fmul_w_ARG2 %2 = tail call <4 x float> @llvm.mips.fmul.w(<4 x float> %0, <4 x float> %1) store <4 x float> %2, <4 x float>* @llvm_mips_fmul_w_RES ret void @@ -359,8 +359,8 @@ declare <4 x float> @llvm.mips.fmul.w(<4 x float>, <4 x float>) nounwind define void @llvm_mips_fmul_d_test() nounwind { entry: - %0 = load <2 x double>* @llvm_mips_fmul_d_ARG1 - %1 = load <2 x double>* @llvm_mips_fmul_d_ARG2 + %0 = load <2 x double>, <2 x double>* @llvm_mips_fmul_d_ARG1 + %1 = load <2 x double>, <2 x double>* @llvm_mips_fmul_d_ARG2 %2 = tail call <2 x double> @llvm.mips.fmul.d(<2 x double> %0, <2 x double> %1) store <2 x double> %2, <2 x double>* @llvm_mips_fmul_d_RES ret void @@ -377,8 +377,8 @@ declare <2 x double> @llvm.mips.fmul.d(<2 x double>, <2 x double>) nounwind define void @fmul_w_test() nounwind { entry: - %0 = load <4 x float>* @llvm_mips_fmul_w_ARG1 - %1 = load <4 x float>* @llvm_mips_fmul_w_ARG2 + %0 = load <4 x float>, <4 x float>* @llvm_mips_fmul_w_ARG1 + %1 = load <4 x float>, <4 x float>* @llvm_mips_fmul_w_ARG2 %2 = fmul <4 x float> %0, %1 store <4 x float> %2, <4 x float>* @llvm_mips_fmul_w_RES ret void @@ -393,8 +393,8 @@ entry: define void @fmul_d_test() nounwind { entry: - %0 = load <2 x double>* @llvm_mips_fmul_d_ARG1 - %1 = load <2 x double>* @llvm_mips_fmul_d_ARG2 + %0 = load <2 x double>, <2 x double>* @llvm_mips_fmul_d_ARG1 + %1 = load <2 x double>, <2 x double>* @llvm_mips_fmul_d_ARG2 %2 = fmul <2 x double> %0, %1 store <2 x double> %2, <2 x double>* @llvm_mips_fmul_d_RES ret void @@ -413,8 +413,8 @@ entry: define void @llvm_mips_fsub_w_test() nounwind { entry: - %0 = load <4 x float>* @llvm_mips_fsub_w_ARG1 - %1 = load <4 x float>* @llvm_mips_fsub_w_ARG2 + %0 = load <4 x float>, <4 x float>* @llvm_mips_fsub_w_ARG1 + %1 = load <4 x float>, <4 x float>* @llvm_mips_fsub_w_ARG2 %2 = tail call <4 x float> @llvm.mips.fsub.w(<4 x float> %0, <4 x float> %1) store <4 x float> %2, <4 x float>* @llvm_mips_fsub_w_RES ret void @@ -435,8 +435,8 @@ declare <4 x float> @llvm.mips.fsub.w(<4 x float>, <4 x float>) nounwind define void @llvm_mips_fsub_d_test() nounwind { entry: - %0 = load <2 x double>* @llvm_mips_fsub_d_ARG1 - %1 = load <2 x double>* @llvm_mips_fsub_d_ARG2 + %0 = load <2 x double>, <2 x double>* @llvm_mips_fsub_d_ARG1 + %1 = load <2 x double>, <2 x double>* @llvm_mips_fsub_d_ARG2 %2 = tail call <2 x double> @llvm.mips.fsub.d(<2 x double> %0, <2 x double> %1) store <2 x double> %2, <2 x double>* @llvm_mips_fsub_d_RES ret void @@ -454,8 +454,8 @@ declare <2 x double> @llvm.mips.fsub.d(<2 x double>, <2 x double>) nounwind define void @fsub_w_test() nounwind { entry: - %0 = load <4 x float>* @llvm_mips_fsub_w_ARG1 - %1 = load <4 x float>* @llvm_mips_fsub_w_ARG2 + %0 = load <4 x float>, <4 x float>* @llvm_mips_fsub_w_ARG1 + %1 = load <4 x float>, <4 x float>* @llvm_mips_fsub_w_ARG2 %2 = fsub <4 x float> %0, %1 store <4 x float> %2, <4 x float>* @llvm_mips_fsub_w_RES ret void @@ -470,8 +470,8 @@ entry: define void @fsub_d_test() nounwind { entry: - %0 = load <2 x double>* @llvm_mips_fsub_d_ARG1 - %1 = load <2 x double>* @llvm_mips_fsub_d_ARG2 + %0 = load <2 x double>, <2 x double>* @llvm_mips_fsub_d_ARG1 + %1 = load <2 x double>, <2 x double>* @llvm_mips_fsub_d_ARG2 %2 = fsub <2 x double> %0, %1 store <2 x double> %2, <2 x double>* @llvm_mips_fsub_d_RES ret void diff --git a/test/CodeGen/Mips/msa/3rf_4rf.ll b/test/CodeGen/Mips/msa/3rf_4rf.ll index 67ef7fd2bae15..f1a3002e8179a 100644 --- a/test/CodeGen/Mips/msa/3rf_4rf.ll +++ b/test/CodeGen/Mips/msa/3rf_4rf.ll @@ -11,9 +11,9 @@ define void @llvm_mips_fmadd_w_test() nounwind { entry: - %0 = load <4 x float>* @llvm_mips_fmadd_w_ARG1 - %1 = load <4 x float>* @llvm_mips_fmadd_w_ARG2 - %2 = load <4 x float>* @llvm_mips_fmadd_w_ARG3 + %0 = load <4 x float>, <4 x float>* @llvm_mips_fmadd_w_ARG1 + %1 = load <4 x float>, <4 x float>* @llvm_mips_fmadd_w_ARG2 + %2 = load <4 x float>, <4 x float>* @llvm_mips_fmadd_w_ARG3 %3 = tail call <4 x float> @llvm.mips.fmadd.w(<4 x float> %0, <4 x float> %1, <4 x float> %2) store <4 x float> %3, <4 x float>* @llvm_mips_fmadd_w_RES ret void @@ -36,9 +36,9 @@ declare <4 x float> @llvm.mips.fmadd.w(<4 x float>, <4 x float>, <4 x float>) no define void @llvm_mips_fmadd_d_test() nounwind { entry: - %0 = load <2 x double>* @llvm_mips_fmadd_d_ARG1 - %1 = load <2 x double>* @llvm_mips_fmadd_d_ARG2 - %2 = load <2 x double>* @llvm_mips_fmadd_d_ARG3 + %0 = load <2 x double>, <2 x double>* @llvm_mips_fmadd_d_ARG1 + %1 = load <2 x double>, <2 x double>* @llvm_mips_fmadd_d_ARG2 + %2 = load <2 x double>, <2 x double>* @llvm_mips_fmadd_d_ARG3 %3 = tail call <2 x double> @llvm.mips.fmadd.d(<2 x double> %0, <2 x double> %1, <2 x double> %2) store <2 x double> %3, <2 x double>* @llvm_mips_fmadd_d_RES ret void @@ -61,9 +61,9 @@ declare <2 x double> @llvm.mips.fmadd.d(<2 x double>, <2 x double>, <2 x double> define void @llvm_mips_fmsub_w_test() nounwind { entry: - %0 = load <4 x float>* @llvm_mips_fmsub_w_ARG1 - %1 = load <4 x float>* @llvm_mips_fmsub_w_ARG2 - %2 = load <4 x float>* @llvm_mips_fmsub_w_ARG3 + %0 = load <4 x float>, <4 x float>* @llvm_mips_fmsub_w_ARG1 + %1 = load <4 x float>, <4 x float>* @llvm_mips_fmsub_w_ARG2 + %2 = load <4 x float>, <4 x float>* @llvm_mips_fmsub_w_ARG3 %3 = tail call <4 x float> @llvm.mips.fmsub.w(<4 x float> %0, <4 x float> %1, <4 x float> %2) store <4 x float> %3, <4 x float>* @llvm_mips_fmsub_w_RES ret void @@ -86,9 +86,9 @@ declare <4 x float> @llvm.mips.fmsub.w(<4 x float>, <4 x float>, <4 x float>) no define void @llvm_mips_fmsub_d_test() nounwind { entry: - %0 = load <2 x double>* @llvm_mips_fmsub_d_ARG1 - %1 = load <2 x double>* @llvm_mips_fmsub_d_ARG2 - %2 = load <2 x double>* @llvm_mips_fmsub_d_ARG3 + %0 = load <2 x double>, <2 x double>* @llvm_mips_fmsub_d_ARG1 + %1 = load <2 x double>, <2 x double>* @llvm_mips_fmsub_d_ARG2 + %2 = load <2 x double>, <2 x double>* @llvm_mips_fmsub_d_ARG3 %3 = tail call <2 x double> @llvm.mips.fmsub.d(<2 x double> %0, <2 x double> %1, <2 x double> %2) store <2 x double> %3, <2 x double>* @llvm_mips_fmsub_d_RES ret void diff --git a/test/CodeGen/Mips/msa/3rf_4rf_q.ll b/test/CodeGen/Mips/msa/3rf_4rf_q.ll index de28be0b1c224..704c4b7e7cbc6 100644 --- a/test/CodeGen/Mips/msa/3rf_4rf_q.ll +++ b/test/CodeGen/Mips/msa/3rf_4rf_q.ll @@ -11,9 +11,9 @@ define void @llvm_mips_madd_q_h_test() nounwind { entry: - %0 = load <8 x i16>* @llvm_mips_madd_q_h_ARG1 - %1 = load <8 x i16>* @llvm_mips_madd_q_h_ARG2 - %2 = load <8 x i16>* @llvm_mips_madd_q_h_ARG3 + %0 = load <8 x i16>, <8 x i16>* @llvm_mips_madd_q_h_ARG1 + %1 = load <8 x i16>, <8 x i16>* @llvm_mips_madd_q_h_ARG2 + %2 = load <8 x i16>, <8 x i16>* @llvm_mips_madd_q_h_ARG3 %3 = tail call <8 x i16> @llvm.mips.madd.q.h(<8 x i16> %0, <8 x i16> %1, <8 x i16> %2) store <8 x i16> %3, <8 x i16>* @llvm_mips_madd_q_h_RES ret void @@ -36,9 +36,9 @@ declare <8 x i16> @llvm.mips.madd.q.h(<8 x i16>, <8 x i16>, <8 x i16>) nounwind define void @llvm_mips_madd_q_w_test() nounwind { entry: - %0 = load <4 x i32>* @llvm_mips_madd_q_w_ARG1 - %1 = load <4 x i32>* @llvm_mips_madd_q_w_ARG2 - %2 = load <4 x i32>* @llvm_mips_madd_q_w_ARG3 + %0 = load <4 x i32>, <4 x i32>* @llvm_mips_madd_q_w_ARG1 + %1 = load <4 x i32>, <4 x i32>* @llvm_mips_madd_q_w_ARG2 + %2 = load <4 x i32>, <4 x i32>* @llvm_mips_madd_q_w_ARG3 %3 = tail call <4 x i32> @llvm.mips.madd.q.w(<4 x i32> %0, <4 x i32> %1, <4 x i32> %2) store <4 x i32> %3, <4 x i32>* @llvm_mips_madd_q_w_RES ret void @@ -61,9 +61,9 @@ declare <4 x i32> @llvm.mips.madd.q.w(<4 x i32>, <4 x i32>, <4 x i32>) nounwind define void @llvm_mips_maddr_q_h_test() nounwind { entry: - %0 = load <8 x i16>* @llvm_mips_maddr_q_h_ARG1 - %1 = load <8 x i16>* @llvm_mips_maddr_q_h_ARG2 - %2 = load <8 x i16>* @llvm_mips_maddr_q_h_ARG3 + %0 = load <8 x i16>, <8 x i16>* @llvm_mips_maddr_q_h_ARG1 + %1 = load <8 x i16>, <8 x i16>* @llvm_mips_maddr_q_h_ARG2 + %2 = load <8 x i16>, <8 x i16>* @llvm_mips_maddr_q_h_ARG3 %3 = tail call <8 x i16> @llvm.mips.maddr.q.h(<8 x i16> %0, <8 x i16> %1, <8 x i16> %2) store <8 x i16> %3, <8 x i16>* @llvm_mips_maddr_q_h_RES ret void @@ -86,9 +86,9 @@ declare <8 x i16> @llvm.mips.maddr.q.h(<8 x i16>, <8 x i16>, <8 x i16>) nounwind define void @llvm_mips_maddr_q_w_test() nounwind { entry: - %0 = load <4 x i32>* @llvm_mips_maddr_q_w_ARG1 - %1 = load <4 x i32>* @llvm_mips_maddr_q_w_ARG2 - %2 = load <4 x i32>* @llvm_mips_maddr_q_w_ARG3 + %0 = load <4 x i32>, <4 x i32>* @llvm_mips_maddr_q_w_ARG1 + %1 = load <4 x i32>, <4 x i32>* @llvm_mips_maddr_q_w_ARG2 + %2 = load <4 x i32>, <4 x i32>* @llvm_mips_maddr_q_w_ARG3 %3 = tail call <4 x i32> @llvm.mips.maddr.q.w(<4 x i32> %0, <4 x i32> %1, <4 x i32> %2) store <4 x i32> %3, <4 x i32>* @llvm_mips_maddr_q_w_RES ret void @@ -111,9 +111,9 @@ declare <4 x i32> @llvm.mips.maddr.q.w(<4 x i32>, <4 x i32>, <4 x i32>) nounwind define void @llvm_mips_msub_q_h_test() nounwind { entry: - %0 = load <8 x i16>* @llvm_mips_msub_q_h_ARG1 - %1 = load <8 x i16>* @llvm_mips_msub_q_h_ARG2 - %2 = load <8 x i16>* @llvm_mips_msub_q_h_ARG3 + %0 = load <8 x i16>, <8 x i16>* @llvm_mips_msub_q_h_ARG1 + %1 = load <8 x i16>, <8 x i16>* @llvm_mips_msub_q_h_ARG2 + %2 = load <8 x i16>, <8 x i16>* @llvm_mips_msub_q_h_ARG3 %3 = tail call <8 x i16> @llvm.mips.msub.q.h(<8 x i16> %0, <8 x i16> %1, <8 x i16> %2) store <8 x i16> %3, <8 x i16>* @llvm_mips_msub_q_h_RES ret void @@ -136,9 +136,9 @@ declare <8 x i16> @llvm.mips.msub.q.h(<8 x i16>, <8 x i16>, <8 x i16>) nounwind define void @llvm_mips_msub_q_w_test() nounwind { entry: - %0 = load <4 x i32>* @llvm_mips_msub_q_w_ARG1 - %1 = load <4 x i32>* @llvm_mips_msub_q_w_ARG2 - %2 = load <4 x i32>* @llvm_mips_msub_q_w_ARG3 + %0 = load <4 x i32>, <4 x i32>* @llvm_mips_msub_q_w_ARG1 + %1 = load <4 x i32>, <4 x i32>* @llvm_mips_msub_q_w_ARG2 + %2 = load <4 x i32>, <4 x i32>* @llvm_mips_msub_q_w_ARG3 %3 = tail call <4 x i32> @llvm.mips.msub.q.w(<4 x i32> %0, <4 x i32> %1, <4 x i32> %2) store <4 x i32> %3, <4 x i32>* @llvm_mips_msub_q_w_RES ret void @@ -161,9 +161,9 @@ declare <4 x i32> @llvm.mips.msub.q.w(<4 x i32>, <4 x i32>, <4 x i32>) nounwind define void @llvm_mips_msubr_q_h_test() nounwind { entry: - %0 = load <8 x i16>* @llvm_mips_msubr_q_h_ARG1 - %1 = load <8 x i16>* @llvm_mips_msubr_q_h_ARG2 - %2 = load <8 x i16>* @llvm_mips_msubr_q_h_ARG3 + %0 = load <8 x i16>, <8 x i16>* @llvm_mips_msubr_q_h_ARG1 + %1 = load <8 x i16>, <8 x i16>* @llvm_mips_msubr_q_h_ARG2 + %2 = load <8 x i16>, <8 x i16>* @llvm_mips_msubr_q_h_ARG3 %3 = tail call <8 x i16> @llvm.mips.msubr.q.h(<8 x i16> %0, <8 x i16> %1, <8 x i16> %2) store <8 x i16> %3, <8 x i16>* @llvm_mips_msubr_q_h_RES ret void @@ -186,9 +186,9 @@ declare <8 x i16> @llvm.mips.msubr.q.h(<8 x i16>, <8 x i16>, <8 x i16>) nounwind define void @llvm_mips_msubr_q_w_test() nounwind { entry: - %0 = load <4 x i32>* @llvm_mips_msubr_q_w_ARG1 - %1 = load <4 x i32>* @llvm_mips_msubr_q_w_ARG2 - %2 = load <4 x i32>* @llvm_mips_msubr_q_w_ARG3 + %0 = load <4 x i32>, <4 x i32>* @llvm_mips_msubr_q_w_ARG1 + %1 = load <4 x i32>, <4 x i32>* @llvm_mips_msubr_q_w_ARG2 + %2 = load <4 x i32>, <4 x i32>* @llvm_mips_msubr_q_w_ARG3 %3 = tail call <4 x i32> @llvm.mips.msubr.q.w(<4 x i32> %0, <4 x i32> %1, <4 x i32> %2) store <4 x i32> %3, <4 x i32>* @llvm_mips_msubr_q_w_RES ret void diff --git a/test/CodeGen/Mips/msa/3rf_exdo.ll b/test/CodeGen/Mips/msa/3rf_exdo.ll index 8a7f268a5069c..1b1b2e9243ec6 100644 --- a/test/CodeGen/Mips/msa/3rf_exdo.ll +++ b/test/CodeGen/Mips/msa/3rf_exdo.ll @@ -10,8 +10,8 @@ define void @llvm_mips_fexdo_h_test() nounwind { entry: - %0 = load <4 x float>* @llvm_mips_fexdo_h_ARG1 - %1 = load <4 x float>* @llvm_mips_fexdo_h_ARG2 + %0 = load <4 x float>, <4 x float>* @llvm_mips_fexdo_h_ARG1 + %1 = load <4 x float>, <4 x float>* @llvm_mips_fexdo_h_ARG2 %2 = tail call <8 x half> @llvm.mips.fexdo.h(<4 x float> %0, <4 x float> %1) store <8 x half> %2, <8 x half>* @llvm_mips_fexdo_h_RES ret void @@ -32,8 +32,8 @@ declare <8 x half> @llvm.mips.fexdo.h(<4 x float>, <4 x float>) nounwind define void @llvm_mips_fexdo_w_test() nounwind { entry: - %0 = load <2 x double>* @llvm_mips_fexdo_w_ARG1 - %1 = load <2 x double>* @llvm_mips_fexdo_w_ARG2 + %0 = load <2 x double>, <2 x double>* @llvm_mips_fexdo_w_ARG1 + %1 = load <2 x double>, <2 x double>* @llvm_mips_fexdo_w_ARG2 %2 = tail call <4 x float> @llvm.mips.fexdo.w(<2 x double> %0, <2 x double> %1) store <4 x float> %2, <4 x float>* @llvm_mips_fexdo_w_RES ret void diff --git a/test/CodeGen/Mips/msa/3rf_float_int.ll b/test/CodeGen/Mips/msa/3rf_float_int.ll index 7b01e1721db90..2bd056d3cc8d1 100644 --- a/test/CodeGen/Mips/msa/3rf_float_int.ll +++ b/test/CodeGen/Mips/msa/3rf_float_int.ll @@ -10,8 +10,8 @@ define void @llvm_mips_fexp2_w_test() nounwind { entry: - %0 = load <4 x float>* @llvm_mips_fexp2_w_ARG1 - %1 = load <4 x i32>* @llvm_mips_fexp2_w_ARG2 + %0 = load <4 x float>, <4 x float>* @llvm_mips_fexp2_w_ARG1 + %1 = load <4 x i32>, <4 x i32>* @llvm_mips_fexp2_w_ARG2 %2 = tail call <4 x float> @llvm.mips.fexp2.w(<4 x float> %0, <4 x i32> %1) store <4 x float> %2, <4 x float>* @llvm_mips_fexp2_w_RES ret void @@ -32,8 +32,8 @@ declare <4 x float> @llvm.mips.fexp2.w(<4 x float>, <4 x i32>) nounwind define void @llvm_mips_fexp2_d_test() nounwind { entry: - %0 = load <2 x double>* @llvm_mips_fexp2_d_ARG1 - %1 = load <2 x i64>* @llvm_mips_fexp2_d_ARG2 + %0 = load <2 x double>, <2 x double>* @llvm_mips_fexp2_d_ARG1 + %1 = load <2 x i64>, <2 x i64>* @llvm_mips_fexp2_d_ARG2 %2 = tail call <2 x double> @llvm.mips.fexp2.d(<2 x double> %0, <2 x i64> %1) store <2 x double> %2, <2 x double>* @llvm_mips_fexp2_d_RES ret void diff --git a/test/CodeGen/Mips/msa/3rf_int_float.ll b/test/CodeGen/Mips/msa/3rf_int_float.ll index 5624771b8357f..545e5435d643d 100644 --- a/test/CodeGen/Mips/msa/3rf_int_float.ll +++ b/test/CodeGen/Mips/msa/3rf_int_float.ll @@ -10,8 +10,8 @@ define void @llvm_mips_fcaf_w_test() nounwind { entry: - %0 = load <4 x float>* @llvm_mips_fcaf_w_ARG1 - %1 = load <4 x float>* @llvm_mips_fcaf_w_ARG2 + %0 = load <4 x float>, <4 x float>* @llvm_mips_fcaf_w_ARG1 + %1 = load <4 x float>, <4 x float>* @llvm_mips_fcaf_w_ARG2 %2 = tail call <4 x i32> @llvm.mips.fcaf.w(<4 x float> %0, <4 x float> %1) store <4 x i32> %2, <4 x i32>* @llvm_mips_fcaf_w_RES ret void @@ -32,8 +32,8 @@ declare <4 x i32> @llvm.mips.fcaf.w(<4 x float>, <4 x float>) nounwind define void @llvm_mips_fcaf_d_test() nounwind { entry: - %0 = load <2 x double>* @llvm_mips_fcaf_d_ARG1 - %1 = load <2 x double>* @llvm_mips_fcaf_d_ARG2 + %0 = load <2 x double>, <2 x double>* @llvm_mips_fcaf_d_ARG1 + %1 = load <2 x double>, <2 x double>* @llvm_mips_fcaf_d_ARG2 %2 = tail call <2 x i64> @llvm.mips.fcaf.d(<2 x double> %0, <2 x double> %1) store <2 x i64> %2, <2 x i64>* @llvm_mips_fcaf_d_RES ret void @@ -54,8 +54,8 @@ declare <2 x i64> @llvm.mips.fcaf.d(<2 x double>, <2 x double>) nounwind define void @llvm_mips_fceq_w_test() nounwind { entry: - %0 = load <4 x float>* @llvm_mips_fceq_w_ARG1 - %1 = load <4 x float>* @llvm_mips_fceq_w_ARG2 + %0 = load <4 x float>, <4 x float>* @llvm_mips_fceq_w_ARG1 + %1 = load <4 x float>, <4 x float>* @llvm_mips_fceq_w_ARG2 %2 = tail call <4 x i32> @llvm.mips.fceq.w(<4 x float> %0, <4 x float> %1) store <4 x i32> %2, <4 x i32>* @llvm_mips_fceq_w_RES ret void @@ -76,8 +76,8 @@ declare <4 x i32> @llvm.mips.fceq.w(<4 x float>, <4 x float>) nounwind define void @llvm_mips_fceq_d_test() nounwind { entry: - %0 = load <2 x double>* @llvm_mips_fceq_d_ARG1 - %1 = load <2 x double>* @llvm_mips_fceq_d_ARG2 + %0 = load <2 x double>, <2 x double>* @llvm_mips_fceq_d_ARG1 + %1 = load <2 x double>, <2 x double>* @llvm_mips_fceq_d_ARG2 %2 = tail call <2 x i64> @llvm.mips.fceq.d(<2 x double> %0, <2 x double> %1) store <2 x i64> %2, <2 x i64>* @llvm_mips_fceq_d_RES ret void @@ -98,8 +98,8 @@ declare <2 x i64> @llvm.mips.fceq.d(<2 x double>, <2 x double>) nounwind define void @llvm_mips_fcle_w_test() nounwind { entry: - %0 = load <4 x float>* @llvm_mips_fcle_w_ARG1 - %1 = load <4 x float>* @llvm_mips_fcle_w_ARG2 + %0 = load <4 x float>, <4 x float>* @llvm_mips_fcle_w_ARG1 + %1 = load <4 x float>, <4 x float>* @llvm_mips_fcle_w_ARG2 %2 = tail call <4 x i32> @llvm.mips.fcle.w(<4 x float> %0, <4 x float> %1) store <4 x i32> %2, <4 x i32>* @llvm_mips_fcle_w_RES ret void @@ -120,8 +120,8 @@ declare <4 x i32> @llvm.mips.fcle.w(<4 x float>, <4 x float>) nounwind define void @llvm_mips_fcle_d_test() nounwind { entry: - %0 = load <2 x double>* @llvm_mips_fcle_d_ARG1 - %1 = load <2 x double>* @llvm_mips_fcle_d_ARG2 + %0 = load <2 x double>, <2 x double>* @llvm_mips_fcle_d_ARG1 + %1 = load <2 x double>, <2 x double>* @llvm_mips_fcle_d_ARG2 %2 = tail call <2 x i64> @llvm.mips.fcle.d(<2 x double> %0, <2 x double> %1) store <2 x i64> %2, <2 x i64>* @llvm_mips_fcle_d_RES ret void @@ -142,8 +142,8 @@ declare <2 x i64> @llvm.mips.fcle.d(<2 x double>, <2 x double>) nounwind define void @llvm_mips_fclt_w_test() nounwind { entry: - %0 = load <4 x float>* @llvm_mips_fclt_w_ARG1 - %1 = load <4 x float>* @llvm_mips_fclt_w_ARG2 + %0 = load <4 x float>, <4 x float>* @llvm_mips_fclt_w_ARG1 + %1 = load <4 x float>, <4 x float>* @llvm_mips_fclt_w_ARG2 %2 = tail call <4 x i32> @llvm.mips.fclt.w(<4 x float> %0, <4 x float> %1) store <4 x i32> %2, <4 x i32>* @llvm_mips_fclt_w_RES ret void @@ -164,8 +164,8 @@ declare <4 x i32> @llvm.mips.fclt.w(<4 x float>, <4 x float>) nounwind define void @llvm_mips_fclt_d_test() nounwind { entry: - %0 = load <2 x double>* @llvm_mips_fclt_d_ARG1 - %1 = load <2 x double>* @llvm_mips_fclt_d_ARG2 + %0 = load <2 x double>, <2 x double>* @llvm_mips_fclt_d_ARG1 + %1 = load <2 x double>, <2 x double>* @llvm_mips_fclt_d_ARG2 %2 = tail call <2 x i64> @llvm.mips.fclt.d(<2 x double> %0, <2 x double> %1) store <2 x i64> %2, <2 x i64>* @llvm_mips_fclt_d_RES ret void @@ -186,8 +186,8 @@ declare <2 x i64> @llvm.mips.fclt.d(<2 x double>, <2 x double>) nounwind define void @llvm_mips_fcor_w_test() nounwind { entry: - %0 = load <4 x float>* @llvm_mips_fcor_w_ARG1 - %1 = load <4 x float>* @llvm_mips_fcor_w_ARG2 + %0 = load <4 x float>, <4 x float>* @llvm_mips_fcor_w_ARG1 + %1 = load <4 x float>, <4 x float>* @llvm_mips_fcor_w_ARG2 %2 = tail call <4 x i32> @llvm.mips.fcor.w(<4 x float> %0, <4 x float> %1) store <4 x i32> %2, <4 x i32>* @llvm_mips_fcor_w_RES ret void @@ -208,8 +208,8 @@ declare <4 x i32> @llvm.mips.fcor.w(<4 x float>, <4 x float>) nounwind define void @llvm_mips_fcor_d_test() nounwind { entry: - %0 = load <2 x double>* @llvm_mips_fcor_d_ARG1 - %1 = load <2 x double>* @llvm_mips_fcor_d_ARG2 + %0 = load <2 x double>, <2 x double>* @llvm_mips_fcor_d_ARG1 + %1 = load <2 x double>, <2 x double>* @llvm_mips_fcor_d_ARG2 %2 = tail call <2 x i64> @llvm.mips.fcor.d(<2 x double> %0, <2 x double> %1) store <2 x i64> %2, <2 x i64>* @llvm_mips_fcor_d_RES ret void @@ -230,8 +230,8 @@ declare <2 x i64> @llvm.mips.fcor.d(<2 x double>, <2 x double>) nounwind define void @llvm_mips_fcne_w_test() nounwind { entry: - %0 = load <4 x float>* @llvm_mips_fcne_w_ARG1 - %1 = load <4 x float>* @llvm_mips_fcne_w_ARG2 + %0 = load <4 x float>, <4 x float>* @llvm_mips_fcne_w_ARG1 + %1 = load <4 x float>, <4 x float>* @llvm_mips_fcne_w_ARG2 %2 = tail call <4 x i32> @llvm.mips.fcne.w(<4 x float> %0, <4 x float> %1) store <4 x i32> %2, <4 x i32>* @llvm_mips_fcne_w_RES ret void @@ -252,8 +252,8 @@ declare <4 x i32> @llvm.mips.fcne.w(<4 x float>, <4 x float>) nounwind define void @llvm_mips_fcne_d_test() nounwind { entry: - %0 = load <2 x double>* @llvm_mips_fcne_d_ARG1 - %1 = load <2 x double>* @llvm_mips_fcne_d_ARG2 + %0 = load <2 x double>, <2 x double>* @llvm_mips_fcne_d_ARG1 + %1 = load <2 x double>, <2 x double>* @llvm_mips_fcne_d_ARG2 %2 = tail call <2 x i64> @llvm.mips.fcne.d(<2 x double> %0, <2 x double> %1) store <2 x i64> %2, <2 x i64>* @llvm_mips_fcne_d_RES ret void @@ -274,8 +274,8 @@ declare <2 x i64> @llvm.mips.fcne.d(<2 x double>, <2 x double>) nounwind define void @llvm_mips_fcueq_w_test() nounwind { entry: - %0 = load <4 x float>* @llvm_mips_fcueq_w_ARG1 - %1 = load <4 x float>* @llvm_mips_fcueq_w_ARG2 + %0 = load <4 x float>, <4 x float>* @llvm_mips_fcueq_w_ARG1 + %1 = load <4 x float>, <4 x float>* @llvm_mips_fcueq_w_ARG2 %2 = tail call <4 x i32> @llvm.mips.fcueq.w(<4 x float> %0, <4 x float> %1) store <4 x i32> %2, <4 x i32>* @llvm_mips_fcueq_w_RES ret void @@ -296,8 +296,8 @@ declare <4 x i32> @llvm.mips.fcueq.w(<4 x float>, <4 x float>) nounwind define void @llvm_mips_fcueq_d_test() nounwind { entry: - %0 = load <2 x double>* @llvm_mips_fcueq_d_ARG1 - %1 = load <2 x double>* @llvm_mips_fcueq_d_ARG2 + %0 = load <2 x double>, <2 x double>* @llvm_mips_fcueq_d_ARG1 + %1 = load <2 x double>, <2 x double>* @llvm_mips_fcueq_d_ARG2 %2 = tail call <2 x i64> @llvm.mips.fcueq.d(<2 x double> %0, <2 x double> %1) store <2 x i64> %2, <2 x i64>* @llvm_mips_fcueq_d_RES ret void @@ -318,8 +318,8 @@ declare <2 x i64> @llvm.mips.fcueq.d(<2 x double>, <2 x double>) nounwind define void @llvm_mips_fcult_w_test() nounwind { entry: - %0 = load <4 x float>* @llvm_mips_fcult_w_ARG1 - %1 = load <4 x float>* @llvm_mips_fcult_w_ARG2 + %0 = load <4 x float>, <4 x float>* @llvm_mips_fcult_w_ARG1 + %1 = load <4 x float>, <4 x float>* @llvm_mips_fcult_w_ARG2 %2 = tail call <4 x i32> @llvm.mips.fcult.w(<4 x float> %0, <4 x float> %1) store <4 x i32> %2, <4 x i32>* @llvm_mips_fcult_w_RES ret void @@ -340,8 +340,8 @@ declare <4 x i32> @llvm.mips.fcult.w(<4 x float>, <4 x float>) nounwind define void @llvm_mips_fcult_d_test() nounwind { entry: - %0 = load <2 x double>* @llvm_mips_fcult_d_ARG1 - %1 = load <2 x double>* @llvm_mips_fcult_d_ARG2 + %0 = load <2 x double>, <2 x double>* @llvm_mips_fcult_d_ARG1 + %1 = load <2 x double>, <2 x double>* @llvm_mips_fcult_d_ARG2 %2 = tail call <2 x i64> @llvm.mips.fcult.d(<2 x double> %0, <2 x double> %1) store <2 x i64> %2, <2 x i64>* @llvm_mips_fcult_d_RES ret void @@ -362,8 +362,8 @@ declare <2 x i64> @llvm.mips.fcult.d(<2 x double>, <2 x double>) nounwind define void @llvm_mips_fcule_w_test() nounwind { entry: - %0 = load <4 x float>* @llvm_mips_fcule_w_ARG1 - %1 = load <4 x float>* @llvm_mips_fcule_w_ARG2 + %0 = load <4 x float>, <4 x float>* @llvm_mips_fcule_w_ARG1 + %1 = load <4 x float>, <4 x float>* @llvm_mips_fcule_w_ARG2 %2 = tail call <4 x i32> @llvm.mips.fcule.w(<4 x float> %0, <4 x float> %1) store <4 x i32> %2, <4 x i32>* @llvm_mips_fcule_w_RES ret void @@ -384,8 +384,8 @@ declare <4 x i32> @llvm.mips.fcule.w(<4 x float>, <4 x float>) nounwind define void @llvm_mips_fcule_d_test() nounwind { entry: - %0 = load <2 x double>* @llvm_mips_fcule_d_ARG1 - %1 = load <2 x double>* @llvm_mips_fcule_d_ARG2 + %0 = load <2 x double>, <2 x double>* @llvm_mips_fcule_d_ARG1 + %1 = load <2 x double>, <2 x double>* @llvm_mips_fcule_d_ARG2 %2 = tail call <2 x i64> @llvm.mips.fcule.d(<2 x double> %0, <2 x double> %1) store <2 x i64> %2, <2 x i64>* @llvm_mips_fcule_d_RES ret void @@ -406,8 +406,8 @@ declare <2 x i64> @llvm.mips.fcule.d(<2 x double>, <2 x double>) nounwind define void @llvm_mips_fcun_w_test() nounwind { entry: - %0 = load <4 x float>* @llvm_mips_fcun_w_ARG1 - %1 = load <4 x float>* @llvm_mips_fcun_w_ARG2 + %0 = load <4 x float>, <4 x float>* @llvm_mips_fcun_w_ARG1 + %1 = load <4 x float>, <4 x float>* @llvm_mips_fcun_w_ARG2 %2 = tail call <4 x i32> @llvm.mips.fcun.w(<4 x float> %0, <4 x float> %1) store <4 x i32> %2, <4 x i32>* @llvm_mips_fcun_w_RES ret void @@ -428,8 +428,8 @@ declare <4 x i32> @llvm.mips.fcun.w(<4 x float>, <4 x float>) nounwind define void @llvm_mips_fcun_d_test() nounwind { entry: - %0 = load <2 x double>* @llvm_mips_fcun_d_ARG1 - %1 = load <2 x double>* @llvm_mips_fcun_d_ARG2 + %0 = load <2 x double>, <2 x double>* @llvm_mips_fcun_d_ARG1 + %1 = load <2 x double>, <2 x double>* @llvm_mips_fcun_d_ARG2 %2 = tail call <2 x i64> @llvm.mips.fcun.d(<2 x double> %0, <2 x double> %1) store <2 x i64> %2, <2 x i64>* @llvm_mips_fcun_d_RES ret void @@ -450,8 +450,8 @@ declare <2 x i64> @llvm.mips.fcun.d(<2 x double>, <2 x double>) nounwind define void @llvm_mips_fcune_w_test() nounwind { entry: - %0 = load <4 x float>* @llvm_mips_fcune_w_ARG1 - %1 = load <4 x float>* @llvm_mips_fcune_w_ARG2 + %0 = load <4 x float>, <4 x float>* @llvm_mips_fcune_w_ARG1 + %1 = load <4 x float>, <4 x float>* @llvm_mips_fcune_w_ARG2 %2 = tail call <4 x i32> @llvm.mips.fcune.w(<4 x float> %0, <4 x float> %1) store <4 x i32> %2, <4 x i32>* @llvm_mips_fcune_w_RES ret void @@ -472,8 +472,8 @@ declare <4 x i32> @llvm.mips.fcune.w(<4 x float>, <4 x float>) nounwind define void @llvm_mips_fcune_d_test() nounwind { entry: - %0 = load <2 x double>* @llvm_mips_fcune_d_ARG1 - %1 = load <2 x double>* @llvm_mips_fcune_d_ARG2 + %0 = load <2 x double>, <2 x double>* @llvm_mips_fcune_d_ARG1 + %1 = load <2 x double>, <2 x double>* @llvm_mips_fcune_d_ARG2 %2 = tail call <2 x i64> @llvm.mips.fcune.d(<2 x double> %0, <2 x double> %1) store <2 x i64> %2, <2 x i64>* @llvm_mips_fcune_d_RES ret void @@ -494,8 +494,8 @@ declare <2 x i64> @llvm.mips.fcune.d(<2 x double>, <2 x double>) nounwind define void @llvm_mips_fsaf_w_test() nounwind { entry: - %0 = load <4 x float>* @llvm_mips_fsaf_w_ARG1 - %1 = load <4 x float>* @llvm_mips_fsaf_w_ARG2 + %0 = load <4 x float>, <4 x float>* @llvm_mips_fsaf_w_ARG1 + %1 = load <4 x float>, <4 x float>* @llvm_mips_fsaf_w_ARG2 %2 = tail call <4 x i32> @llvm.mips.fsaf.w(<4 x float> %0, <4 x float> %1) store <4 x i32> %2, <4 x i32>* @llvm_mips_fsaf_w_RES ret void @@ -516,8 +516,8 @@ declare <4 x i32> @llvm.mips.fsaf.w(<4 x float>, <4 x float>) nounwind define void @llvm_mips_fsaf_d_test() nounwind { entry: - %0 = load <2 x double>* @llvm_mips_fsaf_d_ARG1 - %1 = load <2 x double>* @llvm_mips_fsaf_d_ARG2 + %0 = load <2 x double>, <2 x double>* @llvm_mips_fsaf_d_ARG1 + %1 = load <2 x double>, <2 x double>* @llvm_mips_fsaf_d_ARG2 %2 = tail call <2 x i64> @llvm.mips.fsaf.d(<2 x double> %0, <2 x double> %1) store <2 x i64> %2, <2 x i64>* @llvm_mips_fsaf_d_RES ret void @@ -538,8 +538,8 @@ declare <2 x i64> @llvm.mips.fsaf.d(<2 x double>, <2 x double>) nounwind define void @llvm_mips_fseq_w_test() nounwind { entry: - %0 = load <4 x float>* @llvm_mips_fseq_w_ARG1 - %1 = load <4 x float>* @llvm_mips_fseq_w_ARG2 + %0 = load <4 x float>, <4 x float>* @llvm_mips_fseq_w_ARG1 + %1 = load <4 x float>, <4 x float>* @llvm_mips_fseq_w_ARG2 %2 = tail call <4 x i32> @llvm.mips.fseq.w(<4 x float> %0, <4 x float> %1) store <4 x i32> %2, <4 x i32>* @llvm_mips_fseq_w_RES ret void @@ -560,8 +560,8 @@ declare <4 x i32> @llvm.mips.fseq.w(<4 x float>, <4 x float>) nounwind define void @llvm_mips_fseq_d_test() nounwind { entry: - %0 = load <2 x double>* @llvm_mips_fseq_d_ARG1 - %1 = load <2 x double>* @llvm_mips_fseq_d_ARG2 + %0 = load <2 x double>, <2 x double>* @llvm_mips_fseq_d_ARG1 + %1 = load <2 x double>, <2 x double>* @llvm_mips_fseq_d_ARG2 %2 = tail call <2 x i64> @llvm.mips.fseq.d(<2 x double> %0, <2 x double> %1) store <2 x i64> %2, <2 x i64>* @llvm_mips_fseq_d_RES ret void @@ -582,8 +582,8 @@ declare <2 x i64> @llvm.mips.fseq.d(<2 x double>, <2 x double>) nounwind define void @llvm_mips_fsle_w_test() nounwind { entry: - %0 = load <4 x float>* @llvm_mips_fsle_w_ARG1 - %1 = load <4 x float>* @llvm_mips_fsle_w_ARG2 + %0 = load <4 x float>, <4 x float>* @llvm_mips_fsle_w_ARG1 + %1 = load <4 x float>, <4 x float>* @llvm_mips_fsle_w_ARG2 %2 = tail call <4 x i32> @llvm.mips.fsle.w(<4 x float> %0, <4 x float> %1) store <4 x i32> %2, <4 x i32>* @llvm_mips_fsle_w_RES ret void @@ -604,8 +604,8 @@ declare <4 x i32> @llvm.mips.fsle.w(<4 x float>, <4 x float>) nounwind define void @llvm_mips_fsle_d_test() nounwind { entry: - %0 = load <2 x double>* @llvm_mips_fsle_d_ARG1 - %1 = load <2 x double>* @llvm_mips_fsle_d_ARG2 + %0 = load <2 x double>, <2 x double>* @llvm_mips_fsle_d_ARG1 + %1 = load <2 x double>, <2 x double>* @llvm_mips_fsle_d_ARG2 %2 = tail call <2 x i64> @llvm.mips.fsle.d(<2 x double> %0, <2 x double> %1) store <2 x i64> %2, <2 x i64>* @llvm_mips_fsle_d_RES ret void @@ -626,8 +626,8 @@ declare <2 x i64> @llvm.mips.fsle.d(<2 x double>, <2 x double>) nounwind define void @llvm_mips_fslt_w_test() nounwind { entry: - %0 = load <4 x float>* @llvm_mips_fslt_w_ARG1 - %1 = load <4 x float>* @llvm_mips_fslt_w_ARG2 + %0 = load <4 x float>, <4 x float>* @llvm_mips_fslt_w_ARG1 + %1 = load <4 x float>, <4 x float>* @llvm_mips_fslt_w_ARG2 %2 = tail call <4 x i32> @llvm.mips.fslt.w(<4 x float> %0, <4 x float> %1) store <4 x i32> %2, <4 x i32>* @llvm_mips_fslt_w_RES ret void @@ -648,8 +648,8 @@ declare <4 x i32> @llvm.mips.fslt.w(<4 x float>, <4 x float>) nounwind define void @llvm_mips_fslt_d_test() nounwind { entry: - %0 = load <2 x double>* @llvm_mips_fslt_d_ARG1 - %1 = load <2 x double>* @llvm_mips_fslt_d_ARG2 + %0 = load <2 x double>, <2 x double>* @llvm_mips_fslt_d_ARG1 + %1 = load <2 x double>, <2 x double>* @llvm_mips_fslt_d_ARG2 %2 = tail call <2 x i64> @llvm.mips.fslt.d(<2 x double> %0, <2 x double> %1) store <2 x i64> %2, <2 x i64>* @llvm_mips_fslt_d_RES ret void @@ -670,8 +670,8 @@ declare <2 x i64> @llvm.mips.fslt.d(<2 x double>, <2 x double>) nounwind define void @llvm_mips_fsor_w_test() nounwind { entry: - %0 = load <4 x float>* @llvm_mips_fsor_w_ARG1 - %1 = load <4 x float>* @llvm_mips_fsor_w_ARG2 + %0 = load <4 x float>, <4 x float>* @llvm_mips_fsor_w_ARG1 + %1 = load <4 x float>, <4 x float>* @llvm_mips_fsor_w_ARG2 %2 = tail call <4 x i32> @llvm.mips.fsor.w(<4 x float> %0, <4 x float> %1) store <4 x i32> %2, <4 x i32>* @llvm_mips_fsor_w_RES ret void @@ -692,8 +692,8 @@ declare <4 x i32> @llvm.mips.fsor.w(<4 x float>, <4 x float>) nounwind define void @llvm_mips_fsor_d_test() nounwind { entry: - %0 = load <2 x double>* @llvm_mips_fsor_d_ARG1 - %1 = load <2 x double>* @llvm_mips_fsor_d_ARG2 + %0 = load <2 x double>, <2 x double>* @llvm_mips_fsor_d_ARG1 + %1 = load <2 x double>, <2 x double>* @llvm_mips_fsor_d_ARG2 %2 = tail call <2 x i64> @llvm.mips.fsor.d(<2 x double> %0, <2 x double> %1) store <2 x i64> %2, <2 x i64>* @llvm_mips_fsor_d_RES ret void @@ -714,8 +714,8 @@ declare <2 x i64> @llvm.mips.fsor.d(<2 x double>, <2 x double>) nounwind define void @llvm_mips_fsne_w_test() nounwind { entry: - %0 = load <4 x float>* @llvm_mips_fsne_w_ARG1 - %1 = load <4 x float>* @llvm_mips_fsne_w_ARG2 + %0 = load <4 x float>, <4 x float>* @llvm_mips_fsne_w_ARG1 + %1 = load <4 x float>, <4 x float>* @llvm_mips_fsne_w_ARG2 %2 = tail call <4 x i32> @llvm.mips.fsne.w(<4 x float> %0, <4 x float> %1) store <4 x i32> %2, <4 x i32>* @llvm_mips_fsne_w_RES ret void @@ -736,8 +736,8 @@ declare <4 x i32> @llvm.mips.fsne.w(<4 x float>, <4 x float>) nounwind define void @llvm_mips_fsne_d_test() nounwind { entry: - %0 = load <2 x double>* @llvm_mips_fsne_d_ARG1 - %1 = load <2 x double>* @llvm_mips_fsne_d_ARG2 + %0 = load <2 x double>, <2 x double>* @llvm_mips_fsne_d_ARG1 + %1 = load <2 x double>, <2 x double>* @llvm_mips_fsne_d_ARG2 %2 = tail call <2 x i64> @llvm.mips.fsne.d(<2 x double> %0, <2 x double> %1) store <2 x i64> %2, <2 x i64>* @llvm_mips_fsne_d_RES ret void @@ -758,8 +758,8 @@ declare <2 x i64> @llvm.mips.fsne.d(<2 x double>, <2 x double>) nounwind define void @llvm_mips_fsueq_w_test() nounwind { entry: - %0 = load <4 x float>* @llvm_mips_fsueq_w_ARG1 - %1 = load <4 x float>* @llvm_mips_fsueq_w_ARG2 + %0 = load <4 x float>, <4 x float>* @llvm_mips_fsueq_w_ARG1 + %1 = load <4 x float>, <4 x float>* @llvm_mips_fsueq_w_ARG2 %2 = tail call <4 x i32> @llvm.mips.fsueq.w(<4 x float> %0, <4 x float> %1) store <4 x i32> %2, <4 x i32>* @llvm_mips_fsueq_w_RES ret void @@ -780,8 +780,8 @@ declare <4 x i32> @llvm.mips.fsueq.w(<4 x float>, <4 x float>) nounwind define void @llvm_mips_fsueq_d_test() nounwind { entry: - %0 = load <2 x double>* @llvm_mips_fsueq_d_ARG1 - %1 = load <2 x double>* @llvm_mips_fsueq_d_ARG2 + %0 = load <2 x double>, <2 x double>* @llvm_mips_fsueq_d_ARG1 + %1 = load <2 x double>, <2 x double>* @llvm_mips_fsueq_d_ARG2 %2 = tail call <2 x i64> @llvm.mips.fsueq.d(<2 x double> %0, <2 x double> %1) store <2 x i64> %2, <2 x i64>* @llvm_mips_fsueq_d_RES ret void @@ -802,8 +802,8 @@ declare <2 x i64> @llvm.mips.fsueq.d(<2 x double>, <2 x double>) nounwind define void @llvm_mips_fsult_w_test() nounwind { entry: - %0 = load <4 x float>* @llvm_mips_fsult_w_ARG1 - %1 = load <4 x float>* @llvm_mips_fsult_w_ARG2 + %0 = load <4 x float>, <4 x float>* @llvm_mips_fsult_w_ARG1 + %1 = load <4 x float>, <4 x float>* @llvm_mips_fsult_w_ARG2 %2 = tail call <4 x i32> @llvm.mips.fsult.w(<4 x float> %0, <4 x float> %1) store <4 x i32> %2, <4 x i32>* @llvm_mips_fsult_w_RES ret void @@ -824,8 +824,8 @@ declare <4 x i32> @llvm.mips.fsult.w(<4 x float>, <4 x float>) nounwind define void @llvm_mips_fsult_d_test() nounwind { entry: - %0 = load <2 x double>* @llvm_mips_fsult_d_ARG1 - %1 = load <2 x double>* @llvm_mips_fsult_d_ARG2 + %0 = load <2 x double>, <2 x double>* @llvm_mips_fsult_d_ARG1 + %1 = load <2 x double>, <2 x double>* @llvm_mips_fsult_d_ARG2 %2 = tail call <2 x i64> @llvm.mips.fsult.d(<2 x double> %0, <2 x double> %1) store <2 x i64> %2, <2 x i64>* @llvm_mips_fsult_d_RES ret void @@ -846,8 +846,8 @@ declare <2 x i64> @llvm.mips.fsult.d(<2 x double>, <2 x double>) nounwind define void @llvm_mips_fsule_w_test() nounwind { entry: - %0 = load <4 x float>* @llvm_mips_fsule_w_ARG1 - %1 = load <4 x float>* @llvm_mips_fsule_w_ARG2 + %0 = load <4 x float>, <4 x float>* @llvm_mips_fsule_w_ARG1 + %1 = load <4 x float>, <4 x float>* @llvm_mips_fsule_w_ARG2 %2 = tail call <4 x i32> @llvm.mips.fsule.w(<4 x float> %0, <4 x float> %1) store <4 x i32> %2, <4 x i32>* @llvm_mips_fsule_w_RES ret void @@ -868,8 +868,8 @@ declare <4 x i32> @llvm.mips.fsule.w(<4 x float>, <4 x float>) nounwind define void @llvm_mips_fsule_d_test() nounwind { entry: - %0 = load <2 x double>* @llvm_mips_fsule_d_ARG1 - %1 = load <2 x double>* @llvm_mips_fsule_d_ARG2 + %0 = load <2 x double>, <2 x double>* @llvm_mips_fsule_d_ARG1 + %1 = load <2 x double>, <2 x double>* @llvm_mips_fsule_d_ARG2 %2 = tail call <2 x i64> @llvm.mips.fsule.d(<2 x double> %0, <2 x double> %1) store <2 x i64> %2, <2 x i64>* @llvm_mips_fsule_d_RES ret void @@ -890,8 +890,8 @@ declare <2 x i64> @llvm.mips.fsule.d(<2 x double>, <2 x double>) nounwind define void @llvm_mips_fsun_w_test() nounwind { entry: - %0 = load <4 x float>* @llvm_mips_fsun_w_ARG1 - %1 = load <4 x float>* @llvm_mips_fsun_w_ARG2 + %0 = load <4 x float>, <4 x float>* @llvm_mips_fsun_w_ARG1 + %1 = load <4 x float>, <4 x float>* @llvm_mips_fsun_w_ARG2 %2 = tail call <4 x i32> @llvm.mips.fsun.w(<4 x float> %0, <4 x float> %1) store <4 x i32> %2, <4 x i32>* @llvm_mips_fsun_w_RES ret void @@ -912,8 +912,8 @@ declare <4 x i32> @llvm.mips.fsun.w(<4 x float>, <4 x float>) nounwind define void @llvm_mips_fsun_d_test() nounwind { entry: - %0 = load <2 x double>* @llvm_mips_fsun_d_ARG1 - %1 = load <2 x double>* @llvm_mips_fsun_d_ARG2 + %0 = load <2 x double>, <2 x double>* @llvm_mips_fsun_d_ARG1 + %1 = load <2 x double>, <2 x double>* @llvm_mips_fsun_d_ARG2 %2 = tail call <2 x i64> @llvm.mips.fsun.d(<2 x double> %0, <2 x double> %1) store <2 x i64> %2, <2 x i64>* @llvm_mips_fsun_d_RES ret void @@ -934,8 +934,8 @@ declare <2 x i64> @llvm.mips.fsun.d(<2 x double>, <2 x double>) nounwind define void @llvm_mips_fsune_w_test() nounwind { entry: - %0 = load <4 x float>* @llvm_mips_fsune_w_ARG1 - %1 = load <4 x float>* @llvm_mips_fsune_w_ARG2 + %0 = load <4 x float>, <4 x float>* @llvm_mips_fsune_w_ARG1 + %1 = load <4 x float>, <4 x float>* @llvm_mips_fsune_w_ARG2 %2 = tail call <4 x i32> @llvm.mips.fsune.w(<4 x float> %0, <4 x float> %1) store <4 x i32> %2, <4 x i32>* @llvm_mips_fsune_w_RES ret void @@ -956,8 +956,8 @@ declare <4 x i32> @llvm.mips.fsune.w(<4 x float>, <4 x float>) nounwind define void @llvm_mips_fsune_d_test() nounwind { entry: - %0 = load <2 x double>* @llvm_mips_fsune_d_ARG1 - %1 = load <2 x double>* @llvm_mips_fsune_d_ARG2 + %0 = load <2 x double>, <2 x double>* @llvm_mips_fsune_d_ARG1 + %1 = load <2 x double>, <2 x double>* @llvm_mips_fsune_d_ARG2 %2 = tail call <2 x i64> @llvm.mips.fsune.d(<2 x double> %0, <2 x double> %1) store <2 x i64> %2, <2 x i64>* @llvm_mips_fsune_d_RES ret void diff --git a/test/CodeGen/Mips/msa/3rf_q.ll b/test/CodeGen/Mips/msa/3rf_q.ll index f7000ee913a66..c8b0a5000021d 100644 --- a/test/CodeGen/Mips/msa/3rf_q.ll +++ b/test/CodeGen/Mips/msa/3rf_q.ll @@ -10,8 +10,8 @@ define void @llvm_mips_mul_q_h_test() nounwind { entry: - %0 = load <8 x i16>* @llvm_mips_mul_q_h_ARG1 - %1 = load <8 x i16>* @llvm_mips_mul_q_h_ARG2 + %0 = load <8 x i16>, <8 x i16>* @llvm_mips_mul_q_h_ARG1 + %1 = load <8 x i16>, <8 x i16>* @llvm_mips_mul_q_h_ARG2 %2 = tail call <8 x i16> @llvm.mips.mul.q.h(<8 x i16> %0, <8 x i16> %1) store <8 x i16> %2, <8 x i16>* @llvm_mips_mul_q_h_RES ret void @@ -32,8 +32,8 @@ declare <8 x i16> @llvm.mips.mul.q.h(<8 x i16>, <8 x i16>) nounwind define void @llvm_mips_mul_q_w_test() nounwind { entry: - %0 = load <4 x i32>* @llvm_mips_mul_q_w_ARG1 - %1 = load <4 x i32>* @llvm_mips_mul_q_w_ARG2 + %0 = load <4 x i32>, <4 x i32>* @llvm_mips_mul_q_w_ARG1 + %1 = load <4 x i32>, <4 x i32>* @llvm_mips_mul_q_w_ARG2 %2 = tail call <4 x i32> @llvm.mips.mul.q.w(<4 x i32> %0, <4 x i32> %1) store <4 x i32> %2, <4 x i32>* @llvm_mips_mul_q_w_RES ret void @@ -54,8 +54,8 @@ declare <4 x i32> @llvm.mips.mul.q.w(<4 x i32>, <4 x i32>) nounwind define void @llvm_mips_mulr_q_h_test() nounwind { entry: - %0 = load <8 x i16>* @llvm_mips_mulr_q_h_ARG1 - %1 = load <8 x i16>* @llvm_mips_mulr_q_h_ARG2 + %0 = load <8 x i16>, <8 x i16>* @llvm_mips_mulr_q_h_ARG1 + %1 = load <8 x i16>, <8 x i16>* @llvm_mips_mulr_q_h_ARG2 %2 = tail call <8 x i16> @llvm.mips.mulr.q.h(<8 x i16> %0, <8 x i16> %1) store <8 x i16> %2, <8 x i16>* @llvm_mips_mulr_q_h_RES ret void @@ -76,8 +76,8 @@ declare <8 x i16> @llvm.mips.mulr.q.h(<8 x i16>, <8 x i16>) nounwind define void @llvm_mips_mulr_q_w_test() nounwind { entry: - %0 = load <4 x i32>* @llvm_mips_mulr_q_w_ARG1 - %1 = load <4 x i32>* @llvm_mips_mulr_q_w_ARG2 + %0 = load <4 x i32>, <4 x i32>* @llvm_mips_mulr_q_w_ARG1 + %1 = load <4 x i32>, <4 x i32>* @llvm_mips_mulr_q_w_ARG2 %2 = tail call <4 x i32> @llvm.mips.mulr.q.w(<4 x i32> %0, <4 x i32> %1) store <4 x i32> %2, <4 x i32>* @llvm_mips_mulr_q_w_RES ret void diff --git a/test/CodeGen/Mips/msa/arithmetic.ll b/test/CodeGen/Mips/msa/arithmetic.ll index 09ee5023c7b19..3ecd0e43589f8 100644 --- a/test/CodeGen/Mips/msa/arithmetic.ll +++ b/test/CodeGen/Mips/msa/arithmetic.ll @@ -4,9 +4,9 @@ define void @add_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind { ; CHECK: add_v16i8: - %1 = load <16 x i8>* %a + %1 = load <16 x i8>, <16 x i8>* %a ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5) - %2 = load <16 x i8>* %b + %2 = load <16 x i8>, <16 x i8>* %b ; CHECK-DAG: ld.b [[R2:\$w[0-9]+]], 0($6) %3 = add <16 x i8> %1, %2 ; CHECK-DAG: addv.b [[R3:\$w[0-9]+]], [[R1]], [[R2]] @@ -20,9 +20,9 @@ define void @add_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind { define void @add_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind { ; CHECK: add_v8i16: - %1 = load <8 x i16>* %a + %1 = load <8 x i16>, <8 x i16>* %a ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5) - %2 = load <8 x i16>* %b + %2 = load <8 x i16>, <8 x i16>* %b ; CHECK-DAG: ld.h [[R2:\$w[0-9]+]], 0($6) %3 = add <8 x i16> %1, %2 ; CHECK-DAG: addv.h [[R3:\$w[0-9]+]], [[R1]], [[R2]] @@ -36,9 +36,9 @@ define void @add_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind { define void @add_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind { ; CHECK: add_v4i32: - %1 = load <4 x i32>* %a + %1 = load <4 x i32>, <4 x i32>* %a ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5) - %2 = load <4 x i32>* %b + %2 = load <4 x i32>, <4 x i32>* %b ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6) %3 = add <4 x i32> %1, %2 ; CHECK-DAG: addv.w [[R3:\$w[0-9]+]], [[R1]], [[R2]] @@ -52,9 +52,9 @@ define void @add_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind { define void @add_v2i64(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind { ; CHECK: add_v2i64: - %1 = load <2 x i64>* %a + %1 = load <2 x i64>, <2 x i64>* %a ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5) - %2 = load <2 x i64>* %b + %2 = load <2 x i64>, <2 x i64>* %b ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6) %3 = add <2 x i64> %1, %2 ; CHECK-DAG: addv.d [[R3:\$w[0-9]+]], [[R1]], [[R2]] @@ -68,7 +68,7 @@ define void @add_v2i64(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind { define void @add_v16i8_i(<16 x i8>* %c, <16 x i8>* %a) nounwind { ; CHECK: add_v16i8_i: - %1 = load <16 x i8>* %a + %1 = load <16 x i8>, <16 x i8>* %a ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5) %2 = add <16 x i8> %1, <i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1> @@ -83,7 +83,7 @@ define void @add_v16i8_i(<16 x i8>* %c, <16 x i8>* %a) nounwind { define void @add_v8i16_i(<8 x i16>* %c, <8 x i16>* %a) nounwind { ; CHECK: add_v8i16_i: - %1 = load <8 x i16>* %a + %1 = load <8 x i16>, <8 x i16>* %a ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5) %2 = add <8 x i16> %1, <i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1> @@ -98,7 +98,7 @@ define void @add_v8i16_i(<8 x i16>* %c, <8 x i16>* %a) nounwind { define void @add_v4i32_i(<4 x i32>* %c, <4 x i32>* %a) nounwind { ; CHECK: add_v4i32_i: - %1 = load <4 x i32>* %a + %1 = load <4 x i32>, <4 x i32>* %a ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5) %2 = add <4 x i32> %1, <i32 1, i32 1, i32 1, i32 1> ; CHECK-DAG: addvi.w [[R3:\$w[0-9]+]], [[R1]], 1 @@ -112,7 +112,7 @@ define void @add_v4i32_i(<4 x i32>* %c, <4 x i32>* %a) nounwind { define void @add_v2i64_i(<2 x i64>* %c, <2 x i64>* %a) nounwind { ; CHECK: add_v2i64_i: - %1 = load <2 x i64>* %a + %1 = load <2 x i64>, <2 x i64>* %a ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5) %2 = add <2 x i64> %1, <i64 1, i64 1> ; CHECK-DAG: addvi.d [[R3:\$w[0-9]+]], [[R1]], 1 @@ -126,9 +126,9 @@ define void @add_v2i64_i(<2 x i64>* %c, <2 x i64>* %a) nounwind { define void @sub_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind { ; CHECK: sub_v16i8: - %1 = load <16 x i8>* %a + %1 = load <16 x i8>, <16 x i8>* %a ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5) - %2 = load <16 x i8>* %b + %2 = load <16 x i8>, <16 x i8>* %b ; CHECK-DAG: ld.b [[R2:\$w[0-9]+]], 0($6) %3 = sub <16 x i8> %1, %2 ; CHECK-DAG: subv.b [[R3:\$w[0-9]+]], [[R1]], [[R2]] @@ -142,9 +142,9 @@ define void @sub_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind { define void @sub_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind { ; CHECK: sub_v8i16: - %1 = load <8 x i16>* %a + %1 = load <8 x i16>, <8 x i16>* %a ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5) - %2 = load <8 x i16>* %b + %2 = load <8 x i16>, <8 x i16>* %b ; CHECK-DAG: ld.h [[R2:\$w[0-9]+]], 0($6) %3 = sub <8 x i16> %1, %2 ; CHECK-DAG: subv.h [[R3:\$w[0-9]+]], [[R1]], [[R2]] @@ -158,9 +158,9 @@ define void @sub_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind { define void @sub_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind { ; CHECK: sub_v4i32: - %1 = load <4 x i32>* %a + %1 = load <4 x i32>, <4 x i32>* %a ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5) - %2 = load <4 x i32>* %b + %2 = load <4 x i32>, <4 x i32>* %b ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6) %3 = sub <4 x i32> %1, %2 ; CHECK-DAG: subv.w [[R3:\$w[0-9]+]], [[R1]], [[R2]] @@ -174,9 +174,9 @@ define void @sub_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind { define void @sub_v2i64(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind { ; CHECK: sub_v2i64: - %1 = load <2 x i64>* %a + %1 = load <2 x i64>, <2 x i64>* %a ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5) - %2 = load <2 x i64>* %b + %2 = load <2 x i64>, <2 x i64>* %b ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6) %3 = sub <2 x i64> %1, %2 ; CHECK-DAG: subv.d [[R3:\$w[0-9]+]], [[R1]], [[R2]] @@ -190,7 +190,7 @@ define void @sub_v2i64(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind { define void @sub_v16i8_i(<16 x i8>* %c, <16 x i8>* %a) nounwind { ; CHECK: sub_v16i8_i: - %1 = load <16 x i8>* %a + %1 = load <16 x i8>, <16 x i8>* %a ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5) %2 = sub <16 x i8> %1, <i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1> @@ -205,7 +205,7 @@ define void @sub_v16i8_i(<16 x i8>* %c, <16 x i8>* %a) nounwind { define void @sub_v8i16_i(<8 x i16>* %c, <8 x i16>* %a) nounwind { ; CHECK: sub_v8i16_i: - %1 = load <8 x i16>* %a + %1 = load <8 x i16>, <8 x i16>* %a ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5) %2 = sub <8 x i16> %1, <i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1> @@ -220,7 +220,7 @@ define void @sub_v8i16_i(<8 x i16>* %c, <8 x i16>* %a) nounwind { define void @sub_v4i32_i(<4 x i32>* %c, <4 x i32>* %a) nounwind { ; CHECK: sub_v4i32_i: - %1 = load <4 x i32>* %a + %1 = load <4 x i32>, <4 x i32>* %a ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5) %2 = sub <4 x i32> %1, <i32 1, i32 1, i32 1, i32 1> ; CHECK-DAG: subvi.w [[R3:\$w[0-9]+]], [[R1]], 1 @@ -234,7 +234,7 @@ define void @sub_v4i32_i(<4 x i32>* %c, <4 x i32>* %a) nounwind { define void @sub_v2i64_i(<2 x i64>* %c, <2 x i64>* %a) nounwind { ; CHECK: sub_v2i64_i: - %1 = load <2 x i64>* %a + %1 = load <2 x i64>, <2 x i64>* %a ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5) %2 = sub <2 x i64> %1, <i64 1, i64 1> ; CHECK-DAG: subvi.d [[R3:\$w[0-9]+]], [[R1]], 1 @@ -248,9 +248,9 @@ define void @sub_v2i64_i(<2 x i64>* %c, <2 x i64>* %a) nounwind { define void @mul_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind { ; CHECK: mul_v16i8: - %1 = load <16 x i8>* %a + %1 = load <16 x i8>, <16 x i8>* %a ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5) - %2 = load <16 x i8>* %b + %2 = load <16 x i8>, <16 x i8>* %b ; CHECK-DAG: ld.b [[R2:\$w[0-9]+]], 0($6) %3 = mul <16 x i8> %1, %2 ; CHECK-DAG: mulv.b [[R3:\$w[0-9]+]], [[R1]], [[R2]] @@ -264,9 +264,9 @@ define void @mul_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind { define void @mul_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind { ; CHECK: mul_v8i16: - %1 = load <8 x i16>* %a + %1 = load <8 x i16>, <8 x i16>* %a ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5) - %2 = load <8 x i16>* %b + %2 = load <8 x i16>, <8 x i16>* %b ; CHECK-DAG: ld.h [[R2:\$w[0-9]+]], 0($6) %3 = mul <8 x i16> %1, %2 ; CHECK-DAG: mulv.h [[R3:\$w[0-9]+]], [[R1]], [[R2]] @@ -280,9 +280,9 @@ define void @mul_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind { define void @mul_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind { ; CHECK: mul_v4i32: - %1 = load <4 x i32>* %a + %1 = load <4 x i32>, <4 x i32>* %a ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5) - %2 = load <4 x i32>* %b + %2 = load <4 x i32>, <4 x i32>* %b ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6) %3 = mul <4 x i32> %1, %2 ; CHECK-DAG: mulv.w [[R3:\$w[0-9]+]], [[R1]], [[R2]] @@ -296,9 +296,9 @@ define void @mul_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind { define void @mul_v2i64(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind { ; CHECK: mul_v2i64: - %1 = load <2 x i64>* %a + %1 = load <2 x i64>, <2 x i64>* %a ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5) - %2 = load <2 x i64>* %b + %2 = load <2 x i64>, <2 x i64>* %b ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6) %3 = mul <2 x i64> %1, %2 ; CHECK-DAG: mulv.d [[R3:\$w[0-9]+]], [[R1]], [[R2]] @@ -313,11 +313,11 @@ define void @maddv_v16i8(<16 x i8>* %d, <16 x i8>* %a, <16 x i8>* %b, <16 x i8>* %c) nounwind { ; CHECK: maddv_v16i8: - %1 = load <16 x i8>* %a + %1 = load <16 x i8>, <16 x i8>* %a ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5) - %2 = load <16 x i8>* %b + %2 = load <16 x i8>, <16 x i8>* %b ; CHECK-DAG: ld.b [[R2:\$w[0-9]+]], 0($6) - %3 = load <16 x i8>* %c + %3 = load <16 x i8>, <16 x i8>* %c ; CHECK-DAG: ld.b [[R3:\$w[0-9]+]], 0($7) %4 = mul <16 x i8> %2, %3 %5 = add <16 x i8> %4, %1 @@ -333,11 +333,11 @@ define void @maddv_v8i16(<8 x i16>* %d, <8 x i16>* %a, <8 x i16>* %b, <8 x i16>* %c) nounwind { ; CHECK: maddv_v8i16: - %1 = load <8 x i16>* %a + %1 = load <8 x i16>, <8 x i16>* %a ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5) - %2 = load <8 x i16>* %b + %2 = load <8 x i16>, <8 x i16>* %b ; CHECK-DAG: ld.h [[R2:\$w[0-9]+]], 0($6) - %3 = load <8 x i16>* %c + %3 = load <8 x i16>, <8 x i16>* %c ; CHECK-DAG: ld.h [[R3:\$w[0-9]+]], 0($7) %4 = mul <8 x i16> %2, %3 %5 = add <8 x i16> %4, %1 @@ -353,11 +353,11 @@ define void @maddv_v4i32(<4 x i32>* %d, <4 x i32>* %a, <4 x i32>* %b, <4 x i32>* %c) nounwind { ; CHECK: maddv_v4i32: - %1 = load <4 x i32>* %a + %1 = load <4 x i32>, <4 x i32>* %a ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5) - %2 = load <4 x i32>* %b + %2 = load <4 x i32>, <4 x i32>* %b ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6) - %3 = load <4 x i32>* %c + %3 = load <4 x i32>, <4 x i32>* %c ; CHECK-DAG: ld.w [[R3:\$w[0-9]+]], 0($7) %4 = mul <4 x i32> %2, %3 %5 = add <4 x i32> %4, %1 @@ -373,11 +373,11 @@ define void @maddv_v2i64(<2 x i64>* %d, <2 x i64>* %a, <2 x i64>* %b, <2 x i64>* %c) nounwind { ; CHECK: maddv_v2i64: - %1 = load <2 x i64>* %a + %1 = load <2 x i64>, <2 x i64>* %a ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5) - %2 = load <2 x i64>* %b + %2 = load <2 x i64>, <2 x i64>* %b ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6) - %3 = load <2 x i64>* %c + %3 = load <2 x i64>, <2 x i64>* %c ; CHECK-DAG: ld.d [[R3:\$w[0-9]+]], 0($7) %4 = mul <2 x i64> %2, %3 %5 = add <2 x i64> %4, %1 @@ -393,11 +393,11 @@ define void @msubv_v16i8(<16 x i8>* %d, <16 x i8>* %a, <16 x i8>* %b, <16 x i8>* %c) nounwind { ; CHECK: msubv_v16i8: - %1 = load <16 x i8>* %a + %1 = load <16 x i8>, <16 x i8>* %a ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5) - %2 = load <16 x i8>* %b + %2 = load <16 x i8>, <16 x i8>* %b ; CHECK-DAG: ld.b [[R2:\$w[0-9]+]], 0($6) - %3 = load <16 x i8>* %c + %3 = load <16 x i8>, <16 x i8>* %c ; CHECK-DAG: ld.b [[R3:\$w[0-9]+]], 0($7) %4 = mul <16 x i8> %2, %3 %5 = sub <16 x i8> %1, %4 @@ -413,11 +413,11 @@ define void @msubv_v8i16(<8 x i16>* %d, <8 x i16>* %a, <8 x i16>* %b, <8 x i16>* %c) nounwind { ; CHECK: msubv_v8i16: - %1 = load <8 x i16>* %a + %1 = load <8 x i16>, <8 x i16>* %a ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5) - %2 = load <8 x i16>* %b + %2 = load <8 x i16>, <8 x i16>* %b ; CHECK-DAG: ld.h [[R2:\$w[0-9]+]], 0($6) - %3 = load <8 x i16>* %c + %3 = load <8 x i16>, <8 x i16>* %c ; CHECK-DAG: ld.h [[R3:\$w[0-9]+]], 0($7) %4 = mul <8 x i16> %2, %3 %5 = sub <8 x i16> %1, %4 @@ -433,11 +433,11 @@ define void @msubv_v4i32(<4 x i32>* %d, <4 x i32>* %a, <4 x i32>* %b, <4 x i32>* %c) nounwind { ; CHECK: msubv_v4i32: - %1 = load <4 x i32>* %a + %1 = load <4 x i32>, <4 x i32>* %a ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5) - %2 = load <4 x i32>* %b + %2 = load <4 x i32>, <4 x i32>* %b ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6) - %3 = load <4 x i32>* %c + %3 = load <4 x i32>, <4 x i32>* %c ; CHECK-DAG: ld.w [[R3:\$w[0-9]+]], 0($7) %4 = mul <4 x i32> %2, %3 %5 = sub <4 x i32> %1, %4 @@ -453,11 +453,11 @@ define void @msubv_v2i64(<2 x i64>* %d, <2 x i64>* %a, <2 x i64>* %b, <2 x i64>* %c) nounwind { ; CHECK: msubv_v2i64: - %1 = load <2 x i64>* %a + %1 = load <2 x i64>, <2 x i64>* %a ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5) - %2 = load <2 x i64>* %b + %2 = load <2 x i64>, <2 x i64>* %b ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6) - %3 = load <2 x i64>* %c + %3 = load <2 x i64>, <2 x i64>* %c ; CHECK-DAG: ld.d [[R3:\$w[0-9]+]], 0($7) %4 = mul <2 x i64> %2, %3 %5 = sub <2 x i64> %1, %4 @@ -472,9 +472,9 @@ define void @msubv_v2i64(<2 x i64>* %d, <2 x i64>* %a, <2 x i64>* %b, define void @div_s_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind { ; CHECK: div_s_v16i8: - %1 = load <16 x i8>* %a + %1 = load <16 x i8>, <16 x i8>* %a ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5) - %2 = load <16 x i8>* %b + %2 = load <16 x i8>, <16 x i8>* %b ; CHECK-DAG: ld.b [[R2:\$w[0-9]+]], 0($6) %3 = sdiv <16 x i8> %1, %2 ; CHECK-DAG: div_s.b [[R3:\$w[0-9]+]], [[R1]], [[R2]] @@ -488,9 +488,9 @@ define void @div_s_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind { define void @div_s_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind { ; CHECK: div_s_v8i16: - %1 = load <8 x i16>* %a + %1 = load <8 x i16>, <8 x i16>* %a ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5) - %2 = load <8 x i16>* %b + %2 = load <8 x i16>, <8 x i16>* %b ; CHECK-DAG: ld.h [[R2:\$w[0-9]+]], 0($6) %3 = sdiv <8 x i16> %1, %2 ; CHECK-DAG: div_s.h [[R3:\$w[0-9]+]], [[R1]], [[R2]] @@ -504,9 +504,9 @@ define void @div_s_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind { define void @div_s_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind { ; CHECK: div_s_v4i32: - %1 = load <4 x i32>* %a + %1 = load <4 x i32>, <4 x i32>* %a ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5) - %2 = load <4 x i32>* %b + %2 = load <4 x i32>, <4 x i32>* %b ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6) %3 = sdiv <4 x i32> %1, %2 ; CHECK-DAG: div_s.w [[R3:\$w[0-9]+]], [[R1]], [[R2]] @@ -520,9 +520,9 @@ define void @div_s_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind { define void @div_s_v2i64(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind { ; CHECK: div_s_v2i64: - %1 = load <2 x i64>* %a + %1 = load <2 x i64>, <2 x i64>* %a ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5) - %2 = load <2 x i64>* %b + %2 = load <2 x i64>, <2 x i64>* %b ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6) %3 = sdiv <2 x i64> %1, %2 ; CHECK-DAG: div_s.d [[R3:\$w[0-9]+]], [[R1]], [[R2]] @@ -536,9 +536,9 @@ define void @div_s_v2i64(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind { define void @div_u_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind { ; CHECK: div_u_v16i8: - %1 = load <16 x i8>* %a + %1 = load <16 x i8>, <16 x i8>* %a ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5) - %2 = load <16 x i8>* %b + %2 = load <16 x i8>, <16 x i8>* %b ; CHECK-DAG: ld.b [[R2:\$w[0-9]+]], 0($6) %3 = udiv <16 x i8> %1, %2 ; CHECK-DAG: div_u.b [[R3:\$w[0-9]+]], [[R1]], [[R2]] @@ -552,9 +552,9 @@ define void @div_u_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind { define void @div_u_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind { ; CHECK: div_u_v8i16: - %1 = load <8 x i16>* %a + %1 = load <8 x i16>, <8 x i16>* %a ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5) - %2 = load <8 x i16>* %b + %2 = load <8 x i16>, <8 x i16>* %b ; CHECK-DAG: ld.h [[R2:\$w[0-9]+]], 0($6) %3 = udiv <8 x i16> %1, %2 ; CHECK-DAG: div_u.h [[R3:\$w[0-9]+]], [[R1]], [[R2]] @@ -568,9 +568,9 @@ define void @div_u_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind { define void @div_u_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind { ; CHECK: div_u_v4i32: - %1 = load <4 x i32>* %a + %1 = load <4 x i32>, <4 x i32>* %a ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5) - %2 = load <4 x i32>* %b + %2 = load <4 x i32>, <4 x i32>* %b ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6) %3 = udiv <4 x i32> %1, %2 ; CHECK-DAG: div_u.w [[R3:\$w[0-9]+]], [[R1]], [[R2]] @@ -584,9 +584,9 @@ define void @div_u_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind { define void @div_u_v2i64(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind { ; CHECK: div_u_v2i64: - %1 = load <2 x i64>* %a + %1 = load <2 x i64>, <2 x i64>* %a ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5) - %2 = load <2 x i64>* %b + %2 = load <2 x i64>, <2 x i64>* %b ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6) %3 = udiv <2 x i64> %1, %2 ; CHECK-DAG: div_u.d [[R3:\$w[0-9]+]], [[R1]], [[R2]] @@ -600,9 +600,9 @@ define void @div_u_v2i64(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind { define void @mod_s_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind { ; CHECK: mod_s_v16i8: - %1 = load <16 x i8>* %a + %1 = load <16 x i8>, <16 x i8>* %a ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5) - %2 = load <16 x i8>* %b + %2 = load <16 x i8>, <16 x i8>* %b ; CHECK-DAG: ld.b [[R2:\$w[0-9]+]], 0($6) %3 = srem <16 x i8> %1, %2 ; CHECK-DAG: mod_s.b [[R3:\$w[0-9]+]], [[R1]], [[R2]] @@ -616,9 +616,9 @@ define void @mod_s_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind { define void @mod_s_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind { ; CHECK: mod_s_v8i16: - %1 = load <8 x i16>* %a + %1 = load <8 x i16>, <8 x i16>* %a ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5) - %2 = load <8 x i16>* %b + %2 = load <8 x i16>, <8 x i16>* %b ; CHECK-DAG: ld.h [[R2:\$w[0-9]+]], 0($6) %3 = srem <8 x i16> %1, %2 ; CHECK-DAG: mod_s.h [[R3:\$w[0-9]+]], [[R1]], [[R2]] @@ -632,9 +632,9 @@ define void @mod_s_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind { define void @mod_s_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind { ; CHECK: mod_s_v4i32: - %1 = load <4 x i32>* %a + %1 = load <4 x i32>, <4 x i32>* %a ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5) - %2 = load <4 x i32>* %b + %2 = load <4 x i32>, <4 x i32>* %b ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6) %3 = srem <4 x i32> %1, %2 ; CHECK-DAG: mod_s.w [[R3:\$w[0-9]+]], [[R1]], [[R2]] @@ -648,9 +648,9 @@ define void @mod_s_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind { define void @mod_s_v2i64(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind { ; CHECK: mod_s_v2i64: - %1 = load <2 x i64>* %a + %1 = load <2 x i64>, <2 x i64>* %a ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5) - %2 = load <2 x i64>* %b + %2 = load <2 x i64>, <2 x i64>* %b ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6) %3 = srem <2 x i64> %1, %2 ; CHECK-DAG: mod_s.d [[R3:\$w[0-9]+]], [[R1]], [[R2]] @@ -664,9 +664,9 @@ define void @mod_s_v2i64(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind { define void @mod_u_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind { ; CHECK: mod_u_v16i8: - %1 = load <16 x i8>* %a + %1 = load <16 x i8>, <16 x i8>* %a ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5) - %2 = load <16 x i8>* %b + %2 = load <16 x i8>, <16 x i8>* %b ; CHECK-DAG: ld.b [[R2:\$w[0-9]+]], 0($6) %3 = urem <16 x i8> %1, %2 ; CHECK-DAG: mod_u.b [[R3:\$w[0-9]+]], [[R1]], [[R2]] @@ -680,9 +680,9 @@ define void @mod_u_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind { define void @mod_u_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind { ; CHECK: mod_u_v8i16: - %1 = load <8 x i16>* %a + %1 = load <8 x i16>, <8 x i16>* %a ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5) - %2 = load <8 x i16>* %b + %2 = load <8 x i16>, <8 x i16>* %b ; CHECK-DAG: ld.h [[R2:\$w[0-9]+]], 0($6) %3 = urem <8 x i16> %1, %2 ; CHECK-DAG: mod_u.h [[R3:\$w[0-9]+]], [[R1]], [[R2]] @@ -696,9 +696,9 @@ define void @mod_u_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind { define void @mod_u_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind { ; CHECK: mod_u_v4i32: - %1 = load <4 x i32>* %a + %1 = load <4 x i32>, <4 x i32>* %a ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5) - %2 = load <4 x i32>* %b + %2 = load <4 x i32>, <4 x i32>* %b ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6) %3 = urem <4 x i32> %1, %2 ; CHECK-DAG: mod_u.w [[R3:\$w[0-9]+]], [[R1]], [[R2]] @@ -712,9 +712,9 @@ define void @mod_u_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind { define void @mod_u_v2i64(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind { ; CHECK: mod_u_v2i64: - %1 = load <2 x i64>* %a + %1 = load <2 x i64>, <2 x i64>* %a ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5) - %2 = load <2 x i64>* %b + %2 = load <2 x i64>, <2 x i64>* %b ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6) %3 = urem <2 x i64> %1, %2 ; CHECK-DAG: mod_u.d [[R3:\$w[0-9]+]], [[R1]], [[R2]] diff --git a/test/CodeGen/Mips/msa/arithmetic_float.ll b/test/CodeGen/Mips/msa/arithmetic_float.ll index 9aae284fe5351..d2ead536804ab 100644 --- a/test/CodeGen/Mips/msa/arithmetic_float.ll +++ b/test/CodeGen/Mips/msa/arithmetic_float.ll @@ -4,9 +4,9 @@ define void @add_v4f32(<4 x float>* %c, <4 x float>* %a, <4 x float>* %b) nounwind { ; CHECK: add_v4f32: - %1 = load <4 x float>* %a + %1 = load <4 x float>, <4 x float>* %a ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5) - %2 = load <4 x float>* %b + %2 = load <4 x float>, <4 x float>* %b ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6) %3 = fadd <4 x float> %1, %2 ; CHECK-DAG: fadd.w [[R3:\$w[0-9]+]], [[R1]], [[R2]] @@ -20,9 +20,9 @@ define void @add_v4f32(<4 x float>* %c, <4 x float>* %a, <4 x float>* %b) nounwi define void @add_v2f64(<2 x double>* %c, <2 x double>* %a, <2 x double>* %b) nounwind { ; CHECK: add_v2f64: - %1 = load <2 x double>* %a + %1 = load <2 x double>, <2 x double>* %a ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5) - %2 = load <2 x double>* %b + %2 = load <2 x double>, <2 x double>* %b ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6) %3 = fadd <2 x double> %1, %2 ; CHECK-DAG: fadd.d [[R3:\$w[0-9]+]], [[R1]], [[R2]] @@ -36,9 +36,9 @@ define void @add_v2f64(<2 x double>* %c, <2 x double>* %a, <2 x double>* %b) nou define void @sub_v4f32(<4 x float>* %c, <4 x float>* %a, <4 x float>* %b) nounwind { ; CHECK: sub_v4f32: - %1 = load <4 x float>* %a + %1 = load <4 x float>, <4 x float>* %a ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5) - %2 = load <4 x float>* %b + %2 = load <4 x float>, <4 x float>* %b ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6) %3 = fsub <4 x float> %1, %2 ; CHECK-DAG: fsub.w [[R3:\$w[0-9]+]], [[R1]], [[R2]] @@ -52,9 +52,9 @@ define void @sub_v4f32(<4 x float>* %c, <4 x float>* %a, <4 x float>* %b) nounwi define void @sub_v2f64(<2 x double>* %c, <2 x double>* %a, <2 x double>* %b) nounwind { ; CHECK: sub_v2f64: - %1 = load <2 x double>* %a + %1 = load <2 x double>, <2 x double>* %a ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5) - %2 = load <2 x double>* %b + %2 = load <2 x double>, <2 x double>* %b ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6) %3 = fsub <2 x double> %1, %2 ; CHECK-DAG: fsub.d [[R3:\$w[0-9]+]], [[R1]], [[R2]] @@ -68,9 +68,9 @@ define void @sub_v2f64(<2 x double>* %c, <2 x double>* %a, <2 x double>* %b) nou define void @mul_v4f32(<4 x float>* %c, <4 x float>* %a, <4 x float>* %b) nounwind { ; CHECK: mul_v4f32: - %1 = load <4 x float>* %a + %1 = load <4 x float>, <4 x float>* %a ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5) - %2 = load <4 x float>* %b + %2 = load <4 x float>, <4 x float>* %b ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6) %3 = fmul <4 x float> %1, %2 ; CHECK-DAG: fmul.w [[R3:\$w[0-9]+]], [[R1]], [[R2]] @@ -84,9 +84,9 @@ define void @mul_v4f32(<4 x float>* %c, <4 x float>* %a, <4 x float>* %b) nounwi define void @mul_v2f64(<2 x double>* %c, <2 x double>* %a, <2 x double>* %b) nounwind { ; CHECK: mul_v2f64: - %1 = load <2 x double>* %a + %1 = load <2 x double>, <2 x double>* %a ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5) - %2 = load <2 x double>* %b + %2 = load <2 x double>, <2 x double>* %b ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6) %3 = fmul <2 x double> %1, %2 ; CHECK-DAG: fmul.d [[R3:\$w[0-9]+]], [[R1]], [[R2]] @@ -101,11 +101,11 @@ define void @fma_v4f32(<4 x float>* %d, <4 x float>* %a, <4 x float>* %b, <4 x float>* %c) nounwind { ; CHECK: fma_v4f32: - %1 = load <4 x float>* %a + %1 = load <4 x float>, <4 x float>* %a ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5) - %2 = load <4 x float>* %b + %2 = load <4 x float>, <4 x float>* %b ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6) - %3 = load <4 x float>* %c + %3 = load <4 x float>, <4 x float>* %c ; CHECK-DAG: ld.w [[R3:\$w[0-9]+]], 0($7) %4 = tail call <4 x float> @llvm.fma.v4f32 (<4 x float> %1, <4 x float> %2, <4 x float> %3) @@ -121,11 +121,11 @@ define void @fma_v2f64(<2 x double>* %d, <2 x double>* %a, <2 x double>* %b, <2 x double>* %c) nounwind { ; CHECK: fma_v2f64: - %1 = load <2 x double>* %a + %1 = load <2 x double>, <2 x double>* %a ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5) - %2 = load <2 x double>* %b + %2 = load <2 x double>, <2 x double>* %b ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6) - %3 = load <2 x double>* %c + %3 = load <2 x double>, <2 x double>* %c ; CHECK-DAG: ld.d [[R3:\$w[0-9]+]], 0($7) %4 = tail call <2 x double> @llvm.fma.v2f64 (<2 x double> %1, <2 x double> %2, <2 x double> %3) @@ -141,11 +141,11 @@ define void @fmsub_v4f32(<4 x float>* %d, <4 x float>* %a, <4 x float>* %b, <4 x float>* %c) nounwind { ; CHECK: fmsub_v4f32: - %1 = load <4 x float>* %a + %1 = load <4 x float>, <4 x float>* %a ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5) - %2 = load <4 x float>* %b + %2 = load <4 x float>, <4 x float>* %b ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6) - %3 = load <4 x float>* %c + %3 = load <4 x float>, <4 x float>* %c ; CHECK-DAG: ld.w [[R3:\$w[0-9]+]], 0($7) %4 = fmul <4 x float> %2, %3 %5 = fsub <4 x float> %1, %4 @@ -161,11 +161,11 @@ define void @fmsub_v2f64(<2 x double>* %d, <2 x double>* %a, <2 x double>* %b, <2 x double>* %c) nounwind { ; CHECK: fmsub_v2f64: - %1 = load <2 x double>* %a + %1 = load <2 x double>, <2 x double>* %a ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5) - %2 = load <2 x double>* %b + %2 = load <2 x double>, <2 x double>* %b ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6) - %3 = load <2 x double>* %c + %3 = load <2 x double>, <2 x double>* %c ; CHECK-DAG: ld.d [[R3:\$w[0-9]+]], 0($7) %4 = fmul <2 x double> %2, %3 %5 = fsub <2 x double> %1, %4 @@ -180,9 +180,9 @@ define void @fmsub_v2f64(<2 x double>* %d, <2 x double>* %a, <2 x double>* %b, define void @fdiv_v4f32(<4 x float>* %c, <4 x float>* %a, <4 x float>* %b) nounwind { ; CHECK: fdiv_v4f32: - %1 = load <4 x float>* %a + %1 = load <4 x float>, <4 x float>* %a ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5) - %2 = load <4 x float>* %b + %2 = load <4 x float>, <4 x float>* %b ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6) %3 = fdiv <4 x float> %1, %2 ; CHECK-DAG: fdiv.w [[R3:\$w[0-9]+]], [[R1]], [[R2]] @@ -196,9 +196,9 @@ define void @fdiv_v4f32(<4 x float>* %c, <4 x float>* %a, <4 x float>* %b) nounw define void @fdiv_v2f64(<2 x double>* %c, <2 x double>* %a, <2 x double>* %b) nounwind { ; CHECK: fdiv_v2f64: - %1 = load <2 x double>* %a + %1 = load <2 x double>, <2 x double>* %a ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5) - %2 = load <2 x double>* %b + %2 = load <2 x double>, <2 x double>* %b ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6) %3 = fdiv <2 x double> %1, %2 ; CHECK-DAG: fdiv.d [[R3:\$w[0-9]+]], [[R1]], [[R2]] @@ -212,7 +212,7 @@ define void @fdiv_v2f64(<2 x double>* %c, <2 x double>* %a, <2 x double>* %b) no define void @fabs_v4f32(<4 x float>* %c, <4 x float>* %a) nounwind { ; CHECK: fabs_v4f32: - %1 = load <4 x float>* %a + %1 = load <4 x float>, <4 x float>* %a ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5) %2 = tail call <4 x float> @llvm.fabs.v4f32 (<4 x float> %1) ; CHECK-DAG: fmax_a.w [[R3:\$w[0-9]+]], [[R1]], [[R1]] @@ -226,7 +226,7 @@ define void @fabs_v4f32(<4 x float>* %c, <4 x float>* %a) nounwind { define void @fabs_v2f64(<2 x double>* %c, <2 x double>* %a) nounwind { ; CHECK: fabs_v2f64: - %1 = load <2 x double>* %a + %1 = load <2 x double>, <2 x double>* %a ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5) %2 = tail call <2 x double> @llvm.fabs.v2f64 (<2 x double> %1) ; CHECK-DAG: fmax_a.d [[R3:\$w[0-9]+]], [[R1]], [[R1]] @@ -240,7 +240,7 @@ define void @fabs_v2f64(<2 x double>* %c, <2 x double>* %a) nounwind { define void @fexp2_v4f32(<4 x float>* %c, <4 x float>* %a) nounwind { ; CHECK: fexp2_v4f32: - %1 = load <4 x float>* %a + %1 = load <4 x float>, <4 x float>* %a ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5) %2 = tail call <4 x float> @llvm.exp2.v4f32 (<4 x float> %1) ; CHECK-DAG: ldi.w [[R3:\$w[0-9]+]], 1 @@ -256,7 +256,7 @@ define void @fexp2_v4f32(<4 x float>* %c, <4 x float>* %a) nounwind { define void @fexp2_v2f64(<2 x double>* %c, <2 x double>* %a) nounwind { ; CHECK: fexp2_v2f64: - %1 = load <2 x double>* %a + %1 = load <2 x double>, <2 x double>* %a ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5) %2 = tail call <2 x double> @llvm.exp2.v2f64 (<2 x double> %1) ; CHECK-DAG: ldi.d [[R3:\$w[0-9]+]], 1 @@ -272,7 +272,7 @@ define void @fexp2_v2f64(<2 x double>* %c, <2 x double>* %a) nounwind { define void @fexp2_v4f32_2(<4 x float>* %c, <4 x float>* %a) nounwind { ; CHECK: fexp2_v4f32_2: - %1 = load <4 x float>* %a + %1 = load <4 x float>, <4 x float>* %a ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5) %2 = tail call <4 x float> @llvm.exp2.v4f32 (<4 x float> %1) %3 = fmul <4 x float> <float 2.0, float 2.0, float 2.0, float 2.0>, %2 @@ -289,7 +289,7 @@ define void @fexp2_v4f32_2(<4 x float>* %c, <4 x float>* %a) nounwind { define void @fexp2_v2f64_2(<2 x double>* %c, <2 x double>* %a) nounwind { ; CHECK: fexp2_v2f64_2: - %1 = load <2 x double>* %a + %1 = load <2 x double>, <2 x double>* %a ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5) %2 = tail call <2 x double> @llvm.exp2.v2f64 (<2 x double> %1) %3 = fmul <2 x double> <double 2.0, double 2.0>, %2 @@ -306,7 +306,7 @@ define void @fexp2_v2f64_2(<2 x double>* %c, <2 x double>* %a) nounwind { define void @fsqrt_v4f32(<4 x float>* %c, <4 x float>* %a) nounwind { ; CHECK: fsqrt_v4f32: - %1 = load <4 x float>* %a + %1 = load <4 x float>, <4 x float>* %a ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5) %2 = tail call <4 x float> @llvm.sqrt.v4f32 (<4 x float> %1) ; CHECK-DAG: fsqrt.w [[R3:\$w[0-9]+]], [[R1]] @@ -320,7 +320,7 @@ define void @fsqrt_v4f32(<4 x float>* %c, <4 x float>* %a) nounwind { define void @fsqrt_v2f64(<2 x double>* %c, <2 x double>* %a) nounwind { ; CHECK: fsqrt_v2f64: - %1 = load <2 x double>* %a + %1 = load <2 x double>, <2 x double>* %a ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5) %2 = tail call <2 x double> @llvm.sqrt.v2f64 (<2 x double> %1) ; CHECK-DAG: fsqrt.d [[R3:\$w[0-9]+]], [[R1]] @@ -334,7 +334,7 @@ define void @fsqrt_v2f64(<2 x double>* %c, <2 x double>* %a) nounwind { define void @ffint_u_v4f32(<4 x float>* %c, <4 x i32>* %a) nounwind { ; CHECK: ffint_u_v4f32: - %1 = load <4 x i32>* %a + %1 = load <4 x i32>, <4 x i32>* %a ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5) %2 = uitofp <4 x i32> %1 to <4 x float> ; CHECK-DAG: ffint_u.w [[R3:\$w[0-9]+]], [[R1]] @@ -348,7 +348,7 @@ define void @ffint_u_v4f32(<4 x float>* %c, <4 x i32>* %a) nounwind { define void @ffint_u_v2f64(<2 x double>* %c, <2 x i64>* %a) nounwind { ; CHECK: ffint_u_v2f64: - %1 = load <2 x i64>* %a + %1 = load <2 x i64>, <2 x i64>* %a ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5) %2 = uitofp <2 x i64> %1 to <2 x double> ; CHECK-DAG: ffint_u.d [[R3:\$w[0-9]+]], [[R1]] @@ -362,7 +362,7 @@ define void @ffint_u_v2f64(<2 x double>* %c, <2 x i64>* %a) nounwind { define void @ffint_s_v4f32(<4 x float>* %c, <4 x i32>* %a) nounwind { ; CHECK: ffint_s_v4f32: - %1 = load <4 x i32>* %a + %1 = load <4 x i32>, <4 x i32>* %a ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5) %2 = sitofp <4 x i32> %1 to <4 x float> ; CHECK-DAG: ffint_s.w [[R3:\$w[0-9]+]], [[R1]] @@ -376,7 +376,7 @@ define void @ffint_s_v4f32(<4 x float>* %c, <4 x i32>* %a) nounwind { define void @ffint_s_v2f64(<2 x double>* %c, <2 x i64>* %a) nounwind { ; CHECK: ffint_s_v2f64: - %1 = load <2 x i64>* %a + %1 = load <2 x i64>, <2 x i64>* %a ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5) %2 = sitofp <2 x i64> %1 to <2 x double> ; CHECK-DAG: ffint_s.d [[R3:\$w[0-9]+]], [[R1]] @@ -390,7 +390,7 @@ define void @ffint_s_v2f64(<2 x double>* %c, <2 x i64>* %a) nounwind { define void @ftrunc_u_v4f32(<4 x i32>* %c, <4 x float>* %a) nounwind { ; CHECK: ftrunc_u_v4f32: - %1 = load <4 x float>* %a + %1 = load <4 x float>, <4 x float>* %a ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5) %2 = fptoui <4 x float> %1 to <4 x i32> ; CHECK-DAG: ftrunc_u.w [[R3:\$w[0-9]+]], [[R1]] @@ -404,7 +404,7 @@ define void @ftrunc_u_v4f32(<4 x i32>* %c, <4 x float>* %a) nounwind { define void @ftrunc_u_v2f64(<2 x i64>* %c, <2 x double>* %a) nounwind { ; CHECK: ftrunc_u_v2f64: - %1 = load <2 x double>* %a + %1 = load <2 x double>, <2 x double>* %a ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5) %2 = fptoui <2 x double> %1 to <2 x i64> ; CHECK-DAG: ftrunc_u.d [[R3:\$w[0-9]+]], [[R1]] @@ -418,7 +418,7 @@ define void @ftrunc_u_v2f64(<2 x i64>* %c, <2 x double>* %a) nounwind { define void @ftrunc_s_v4f32(<4 x i32>* %c, <4 x float>* %a) nounwind { ; CHECK: ftrunc_s_v4f32: - %1 = load <4 x float>* %a + %1 = load <4 x float>, <4 x float>* %a ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5) %2 = fptosi <4 x float> %1 to <4 x i32> ; CHECK-DAG: ftrunc_s.w [[R3:\$w[0-9]+]], [[R1]] @@ -432,7 +432,7 @@ define void @ftrunc_s_v4f32(<4 x i32>* %c, <4 x float>* %a) nounwind { define void @ftrunc_s_v2f64(<2 x i64>* %c, <2 x double>* %a) nounwind { ; CHECK: ftrunc_s_v2f64: - %1 = load <2 x double>* %a + %1 = load <2 x double>, <2 x double>* %a ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5) %2 = fptosi <2 x double> %1 to <2 x i64> ; CHECK-DAG: ftrunc_s.d [[R3:\$w[0-9]+]], [[R1]] diff --git a/test/CodeGen/Mips/msa/basic_operations.ll b/test/CodeGen/Mips/msa/basic_operations.ll index dbdf42be49ca4..2efec2911935f 100644 --- a/test/CodeGen/Mips/msa/basic_operations.ll +++ b/test/CodeGen/Mips/msa/basic_operations.ll @@ -1,5 +1,9 @@ -; RUN: llc -march=mips -mattr=+msa,+fp64 < %s | FileCheck -check-prefix=MIPS32-AE -check-prefix=MIPS32-BE %s -; RUN: llc -march=mipsel -mattr=+msa,+fp64 < %s | FileCheck -check-prefix=MIPS32-AE -check-prefix=MIPS32-LE %s +; RUN: llc -march=mips -mattr=+msa,+fp64 < %s | FileCheck -check-prefix=ALL -check-prefix=O32 -check-prefix=MIPS32 -check-prefix=ALL-BE %s +; RUN: llc -march=mipsel -mattr=+msa,+fp64 < %s | FileCheck -check-prefix=ALL -check-prefix=O32 -check-prefix=MIPS32 -check-prefix=ALL-LE %s +; RUN: llc -march=mips64 -target-abi n32 -mattr=+msa,+fp64 < %s | FileCheck -check-prefix=ALL -check-prefix=N32 -check-prefix=MIPS64 -check-prefix=ALL-BE %s +; RUN: llc -march=mips64el -target-abi n32 -mattr=+msa,+fp64 < %s | FileCheck -check-prefix=ALL -check-prefix=N32 -check-prefix=MIPS64 -check-prefix=ALL-LE %s +; RUN: llc -march=mips64 -mattr=+msa,+fp64 < %s | FileCheck -check-prefix=ALL -check-prefix=N64 -check-prefix=MIPS64 -check-prefix=ALL-BE %s +; RUN: llc -march=mips64el -mattr=+msa,+fp64 < %s | FileCheck -check-prefix=ALL -check-prefix=N64 -check-prefix=MIPS64 -check-prefix=ALL-LE %s @v4i8 = global <4 x i8> <i8 0, i8 0, i8 0, i8 0> @v16i8 = global <16 x i8> <i8 0, i8 0, i8 0, i8 0, i8 0, i8 0, i8 0, i8 0, i8 0, i8 0, i8 0, i8 0, i8 0, i8 0, i8 0, i8 0> @@ -10,135 +14,153 @@ @i64 = global i64 0 define void @const_v16i8() nounwind { - ; MIPS32-AE-LABEL: const_v16i8: + ; ALL-LABEL: const_v16i8: store volatile <16 x i8> <i8 0, i8 0, i8 0, i8 0, i8 0, i8 0, i8 0, i8 0, i8 0, i8 0, i8 0, i8 0, i8 0, i8 0, i8 0, i8 0>, <16 x i8>*@v16i8 - ; MIPS32-AE: ldi.b [[R1:\$w[0-9]+]], 0 + ; ALL: ldi.b [[R1:\$w[0-9]+]], 0 store volatile <16 x i8> <i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1>, <16 x i8>*@v16i8 - ; MIPS32-AE: ldi.b [[R1:\$w[0-9]+]], 1 + ; ALL: ldi.b [[R1:\$w[0-9]+]], 1 store volatile <16 x i8> <i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 31>, <16 x i8>*@v16i8 - ; MIPS32-AE: addiu [[G_PTR:\$[0-9]+]], {{.*}}, %lo($ - ; MIPS32-AE: ld.b [[R1:\$w[0-9]+]], 0([[G_PTR]]) + ; O32: addiu [[G_PTR:\$[0-9]+]], {{.*}}, %lo($ + ; N32: addiu [[G_PTR:\$[0-9]+]], {{.*}}, %got_ofst($ + ; N64: daddiu [[G_PTR:\$[0-9]+]], {{.*}}, %got_ofst($ + ; ALL: ld.b [[R1:\$w[0-9]+]], 0([[G_PTR]]) store volatile <16 x i8> <i8 1, i8 2, i8 3, i8 4, i8 5, i8 6, i8 7, i8 8, i8 9, i8 0, i8 1, i8 2, i8 3, i8 4, i8 5, i8 6>, <16 x i8>*@v16i8 - ; MIPS32-AE: addiu [[G_PTR:\$[0-9]+]], {{.*}}, %lo($ - ; MIPS32-AE: ld.b [[R1:\$w[0-9]+]], 0([[G_PTR]]) + ; O32: addiu [[G_PTR:\$[0-9]+]], {{.*}}, %lo($ + ; N32: addiu [[G_PTR:\$[0-9]+]], {{.*}}, %got_ofst($ + ; N64: daddiu [[G_PTR:\$[0-9]+]], {{.*}}, %got_ofst($ + ; ALL: ld.b [[R1:\$w[0-9]+]], 0([[G_PTR]]) store volatile <16 x i8> <i8 1, i8 0, i8 1, i8 0, i8 1, i8 0, i8 1, i8 0, i8 1, i8 0, i8 1, i8 0, i8 1, i8 0, i8 1, i8 0>, <16 x i8>*@v16i8 - ; MIPS32-BE: ldi.h [[R1:\$w[0-9]+]], 256 - ; MIPS32-LE: ldi.h [[R1:\$w[0-9]+]], 1 + ; ALL-BE: ldi.h [[R1:\$w[0-9]+]], 256 + ; ALL-LE: ldi.h [[R1:\$w[0-9]+]], 1 store volatile <16 x i8> <i8 1, i8 2, i8 3, i8 4, i8 1, i8 2, i8 3, i8 4, i8 1, i8 2, i8 3, i8 4, i8 1, i8 2, i8 3, i8 4>, <16 x i8>*@v16i8 - ; MIPS32-BE-DAG: lui [[R2:\$[0-9]+]], 258 - ; MIPS32-LE-DAG: lui [[R2:\$[0-9]+]], 1027 - ; MIPS32-BE-DAG: ori [[R2]], [[R2]], 772 - ; MIPS32-LE-DAG: ori [[R2]], [[R2]], 513 - ; MIPS32-AE-DAG: fill.w [[R1:\$w[0-9]+]], [[R2]] + ; ALL-BE-DAG: lui [[R2:\$[0-9]+]], 258 + ; ALL-LE-DAG: lui [[R2:\$[0-9]+]], 1027 + ; ALL-BE-DAG: ori [[R2]], [[R2]], 772 + ; ALL-LE-DAG: ori [[R2]], [[R2]], 513 + ; ALL-DAG: fill.w [[R1:\$w[0-9]+]], [[R2]] store volatile <16 x i8> <i8 1, i8 2, i8 3, i8 4, i8 5, i8 6, i8 7, i8 8, i8 1, i8 2, i8 3, i8 4, i8 5, i8 6, i8 7, i8 8>, <16 x i8>*@v16i8 - ; MIPS32-AE: addiu [[G_PTR:\$[0-9]+]], {{.*}}, %lo($ - ; MIPS32-AE: ld.b [[R1:\$w[0-9]+]], 0([[G_PTR]]) + ; O32: addiu [[G_PTR:\$[0-9]+]], {{.*}}, %lo($ + ; N32: addiu [[G_PTR:\$[0-9]+]], {{.*}}, %got_ofst($ + ; N64: daddiu [[G_PTR:\$[0-9]+]], {{.*}}, %got_ofst($ + ; ALL: ld.b [[R1:\$w[0-9]+]], 0([[G_PTR]]) ret void - ; MIPS32-AE: .size const_v16i8 } define void @const_v8i16() nounwind { - ; MIPS32-AE-LABEL: const_v8i16: + ; ALL-LABEL: const_v8i16: store volatile <8 x i16> <i16 0, i16 0, i16 0, i16 0, i16 0, i16 0, i16 0, i16 0>, <8 x i16>*@v8i16 - ; MIPS32-AE: ldi.b [[R1:\$w[0-9]+]], 0 + ; ALL: ldi.b [[R1:\$w[0-9]+]], 0 store volatile <8 x i16> <i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1>, <8 x i16>*@v8i16 - ; MIPS32-AE: ldi.h [[R1:\$w[0-9]+]], 1 + ; ALL: ldi.h [[R1:\$w[0-9]+]], 1 store volatile <8 x i16> <i16 1, i16 1, i16 1, i16 2, i16 1, i16 1, i16 1, i16 31>, <8 x i16>*@v8i16 - ; MIPS32-AE: addiu [[G_PTR:\$[0-9]+]], {{.*}}, %lo($ - ; MIPS32-AE: ld.h [[R1:\$w[0-9]+]], 0([[G_PTR]]) + ; O32: addiu [[G_PTR:\$[0-9]+]], {{.*}}, %lo($ + ; N32: addiu [[G_PTR:\$[0-9]+]], {{.*}}, %got_ofst($ + ; N64: daddiu [[G_PTR:\$[0-9]+]], {{.*}}, %got_ofst($ + ; ALL: ld.h [[R1:\$w[0-9]+]], 0([[G_PTR]]) store volatile <8 x i16> <i16 1028, i16 1028, i16 1028, i16 1028, i16 1028, i16 1028, i16 1028, i16 1028>, <8 x i16>*@v8i16 - ; MIPS32-AE: ldi.b [[R1:\$w[0-9]+]], 4 + ; ALL: ldi.b [[R1:\$w[0-9]+]], 4 store volatile <8 x i16> <i16 1, i16 2, i16 1, i16 2, i16 1, i16 2, i16 1, i16 2>, <8 x i16>*@v8i16 - ; MIPS32-BE-DAG: lui [[R2:\$[0-9]+]], 1 - ; MIPS32-LE-DAG: lui [[R2:\$[0-9]+]], 2 - ; MIPS32-BE-DAG: ori [[R2]], [[R2]], 2 - ; MIPS32-LE-DAG: ori [[R2]], [[R2]], 1 - ; MIPS32-AE-DAG: fill.w [[R1:\$w[0-9]+]], [[R2]] + ; ALL-BE-DAG: lui [[R2:\$[0-9]+]], 1 + ; ALL-LE-DAG: lui [[R2:\$[0-9]+]], 2 + ; ALL-BE-DAG: ori [[R2]], [[R2]], 2 + ; ALL-LE-DAG: ori [[R2]], [[R2]], 1 + ; ALL-DAG: fill.w [[R1:\$w[0-9]+]], [[R2]] store volatile <8 x i16> <i16 1, i16 2, i16 3, i16 4, i16 1, i16 2, i16 3, i16 4>, <8 x i16>*@v8i16 - ; MIPS32-AE: addiu [[G_PTR:\$[0-9]+]], {{.*}}, %lo($ - ; MIPS32-AE: ld.h [[R1:\$w[0-9]+]], 0([[G_PTR]]) + ; O32: addiu [[G_PTR:\$[0-9]+]], {{.*}}, %lo($ + ; N32: addiu [[G_PTR:\$[0-9]+]], {{.*}}, %got_ofst($ + ; N64: daddiu [[G_PTR:\$[0-9]+]], {{.*}}, %got_ofst($ + ; ALL: ld.h [[R1:\$w[0-9]+]], 0([[G_PTR]]) ret void - ; MIPS32-AE: .size const_v8i16 } define void @const_v4i32() nounwind { - ; MIPS32-AE-LABEL: const_v4i32: + ; ALL-LABEL: const_v4i32: store volatile <4 x i32> <i32 0, i32 0, i32 0, i32 0>, <4 x i32>*@v4i32 - ; MIPS32-AE: ldi.b [[R1:\$w[0-9]+]], 0 + ; ALL: ldi.b [[R1:\$w[0-9]+]], 0 store volatile <4 x i32> <i32 1, i32 1, i32 1, i32 1>, <4 x i32>*@v4i32 - ; MIPS32-AE: ldi.w [[R1:\$w[0-9]+]], 1 + ; ALL: ldi.w [[R1:\$w[0-9]+]], 1 store volatile <4 x i32> <i32 1, i32 1, i32 1, i32 31>, <4 x i32>*@v4i32 - ; MIPS32-AE: addiu [[G_PTR:\$[0-9]+]], {{.*}}, %lo($ - ; MIPS32-AE: ld.w [[R1:\$w[0-9]+]], 0([[G_PTR]]) + ; O32: addiu [[G_PTR:\$[0-9]+]], {{.*}}, %lo($ + ; N32: addiu [[G_PTR:\$[0-9]+]], {{.*}}, %got_ofst($ + ; N64: daddiu [[G_PTR:\$[0-9]+]], {{.*}}, %got_ofst($ + ; ALL: ld.w [[R1:\$w[0-9]+]], 0([[G_PTR]]) store volatile <4 x i32> <i32 16843009, i32 16843009, i32 16843009, i32 16843009>, <4 x i32>*@v4i32 - ; MIPS32-AE: ldi.b [[R1:\$w[0-9]+]], 1 + ; ALL: ldi.b [[R1:\$w[0-9]+]], 1 store volatile <4 x i32> <i32 65537, i32 65537, i32 65537, i32 65537>, <4 x i32>*@v4i32 - ; MIPS32-AE: ldi.h [[R1:\$w[0-9]+]], 1 + ; ALL: ldi.h [[R1:\$w[0-9]+]], 1 store volatile <4 x i32> <i32 1, i32 2, i32 1, i32 2>, <4 x i32>*@v4i32 - ; MIPS32-AE: addiu [[G_PTR:\$[0-9]+]], {{.*}}, %lo($ - ; MIPS32-AE: ld.w [[R1:\$w[0-9]+]], 0([[G_PTR]]) + ; O32: addiu [[G_PTR:\$[0-9]+]], {{.*}}, %lo($ + ; N32: addiu [[G_PTR:\$[0-9]+]], {{.*}}, %got_ofst($ + ; N64: daddiu [[G_PTR:\$[0-9]+]], {{.*}}, %got_ofst($ + ; ALL: ld.w [[R1:\$w[0-9]+]], 0([[G_PTR]]) store volatile <4 x i32> <i32 3, i32 4, i32 5, i32 6>, <4 x i32>*@v4i32 - ; MIPS32-AE: addiu [[G_PTR:\$[0-9]+]], {{.*}}, %lo($ - ; MIPS32-AE: ld.w [[R1:\$w[0-9]+]], 0([[G_PTR]]) + ; O32: addiu [[G_PTR:\$[0-9]+]], {{.*}}, %lo($ + ; N32: addiu [[G_PTR:\$[0-9]+]], {{.*}}, %got_ofst($ + ; N64: daddiu [[G_PTR:\$[0-9]+]], {{.*}}, %got_ofst($ + ; ALL: ld.w [[R1:\$w[0-9]+]], 0([[G_PTR]]) ret void - ; MIPS32-AE: .size const_v4i32 } define void @const_v2i64() nounwind { - ; MIPS32-AE-LABEL: const_v2i64: + ; ALL-LABEL: const_v2i64: store volatile <2 x i64> <i64 0, i64 0>, <2 x i64>*@v2i64 - ; MIPS32-AE: ldi.b [[R1:\$w[0-9]+]], 0 + ; ALL: ldi.b [[R1:\$w[0-9]+]], 0 store volatile <2 x i64> <i64 72340172838076673, i64 72340172838076673>, <2 x i64>*@v2i64 - ; MIPS32-AE: ldi.b [[R1:\$w[0-9]+]], 1 + ; ALL: ldi.b [[R1:\$w[0-9]+]], 1 store volatile <2 x i64> <i64 281479271743489, i64 281479271743489>, <2 x i64>*@v2i64 - ; MIPS32-AE: ldi.h [[R1:\$w[0-9]+]], 1 + ; ALL: ldi.h [[R1:\$w[0-9]+]], 1 store volatile <2 x i64> <i64 4294967297, i64 4294967297>, <2 x i64>*@v2i64 - ; MIPS32-AE: ldi.w [[R1:\$w[0-9]+]], 1 + ; ALL: ldi.w [[R1:\$w[0-9]+]], 1 store volatile <2 x i64> <i64 1, i64 1>, <2 x i64>*@v2i64 - ; MIPS32-AE: ldi.d [[R1:\$w[0-9]+]], 1 + ; ALL: ldi.d [[R1:\$w[0-9]+]], 1 store volatile <2 x i64> <i64 1, i64 31>, <2 x i64>*@v2i64 - ; MIPS32-AE: addiu [[G_PTR:\$[0-9]+]], {{.*}}, %lo($ - ; MIPS32-AE: ld.w [[R1:\$w[0-9]+]], 0([[G_PTR]]) + ; O32: addiu [[G_PTR:\$[0-9]+]], {{.*}}, %lo($ + ; N32: addiu [[G_PTR:\$[0-9]+]], {{.*}}, %got_ofst($ + ; N64: daddiu [[G_PTR:\$[0-9]+]], {{.*}}, %got_ofst($ + ; MIPS32: ld.w [[R1:\$w[0-9]+]], 0([[G_PTR]]) + ; MIPS64: ld.d [[R1:\$w[0-9]+]], 0([[G_PTR]]) store volatile <2 x i64> <i64 3, i64 4>, <2 x i64>*@v2i64 - ; MIPS32-AE: addiu [[G_PTR:\$[0-9]+]], {{.*}}, %lo($ - ; MIPS32-AE: ld.w [[R1:\$w[0-9]+]], 0([[G_PTR]]) + ; O32: addiu [[G_PTR:\$[0-9]+]], {{.*}}, %lo($ + ; N32: addiu [[G_PTR:\$[0-9]+]], {{.*}}, %got_ofst($ + ; N64: daddiu [[G_PTR:\$[0-9]+]], {{.*}}, %got_ofst($ + ; MIPS32: ld.w [[R1:\$w[0-9]+]], 0([[G_PTR]]) + ; MIPS64: ld.d [[R1:\$w[0-9]+]], 0([[G_PTR]]) ret void - ; MIPS32-AE: .size const_v2i64 } -define void @nonconst_v16i8(i8 %a, i8 %b, i8 %c, i8 %d, i8 %e, i8 %f, i8 %g, i8 %h) nounwind { - ; MIPS32-AE-LABEL: nonconst_v16i8: +define void @nonconst_v16i8(i8 signext %a, i8 signext %b, i8 signext %c, i8 signext %d, i8 signext %e, i8 signext %f, i8 signext %g, i8 signext %h) nounwind { + ; ALL-LABEL: nonconst_v16i8: %1 = insertelement <16 x i8> undef, i8 %a, i32 0 %2 = insertelement <16 x i8> %1, i8 %b, i32 1 @@ -156,39 +178,38 @@ define void @nonconst_v16i8(i8 %a, i8 %b, i8 %c, i8 %d, i8 %e, i8 %f, i8 %g, i8 %14 = insertelement <16 x i8> %13, i8 %h, i32 13 %15 = insertelement <16 x i8> %14, i8 %h, i32 14 %16 = insertelement <16 x i8> %15, i8 %h, i32 15 - ; MIPS32-AE-DAG: insert.b [[R1:\$w[0-9]+]][0], $4 - ; MIPS32-AE-DAG: insert.b [[R1]][1], $5 - ; MIPS32-AE-DAG: insert.b [[R1]][2], $6 - ; MIPS32-AE-DAG: insert.b [[R1]][3], $7 - ; MIPS32-BE-DAG: lbu [[R2:\$[0-9]+]], 19($sp) - ; MIPS32-LE-DAG: lbu [[R2:\$[0-9]+]], 16($sp) - ; MIPS32-AE-DAG: insert.b [[R1]][4], [[R2]] - ; MIPS32-BE-DAG: lbu [[R3:\$[0-9]+]], 23($sp) - ; MIPS32-LE-DAG: lbu [[R3:\$[0-9]+]], 20($sp) - ; MIPS32-AE-DAG: insert.b [[R1]][5], [[R3]] - ; MIPS32-BE-DAG: lbu [[R4:\$[0-9]+]], 27($sp) - ; MIPS32-LE-DAG: lbu [[R4:\$[0-9]+]], 24($sp) - ; MIPS32-AE-DAG: insert.b [[R1]][6], [[R4]] - ; MIPS32-BE-DAG: lbu [[R5:\$[0-9]+]], 31($sp) - ; MIPS32-LE-DAG: lbu [[R5:\$[0-9]+]], 28($sp) - ; MIPS32-AE-DAG: insert.b [[R1]][7], [[R5]] - ; MIPS32-AE-DAG: insert.b [[R1]][8], [[R5]] - ; MIPS32-AE-DAG: insert.b [[R1]][9], [[R5]] - ; MIPS32-AE-DAG: insert.b [[R1]][10], [[R5]] - ; MIPS32-AE-DAG: insert.b [[R1]][11], [[R5]] - ; MIPS32-AE-DAG: insert.b [[R1]][12], [[R5]] - ; MIPS32-AE-DAG: insert.b [[R1]][13], [[R5]] - ; MIPS32-AE-DAG: insert.b [[R1]][14], [[R5]] - ; MIPS32-AE-DAG: insert.b [[R1]][15], [[R5]] + ; ALL-DAG: insert.b [[R1:\$w[0-9]+]][0], $4 + ; ALL-DAG: insert.b [[R1]][1], $5 + ; ALL-DAG: insert.b [[R1]][2], $6 + ; ALL-DAG: insert.b [[R1]][3], $7 + ; MIPS32-DAG: lw [[R2:\$[0-9]+]], 16($sp) + ; MIPS32-DAG: insert.b [[R1]][4], [[R2]] + ; MIPS64-DAG: insert.b [[R1]][4], $8 + ; MIPS32-DAG: lw [[R3:\$[0-9]+]], 20($sp) + ; MIPS32-DAG: insert.b [[R1]][5], [[R3]] + ; MIPS64-DAG: insert.b [[R1]][5], $9 + ; MIPS32-DAG: lw [[R4:\$[0-9]+]], 24($sp) + ; MIPS32-DAG: insert.b [[R1]][6], [[R4]] + ; MIPS64-DAG: insert.b [[R1]][6], $10 + ; MIPS32-DAG: lw [[R5:\$[0-9]+]], 28($sp) + ; MIPS32-DAG: insert.b [[R1]][7], [[R5]] + ; MIPS64-DAG: insert.b [[R1]][7], [[R5:\$11]] + ; ALL-DAG: insert.b [[R1]][8], [[R5]] + ; ALL-DAG: insert.b [[R1]][9], [[R5]] + ; ALL-DAG: insert.b [[R1]][10], [[R5]] + ; ALL-DAG: insert.b [[R1]][11], [[R5]] + ; ALL-DAG: insert.b [[R1]][12], [[R5]] + ; ALL-DAG: insert.b [[R1]][13], [[R5]] + ; ALL-DAG: insert.b [[R1]][14], [[R5]] + ; ALL-DAG: insert.b [[R1]][15], [[R5]] store volatile <16 x i8> %16, <16 x i8>*@v16i8 ret void - ; MIPS32-AE: .size nonconst_v16i8 } -define void @nonconst_v8i16(i16 %a, i16 %b, i16 %c, i16 %d, i16 %e, i16 %f, i16 %g, i16 %h) nounwind { - ; MIPS32-AE-LABEL: nonconst_v8i16: +define void @nonconst_v8i16(i16 signext %a, i16 signext %b, i16 signext %c, i16 signext %d, i16 signext %e, i16 signext %f, i16 signext %g, i16 signext %h) nounwind { + ; ALL-LABEL: nonconst_v8i16: %1 = insertelement <8 x i16> undef, i16 %a, i32 0 %2 = insertelement <8 x i16> %1, i16 %b, i32 1 @@ -198,610 +219,648 @@ define void @nonconst_v8i16(i16 %a, i16 %b, i16 %c, i16 %d, i16 %e, i16 %f, i16 %6 = insertelement <8 x i16> %5, i16 %f, i32 5 %7 = insertelement <8 x i16> %6, i16 %g, i32 6 %8 = insertelement <8 x i16> %7, i16 %h, i32 7 - ; MIPS32-AE-DAG: insert.h [[R1:\$w[0-9]+]][0], $4 - ; MIPS32-AE-DAG: insert.h [[R1]][1], $5 - ; MIPS32-AE-DAG: insert.h [[R1]][2], $6 - ; MIPS32-AE-DAG: insert.h [[R1]][3], $7 - ; MIPS32-BE-DAG: lhu [[R2:\$[0-9]+]], 18($sp) - ; MIPS32-LE-DAG: lhu [[R2:\$[0-9]+]], 16($sp) - ; MIPS32-AE-DAG: insert.h [[R1]][4], [[R2]] - ; MIPS32-BE-DAG: lhu [[R2:\$[0-9]+]], 22($sp) - ; MIPS32-LE-DAG: lhu [[R2:\$[0-9]+]], 20($sp) - ; MIPS32-AE-DAG: insert.h [[R1]][5], [[R2]] - ; MIPS32-BE-DAG: lhu [[R2:\$[0-9]+]], 26($sp) - ; MIPS32-LE-DAG: lhu [[R2:\$[0-9]+]], 24($sp) - ; MIPS32-AE-DAG: insert.h [[R1]][6], [[R2]] - ; MIPS32-BE-DAG: lhu [[R2:\$[0-9]+]], 30($sp) - ; MIPS32-LE-DAG: lhu [[R2:\$[0-9]+]], 28($sp) - ; MIPS32-AE-DAG: insert.h [[R1]][7], [[R2]] + ; ALL-DAG: insert.h [[R1:\$w[0-9]+]][0], $4 + ; ALL-DAG: insert.h [[R1]][1], $5 + ; ALL-DAG: insert.h [[R1]][2], $6 + ; ALL-DAG: insert.h [[R1]][3], $7 + ; MIPS32-DAG: lw [[R2:\$[0-9]+]], 16($sp) + ; MIPS32-DAG: insert.h [[R1]][4], [[R2]] + ; MIPS64-DAG: insert.h [[R1]][4], $8 + ; MIPS32-DAG: lw [[R2:\$[0-9]+]], 20($sp) + ; MIPS32-DAG: insert.h [[R1]][5], [[R2]] + ; MIPS64-DAG: insert.h [[R1]][5], $9 + ; MIPS32-DAG: lw [[R2:\$[0-9]+]], 24($sp) + ; MIPS32-DAG: insert.h [[R1]][6], [[R2]] + ; MIPS64-DAG: insert.h [[R1]][6], $10 + ; MIPS32-DAG: lw [[R2:\$[0-9]+]], 28($sp) + ; MIPS32-DAG: insert.h [[R1]][7], [[R2]] + ; MIPS64-DAG: insert.h [[R1]][7], $11 store volatile <8 x i16> %8, <8 x i16>*@v8i16 ret void - ; MIPS32-AE: .size nonconst_v8i16 } -define void @nonconst_v4i32(i32 %a, i32 %b, i32 %c, i32 %d) nounwind { - ; MIPS32-AE-LABEL: nonconst_v4i32: +define void @nonconst_v4i32(i32 signext %a, i32 signext %b, i32 signext %c, i32 signext %d) nounwind { + ; ALL-LABEL: nonconst_v4i32: %1 = insertelement <4 x i32> undef, i32 %a, i32 0 %2 = insertelement <4 x i32> %1, i32 %b, i32 1 %3 = insertelement <4 x i32> %2, i32 %c, i32 2 %4 = insertelement <4 x i32> %3, i32 %d, i32 3 - ; MIPS32-AE: insert.w [[R1:\$w[0-9]+]][0], $4 - ; MIPS32-AE: insert.w [[R1]][1], $5 - ; MIPS32-AE: insert.w [[R1]][2], $6 - ; MIPS32-AE: insert.w [[R1]][3], $7 + ; ALL: insert.w [[R1:\$w[0-9]+]][0], $4 + ; ALL: insert.w [[R1]][1], $5 + ; ALL: insert.w [[R1]][2], $6 + ; ALL: insert.w [[R1]][3], $7 store volatile <4 x i32> %4, <4 x i32>*@v4i32 ret void - ; MIPS32-AE: .size nonconst_v4i32 } -define void @nonconst_v2i64(i64 %a, i64 %b) nounwind { - ; MIPS32-AE-LABEL: nonconst_v2i64: +define void @nonconst_v2i64(i64 signext %a, i64 signext %b) nounwind { + ; ALL-LABEL: nonconst_v2i64: %1 = insertelement <2 x i64> undef, i64 %a, i32 0 %2 = insertelement <2 x i64> %1, i64 %b, i32 1 - ; MIPS32-AE: insert.w [[R1:\$w[0-9]+]][0], $4 - ; MIPS32-AE: insert.w [[R1]][1], $5 - ; MIPS32-AE: insert.w [[R1]][2], $6 - ; MIPS32-AE: insert.w [[R1]][3], $7 + ; MIPS32: insert.w [[R1:\$w[0-9]+]][0], $4 + ; MIPS32: insert.w [[R1]][1], $5 + ; MIPS32: insert.w [[R1]][2], $6 + ; MIPS32: insert.w [[R1]][3], $7 + ; MIPS64: insert.d [[R1:\$w[0-9]+]][0], $4 + ; MIPS64: insert.d [[R1]][1], $5 store volatile <2 x i64> %2, <2 x i64>*@v2i64 ret void - ; MIPS32-AE: .size nonconst_v2i64 } define i32 @extract_sext_v16i8() nounwind { - ; MIPS32-AE-LABEL: extract_sext_v16i8: + ; ALL-LABEL: extract_sext_v16i8: - %1 = load <16 x i8>* @v16i8 - ; MIPS32-AE-DAG: ld.b [[R1:\$w[0-9]+]], + %1 = load <16 x i8>, <16 x i8>* @v16i8 + ; ALL-DAG: ld.b [[R1:\$w[0-9]+]], %2 = add <16 x i8> %1, %1 - ; MIPS32-AE-DAG: addv.b [[R2:\$w[0-9]+]], [[R1]], [[R1]] + ; ALL-DAG: addv.b [[R2:\$w[0-9]+]], [[R1]], [[R1]] %3 = extractelement <16 x i8> %2, i32 1 %4 = sext i8 %3 to i32 - ; MIPS32-AE-DAG: copy_s.b [[R3:\$[0-9]+]], [[R1]][1] - ; MIPS32-AE-NOT: sll - ; MIPS32-AE-NOT: sra + ; ALL-DAG: copy_s.b [[R3:\$[0-9]+]], [[R1]][1] + ; ALL-NOT: sll + ; ALL-NOT: sra ret i32 %4 - ; MIPS32-AE: .size extract_sext_v16i8 } define i32 @extract_sext_v8i16() nounwind { - ; MIPS32-AE-LABEL: extract_sext_v8i16: + ; ALL-LABEL: extract_sext_v8i16: - %1 = load <8 x i16>* @v8i16 - ; MIPS32-AE-DAG: ld.h [[R1:\$w[0-9]+]], + %1 = load <8 x i16>, <8 x i16>* @v8i16 + ; ALL-DAG: ld.h [[R1:\$w[0-9]+]], %2 = add <8 x i16> %1, %1 - ; MIPS32-AE-DAG: addv.h [[R2:\$w[0-9]+]], [[R1]], [[R1]] + ; ALL-DAG: addv.h [[R2:\$w[0-9]+]], [[R1]], [[R1]] %3 = extractelement <8 x i16> %2, i32 1 %4 = sext i16 %3 to i32 - ; MIPS32-AE-DAG: copy_s.h [[R3:\$[0-9]+]], [[R1]][1] - ; MIPS32-AE-NOT: sll - ; MIPS32-AE-NOT: sra + ; ALL-DAG: copy_s.h [[R3:\$[0-9]+]], [[R1]][1] + ; ALL-NOT: sll + ; ALL-NOT: sra ret i32 %4 - ; MIPS32-AE: .size extract_sext_v8i16 } define i32 @extract_sext_v4i32() nounwind { - ; MIPS32-AE-LABEL: extract_sext_v4i32: + ; ALL-LABEL: extract_sext_v4i32: - %1 = load <4 x i32>* @v4i32 - ; MIPS32-AE-DAG: ld.w [[R1:\$w[0-9]+]], + %1 = load <4 x i32>, <4 x i32>* @v4i32 + ; ALL-DAG: ld.w [[R1:\$w[0-9]+]], %2 = add <4 x i32> %1, %1 - ; MIPS32-AE-DAG: addv.w [[R2:\$w[0-9]+]], [[R1]], [[R1]] + ; ALL-DAG: addv.w [[R2:\$w[0-9]+]], [[R1]], [[R1]] %3 = extractelement <4 x i32> %2, i32 1 - ; MIPS32-AE-DAG: copy_s.w [[R3:\$[0-9]+]], [[R1]][1] + ; ALL-DAG: copy_s.w [[R3:\$[0-9]+]], [[R1]][1] ret i32 %3 - ; MIPS32-AE: .size extract_sext_v4i32 } define i64 @extract_sext_v2i64() nounwind { - ; MIPS32-AE-LABEL: extract_sext_v2i64: + ; ALL-LABEL: extract_sext_v2i64: - %1 = load <2 x i64>* @v2i64 - ; MIPS32-AE-DAG: ld.d [[R1:\$w[0-9]+]], + %1 = load <2 x i64>, <2 x i64>* @v2i64 + ; ALL-DAG: ld.d [[R1:\$w[0-9]+]], %2 = add <2 x i64> %1, %1 - ; MIPS32-AE-DAG: addv.d [[R2:\$w[0-9]+]], [[R1]], [[R1]] + ; ALL-DAG: addv.d [[R2:\$w[0-9]+]], [[R1]], [[R1]] %3 = extractelement <2 x i64> %2, i32 1 - ; MIPS32-AE-DAG: copy_s.w [[R3:\$[0-9]+]], [[R1]][2] - ; MIPS32-AE-DAG: copy_s.w [[R4:\$[0-9]+]], [[R1]][3] - ; MIPS32-AE-NOT: sll - ; MIPS32-AE-NOT: sra + ; MIPS32-DAG: copy_s.w [[R3:\$[0-9]+]], [[R1]][2] + ; MIPS32-DAG: copy_s.w [[R4:\$[0-9]+]], [[R1]][3] + ; MIPS64-DAG: copy_s.d [[R3:\$[0-9]+]], [[R1]][1] + ; ALL-NOT: sll + ; ALL-NOT: sra ret i64 %3 - ; MIPS32-AE: .size extract_sext_v2i64 } define i32 @extract_zext_v16i8() nounwind { - ; MIPS32-AE-LABEL: extract_zext_v16i8: + ; ALL-LABEL: extract_zext_v16i8: - %1 = load <16 x i8>* @v16i8 - ; MIPS32-AE-DAG: ld.b [[R1:\$w[0-9]+]], + %1 = load <16 x i8>, <16 x i8>* @v16i8 + ; ALL-DAG: ld.b [[R1:\$w[0-9]+]], %2 = add <16 x i8> %1, %1 - ; MIPS32-AE-DAG: addv.b [[R2:\$w[0-9]+]], [[R1]], [[R1]] + ; ALL-DAG: addv.b [[R2:\$w[0-9]+]], [[R1]], [[R1]] %3 = extractelement <16 x i8> %2, i32 1 %4 = zext i8 %3 to i32 - ; MIPS32-AE-DAG: copy_u.b [[R3:\$[0-9]+]], [[R1]][1] - ; MIPS32-AE-NOT: andi + ; ALL-DAG: copy_u.b [[R3:\$[0-9]+]], [[R1]][1] + ; ALL-NOT: andi ret i32 %4 - ; MIPS32-AE: .size extract_zext_v16i8 } define i32 @extract_zext_v8i16() nounwind { - ; MIPS32-AE-LABEL: extract_zext_v8i16: + ; ALL-LABEL: extract_zext_v8i16: - %1 = load <8 x i16>* @v8i16 - ; MIPS32-AE-DAG: ld.h [[R1:\$w[0-9]+]], + %1 = load <8 x i16>, <8 x i16>* @v8i16 + ; ALL-DAG: ld.h [[R1:\$w[0-9]+]], %2 = add <8 x i16> %1, %1 - ; MIPS32-AE-DAG: addv.h [[R2:\$w[0-9]+]], [[R1]], [[R1]] + ; ALL-DAG: addv.h [[R2:\$w[0-9]+]], [[R1]], [[R1]] %3 = extractelement <8 x i16> %2, i32 1 %4 = zext i16 %3 to i32 - ; MIPS32-AE-DAG: copy_u.h [[R3:\$[0-9]+]], [[R1]][1] - ; MIPS32-AE-NOT: andi + ; ALL-DAG: copy_u.h [[R3:\$[0-9]+]], [[R1]][1] + ; ALL-NOT: andi ret i32 %4 - ; MIPS32-AE: .size extract_zext_v8i16 } define i32 @extract_zext_v4i32() nounwind { - ; MIPS32-AE-LABEL: extract_zext_v4i32: + ; ALL-LABEL: extract_zext_v4i32: - %1 = load <4 x i32>* @v4i32 - ; MIPS32-AE-DAG: ld.w [[R1:\$w[0-9]+]], + %1 = load <4 x i32>, <4 x i32>* @v4i32 + ; ALL-DAG: ld.w [[R1:\$w[0-9]+]], %2 = add <4 x i32> %1, %1 - ; MIPS32-AE-DAG: addv.w [[R2:\$w[0-9]+]], [[R1]], [[R1]] + ; ALL-DAG: addv.w [[R2:\$w[0-9]+]], [[R1]], [[R1]] %3 = extractelement <4 x i32> %2, i32 1 - ; MIPS32-AE-DAG: copy_{{[su]}}.w [[R3:\$[0-9]+]], [[R1]][1] + ; ALL-DAG: copy_{{[su]}}.w [[R3:\$[0-9]+]], [[R1]][1] ret i32 %3 - ; MIPS32-AE: .size extract_zext_v4i32 } define i64 @extract_zext_v2i64() nounwind { - ; MIPS32-AE-LABEL: extract_zext_v2i64: + ; ALL-LABEL: extract_zext_v2i64: - %1 = load <2 x i64>* @v2i64 - ; MIPS32-AE-DAG: ld.d [[R1:\$w[0-9]+]], + %1 = load <2 x i64>, <2 x i64>* @v2i64 + ; ALL-DAG: ld.d [[R1:\$w[0-9]+]], %2 = add <2 x i64> %1, %1 - ; MIPS32-AE-DAG: addv.d [[R2:\$w[0-9]+]], [[R1]], [[R1]] + ; ALL-DAG: addv.d [[R2:\$w[0-9]+]], [[R1]], [[R1]] %3 = extractelement <2 x i64> %2, i32 1 - ; MIPS32-AE-DAG: copy_{{[su]}}.w [[R3:\$[0-9]+]], [[R1]][2] - ; MIPS32-AE-DAG: copy_{{[su]}}.w [[R4:\$[0-9]+]], [[R1]][3] - ; MIPS32-AE-NOT: andi + ; MIPS32-DAG: copy_{{[su]}}.w [[R3:\$[0-9]+]], [[R1]][2] + ; MIPS32-DAG: copy_{{[su]}}.w [[R4:\$[0-9]+]], [[R1]][3] + ; MIPS64-DAG: copy_{{[su]}}.d [[R3:\$[0-9]+]], [[R1]][1] + ; ALL-NOT: andi ret i64 %3 - ; MIPS32-AE: .size extract_zext_v2i64 } define i32 @extract_sext_v16i8_vidx() nounwind { - ; MIPS32-AE-LABEL: extract_sext_v16i8_vidx: + ; ALL-LABEL: extract_sext_v16i8_vidx: - %1 = load <16 x i8>* @v16i8 - ; MIPS32-AE-DAG: lw [[PTR_V:\$[0-9]+]], %got(v16i8)( - ; MIPS32-AE-DAG: ld.b [[R1:\$w[0-9]+]], 0([[PTR_V]]) + %1 = load <16 x i8>, <16 x i8>* @v16i8 + ; O32-DAG: lw [[PTR_V:\$[0-9]+]], %got(v16i8)( + ; N32-DAG: lw [[PTR_V:\$[0-9]+]], %got_disp(v16i8)( + ; N64-DAG: ld [[PTR_V:\$[0-9]+]], %got_disp(v16i8)( + ; ALL-DAG: ld.b [[R1:\$w[0-9]+]], 0([[PTR_V]]) %2 = add <16 x i8> %1, %1 - ; MIPS32-AE-DAG: addv.b [[R2:\$w[0-9]+]], [[R1]], [[R1]] + ; ALL-DAG: addv.b [[R2:\$w[0-9]+]], [[R1]], [[R1]] - %3 = load i32* @i32 - ; MIPS32-AE-DAG: lw [[PTR_I:\$[0-9]+]], %got(i32)( - ; MIPS32-AE-DAG: lw [[IDX:\$[0-9]+]], 0([[PTR_I]]) + %3 = load i32, i32* @i32 + ; O32-DAG: lw [[PTR_I:\$[0-9]+]], %got(i32)( + ; N32-DAG: lw [[PTR_I:\$[0-9]+]], %got_disp(i32)( + ; N64-DAG: ld [[PTR_I:\$[0-9]+]], %got_disp(i32)( + ; ALL-DAG: lw [[IDX:\$[0-9]+]], 0([[PTR_I]]) %4 = extractelement <16 x i8> %2, i32 %3 %5 = sext i8 %4 to i32 - ; MIPS32-AE-DAG: splat.b $w[[R3:[0-9]+]], [[R1]]{{\[}}[[IDX]]] - ; MIPS32-AE-DAG: mfc1 [[R5:\$[0-9]+]], $f[[R3]] - ; MIPS32-AE-DAG: sra [[R6:\$[0-9]+]], [[R5]], 24 + ; ALL-DAG: splat.b $w[[R3:[0-9]+]], [[R1]]{{\[}}[[IDX]]] + ; ALL-DAG: mfc1 [[R5:\$[0-9]+]], $f[[R3]] + ; ALL-DAG: sra [[R6:\$[0-9]+]], [[R5]], 24 ret i32 %5 - ; MIPS32-AE: .size extract_sext_v16i8_vidx } define i32 @extract_sext_v8i16_vidx() nounwind { - ; MIPS32-AE-LABEL: extract_sext_v8i16_vidx: + ; ALL-LABEL: extract_sext_v8i16_vidx: - %1 = load <8 x i16>* @v8i16 - ; MIPS32-AE-DAG: lw [[PTR_V:\$[0-9]+]], %got(v8i16)( - ; MIPS32-AE-DAG: ld.h [[R1:\$w[0-9]+]], 0([[PTR_V]]) + %1 = load <8 x i16>, <8 x i16>* @v8i16 + ; O32-DAG: lw [[PTR_V:\$[0-9]+]], %got(v8i16)( + ; N32-DAG: lw [[PTR_V:\$[0-9]+]], %got_disp(v8i16)( + ; N64-DAG: ld [[PTR_V:\$[0-9]+]], %got_disp(v8i16)( + ; ALL-DAG: ld.h [[R1:\$w[0-9]+]], 0([[PTR_V]]) %2 = add <8 x i16> %1, %1 - ; MIPS32-AE-DAG: addv.h [[R2:\$w[0-9]+]], [[R1]], [[R1]] + ; ALL-DAG: addv.h [[R2:\$w[0-9]+]], [[R1]], [[R1]] - %3 = load i32* @i32 - ; MIPS32-AE-DAG: lw [[PTR_I:\$[0-9]+]], %got(i32)( - ; MIPS32-AE-DAG: lw [[IDX:\$[0-9]+]], 0([[PTR_I]]) + %3 = load i32, i32* @i32 + ; O32-DAG: lw [[PTR_I:\$[0-9]+]], %got(i32)( + ; N32-DAG: lw [[PTR_I:\$[0-9]+]], %got_disp(i32)( + ; N64-DAG: ld [[PTR_I:\$[0-9]+]], %got_disp(i32)( + ; ALL-DAG: lw [[IDX:\$[0-9]+]], 0([[PTR_I]]) %4 = extractelement <8 x i16> %2, i32 %3 %5 = sext i16 %4 to i32 - ; MIPS32-AE-DAG: splat.h $w[[R3:[0-9]+]], [[R1]]{{\[}}[[IDX]]] - ; MIPS32-AE-DAG: mfc1 [[R5:\$[0-9]+]], $f[[R3]] - ; MIPS32-AE-DAG: sra [[R6:\$[0-9]+]], [[R5]], 16 + ; ALL-DAG: splat.h $w[[R3:[0-9]+]], [[R1]]{{\[}}[[IDX]]] + ; ALL-DAG: mfc1 [[R5:\$[0-9]+]], $f[[R3]] + ; ALL-DAG: sra [[R6:\$[0-9]+]], [[R5]], 16 ret i32 %5 - ; MIPS32-AE: .size extract_sext_v8i16_vidx } define i32 @extract_sext_v4i32_vidx() nounwind { - ; MIPS32-AE-LABEL: extract_sext_v4i32_vidx: + ; ALL-LABEL: extract_sext_v4i32_vidx: - %1 = load <4 x i32>* @v4i32 - ; MIPS32-AE-DAG: lw [[PTR_V:\$[0-9]+]], %got(v4i32)( - ; MIPS32-AE-DAG: ld.w [[R1:\$w[0-9]+]], 0([[PTR_V]]) + %1 = load <4 x i32>, <4 x i32>* @v4i32 + ; O32-DAG: lw [[PTR_V:\$[0-9]+]], %got(v4i32)( + ; N32-DAG: lw [[PTR_V:\$[0-9]+]], %got_disp(v4i32)( + ; N64-DAG: ld [[PTR_V:\$[0-9]+]], %got_disp(v4i32)( + ; ALL-DAG: ld.w [[R1:\$w[0-9]+]], 0([[PTR_V]]) %2 = add <4 x i32> %1, %1 - ; MIPS32-AE-DAG: addv.w [[R2:\$w[0-9]+]], [[R1]], [[R1]] + ; ALL-DAG: addv.w [[R2:\$w[0-9]+]], [[R1]], [[R1]] - %3 = load i32* @i32 - ; MIPS32-AE-DAG: lw [[PTR_I:\$[0-9]+]], %got(i32)( - ; MIPS32-AE-DAG: lw [[IDX:\$[0-9]+]], 0([[PTR_I]]) + %3 = load i32, i32* @i32 + ; O32-DAG: lw [[PTR_I:\$[0-9]+]], %got(i32)( + ; N32-DAG: lw [[PTR_I:\$[0-9]+]], %got_disp(i32)( + ; N64-DAG: ld [[PTR_I:\$[0-9]+]], %got_disp(i32)( + ; ALL-DAG: lw [[IDX:\$[0-9]+]], 0([[PTR_I]]) %4 = extractelement <4 x i32> %2, i32 %3 - ; MIPS32-AE-DAG: splat.w $w[[R3:[0-9]+]], [[R1]]{{\[}}[[IDX]]] - ; MIPS32-AE-DAG: mfc1 [[R5:\$[0-9]+]], $f[[R3]] - ; MIPS32-AE-NOT: sra + ; ALL-DAG: splat.w $w[[R3:[0-9]+]], [[R1]]{{\[}}[[IDX]]] + ; ALL-DAG: mfc1 [[R5:\$[0-9]+]], $f[[R3]] + ; ALL-NOT: sra ret i32 %4 - ; MIPS32-AE: .size extract_sext_v4i32_vidx } define i64 @extract_sext_v2i64_vidx() nounwind { - ; MIPS32-AE-LABEL: extract_sext_v2i64_vidx: + ; ALL-LABEL: extract_sext_v2i64_vidx: - %1 = load <2 x i64>* @v2i64 - ; MIPS32-AE-DAG: lw [[PTR_V:\$[0-9]+]], %got(v2i64)( - ; MIPS32-AE-DAG: ld.d [[R1:\$w[0-9]+]], 0([[PTR_V]]) + %1 = load <2 x i64>, <2 x i64>* @v2i64 + ; O32-DAG: lw [[PTR_V:\$[0-9]+]], %got(v2i64)( + ; N32-DAG: lw [[PTR_V:\$[0-9]+]], %got_disp(v2i64)( + ; N64-DAG: ld [[PTR_V:\$[0-9]+]], %got_disp(v2i64)( + ; ALL-DAG: ld.d [[R1:\$w[0-9]+]], 0([[PTR_V]]) %2 = add <2 x i64> %1, %1 - ; MIPS32-AE-DAG: addv.d [[R2:\$w[0-9]+]], [[R1]], [[R1]] + ; ALL-DAG: addv.d [[R2:\$w[0-9]+]], [[R1]], [[R1]] - %3 = load i32* @i32 - ; MIPS32-AE-DAG: lw [[PTR_I:\$[0-9]+]], %got(i32)( - ; MIPS32-AE-DAG: lw [[IDX:\$[0-9]+]], 0([[PTR_I]]) + %3 = load i32, i32* @i32 + ; O32-DAG: lw [[PTR_I:\$[0-9]+]], %got(i32)( + ; N32-DAG: lw [[PTR_I:\$[0-9]+]], %got_disp(i32)( + ; N64-DAG: ld [[PTR_I:\$[0-9]+]], %got_disp(i32)( + ; ALL-DAG: lw [[IDX:\$[0-9]+]], 0([[PTR_I]]) %4 = extractelement <2 x i64> %2, i32 %3 - ; MIPS32-AE-DAG: splat.w $w[[R3:[0-9]+]], [[R1]]{{\[}}[[IDX]]] - ; MIPS32-AE-DAG: mfc1 [[R5:\$[0-9]+]], $f[[R3]] - ; MIPS32-AE-DAG: splat.w $w[[R4:[0-9]+]], [[R1]]{{\[}}[[IDX]]] - ; MIPS32-AE-DAG: mfc1 [[R6:\$[0-9]+]], $f[[R4]] - ; MIPS32-AE-NOT: sra + ; MIPS32-DAG: splat.w $w[[R3:[0-9]+]], [[R1]]{{\[}}[[IDX]]] + ; MIPS32-DAG: mfc1 [[R5:\$[0-9]+]], $f[[R3]] + ; MIPS32-DAG: splat.w $w[[R4:[0-9]+]], [[R1]]{{\[}}[[IDX]]] + ; MIPS32-DAG: mfc1 [[R6:\$[0-9]+]], $f[[R4]] + ; MIPS64-DAG: splat.d $w[[R3:[0-9]+]], [[R1]]{{\[}}[[IDX]]] + ; MIPS64-DAG: dmfc1 [[R5:\$[0-9]+]], $f[[R3]] + ; ALL-NOT: sra ret i64 %4 - ; MIPS32-AE: .size extract_sext_v2i64_vidx } define i32 @extract_zext_v16i8_vidx() nounwind { - ; MIPS32-AE-LABEL: extract_zext_v16i8_vidx: + ; ALL-LABEL: extract_zext_v16i8_vidx: - %1 = load <16 x i8>* @v16i8 - ; MIPS32-AE-DAG: lw [[PTR_V:\$[0-9]+]], %got(v16i8)( - ; MIPS32-AE-DAG: ld.b [[R1:\$w[0-9]+]], 0([[PTR_V]]) + %1 = load <16 x i8>, <16 x i8>* @v16i8 + ; O32-DAG: lw [[PTR_V:\$[0-9]+]], %got(v16i8)( + ; N32-DAG: lw [[PTR_V:\$[0-9]+]], %got_disp(v16i8)( + ; N64-DAG: ld [[PTR_V:\$[0-9]+]], %got_disp(v16i8)( + ; ALL-DAG: ld.b [[R1:\$w[0-9]+]], 0([[PTR_V]]) %2 = add <16 x i8> %1, %1 - ; MIPS32-AE-DAG: addv.b [[R2:\$w[0-9]+]], [[R1]], [[R1]] + ; ALL-DAG: addv.b [[R2:\$w[0-9]+]], [[R1]], [[R1]] - %3 = load i32* @i32 - ; MIPS32-AE-DAG: lw [[PTR_I:\$[0-9]+]], %got(i32)( - ; MIPS32-AE-DAG: lw [[IDX:\$[0-9]+]], 0([[PTR_I]]) + %3 = load i32, i32* @i32 + ; O32-DAG: lw [[PTR_I:\$[0-9]+]], %got(i32)( + ; N32-DAG: lw [[PTR_I:\$[0-9]+]], %got_disp(i32)( + ; N64-DAG: ld [[PTR_I:\$[0-9]+]], %got_disp(i32)( + ; ALL-DAG: lw [[IDX:\$[0-9]+]], 0([[PTR_I]]) %4 = extractelement <16 x i8> %2, i32 %3 %5 = zext i8 %4 to i32 - ; MIPS32-AE-DAG: splat.b $w[[R3:[0-9]+]], [[R1]]{{\[}}[[IDX]]] - ; MIPS32-AE-DAG: mfc1 [[R5:\$[0-9]+]], $f[[R3]] - ; MIPS32-AE-DAG: srl [[R6:\$[0-9]+]], [[R5]], 24 + ; ALL-DAG: splat.b $w[[R3:[0-9]+]], [[R1]]{{\[}}[[IDX]]] + ; ALL-DAG: mfc1 [[R5:\$[0-9]+]], $f[[R3]] + ; ALL-DAG: srl [[R6:\$[0-9]+]], [[R5]], 24 ret i32 %5 - ; MIPS32-AE: .size extract_zext_v16i8_vidx } define i32 @extract_zext_v8i16_vidx() nounwind { - ; MIPS32-AE-LABEL: extract_zext_v8i16_vidx: + ; ALL-LABEL: extract_zext_v8i16_vidx: - %1 = load <8 x i16>* @v8i16 - ; MIPS32-AE-DAG: lw [[PTR_V:\$[0-9]+]], %got(v8i16)( - ; MIPS32-AE-DAG: ld.h [[R1:\$w[0-9]+]], 0([[PTR_V]]) + %1 = load <8 x i16>, <8 x i16>* @v8i16 + ; O32-DAG: lw [[PTR_V:\$[0-9]+]], %got(v8i16)( + ; N32-DAG: lw [[PTR_V:\$[0-9]+]], %got_disp(v8i16)( + ; N64-DAG: ld [[PTR_V:\$[0-9]+]], %got_disp(v8i16)( + ; ALL-DAG: ld.h [[R1:\$w[0-9]+]], 0([[PTR_V]]) %2 = add <8 x i16> %1, %1 - ; MIPS32-AE-DAG: addv.h [[R2:\$w[0-9]+]], [[R1]], [[R1]] + ; ALL-DAG: addv.h [[R2:\$w[0-9]+]], [[R1]], [[R1]] - %3 = load i32* @i32 - ; MIPS32-AE-DAG: lw [[PTR_I:\$[0-9]+]], %got(i32)( - ; MIPS32-AE-DAG: lw [[IDX:\$[0-9]+]], 0([[PTR_I]]) + %3 = load i32, i32* @i32 + ; O32-DAG: lw [[PTR_I:\$[0-9]+]], %got(i32)( + ; N32-DAG: lw [[PTR_I:\$[0-9]+]], %got_disp(i32)( + ; N64-DAG: ld [[PTR_I:\$[0-9]+]], %got_disp(i32)( + ; ALL-DAG: lw [[IDX:\$[0-9]+]], 0([[PTR_I]]) %4 = extractelement <8 x i16> %2, i32 %3 %5 = zext i16 %4 to i32 - ; MIPS32-AE-DAG: splat.h $w[[R3:[0-9]+]], [[R1]]{{\[}}[[IDX]]] - ; MIPS32-AE-DAG: mfc1 [[R5:\$[0-9]+]], $f[[R3]] - ; MIPS32-AE-DAG: srl [[R6:\$[0-9]+]], [[R5]], 16 + ; ALL-DAG: splat.h $w[[R3:[0-9]+]], [[R1]]{{\[}}[[IDX]]] + ; ALL-DAG: mfc1 [[R5:\$[0-9]+]], $f[[R3]] + ; ALL-DAG: srl [[R6:\$[0-9]+]], [[R5]], 16 ret i32 %5 - ; MIPS32-AE: .size extract_zext_v8i16_vidx } define i32 @extract_zext_v4i32_vidx() nounwind { - ; MIPS32-AE-LABEL: extract_zext_v4i32_vidx: + ; ALL-LABEL: extract_zext_v4i32_vidx: - %1 = load <4 x i32>* @v4i32 - ; MIPS32-AE-DAG: lw [[PTR_V:\$[0-9]+]], %got(v4i32)( - ; MIPS32-AE-DAG: ld.w [[R1:\$w[0-9]+]], 0([[PTR_V]]) + %1 = load <4 x i32>, <4 x i32>* @v4i32 + ; O32-DAG: lw [[PTR_V:\$[0-9]+]], %got(v4i32)( + ; N32-DAG: lw [[PTR_V:\$[0-9]+]], %got_disp(v4i32)( + ; N64-DAG: ld [[PTR_V:\$[0-9]+]], %got_disp(v4i32)( + ; ALL-DAG: ld.w [[R1:\$w[0-9]+]], 0([[PTR_V]]) %2 = add <4 x i32> %1, %1 - ; MIPS32-AE-DAG: addv.w [[R2:\$w[0-9]+]], [[R1]], [[R1]] + ; ALL-DAG: addv.w [[R2:\$w[0-9]+]], [[R1]], [[R1]] - %3 = load i32* @i32 - ; MIPS32-AE-DAG: lw [[PTR_I:\$[0-9]+]], %got(i32)( - ; MIPS32-AE-DAG: lw [[IDX:\$[0-9]+]], 0([[PTR_I]]) + %3 = load i32, i32* @i32 + ; O32-DAG: lw [[PTR_I:\$[0-9]+]], %got(i32)( + ; N32-DAG: lw [[PTR_I:\$[0-9]+]], %got_disp(i32)( + ; N64-DAG: ld [[PTR_I:\$[0-9]+]], %got_disp(i32)( + ; ALL-DAG: lw [[IDX:\$[0-9]+]], 0([[PTR_I]]) %4 = extractelement <4 x i32> %2, i32 %3 - ; MIPS32-AE-DAG: splat.w $w[[R3:[0-9]+]], [[R1]]{{\[}}[[IDX]]] - ; MIPS32-AE-DAG: mfc1 [[R5:\$[0-9]+]], $f[[R3]] - ; MIPS32-AE-NOT: srl + ; ALL-DAG: splat.w $w[[R3:[0-9]+]], [[R1]]{{\[}}[[IDX]]] + ; ALL-DAG: mfc1 [[R5:\$[0-9]+]], $f[[R3]] + ; ALL-NOT: srl ret i32 %4 - ; MIPS32-AE: .size extract_zext_v4i32_vidx } define i64 @extract_zext_v2i64_vidx() nounwind { - ; MIPS32-AE-LABEL: extract_zext_v2i64_vidx: + ; ALL-LABEL: extract_zext_v2i64_vidx: - %1 = load <2 x i64>* @v2i64 - ; MIPS32-AE-DAG: lw [[PTR_V:\$[0-9]+]], %got(v2i64)( - ; MIPS32-AE-DAG: ld.d [[R1:\$w[0-9]+]], 0([[PTR_V]]) + %1 = load <2 x i64>, <2 x i64>* @v2i64 + ; O32-DAG: lw [[PTR_V:\$[0-9]+]], %got(v2i64)( + ; N32-DAG: lw [[PTR_V:\$[0-9]+]], %got_disp(v2i64)( + ; N64-DAG: ld [[PTR_V:\$[0-9]+]], %got_disp(v2i64)( + ; ALL-DAG: ld.d [[R1:\$w[0-9]+]], 0([[PTR_V]]) %2 = add <2 x i64> %1, %1 - ; MIPS32-AE-DAG: addv.d [[R2:\$w[0-9]+]], [[R1]], [[R1]] + ; ALL-DAG: addv.d [[R2:\$w[0-9]+]], [[R1]], [[R1]] - %3 = load i32* @i32 - ; MIPS32-AE-DAG: lw [[PTR_I:\$[0-9]+]], %got(i32)( - ; MIPS32-AE-DAG: lw [[IDX:\$[0-9]+]], 0([[PTR_I]]) + %3 = load i32, i32* @i32 + ; O32-DAG: lw [[PTR_I:\$[0-9]+]], %got(i32)( + ; N32-DAG: lw [[PTR_I:\$[0-9]+]], %got_disp(i32)( + ; N64-DAG: ld [[PTR_I:\$[0-9]+]], %got_disp(i32)( + ; ALL-DAG: lw [[IDX:\$[0-9]+]], 0([[PTR_I]]) %4 = extractelement <2 x i64> %2, i32 %3 - ; MIPS32-AE-DAG: splat.w $w[[R3:[0-9]+]], [[R1]]{{\[}}[[IDX]]] - ; MIPS32-AE-DAG: mfc1 [[R5:\$[0-9]+]], $f[[R3]] - ; MIPS32-AE-DAG: splat.w $w[[R4:[0-9]+]], [[R1]]{{\[}}[[IDX]]] - ; MIPS32-AE-DAG: mfc1 [[R6:\$[0-9]+]], $f[[R4]] - ; MIPS32-AE-NOT: srl + ; MIPS32-DAG: splat.w $w[[R3:[0-9]+]], [[R1]]{{\[}}[[IDX]]] + ; MIPS32-DAG: mfc1 [[R5:\$[0-9]+]], $f[[R3]] + ; MIPS32-DAG: splat.w $w[[R4:[0-9]+]], [[R1]]{{\[}}[[IDX]]] + ; MIPS32-DAG: mfc1 [[R6:\$[0-9]+]], $f[[R4]] + ; MIPS64-DAG: splat.d $w[[R3:[0-9]+]], [[R1]]{{\[}}[[IDX]]] + ; MIPS64-DAG: dmfc1 [[R5:\$[0-9]+]], $f[[R3]] + ; ALL-NOT: srl ret i64 %4 - ; MIPS32-AE: .size extract_zext_v2i64_vidx } -define void @insert_v16i8(i32 %a) nounwind { - ; MIPS32-AE-LABEL: insert_v16i8: +define void @insert_v16i8(i32 signext %a) nounwind { + ; ALL-LABEL: insert_v16i8: - %1 = load <16 x i8>* @v16i8 - ; MIPS32-AE-DAG: ld.b [[R1:\$w[0-9]+]], + %1 = load <16 x i8>, <16 x i8>* @v16i8 + ; ALL-DAG: ld.b [[R1:\$w[0-9]+]], %a2 = trunc i32 %a to i8 %a3 = sext i8 %a2 to i32 %a4 = trunc i32 %a3 to i8 - ; MIPS32-AE-NOT: andi - ; MIPS32-AE-NOT: sra + ; ALL-NOT: andi + ; ALL-NOT: sra %2 = insertelement <16 x i8> %1, i8 %a4, i32 1 - ; MIPS32-AE-DAG: insert.b [[R1]][1], $4 + ; ALL-DAG: insert.b [[R1]][1], $4 store <16 x i8> %2, <16 x i8>* @v16i8 - ; MIPS32-AE-DAG: st.b [[R1]] + ; ALL-DAG: st.b [[R1]] ret void - ; MIPS32-AE: .size insert_v16i8 } -define void @insert_v8i16(i32 %a) nounwind { - ; MIPS32-AE-LABEL: insert_v8i16: +define void @insert_v8i16(i32 signext %a) nounwind { + ; ALL-LABEL: insert_v8i16: - %1 = load <8 x i16>* @v8i16 - ; MIPS32-AE-DAG: ld.h [[R1:\$w[0-9]+]], + %1 = load <8 x i16>, <8 x i16>* @v8i16 + ; ALL-DAG: ld.h [[R1:\$w[0-9]+]], %a2 = trunc i32 %a to i16 %a3 = sext i16 %a2 to i32 %a4 = trunc i32 %a3 to i16 - ; MIPS32-AE-NOT: andi - ; MIPS32-AE-NOT: sra + ; ALL-NOT: andi + ; ALL-NOT: sra %2 = insertelement <8 x i16> %1, i16 %a4, i32 1 - ; MIPS32-AE-DAG: insert.h [[R1]][1], $4 + ; ALL-DAG: insert.h [[R1]][1], $4 store <8 x i16> %2, <8 x i16>* @v8i16 - ; MIPS32-AE-DAG: st.h [[R1]] + ; ALL-DAG: st.h [[R1]] ret void - ; MIPS32-AE: .size insert_v8i16 } -define void @insert_v4i32(i32 %a) nounwind { - ; MIPS32-AE-LABEL: insert_v4i32: +define void @insert_v4i32(i32 signext %a) nounwind { + ; ALL-LABEL: insert_v4i32: - %1 = load <4 x i32>* @v4i32 - ; MIPS32-AE-DAG: ld.w [[R1:\$w[0-9]+]], + %1 = load <4 x i32>, <4 x i32>* @v4i32 + ; ALL-DAG: ld.w [[R1:\$w[0-9]+]], - ; MIPS32-AE-NOT: andi - ; MIPS32-AE-NOT: sra + ; ALL-NOT: andi + ; ALL-NOT: sra %2 = insertelement <4 x i32> %1, i32 %a, i32 1 - ; MIPS32-AE-DAG: insert.w [[R1]][1], $4 + ; ALL-DAG: insert.w [[R1]][1], $4 store <4 x i32> %2, <4 x i32>* @v4i32 - ; MIPS32-AE-DAG: st.w [[R1]] + ; ALL-DAG: st.w [[R1]] ret void - ; MIPS32-AE: .size insert_v4i32 } -define void @insert_v2i64(i64 %a) nounwind { - ; MIPS32-AE-LABEL: insert_v2i64: +define void @insert_v2i64(i64 signext %a) nounwind { + ; ALL-LABEL: insert_v2i64: - %1 = load <2 x i64>* @v2i64 - ; MIPS32-AE-DAG: ld.w [[R1:\$w[0-9]+]], + %1 = load <2 x i64>, <2 x i64>* @v2i64 + ; MIPS32-DAG: ld.w [[R1:\$w[0-9]+]], + ; MIPS64-DAG: ld.d [[R1:\$w[0-9]+]], - ; MIPS32-AE-NOT: andi - ; MIPS32-AE-NOT: sra + ; ALL-NOT: andi + ; ALL-NOT: sra %2 = insertelement <2 x i64> %1, i64 %a, i32 1 - ; MIPS32-AE-DAG: insert.w [[R1]][2], $4 - ; MIPS32-AE-DAG: insert.w [[R1]][3], $5 + ; MIPS32-DAG: insert.w [[R1]][2], $4 + ; MIPS32-DAG: insert.w [[R1]][3], $5 + ; MIPS64-DAG: insert.d [[R1]][1], $4 store <2 x i64> %2, <2 x i64>* @v2i64 - ; MIPS32-AE-DAG: st.w [[R1]] + ; MIPS32-DAG: st.w [[R1]] + ; MIPS64-DAG: st.d [[R1]] ret void - ; MIPS32-AE: .size insert_v2i64 } -define void @insert_v16i8_vidx(i32 %a) nounwind { - ; MIPS32-AE: insert_v16i8_vidx: +define void @insert_v16i8_vidx(i32 signext %a) nounwind { + ; ALL-LABEL: insert_v16i8_vidx: - %1 = load <16 x i8>* @v16i8 - ; MIPS32-AE-DAG: ld.b [[R1:\$w[0-9]+]], + %1 = load <16 x i8>, <16 x i8>* @v16i8 + ; ALL-DAG: ld.b [[R1:\$w[0-9]+]], - %2 = load i32* @i32 - ; MIPS32-AE-DAG: lw [[PTR_I:\$[0-9]+]], %got(i32)( - ; MIPS32-AE-DAG: lw [[IDX:\$[0-9]+]], 0([[PTR_I]]) + %2 = load i32, i32* @i32 + ; O32-DAG: lw [[PTR_I:\$[0-9]+]], %got(i32)( + ; N32-DAG: lw [[PTR_I:\$[0-9]+]], %got_disp(i32)( + ; N64-DAG: ld [[PTR_I:\$[0-9]+]], %got_disp(i32)( + ; ALL-DAG: lw [[IDX:\$[0-9]+]], 0([[PTR_I]]) %a2 = trunc i32 %a to i8 %a3 = sext i8 %a2 to i32 %a4 = trunc i32 %a3 to i8 - ; MIPS32-AE-NOT: andi - ; MIPS32-AE-NOT: sra + ; ALL-NOT: andi + ; ALL-NOT: sra %3 = insertelement <16 x i8> %1, i8 %a4, i32 %2 - ; MIPS32-AE-DAG: sld.b [[R1]], [[R1]]{{\[}}[[IDX]]] - ; MIPS32-AE-DAG: insert.b [[R1]][0], $4 - ; MIPS32-AE-DAG: neg [[NIDX:\$[0-9]+]], [[IDX]] - ; MIPS32-AE-DAG: sld.b [[R1]], [[R1]]{{\[}}[[NIDX]]] + ; ALL-DAG: sld.b [[R1]], [[R1]]{{\[}}[[IDX]]] + ; ALL-DAG: insert.b [[R1]][0], $4 + ; O32-DAG: neg [[NIDX:\$[0-9]+]], [[IDX]] + ; N32-DAG: neg [[NIDX:\$[0-9]+]], [[IDX]] + ; N64-DAG: dneg [[NIDX:\$[0-9]+]], [[IDX]] + ; ALL-DAG: sld.b [[R1]], [[R1]]{{\[}}[[NIDX]]] store <16 x i8> %3, <16 x i8>* @v16i8 - ; MIPS32-AE-DAG: st.b [[R1]] + ; ALL-DAG: st.b [[R1]] ret void - ; MIPS32-AE: .size insert_v16i8_vidx } -define void @insert_v8i16_vidx(i32 %a) nounwind { - ; MIPS32-AE: insert_v8i16_vidx: +define void @insert_v8i16_vidx(i32 signext %a) nounwind { + ; ALL-LABEL: insert_v8i16_vidx: - %1 = load <8 x i16>* @v8i16 - ; MIPS32-AE-DAG: ld.h [[R1:\$w[0-9]+]], + %1 = load <8 x i16>, <8 x i16>* @v8i16 + ; ALL-DAG: ld.h [[R1:\$w[0-9]+]], - %2 = load i32* @i32 - ; MIPS32-AE-DAG: lw [[PTR_I:\$[0-9]+]], %got(i32)( - ; MIPS32-AE-DAG: lw [[IDX:\$[0-9]+]], 0([[PTR_I]]) + %2 = load i32, i32* @i32 + ; O32-DAG: lw [[PTR_I:\$[0-9]+]], %got(i32)( + ; N32-DAG: lw [[PTR_I:\$[0-9]+]], %got_disp(i32)( + ; N64-DAG: ld [[PTR_I:\$[0-9]+]], %got_disp(i32)( + ; ALL-DAG: lw [[IDX:\$[0-9]+]], 0([[PTR_I]]) %a2 = trunc i32 %a to i16 %a3 = sext i16 %a2 to i32 %a4 = trunc i32 %a3 to i16 - ; MIPS32-AE-NOT: andi - ; MIPS32-AE-NOT: sra + ; ALL-NOT: andi + ; ALL-NOT: sra %3 = insertelement <8 x i16> %1, i16 %a4, i32 %2 - ; MIPS32-AE-DAG: sll [[BIDX:\$[0-9]+]], [[IDX]], 1 - ; MIPS32-AE-DAG: sld.b [[R1]], [[R1]]{{\[}}[[BIDX]]] - ; MIPS32-AE-DAG: insert.h [[R1]][0], $4 - ; MIPS32-AE-DAG: neg [[NIDX:\$[0-9]+]], [[BIDX]] - ; MIPS32-AE-DAG: sld.b [[R1]], [[R1]]{{\[}}[[NIDX]]] + ; ALL-DAG: sll [[BIDX:\$[0-9]+]], [[IDX]], 1 + ; ALL-DAG: sld.b [[R1]], [[R1]]{{\[}}[[BIDX]]] + ; ALL-DAG: insert.h [[R1]][0], $4 + ; O32-DAG: neg [[NIDX:\$[0-9]+]], [[BIDX]] + ; N32-DAG: neg [[NIDX:\$[0-9]+]], [[BIDX]] + ; N64-DAG: dneg [[NIDX:\$[0-9]+]], [[BIDX]] + ; ALL-DAG: sld.b [[R1]], [[R1]]{{\[}}[[NIDX]]] store <8 x i16> %3, <8 x i16>* @v8i16 - ; MIPS32-AE-DAG: st.h [[R1]] + ; ALL-DAG: st.h [[R1]] ret void - ; MIPS32-AE: .size insert_v8i16_vidx } -define void @insert_v4i32_vidx(i32 %a) nounwind { - ; MIPS32-AE: insert_v4i32_vidx: +define void @insert_v4i32_vidx(i32 signext %a) nounwind { + ; ALL-LABEL: insert_v4i32_vidx: - %1 = load <4 x i32>* @v4i32 - ; MIPS32-AE-DAG: ld.w [[R1:\$w[0-9]+]], + %1 = load <4 x i32>, <4 x i32>* @v4i32 + ; ALL-DAG: ld.w [[R1:\$w[0-9]+]], - %2 = load i32* @i32 - ; MIPS32-AE-DAG: lw [[PTR_I:\$[0-9]+]], %got(i32)( - ; MIPS32-AE-DAG: lw [[IDX:\$[0-9]+]], 0([[PTR_I]]) + %2 = load i32, i32* @i32 + ; O32-DAG: lw [[PTR_I:\$[0-9]+]], %got(i32)( + ; N32-DAG: lw [[PTR_I:\$[0-9]+]], %got_disp(i32)( + ; N64-DAG: ld [[PTR_I:\$[0-9]+]], %got_disp(i32)( + ; ALL-DAG: lw [[IDX:\$[0-9]+]], 0([[PTR_I]]) - ; MIPS32-AE-NOT: andi - ; MIPS32-AE-NOT: sra + ; ALL-NOT: andi + ; ALL-NOT: sra %3 = insertelement <4 x i32> %1, i32 %a, i32 %2 - ; MIPS32-AE-DAG: sll [[BIDX:\$[0-9]+]], [[IDX]], 2 - ; MIPS32-AE-DAG: sld.b [[R1]], [[R1]]{{\[}}[[BIDX]]] - ; MIPS32-AE-DAG: insert.w [[R1]][0], $4 - ; MIPS32-AE-DAG: neg [[NIDX:\$[0-9]+]], [[BIDX]] - ; MIPS32-AE-DAG: sld.b [[R1]], [[R1]]{{\[}}[[NIDX]]] + ; ALL-DAG: sll [[BIDX:\$[0-9]+]], [[IDX]], 2 + ; ALL-DAG: sld.b [[R1]], [[R1]]{{\[}}[[BIDX]]] + ; ALL-DAG: insert.w [[R1]][0], $4 + ; O32-DAG: neg [[NIDX:\$[0-9]+]], [[BIDX]] + ; N32-DAG: neg [[NIDX:\$[0-9]+]], [[BIDX]] + ; N64-DAG: dneg [[NIDX:\$[0-9]+]], [[BIDX]] + ; ALL-DAG: sld.b [[R1]], [[R1]]{{\[}}[[NIDX]]] store <4 x i32> %3, <4 x i32>* @v4i32 - ; MIPS32-AE-DAG: st.w [[R1]] + ; ALL-DAG: st.w [[R1]] ret void - ; MIPS32-AE: .size insert_v4i32_vidx } -define void @insert_v2i64_vidx(i64 %a) nounwind { - ; MIPS32-AE: insert_v2i64_vidx: +define void @insert_v2i64_vidx(i64 signext %a) nounwind { + ; ALL-LABEL: insert_v2i64_vidx: - %1 = load <2 x i64>* @v2i64 - ; MIPS32-AE-DAG: ld.w [[R1:\$w[0-9]+]], + %1 = load <2 x i64>, <2 x i64>* @v2i64 + ; MIPS32-DAG: ld.w [[R1:\$w[0-9]+]], + ; MIPS64-DAG: ld.d [[R1:\$w[0-9]+]], - %2 = load i32* @i32 - ; MIPS32-AE-DAG: lw [[PTR_I:\$[0-9]+]], %got(i32)( - ; MIPS32-AE-DAG: lw [[IDX:\$[0-9]+]], 0([[PTR_I]]) + %2 = load i32, i32* @i32 + ; O32-DAG: lw [[PTR_I:\$[0-9]+]], %got(i32)( + ; N32-DAG: lw [[PTR_I:\$[0-9]+]], %got_disp(i32)( + ; N64-DAG: ld [[PTR_I:\$[0-9]+]], %got_disp(i32)( + ; ALL-DAG: lw [[IDX:\$[0-9]+]], 0([[PTR_I]]) - ; MIPS32-AE-NOT: andi - ; MIPS32-AE-NOT: sra + ; ALL-NOT: andi + ; ALL-NOT: sra %3 = insertelement <2 x i64> %1, i64 %a, i32 %2 ; TODO: This code could be a lot better but it works. The legalizer splits ; 64-bit inserts into two 32-bit inserts because there is no i64 type on ; MIPS32. The obvious optimisation is to perform both insert.w's at once while ; the vector is rotated. - ; MIPS32-AE-DAG: sll [[BIDX:\$[0-9]+]], [[IDX]], 2 - ; MIPS32-AE-DAG: sld.b [[R1]], [[R1]]{{\[}}[[BIDX]]] - ; MIPS32-AE-DAG: insert.w [[R1]][0], $4 - ; MIPS32-AE-DAG: neg [[NIDX:\$[0-9]+]], [[BIDX]] - ; MIPS32-AE-DAG: sld.b [[R1]], [[R1]]{{\[}}[[NIDX]]] - ; MIPS32-AE-DAG: addiu [[IDX2:\$[0-9]+]], [[IDX]], 1 - ; MIPS32-AE-DAG: sll [[BIDX:\$[0-9]+]], [[IDX2]], 2 - ; MIPS32-AE-DAG: sld.b [[R1]], [[R1]]{{\[}}[[BIDX]]] - ; MIPS32-AE-DAG: insert.w [[R1]][0], $5 - ; MIPS32-AE-DAG: neg [[NIDX:\$[0-9]+]], [[BIDX]] - ; MIPS32-AE-DAG: sld.b [[R1]], [[R1]]{{\[}}[[NIDX]]] + ; MIPS32-DAG: sll [[BIDX:\$[0-9]+]], [[IDX]], 2 + ; MIPS32-DAG: sld.b [[R1]], [[R1]]{{\[}}[[BIDX]]] + ; MIPS32-DAG: insert.w [[R1]][0], $4 + ; MIPS32-DAG: neg [[NIDX:\$[0-9]+]], [[BIDX]] + ; MIPS32-DAG: sld.b [[R1]], [[R1]]{{\[}}[[NIDX]]] + ; MIPS32-DAG: addiu [[IDX2:\$[0-9]+]], [[IDX]], 1 + ; MIPS32-DAG: sll [[BIDX:\$[0-9]+]], [[IDX2]], 2 + ; MIPS32-DAG: sld.b [[R1]], [[R1]]{{\[}}[[BIDX]]] + ; MIPS32-DAG: insert.w [[R1]][0], $5 + ; MIPS32-DAG: neg [[NIDX:\$[0-9]+]], [[BIDX]] + ; MIPS32-DAG: sld.b [[R1]], [[R1]]{{\[}}[[NIDX]]] + + ; MIPS64-DAG: sll [[BIDX:\$[0-9]+]], [[IDX]], 3 + ; MIPS64-DAG: sld.b [[R1]], [[R1]]{{\[}}[[BIDX]]] + ; MIPS64-DAG: insert.d [[R1]][0], $4 + ; N32-DAG: neg [[NIDX:\$[0-9]+]], [[BIDX]] + ; N64-DAG: dneg [[NIDX:\$[0-9]+]], [[BIDX]] + ; MIPS64-DAG: sld.b [[R1]], [[R1]]{{\[}}[[NIDX]]] store <2 x i64> %3, <2 x i64>* @v2i64 - ; MIPS32-AE-DAG: st.w [[R1]] + ; MIPS32-DAG: st.w [[R1]] + ; MIPS64-DAG: st.d [[R1]] ret void - ; MIPS32-AE: .size insert_v2i64_vidx } define void @truncstore() nounwind { - ; MIPS32-AE-LABEL: truncstore: + ; ALL-LABEL: truncstore: store volatile <4 x i8> <i8 -1, i8 -1, i8 -1, i8 -1>, <4 x i8>*@v4i8 ; TODO: What code should be emitted? ret void - ; MIPS32-AE: .size truncstore } diff --git a/test/CodeGen/Mips/msa/basic_operations_float.ll b/test/CodeGen/Mips/msa/basic_operations_float.ll index a0c9d29e231ae..f19cb9b7c2e55 100644 --- a/test/CodeGen/Mips/msa/basic_operations_float.ll +++ b/test/CodeGen/Mips/msa/basic_operations_float.ll @@ -1,5 +1,9 @@ -; RUN: llc -march=mips -mattr=+msa,+fp64 < %s | FileCheck -check-prefix=MIPS32 %s -; RUN: llc -march=mipsel -mattr=+msa,+fp64 < %s | FileCheck -check-prefix=MIPS32 %s +; RUN: llc -march=mips -mattr=+msa,+fp64 < %s | FileCheck -check-prefix=ALL -check-prefix=O32 %s +; RUN: llc -march=mipsel -mattr=+msa,+fp64 < %s | FileCheck -check-prefix=ALL -check-prefix=O32 %s +; RUN: llc -march=mips64 -target-abi=n32 -mattr=+msa,+fp64 < %s | FileCheck -check-prefix=ALL -check-prefix=N32 %s +; RUN: llc -march=mips64el -target-abi=n32 -mattr=+msa,+fp64 < %s | FileCheck -check-prefix=ALL -check-prefix=N32 %s +; RUN: llc -march=mips64 -mattr=+msa,+fp64 < %s | FileCheck -check-prefix=ALL -check-prefix=N64 %s +; RUN: llc -march=mips64el -mattr=+msa,+fp64 < %s | FileCheck -check-prefix=ALL -check-prefix=N64 %s @v4f32 = global <4 x float> <float 0.0, float 0.0, float 0.0, float 0.0> @v2f64 = global <2 x double> <double 0.0, double 0.0> @@ -8,322 +12,341 @@ @f64 = global double 0.0 define void @const_v4f32() nounwind { - ; MIPS32-LABEL: const_v4f32: + ; ALL-LABEL: const_v4f32: store volatile <4 x float> <float 0.0, float 0.0, float 0.0, float 0.0>, <4 x float>*@v4f32 - ; MIPS32: ldi.b [[R1:\$w[0-9]+]], 0 + ; ALL: ldi.b [[R1:\$w[0-9]+]], 0 store volatile <4 x float> <float 1.0, float 1.0, float 1.0, float 1.0>, <4 x float>*@v4f32 - ; MIPS32: lui [[R1:\$[0-9]+]], 16256 - ; MIPS32: fill.w [[R2:\$w[0-9]+]], [[R1]] + ; ALL: lui [[R1:\$[0-9]+]], 16256 + ; ALL: fill.w [[R2:\$w[0-9]+]], [[R1]] store volatile <4 x float> <float 1.0, float 1.0, float 1.0, float 31.0>, <4 x float>*@v4f32 - ; MIPS32: addiu [[G_PTR:\$[0-9]+]], {{.*}}, %lo($ - ; MIPS32: ld.w [[R1:\$w[0-9]+]], 0([[G_PTR]]) + ; O32: addiu [[G_PTR:\$[0-9]+]], {{.*}}, %lo($ + ; N32: addiu [[G_PTR:\$[0-9]+]], {{.*}}, %got_ofst($ + ; N64: daddiu [[G_PTR:\$[0-9]+]], {{.*}}, %got_ofst($ + ; ALL: ld.w [[R1:\$w[0-9]+]], 0([[G_PTR]]) store volatile <4 x float> <float 65537.0, float 65537.0, float 65537.0, float 65537.0>, <4 x float>*@v4f32 - ; MIPS32: lui [[R1:\$[0-9]+]], 18304 - ; MIPS32: ori [[R2:\$[0-9]+]], [[R1]], 128 - ; MIPS32: fill.w [[R3:\$w[0-9]+]], [[R2]] + ; ALL: lui [[R1:\$[0-9]+]], 18304 + ; ALL: ori [[R2:\$[0-9]+]], [[R1]], 128 + ; ALL: fill.w [[R3:\$w[0-9]+]], [[R2]] store volatile <4 x float> <float 1.0, float 2.0, float 1.0, float 2.0>, <4 x float>*@v4f32 - ; MIPS32: addiu [[G_PTR:\$[0-9]+]], {{.*}}, %lo($ - ; MIPS32: ld.w [[R1:\$w[0-9]+]], 0([[G_PTR]]) + ; O32: addiu [[G_PTR:\$[0-9]+]], {{.*}}, %lo($ + ; N32: addiu [[G_PTR:\$[0-9]+]], {{.*}}, %got_ofst($ + ; N64: daddiu [[G_PTR:\$[0-9]+]], {{.*}}, %got_ofst($ + ; ALL: ld.w [[R1:\$w[0-9]+]], 0([[G_PTR]]) store volatile <4 x float> <float 3.0, float 4.0, float 5.0, float 6.0>, <4 x float>*@v4f32 - ; MIPS32: addiu [[G_PTR:\$[0-9]+]], {{.*}}, %lo($ - ; MIPS32: ld.w [[R1:\$w[0-9]+]], 0([[G_PTR]]) + ; O32: addiu [[G_PTR:\$[0-9]+]], {{.*}}, %lo($ + ; N32: addiu [[G_PTR:\$[0-9]+]], {{.*}}, %got_ofst($ + ; N64: daddiu [[G_PTR:\$[0-9]+]], {{.*}}, %got_ofst($ + ; ALL: ld.w [[R1:\$w[0-9]+]], 0([[G_PTR]]) ret void - ; MIPS32: .size const_v4f32 } define void @const_v2f64() nounwind { - ; MIPS32-LABEL: const_v2f64: + ; ALL-LABEL: const_v2f64: store volatile <2 x double> <double 0.0, double 0.0>, <2 x double>*@v2f64 - ; MIPS32: ldi.b [[R1:\$w[0-9]+]], 0 + ; ALL: ldi.b [[R1:\$w[0-9]+]], 0 store volatile <2 x double> <double 72340172838076673.0, double 72340172838076673.0>, <2 x double>*@v2f64 - ; MIPS32: addiu [[G_PTR:\$[0-9]+]], {{.*}}, %lo($ - ; MIPS32: ld.d [[R1:\$w[0-9]+]], 0([[G_PTR]]) + ; O32: addiu [[G_PTR:\$[0-9]+]], {{.*}}, %lo($ + ; N32: addiu [[G_PTR:\$[0-9]+]], {{.*}}, %got_ofst($ + ; N64: daddiu [[G_PTR:\$[0-9]+]], {{.*}}, %got_ofst($ + ; ALL: ld.d [[R1:\$w[0-9]+]], 0([[G_PTR]]) store volatile <2 x double> <double 281479271743489.0, double 281479271743489.0>, <2 x double>*@v2f64 - ; MIPS32: addiu [[G_PTR:\$[0-9]+]], {{.*}}, %lo($ - ; MIPS32: ld.d [[R1:\$w[0-9]+]], 0([[G_PTR]]) + ; O32: addiu [[G_PTR:\$[0-9]+]], {{.*}}, %lo($ + ; N32: addiu [[G_PTR:\$[0-9]+]], {{.*}}, %got_ofst($ + ; N64: daddiu [[G_PTR:\$[0-9]+]], {{.*}}, %got_ofst($ + ; ALL: ld.d [[R1:\$w[0-9]+]], 0([[G_PTR]]) store volatile <2 x double> <double 4294967297.0, double 4294967297.0>, <2 x double>*@v2f64 - ; MIPS32: addiu [[G_PTR:\$[0-9]+]], {{.*}}, %lo($ - ; MIPS32: ld.d [[R1:\$w[0-9]+]], 0([[G_PTR]]) + ; O32: addiu [[G_PTR:\$[0-9]+]], {{.*}}, %lo($ + ; N32: addiu [[G_PTR:\$[0-9]+]], {{.*}}, %got_ofst($ + ; N64: daddiu [[G_PTR:\$[0-9]+]], {{.*}}, %got_ofst($ + ; ALL: ld.d [[R1:\$w[0-9]+]], 0([[G_PTR]]) store volatile <2 x double> <double 1.0, double 1.0>, <2 x double>*@v2f64 - ; MIPS32: addiu [[G_PTR:\$[0-9]+]], {{.*}}, %lo($ - ; MIPS32: ld.d [[R1:\$w[0-9]+]], 0([[G_PTR]]) + ; O32: addiu [[G_PTR:\$[0-9]+]], {{.*}}, %lo($ + ; N32: addiu [[G_PTR:\$[0-9]+]], {{.*}}, %got_ofst($ + ; N64: daddiu [[G_PTR:\$[0-9]+]], {{.*}}, %got_ofst($ + ; ALL: ld.d [[R1:\$w[0-9]+]], 0([[G_PTR]]) store volatile <2 x double> <double 1.0, double 31.0>, <2 x double>*@v2f64 - ; MIPS32: addiu [[G_PTR:\$[0-9]+]], {{.*}}, %lo($ - ; MIPS32: ld.d [[R1:\$w[0-9]+]], 0([[G_PTR]]) + ; O32: addiu [[G_PTR:\$[0-9]+]], {{.*}}, %lo($ + ; N32: addiu [[G_PTR:\$[0-9]+]], {{.*}}, %got_ofst($ + ; N64: daddiu [[G_PTR:\$[0-9]+]], {{.*}}, %got_ofst($ + ; ALL: ld.d [[R1:\$w[0-9]+]], 0([[G_PTR]]) store volatile <2 x double> <double 3.0, double 4.0>, <2 x double>*@v2f64 - ; MIPS32: addiu [[G_PTR:\$[0-9]+]], {{.*}}, %lo($ - ; MIPS32: ld.d [[R1:\$w[0-9]+]], 0([[G_PTR]]) + ; O32: addiu [[G_PTR:\$[0-9]+]], {{.*}}, %lo($ + ; N32: addiu [[G_PTR:\$[0-9]+]], {{.*}}, %got_ofst($ + ; N64: daddiu [[G_PTR:\$[0-9]+]], {{.*}}, %got_ofst($ + ; ALL: ld.d [[R1:\$w[0-9]+]], 0([[G_PTR]]) ret void - ; MIPS32: .size const_v2f64 } define void @nonconst_v4f32() nounwind { - ; MIPS32-LABEL: nonconst_v4f32: + ; ALL-LABEL: nonconst_v4f32: - %1 = load float *@f32 + %1 = load float , float *@f32 %2 = insertelement <4 x float> undef, float %1, i32 0 %3 = insertelement <4 x float> %2, float %1, i32 1 %4 = insertelement <4 x float> %3, float %1, i32 2 %5 = insertelement <4 x float> %4, float %1, i32 3 store volatile <4 x float> %5, <4 x float>*@v4f32 - ; MIPS32: lwc1 $f[[R1:[0-9]+]], 0( - ; MIPS32: splati.w [[R2:\$w[0-9]+]], $w[[R1]] + ; ALL: lwc1 $f[[R1:[0-9]+]], 0( + ; ALL: splati.w [[R2:\$w[0-9]+]], $w[[R1]] ret void - ; MIPS32: .size nonconst_v4f32 } define void @nonconst_v2f64() nounwind { - ; MIPS32-LABEL: nonconst_v2f64: + ; ALL-LABEL: nonconst_v2f64: - %1 = load double *@f64 + %1 = load double , double *@f64 %2 = insertelement <2 x double> undef, double %1, i32 0 %3 = insertelement <2 x double> %2, double %1, i32 1 store volatile <2 x double> %3, <2 x double>*@v2f64 - ; MIPS32: ldc1 $f[[R1:[0-9]+]], 0( - ; MIPS32: splati.d [[R2:\$w[0-9]+]], $w[[R1]] + ; ALL: ldc1 $f[[R1:[0-9]+]], 0( + ; ALL: splati.d [[R2:\$w[0-9]+]], $w[[R1]] ret void - ; MIPS32: .size nonconst_v2f64 } define float @extract_v4f32() nounwind { - ; MIPS32-LABEL: extract_v4f32: + ; ALL-LABEL: extract_v4f32: - %1 = load <4 x float>* @v4f32 - ; MIPS32-DAG: ld.w [[R1:\$w[0-9]+]], + %1 = load <4 x float>, <4 x float>* @v4f32 + ; ALL-DAG: ld.w [[R1:\$w[0-9]+]], %2 = fadd <4 x float> %1, %1 - ; MIPS32-DAG: fadd.w [[R2:\$w[0-9]+]], [[R1]], [[R1]] + ; ALL-DAG: fadd.w [[R2:\$w[0-9]+]], [[R1]], [[R1]] %3 = extractelement <4 x float> %2, i32 1 ; Element 1 can be obtained by splatting it across the vector and extracting ; $w0:sub_lo - ; MIPS32-DAG: splati.w $w0, [[R1]][1] + ; ALL-DAG: splati.w $w0, [[R1]][1] ret float %3 - ; MIPS32: .size extract_v4f32 } define float @extract_v4f32_elt0() nounwind { - ; MIPS32-LABEL: extract_v4f32_elt0: + ; ALL-LABEL: extract_v4f32_elt0: - %1 = load <4 x float>* @v4f32 - ; MIPS32-DAG: ld.w [[R1:\$w[0-9]+]], + %1 = load <4 x float>, <4 x float>* @v4f32 + ; ALL-DAG: ld.w [[R1:\$w[0-9]+]], %2 = fadd <4 x float> %1, %1 - ; MIPS32-DAG: fadd.w $w0, [[R1]], [[R1]] + ; ALL-DAG: fadd.w $w0, [[R1]], [[R1]] %3 = extractelement <4 x float> %2, i32 0 ; Element 0 can be obtained by extracting $w0:sub_lo ($f0) - ; MIPS32-NOT: copy_u.w - ; MIPS32-NOT: mtc1 + ; ALL-NOT: copy_u.w + ; ALL-NOT: mtc1 ret float %3 - ; MIPS32: .size extract_v4f32_elt0 } define float @extract_v4f32_elt2() nounwind { - ; MIPS32-LABEL: extract_v4f32_elt2: + ; ALL-LABEL: extract_v4f32_elt2: - %1 = load <4 x float>* @v4f32 - ; MIPS32-DAG: ld.w [[R1:\$w[0-9]+]], + %1 = load <4 x float>, <4 x float>* @v4f32 + ; ALL-DAG: ld.w [[R1:\$w[0-9]+]], %2 = fadd <4 x float> %1, %1 - ; MIPS32-DAG: fadd.w [[R2:\$w[0-9]+]], [[R1]], [[R1]] + ; ALL-DAG: fadd.w [[R2:\$w[0-9]+]], [[R1]], [[R1]] %3 = extractelement <4 x float> %2, i32 2 ; Element 2 can be obtained by splatting it across the vector and extracting ; $w0:sub_lo - ; MIPS32-DAG: splati.w $w0, [[R1]][2] + ; ALL-DAG: splati.w $w0, [[R1]][2] ret float %3 - ; MIPS32: .size extract_v4f32_elt2 } define float @extract_v4f32_vidx() nounwind { - ; MIPS32-LABEL: extract_v4f32_vidx: + ; ALL-LABEL: extract_v4f32_vidx: - %1 = load <4 x float>* @v4f32 - ; MIPS32-DAG: lw [[PTR_V:\$[0-9]+]], %got(v4f32)( - ; MIPS32-DAG: ld.w [[R1:\$w[0-9]+]], 0([[PTR_V]]) + %1 = load <4 x float>, <4 x float>* @v4f32 + ; O32-DAG: lw [[PTR_V:\$[0-9]+]], %got(v4f32)( + ; N32-DAG: lw [[PTR_V:\$[0-9]+]], %got_disp(v4f32)( + ; N64-DAG: ld [[PTR_V:\$[0-9]+]], %got_disp(v4f32)( + ; ALL-DAG: ld.w [[R1:\$w[0-9]+]], 0([[PTR_V]]) %2 = fadd <4 x float> %1, %1 - ; MIPS32-DAG: fadd.w [[R2:\$w[0-9]+]], [[R1]], [[R1]] + ; ALL-DAG: fadd.w [[R2:\$w[0-9]+]], [[R1]], [[R1]] - %3 = load i32* @i32 - ; MIPS32-DAG: lw [[PTR_I:\$[0-9]+]], %got(i32)( - ; MIPS32-DAG: lw [[IDX:\$[0-9]+]], 0([[PTR_I]]) + %3 = load i32, i32* @i32 + ; O32-DAG: lw [[PTR_I:\$[0-9]+]], %got(i32)( + ; N32-DAG: lw [[PTR_I:\$[0-9]+]], %got_disp(i32)( + ; N64-DAG: ld [[PTR_I:\$[0-9]+]], %got_disp(i32)( + ; ALL-DAG: lw [[IDX:\$[0-9]+]], 0([[PTR_I]]) %4 = extractelement <4 x float> %2, i32 %3 - ; MIPS32-DAG: splat.w $w0, [[R1]]{{\[}}[[IDX]]] + ; ALL-DAG: splat.w $w0, [[R1]]{{\[}}[[IDX]]] ret float %4 - ; MIPS32: .size extract_v4f32_vidx } define double @extract_v2f64() nounwind { - ; MIPS32-LABEL: extract_v2f64: + ; ALL-LABEL: extract_v2f64: - %1 = load <2 x double>* @v2f64 - ; MIPS32-DAG: ld.d [[R1:\$w[0-9]+]], + %1 = load <2 x double>, <2 x double>* @v2f64 + ; ALL-DAG: ld.d [[R1:\$w[0-9]+]], %2 = fadd <2 x double> %1, %1 - ; MIPS32-DAG: fadd.d [[R2:\$w[0-9]+]], [[R1]], [[R1]] + ; ALL-DAG: fadd.d [[R2:\$w[0-9]+]], [[R1]], [[R1]] %3 = extractelement <2 x double> %2, i32 1 ; Element 1 can be obtained by splatting it across the vector and extracting ; $w0:sub_64 - ; MIPS32-DAG: splati.d $w0, [[R1]][1] - ; MIPS32-NOT: copy_u.w - ; MIPS32-NOT: mtc1 - ; MIPS32-NOT: mthc1 - ; MIPS32-NOT: sll - ; MIPS32-NOT: sra + ; ALL-DAG: splati.d $w0, [[R1]][1] + ; ALL-NOT: copy_u.w + ; ALL-NOT: mtc1 + ; ALL-NOT: mthc1 + ; ALL-NOT: sll + ; ALL-NOT: sra ret double %3 - ; MIPS32: .size extract_v2f64 } define double @extract_v2f64_elt0() nounwind { - ; MIPS32-LABEL: extract_v2f64_elt0: + ; ALL-LABEL: extract_v2f64_elt0: - %1 = load <2 x double>* @v2f64 - ; MIPS32-DAG: ld.d [[R1:\$w[0-9]+]], + %1 = load <2 x double>, <2 x double>* @v2f64 + ; ALL-DAG: ld.d [[R1:\$w[0-9]+]], %2 = fadd <2 x double> %1, %1 - ; MIPS32-DAG: fadd.d $w0, [[R1]], [[R1]] + ; ALL-DAG: fadd.d $w0, [[R1]], [[R1]] %3 = extractelement <2 x double> %2, i32 0 ; Element 0 can be obtained by extracting $w0:sub_64 ($f0) - ; MIPS32-NOT: copy_u.w - ; MIPS32-NOT: mtc1 - ; MIPS32-NOT: mthc1 - ; MIPS32-NOT: sll - ; MIPS32-NOT: sra + ; ALL-NOT: copy_u.w + ; ALL-NOT: mtc1 + ; ALL-NOT: mthc1 + ; ALL-NOT: sll + ; ALL-NOT: sra ret double %3 - ; MIPS32: .size extract_v2f64_elt0 } define double @extract_v2f64_vidx() nounwind { - ; MIPS32-LABEL: extract_v2f64_vidx: + ; ALL-LABEL: extract_v2f64_vidx: - %1 = load <2 x double>* @v2f64 - ; MIPS32-DAG: lw [[PTR_V:\$[0-9]+]], %got(v2f64)( - ; MIPS32-DAG: ld.d [[R1:\$w[0-9]+]], 0([[PTR_V]]) + %1 = load <2 x double>, <2 x double>* @v2f64 + ; O32-DAG: lw [[PTR_V:\$[0-9]+]], %got(v2f64)( + ; N32-DAG: lw [[PTR_V:\$[0-9]+]], %got_disp(v2f64)( + ; N64-DAG: ld [[PTR_V:\$[0-9]+]], %got_disp(v2f64)( + ; ALL-DAG: ld.d [[R1:\$w[0-9]+]], 0([[PTR_V]]) %2 = fadd <2 x double> %1, %1 - ; MIPS32-DAG: fadd.d [[R2:\$w[0-9]+]], [[R1]], [[R1]] + ; ALL-DAG: fadd.d [[R2:\$w[0-9]+]], [[R1]], [[R1]] - %3 = load i32* @i32 - ; MIPS32-DAG: lw [[PTR_I:\$[0-9]+]], %got(i32)( - ; MIPS32-DAG: lw [[IDX:\$[0-9]+]], 0([[PTR_I]]) + %3 = load i32, i32* @i32 + ; O32-DAG: lw [[PTR_I:\$[0-9]+]], %got(i32)( + ; N32-DAG: lw [[PTR_I:\$[0-9]+]], %got_disp(i32)( + ; N64-DAG: ld [[PTR_I:\$[0-9]+]], %got_disp(i32)( + ; ALL-DAG: lw [[IDX:\$[0-9]+]], 0([[PTR_I]]) %4 = extractelement <2 x double> %2, i32 %3 - ; MIPS32-DAG: splat.d $w0, [[R1]]{{\[}}[[IDX]]] + ; ALL-DAG: splat.d $w0, [[R1]]{{\[}}[[IDX]]] ret double %4 - ; MIPS32: .size extract_v2f64_vidx } define void @insert_v4f32(float %a) nounwind { - ; MIPS32-LABEL: insert_v4f32: + ; ALL-LABEL: insert_v4f32: - %1 = load <4 x float>* @v4f32 - ; MIPS32-DAG: ld.w [[R1:\$w[0-9]+]], + %1 = load <4 x float>, <4 x float>* @v4f32 + ; ALL-DAG: ld.w [[R1:\$w[0-9]+]], %2 = insertelement <4 x float> %1, float %a, i32 1 ; float argument passed in $f12 - ; MIPS32-DAG: insve.w [[R1]][1], $w12[0] + ; ALL-DAG: insve.w [[R1]][1], $w12[0] store <4 x float> %2, <4 x float>* @v4f32 - ; MIPS32-DAG: st.w [[R1]] + ; ALL-DAG: st.w [[R1]] ret void - ; MIPS32: .size insert_v4f32 } define void @insert_v2f64(double %a) nounwind { - ; MIPS32-LABEL: insert_v2f64: + ; ALL-LABEL: insert_v2f64: - %1 = load <2 x double>* @v2f64 - ; MIPS32-DAG: ld.d [[R1:\$w[0-9]+]], + %1 = load <2 x double>, <2 x double>* @v2f64 + ; ALL-DAG: ld.d [[R1:\$w[0-9]+]], %2 = insertelement <2 x double> %1, double %a, i32 1 ; double argument passed in $f12 - ; MIPS32-DAG: insve.d [[R1]][1], $w12[0] + ; ALL-DAG: insve.d [[R1]][1], $w12[0] store <2 x double> %2, <2 x double>* @v2f64 - ; MIPS32-DAG: st.d [[R1]] + ; ALL-DAG: st.d [[R1]] ret void - ; MIPS32: .size insert_v2f64 } define void @insert_v4f32_vidx(float %a) nounwind { - ; MIPS32-LABEL: insert_v4f32_vidx: + ; ALL-LABEL: insert_v4f32_vidx: - %1 = load <4 x float>* @v4f32 - ; MIPS32-DAG: lw [[PTR_V:\$[0-9]+]], %got(v4f32)( - ; MIPS32-DAG: ld.w [[R1:\$w[0-9]+]], 0([[PTR_V]]) + %1 = load <4 x float>, <4 x float>* @v4f32 + ; O32-DAG: lw [[PTR_V:\$[0-9]+]], %got(v4f32)( + ; N32-DAG: lw [[PTR_V:\$[0-9]+]], %got_disp(v4f32)( + ; N64-DAG: ld [[PTR_V:\$[0-9]+]], %got_disp(v4f32)( + ; ALL-DAG: ld.w [[R1:\$w[0-9]+]], 0([[PTR_V]]) - %2 = load i32* @i32 - ; MIPS32-DAG: lw [[PTR_I:\$[0-9]+]], %got(i32)( - ; MIPS32-DAG: lw [[IDX:\$[0-9]+]], 0([[PTR_I]]) + %2 = load i32, i32* @i32 + ; O32-DAG: lw [[PTR_I:\$[0-9]+]], %got(i32)( + ; N32-DAG: lw [[PTR_I:\$[0-9]+]], %got_disp(i32)( + ; N64-DAG: ld [[PTR_I:\$[0-9]+]], %got_disp(i32)( + ; ALL-DAG: lw [[IDX:\$[0-9]+]], 0([[PTR_I]]) %3 = insertelement <4 x float> %1, float %a, i32 %2 ; float argument passed in $f12 - ; MIPS32-DAG: sll [[BIDX:\$[0-9]+]], [[IDX]], 2 - ; MIPS32-DAG: sld.b [[R1]], [[R1]]{{\[}}[[BIDX]]] - ; MIPS32-DAG: insve.w [[R1]][0], $w12[0] - ; MIPS32-DAG: neg [[NIDX:\$[0-9]+]], [[BIDX]] - ; MIPS32-DAG: sld.b [[R1]], [[R1]]{{\[}}[[NIDX]]] + ; ALL-DAG: sll [[BIDX:\$[0-9]+]], [[IDX]], 2 + ; ALL-DAG: sld.b [[R1]], [[R1]]{{\[}}[[BIDX]]] + ; ALL-DAG: insve.w [[R1]][0], $w12[0] + ; ALL-DAG: neg [[NIDX:\$[0-9]+]], [[BIDX]] + ; ALL-DAG: sld.b [[R1]], [[R1]]{{\[}}[[NIDX]]] store <4 x float> %3, <4 x float>* @v4f32 - ; MIPS32-DAG: st.w [[R1]] + ; ALL-DAG: st.w [[R1]] ret void - ; MIPS32: .size insert_v4f32_vidx } define void @insert_v2f64_vidx(double %a) nounwind { - ; MIPS32-LABEL: insert_v2f64_vidx: + ; ALL-LABEL: insert_v2f64_vidx: - %1 = load <2 x double>* @v2f64 - ; MIPS32-DAG: lw [[PTR_V:\$[0-9]+]], %got(v2f64)( - ; MIPS32-DAG: ld.d [[R1:\$w[0-9]+]], 0([[PTR_V]]) + %1 = load <2 x double>, <2 x double>* @v2f64 + ; O32-DAG: lw [[PTR_V:\$[0-9]+]], %got(v2f64)( + ; N32-DAG: lw [[PTR_V:\$[0-9]+]], %got_disp(v2f64)( + ; N64-DAG: ld [[PTR_V:\$[0-9]+]], %got_disp(v2f64)( + ; ALL-DAG: ld.d [[R1:\$w[0-9]+]], 0([[PTR_V]]) - %2 = load i32* @i32 - ; MIPS32-DAG: lw [[PTR_I:\$[0-9]+]], %got(i32)( - ; MIPS32-DAG: lw [[IDX:\$[0-9]+]], 0([[PTR_I]]) + %2 = load i32, i32* @i32 + ; O32-DAG: lw [[PTR_I:\$[0-9]+]], %got(i32)( + ; N32-DAG: lw [[PTR_I:\$[0-9]+]], %got_disp(i32)( + ; N64-DAG: ld [[PTR_I:\$[0-9]+]], %got_disp(i32)( + ; ALL-DAG: lw [[IDX:\$[0-9]+]], 0([[PTR_I]]) %3 = insertelement <2 x double> %1, double %a, i32 %2 ; double argument passed in $f12 - ; MIPS32-DAG: sll [[BIDX:\$[0-9]+]], [[IDX]], 3 - ; MIPS32-DAG: sld.b [[R1]], [[R1]]{{\[}}[[BIDX]]] - ; MIPS32-DAG: insve.d [[R1]][0], $w12[0] - ; MIPS32-DAG: neg [[NIDX:\$[0-9]+]], [[BIDX]] - ; MIPS32-DAG: sld.b [[R1]], [[R1]]{{\[}}[[NIDX]]] + ; ALL-DAG: sll [[BIDX:\$[0-9]+]], [[IDX]], 3 + ; ALL-DAG: sld.b [[R1]], [[R1]]{{\[}}[[BIDX]]] + ; ALL-DAG: insve.d [[R1]][0], $w12[0] + ; ALL-DAG: neg [[NIDX:\$[0-9]+]], [[BIDX]] + ; ALL-DAG: sld.b [[R1]], [[R1]]{{\[}}[[NIDX]]] store <2 x double> %3, <2 x double>* @v2f64 - ; MIPS32-DAG: st.d [[R1]] + ; ALL-DAG: st.d [[R1]] ret void - ; MIPS32: .size insert_v2f64_vidx } diff --git a/test/CodeGen/Mips/msa/bit.ll b/test/CodeGen/Mips/msa/bit.ll index 59ddbe17a33f7..f0057307bbf78 100644 --- a/test/CodeGen/Mips/msa/bit.ll +++ b/test/CodeGen/Mips/msa/bit.ll @@ -8,7 +8,7 @@ define void @llvm_mips_sat_s_b_test() nounwind { entry: - %0 = load <16 x i8>* @llvm_mips_sat_s_b_ARG1 + %0 = load <16 x i8>, <16 x i8>* @llvm_mips_sat_s_b_ARG1 %1 = tail call <16 x i8> @llvm.mips.sat.s.b(<16 x i8> %0, i32 7) store <16 x i8> %1, <16 x i8>* @llvm_mips_sat_s_b_RES ret void @@ -27,7 +27,7 @@ declare <16 x i8> @llvm.mips.sat.s.b(<16 x i8>, i32) nounwind define void @llvm_mips_sat_s_h_test() nounwind { entry: - %0 = load <8 x i16>* @llvm_mips_sat_s_h_ARG1 + %0 = load <8 x i16>, <8 x i16>* @llvm_mips_sat_s_h_ARG1 %1 = tail call <8 x i16> @llvm.mips.sat.s.h(<8 x i16> %0, i32 7) store <8 x i16> %1, <8 x i16>* @llvm_mips_sat_s_h_RES ret void @@ -46,7 +46,7 @@ declare <8 x i16> @llvm.mips.sat.s.h(<8 x i16>, i32) nounwind define void @llvm_mips_sat_s_w_test() nounwind { entry: - %0 = load <4 x i32>* @llvm_mips_sat_s_w_ARG1 + %0 = load <4 x i32>, <4 x i32>* @llvm_mips_sat_s_w_ARG1 %1 = tail call <4 x i32> @llvm.mips.sat.s.w(<4 x i32> %0, i32 7) store <4 x i32> %1, <4 x i32>* @llvm_mips_sat_s_w_RES ret void @@ -65,7 +65,7 @@ declare <4 x i32> @llvm.mips.sat.s.w(<4 x i32>, i32) nounwind define void @llvm_mips_sat_s_d_test() nounwind { entry: - %0 = load <2 x i64>* @llvm_mips_sat_s_d_ARG1 + %0 = load <2 x i64>, <2 x i64>* @llvm_mips_sat_s_d_ARG1 %1 = tail call <2 x i64> @llvm.mips.sat.s.d(<2 x i64> %0, i32 7) store <2 x i64> %1, <2 x i64>* @llvm_mips_sat_s_d_RES ret void @@ -84,7 +84,7 @@ declare <2 x i64> @llvm.mips.sat.s.d(<2 x i64>, i32) nounwind define void @llvm_mips_sat_u_b_test() nounwind { entry: - %0 = load <16 x i8>* @llvm_mips_sat_u_b_ARG1 + %0 = load <16 x i8>, <16 x i8>* @llvm_mips_sat_u_b_ARG1 %1 = tail call <16 x i8> @llvm.mips.sat.u.b(<16 x i8> %0, i32 7) store <16 x i8> %1, <16 x i8>* @llvm_mips_sat_u_b_RES ret void @@ -103,7 +103,7 @@ declare <16 x i8> @llvm.mips.sat.u.b(<16 x i8>, i32) nounwind define void @llvm_mips_sat_u_h_test() nounwind { entry: - %0 = load <8 x i16>* @llvm_mips_sat_u_h_ARG1 + %0 = load <8 x i16>, <8 x i16>* @llvm_mips_sat_u_h_ARG1 %1 = tail call <8 x i16> @llvm.mips.sat.u.h(<8 x i16> %0, i32 7) store <8 x i16> %1, <8 x i16>* @llvm_mips_sat_u_h_RES ret void @@ -122,7 +122,7 @@ declare <8 x i16> @llvm.mips.sat.u.h(<8 x i16>, i32) nounwind define void @llvm_mips_sat_u_w_test() nounwind { entry: - %0 = load <4 x i32>* @llvm_mips_sat_u_w_ARG1 + %0 = load <4 x i32>, <4 x i32>* @llvm_mips_sat_u_w_ARG1 %1 = tail call <4 x i32> @llvm.mips.sat.u.w(<4 x i32> %0, i32 7) store <4 x i32> %1, <4 x i32>* @llvm_mips_sat_u_w_RES ret void @@ -141,7 +141,7 @@ declare <4 x i32> @llvm.mips.sat.u.w(<4 x i32>, i32) nounwind define void @llvm_mips_sat_u_d_test() nounwind { entry: - %0 = load <2 x i64>* @llvm_mips_sat_u_d_ARG1 + %0 = load <2 x i64>, <2 x i64>* @llvm_mips_sat_u_d_ARG1 %1 = tail call <2 x i64> @llvm.mips.sat.u.d(<2 x i64> %0, i32 7) store <2 x i64> %1, <2 x i64>* @llvm_mips_sat_u_d_RES ret void @@ -160,7 +160,7 @@ declare <2 x i64> @llvm.mips.sat.u.d(<2 x i64>, i32) nounwind define void @llvm_mips_slli_b_test() nounwind { entry: - %0 = load <16 x i8>* @llvm_mips_slli_b_ARG1 + %0 = load <16 x i8>, <16 x i8>* @llvm_mips_slli_b_ARG1 %1 = tail call <16 x i8> @llvm.mips.slli.b(<16 x i8> %0, i32 7) store <16 x i8> %1, <16 x i8>* @llvm_mips_slli_b_RES ret void @@ -179,7 +179,7 @@ declare <16 x i8> @llvm.mips.slli.b(<16 x i8>, i32) nounwind define void @llvm_mips_slli_h_test() nounwind { entry: - %0 = load <8 x i16>* @llvm_mips_slli_h_ARG1 + %0 = load <8 x i16>, <8 x i16>* @llvm_mips_slli_h_ARG1 %1 = tail call <8 x i16> @llvm.mips.slli.h(<8 x i16> %0, i32 7) store <8 x i16> %1, <8 x i16>* @llvm_mips_slli_h_RES ret void @@ -198,7 +198,7 @@ declare <8 x i16> @llvm.mips.slli.h(<8 x i16>, i32) nounwind define void @llvm_mips_slli_w_test() nounwind { entry: - %0 = load <4 x i32>* @llvm_mips_slli_w_ARG1 + %0 = load <4 x i32>, <4 x i32>* @llvm_mips_slli_w_ARG1 %1 = tail call <4 x i32> @llvm.mips.slli.w(<4 x i32> %0, i32 7) store <4 x i32> %1, <4 x i32>* @llvm_mips_slli_w_RES ret void @@ -217,7 +217,7 @@ declare <4 x i32> @llvm.mips.slli.w(<4 x i32>, i32) nounwind define void @llvm_mips_slli_d_test() nounwind { entry: - %0 = load <2 x i64>* @llvm_mips_slli_d_ARG1 + %0 = load <2 x i64>, <2 x i64>* @llvm_mips_slli_d_ARG1 %1 = tail call <2 x i64> @llvm.mips.slli.d(<2 x i64> %0, i32 7) store <2 x i64> %1, <2 x i64>* @llvm_mips_slli_d_RES ret void @@ -236,7 +236,7 @@ declare <2 x i64> @llvm.mips.slli.d(<2 x i64>, i32) nounwind define void @llvm_mips_srai_b_test() nounwind { entry: - %0 = load <16 x i8>* @llvm_mips_srai_b_ARG1 + %0 = load <16 x i8>, <16 x i8>* @llvm_mips_srai_b_ARG1 %1 = tail call <16 x i8> @llvm.mips.srai.b(<16 x i8> %0, i32 7) store <16 x i8> %1, <16 x i8>* @llvm_mips_srai_b_RES ret void @@ -255,7 +255,7 @@ declare <16 x i8> @llvm.mips.srai.b(<16 x i8>, i32) nounwind define void @llvm_mips_srai_h_test() nounwind { entry: - %0 = load <8 x i16>* @llvm_mips_srai_h_ARG1 + %0 = load <8 x i16>, <8 x i16>* @llvm_mips_srai_h_ARG1 %1 = tail call <8 x i16> @llvm.mips.srai.h(<8 x i16> %0, i32 7) store <8 x i16> %1, <8 x i16>* @llvm_mips_srai_h_RES ret void @@ -274,7 +274,7 @@ declare <8 x i16> @llvm.mips.srai.h(<8 x i16>, i32) nounwind define void @llvm_mips_srai_w_test() nounwind { entry: - %0 = load <4 x i32>* @llvm_mips_srai_w_ARG1 + %0 = load <4 x i32>, <4 x i32>* @llvm_mips_srai_w_ARG1 %1 = tail call <4 x i32> @llvm.mips.srai.w(<4 x i32> %0, i32 7) store <4 x i32> %1, <4 x i32>* @llvm_mips_srai_w_RES ret void @@ -293,7 +293,7 @@ declare <4 x i32> @llvm.mips.srai.w(<4 x i32>, i32) nounwind define void @llvm_mips_srai_d_test() nounwind { entry: - %0 = load <2 x i64>* @llvm_mips_srai_d_ARG1 + %0 = load <2 x i64>, <2 x i64>* @llvm_mips_srai_d_ARG1 %1 = tail call <2 x i64> @llvm.mips.srai.d(<2 x i64> %0, i32 7) store <2 x i64> %1, <2 x i64>* @llvm_mips_srai_d_RES ret void @@ -312,7 +312,7 @@ declare <2 x i64> @llvm.mips.srai.d(<2 x i64>, i32) nounwind define void @llvm_mips_srari_b_test() nounwind { entry: - %0 = load <16 x i8>* @llvm_mips_srari_b_ARG1 + %0 = load <16 x i8>, <16 x i8>* @llvm_mips_srari_b_ARG1 %1 = tail call <16 x i8> @llvm.mips.srari.b(<16 x i8> %0, i32 7) store <16 x i8> %1, <16 x i8>* @llvm_mips_srari_b_RES ret void @@ -331,7 +331,7 @@ declare <16 x i8> @llvm.mips.srari.b(<16 x i8>, i32) nounwind define void @llvm_mips_srari_h_test() nounwind { entry: - %0 = load <8 x i16>* @llvm_mips_srari_h_ARG1 + %0 = load <8 x i16>, <8 x i16>* @llvm_mips_srari_h_ARG1 %1 = tail call <8 x i16> @llvm.mips.srari.h(<8 x i16> %0, i32 7) store <8 x i16> %1, <8 x i16>* @llvm_mips_srari_h_RES ret void @@ -350,7 +350,7 @@ declare <8 x i16> @llvm.mips.srari.h(<8 x i16>, i32) nounwind define void @llvm_mips_srari_w_test() nounwind { entry: - %0 = load <4 x i32>* @llvm_mips_srari_w_ARG1 + %0 = load <4 x i32>, <4 x i32>* @llvm_mips_srari_w_ARG1 %1 = tail call <4 x i32> @llvm.mips.srari.w(<4 x i32> %0, i32 7) store <4 x i32> %1, <4 x i32>* @llvm_mips_srari_w_RES ret void @@ -369,7 +369,7 @@ declare <4 x i32> @llvm.mips.srari.w(<4 x i32>, i32) nounwind define void @llvm_mips_srari_d_test() nounwind { entry: - %0 = load <2 x i64>* @llvm_mips_srari_d_ARG1 + %0 = load <2 x i64>, <2 x i64>* @llvm_mips_srari_d_ARG1 %1 = tail call <2 x i64> @llvm.mips.srari.d(<2 x i64> %0, i32 7) store <2 x i64> %1, <2 x i64>* @llvm_mips_srari_d_RES ret void @@ -388,7 +388,7 @@ declare <2 x i64> @llvm.mips.srari.d(<2 x i64>, i32) nounwind define void @llvm_mips_srli_b_test() nounwind { entry: - %0 = load <16 x i8>* @llvm_mips_srli_b_ARG1 + %0 = load <16 x i8>, <16 x i8>* @llvm_mips_srli_b_ARG1 %1 = tail call <16 x i8> @llvm.mips.srli.b(<16 x i8> %0, i32 7) store <16 x i8> %1, <16 x i8>* @llvm_mips_srli_b_RES ret void @@ -407,7 +407,7 @@ declare <16 x i8> @llvm.mips.srli.b(<16 x i8>, i32) nounwind define void @llvm_mips_srli_h_test() nounwind { entry: - %0 = load <8 x i16>* @llvm_mips_srli_h_ARG1 + %0 = load <8 x i16>, <8 x i16>* @llvm_mips_srli_h_ARG1 %1 = tail call <8 x i16> @llvm.mips.srli.h(<8 x i16> %0, i32 7) store <8 x i16> %1, <8 x i16>* @llvm_mips_srli_h_RES ret void @@ -426,7 +426,7 @@ declare <8 x i16> @llvm.mips.srli.h(<8 x i16>, i32) nounwind define void @llvm_mips_srli_w_test() nounwind { entry: - %0 = load <4 x i32>* @llvm_mips_srli_w_ARG1 + %0 = load <4 x i32>, <4 x i32>* @llvm_mips_srli_w_ARG1 %1 = tail call <4 x i32> @llvm.mips.srli.w(<4 x i32> %0, i32 7) store <4 x i32> %1, <4 x i32>* @llvm_mips_srli_w_RES ret void @@ -445,7 +445,7 @@ declare <4 x i32> @llvm.mips.srli.w(<4 x i32>, i32) nounwind define void @llvm_mips_srli_d_test() nounwind { entry: - %0 = load <2 x i64>* @llvm_mips_srli_d_ARG1 + %0 = load <2 x i64>, <2 x i64>* @llvm_mips_srli_d_ARG1 %1 = tail call <2 x i64> @llvm.mips.srli.d(<2 x i64> %0, i32 7) store <2 x i64> %1, <2 x i64>* @llvm_mips_srli_d_RES ret void @@ -464,7 +464,7 @@ declare <2 x i64> @llvm.mips.srli.d(<2 x i64>, i32) nounwind define void @llvm_mips_srlri_b_test() nounwind { entry: - %0 = load <16 x i8>* @llvm_mips_srlri_b_ARG1 + %0 = load <16 x i8>, <16 x i8>* @llvm_mips_srlri_b_ARG1 %1 = tail call <16 x i8> @llvm.mips.srlri.b(<16 x i8> %0, i32 7) store <16 x i8> %1, <16 x i8>* @llvm_mips_srlri_b_RES ret void @@ -483,7 +483,7 @@ declare <16 x i8> @llvm.mips.srlri.b(<16 x i8>, i32) nounwind define void @llvm_mips_srlri_h_test() nounwind { entry: - %0 = load <8 x i16>* @llvm_mips_srlri_h_ARG1 + %0 = load <8 x i16>, <8 x i16>* @llvm_mips_srlri_h_ARG1 %1 = tail call <8 x i16> @llvm.mips.srlri.h(<8 x i16> %0, i32 7) store <8 x i16> %1, <8 x i16>* @llvm_mips_srlri_h_RES ret void @@ -502,7 +502,7 @@ declare <8 x i16> @llvm.mips.srlri.h(<8 x i16>, i32) nounwind define void @llvm_mips_srlri_w_test() nounwind { entry: - %0 = load <4 x i32>* @llvm_mips_srlri_w_ARG1 + %0 = load <4 x i32>, <4 x i32>* @llvm_mips_srlri_w_ARG1 %1 = tail call <4 x i32> @llvm.mips.srlri.w(<4 x i32> %0, i32 7) store <4 x i32> %1, <4 x i32>* @llvm_mips_srlri_w_RES ret void @@ -521,7 +521,7 @@ declare <4 x i32> @llvm.mips.srlri.w(<4 x i32>, i32) nounwind define void @llvm_mips_srlri_d_test() nounwind { entry: - %0 = load <2 x i64>* @llvm_mips_srlri_d_ARG1 + %0 = load <2 x i64>, <2 x i64>* @llvm_mips_srlri_d_ARG1 %1 = tail call <2 x i64> @llvm.mips.srlri.d(<2 x i64> %0, i32 7) store <2 x i64> %1, <2 x i64>* @llvm_mips_srlri_d_RES ret void diff --git a/test/CodeGen/Mips/msa/bitcast.ll b/test/CodeGen/Mips/msa/bitcast.ll index 8e880ecd9afbc..837cc28aa82d3 100644 --- a/test/CodeGen/Mips/msa/bitcast.ll +++ b/test/CodeGen/Mips/msa/bitcast.ll @@ -5,7 +5,7 @@ define void @v16i8_to_v16i8(<16 x i8>* %src, <16 x i8>* %dst) nounwind { entry: - %0 = load volatile <16 x i8>* %src + %0 = load volatile <16 x i8>, <16 x i8>* %src %1 = tail call <16 x i8> @llvm.mips.addv.b(<16 x i8> %0, <16 x i8> %0) %2 = bitcast <16 x i8> %1 to <16 x i8> %3 = tail call <16 x i8> @llvm.mips.addv.b(<16 x i8> %2, <16 x i8> %2) @@ -29,7 +29,7 @@ entry: define void @v16i8_to_v8i16(<16 x i8>* %src, <8 x i16>* %dst) nounwind { entry: - %0 = load volatile <16 x i8>* %src + %0 = load volatile <16 x i8>, <16 x i8>* %src %1 = tail call <16 x i8> @llvm.mips.addv.b(<16 x i8> %0, <16 x i8> %0) %2 = bitcast <16 x i8> %1 to <8 x i16> %3 = tail call <8 x i16> @llvm.mips.addv.h(<8 x i16> %2, <8 x i16> %2) @@ -56,7 +56,7 @@ entry: ; are no operations for v8f16 to put in the way. define void @v16i8_to_v8f16(<16 x i8>* %src, <8 x half>* %dst) nounwind { entry: - %0 = load volatile <16 x i8>* %src + %0 = load volatile <16 x i8>, <16 x i8>* %src %1 = tail call <16 x i8> @llvm.mips.addv.b(<16 x i8> %0, <16 x i8> %0) %2 = bitcast <16 x i8> %1 to <8 x half> store <8 x half> %2, <8 x half>* %dst @@ -77,7 +77,7 @@ entry: define void @v16i8_to_v4i32(<16 x i8>* %src, <4 x i32>* %dst) nounwind { entry: - %0 = load volatile <16 x i8>* %src + %0 = load volatile <16 x i8>, <16 x i8>* %src %1 = tail call <16 x i8> @llvm.mips.addv.b(<16 x i8> %0, <16 x i8> %0) %2 = bitcast <16 x i8> %1 to <4 x i32> %3 = tail call <4 x i32> @llvm.mips.addv.w(<4 x i32> %2, <4 x i32> %2) @@ -102,7 +102,7 @@ entry: define void @v16i8_to_v4f32(<16 x i8>* %src, <4 x float>* %dst) nounwind { entry: - %0 = load volatile <16 x i8>* %src + %0 = load volatile <16 x i8>, <16 x i8>* %src %1 = tail call <16 x i8> @llvm.mips.addv.b(<16 x i8> %0, <16 x i8> %0) %2 = bitcast <16 x i8> %1 to <4 x float> %3 = tail call <4 x float> @llvm.mips.fadd.w(<4 x float> %2, <4 x float> %2) @@ -127,7 +127,7 @@ entry: define void @v16i8_to_v2i64(<16 x i8>* %src, <2 x i64>* %dst) nounwind { entry: - %0 = load volatile <16 x i8>* %src + %0 = load volatile <16 x i8>, <16 x i8>* %src %1 = tail call <16 x i8> @llvm.mips.addv.b(<16 x i8> %0, <16 x i8> %0) %2 = bitcast <16 x i8> %1 to <2 x i64> %3 = tail call <2 x i64> @llvm.mips.addv.d(<2 x i64> %2, <2 x i64> %2) @@ -153,7 +153,7 @@ entry: define void @v16i8_to_v2f64(<16 x i8>* %src, <2 x double>* %dst) nounwind { entry: - %0 = load volatile <16 x i8>* %src + %0 = load volatile <16 x i8>, <16 x i8>* %src %1 = tail call <16 x i8> @llvm.mips.addv.b(<16 x i8> %0, <16 x i8> %0) %2 = bitcast <16 x i8> %1 to <2 x double> %3 = tail call <2 x double> @llvm.mips.fadd.d(<2 x double> %2, <2 x double> %2) @@ -179,7 +179,7 @@ entry: define void @v8i16_to_v16i8(<8 x i16>* %src, <16 x i8>* %dst) nounwind { entry: - %0 = load volatile <8 x i16>* %src + %0 = load volatile <8 x i16>, <8 x i16>* %src %1 = tail call <8 x i16> @llvm.mips.addv.h(<8 x i16> %0, <8 x i16> %0) %2 = bitcast <8 x i16> %1 to <16 x i8> %3 = tail call <16 x i8> @llvm.mips.addv.b(<16 x i8> %2, <16 x i8> %2) @@ -204,7 +204,7 @@ entry: define void @v8i16_to_v8i16(<8 x i16>* %src, <8 x i16>* %dst) nounwind { entry: - %0 = load volatile <8 x i16>* %src + %0 = load volatile <8 x i16>, <8 x i16>* %src %1 = tail call <8 x i16> @llvm.mips.addv.h(<8 x i16> %0, <8 x i16> %0) %2 = bitcast <8 x i16> %1 to <8 x i16> %3 = tail call <8 x i16> @llvm.mips.addv.h(<8 x i16> %2, <8 x i16> %2) @@ -230,7 +230,7 @@ entry: ; are no operations for v8f16 to put in the way. define void @v8i16_to_v8f16(<8 x i16>* %src, <8 x half>* %dst) nounwind { entry: - %0 = load volatile <8 x i16>* %src + %0 = load volatile <8 x i16>, <8 x i16>* %src %1 = tail call <8 x i16> @llvm.mips.addv.h(<8 x i16> %0, <8 x i16> %0) %2 = bitcast <8 x i16> %1 to <8 x half> store <8 x half> %2, <8 x half>* %dst @@ -251,7 +251,7 @@ entry: define void @v8i16_to_v4i32(<8 x i16>* %src, <4 x i32>* %dst) nounwind { entry: - %0 = load volatile <8 x i16>* %src + %0 = load volatile <8 x i16>, <8 x i16>* %src %1 = tail call <8 x i16> @llvm.mips.addv.h(<8 x i16> %0, <8 x i16> %0) %2 = bitcast <8 x i16> %1 to <4 x i32> %3 = tail call <4 x i32> @llvm.mips.addv.w(<4 x i32> %2, <4 x i32> %2) @@ -276,7 +276,7 @@ entry: define void @v8i16_to_v4f32(<8 x i16>* %src, <4 x float>* %dst) nounwind { entry: - %0 = load volatile <8 x i16>* %src + %0 = load volatile <8 x i16>, <8 x i16>* %src %1 = tail call <8 x i16> @llvm.mips.addv.h(<8 x i16> %0, <8 x i16> %0) %2 = bitcast <8 x i16> %1 to <4 x float> %3 = tail call <4 x float> @llvm.mips.fadd.w(<4 x float> %2, <4 x float> %2) @@ -301,7 +301,7 @@ entry: define void @v8i16_to_v2i64(<8 x i16>* %src, <2 x i64>* %dst) nounwind { entry: - %0 = load volatile <8 x i16>* %src + %0 = load volatile <8 x i16>, <8 x i16>* %src %1 = tail call <8 x i16> @llvm.mips.addv.h(<8 x i16> %0, <8 x i16> %0) %2 = bitcast <8 x i16> %1 to <2 x i64> %3 = tail call <2 x i64> @llvm.mips.addv.d(<2 x i64> %2, <2 x i64> %2) @@ -326,7 +326,7 @@ entry: define void @v8i16_to_v2f64(<8 x i16>* %src, <2 x double>* %dst) nounwind { entry: - %0 = load volatile <8 x i16>* %src + %0 = load volatile <8 x i16>, <8 x i16>* %src %1 = tail call <8 x i16> @llvm.mips.addv.h(<8 x i16> %0, <8 x i16> %0) %2 = bitcast <8 x i16> %1 to <2 x double> %3 = tail call <2 x double> @llvm.mips.fadd.d(<2 x double> %2, <2 x double> %2) @@ -354,7 +354,7 @@ entry: ; are no operations for v8f16 to put in the way. define void @v8f16_to_v16i8(<8 x half>* %src, <16 x i8>* %dst) nounwind { entry: - %0 = load volatile <8 x half>* %src + %0 = load volatile <8 x half>, <8 x half>* %src %1 = bitcast <8 x half> %0 to <16 x i8> %2 = tail call <16 x i8> @llvm.mips.addv.b(<16 x i8> %1, <16 x i8> %1) store <16 x i8> %2, <16 x i8>* %dst @@ -378,7 +378,7 @@ entry: ; are no operations for v8f16 to put in the way. define void @v8f16_to_v8i16(<8 x half>* %src, <8 x i16>* %dst) nounwind { entry: - %0 = load volatile <8 x half>* %src + %0 = load volatile <8 x half>, <8 x half>* %src %1 = bitcast <8 x half> %0 to <8 x i16> %2 = tail call <8 x i16> @llvm.mips.addv.h(<8 x i16> %1, <8 x i16> %1) store <8 x i16> %2, <8 x i16>* %dst @@ -403,7 +403,7 @@ entry: ; are no operations for v8f16 to put in the way. define void @v8f16_to_v8f16(<8 x half>* %src, <8 x half>* %dst) nounwind { entry: - %0 = load volatile <8 x half>* %src + %0 = load volatile <8 x half>, <8 x half>* %src %1 = bitcast <8 x half> %0 to <8 x half> store <8 x half> %1, <8 x half>* %dst ret void @@ -423,7 +423,7 @@ entry: ; are no operations for v8f16 to put in the way. define void @v8f16_to_v4i32(<8 x half>* %src, <4 x i32>* %dst) nounwind { entry: - %0 = load volatile <8 x half>* %src + %0 = load volatile <8 x half>, <8 x half>* %src %1 = bitcast <8 x half> %0 to <4 x i32> %2 = tail call <4 x i32> @llvm.mips.addv.w(<4 x i32> %1, <4 x i32> %1) store <4 x i32> %2, <4 x i32>* %dst @@ -447,7 +447,7 @@ entry: ; are no operations for v8f16 to put in the way. define void @v8f16_to_v4f32(<8 x half>* %src, <4 x float>* %dst) nounwind { entry: - %0 = load volatile <8 x half>* %src + %0 = load volatile <8 x half>, <8 x half>* %src %1 = bitcast <8 x half> %0 to <4 x float> %2 = tail call <4 x float> @llvm.mips.fadd.w(<4 x float> %1, <4 x float> %1) store <4 x float> %2, <4 x float>* %dst @@ -471,7 +471,7 @@ entry: ; are no operations for v8f16 to put in the way. define void @v8f16_to_v2i64(<8 x half>* %src, <2 x i64>* %dst) nounwind { entry: - %0 = load volatile <8 x half>* %src + %0 = load volatile <8 x half>, <8 x half>* %src %1 = bitcast <8 x half> %0 to <2 x i64> %2 = tail call <2 x i64> @llvm.mips.addv.d(<2 x i64> %1, <2 x i64> %1) store <2 x i64> %2, <2 x i64>* %dst @@ -495,7 +495,7 @@ entry: ; are no operations for v8f16 to put in the way. define void @v8f16_to_v2f64(<8 x half>* %src, <2 x double>* %dst) nounwind { entry: - %0 = load volatile <8 x half>* %src + %0 = load volatile <8 x half>, <8 x half>* %src %1 = bitcast <8 x half> %0 to <2 x double> %2 = tail call <2 x double> @llvm.mips.fadd.d(<2 x double> %1, <2 x double> %1) store <2 x double> %2, <2 x double>* %dst @@ -518,7 +518,7 @@ entry: define void @v4i32_to_v16i8(<4 x i32>* %src, <16 x i8>* %dst) nounwind { entry: - %0 = load volatile <4 x i32>* %src + %0 = load volatile <4 x i32>, <4 x i32>* %src %1 = tail call <4 x i32> @llvm.mips.addv.w(<4 x i32> %0, <4 x i32> %0) %2 = bitcast <4 x i32> %1 to <16 x i8> %3 = tail call <16 x i8> @llvm.mips.addv.b(<16 x i8> %2, <16 x i8> %2) @@ -543,7 +543,7 @@ entry: define void @v4i32_to_v8i16(<4 x i32>* %src, <8 x i16>* %dst) nounwind { entry: - %0 = load volatile <4 x i32>* %src + %0 = load volatile <4 x i32>, <4 x i32>* %src %1 = tail call <4 x i32> @llvm.mips.addv.w(<4 x i32> %0, <4 x i32> %0) %2 = bitcast <4 x i32> %1 to <8 x i16> %3 = tail call <8 x i16> @llvm.mips.addv.h(<8 x i16> %2, <8 x i16> %2) @@ -570,7 +570,7 @@ entry: ; are no operations for v8f16 to put in the way. define void @v4i32_to_v8f16(<4 x i32>* %src, <8 x half>* %dst) nounwind { entry: - %0 = load volatile <4 x i32>* %src + %0 = load volatile <4 x i32>, <4 x i32>* %src %1 = tail call <4 x i32> @llvm.mips.addv.w(<4 x i32> %0, <4 x i32> %0) %2 = bitcast <4 x i32> %1 to <8 x half> store <8 x half> %2, <8 x half>* %dst @@ -591,7 +591,7 @@ entry: define void @v4i32_to_v4i32(<4 x i32>* %src, <4 x i32>* %dst) nounwind { entry: - %0 = load volatile <4 x i32>* %src + %0 = load volatile <4 x i32>, <4 x i32>* %src %1 = tail call <4 x i32> @llvm.mips.addv.w(<4 x i32> %0, <4 x i32> %0) %2 = bitcast <4 x i32> %1 to <4 x i32> %3 = tail call <4 x i32> @llvm.mips.addv.w(<4 x i32> %2, <4 x i32> %2) @@ -615,7 +615,7 @@ entry: define void @v4i32_to_v4f32(<4 x i32>* %src, <4 x float>* %dst) nounwind { entry: - %0 = load volatile <4 x i32>* %src + %0 = load volatile <4 x i32>, <4 x i32>* %src %1 = tail call <4 x i32> @llvm.mips.addv.w(<4 x i32> %0, <4 x i32> %0) %2 = bitcast <4 x i32> %1 to <4 x float> %3 = tail call <4 x float> @llvm.mips.fadd.w(<4 x float> %2, <4 x float> %2) @@ -639,7 +639,7 @@ entry: define void @v4i32_to_v2i64(<4 x i32>* %src, <2 x i64>* %dst) nounwind { entry: - %0 = load volatile <4 x i32>* %src + %0 = load volatile <4 x i32>, <4 x i32>* %src %1 = tail call <4 x i32> @llvm.mips.addv.w(<4 x i32> %0, <4 x i32> %0) %2 = bitcast <4 x i32> %1 to <2 x i64> %3 = tail call <2 x i64> @llvm.mips.addv.d(<2 x i64> %2, <2 x i64> %2) @@ -664,7 +664,7 @@ entry: define void @v4i32_to_v2f64(<4 x i32>* %src, <2 x double>* %dst) nounwind { entry: - %0 = load volatile <4 x i32>* %src + %0 = load volatile <4 x i32>, <4 x i32>* %src %1 = tail call <4 x i32> @llvm.mips.addv.w(<4 x i32> %0, <4 x i32> %0) %2 = bitcast <4 x i32> %1 to <2 x double> %3 = tail call <2 x double> @llvm.mips.fadd.d(<2 x double> %2, <2 x double> %2) @@ -689,7 +689,7 @@ entry: define void @v4f32_to_v16i8(<4 x float>* %src, <16 x i8>* %dst) nounwind { entry: - %0 = load volatile <4 x float>* %src + %0 = load volatile <4 x float>, <4 x float>* %src %1 = tail call <4 x float> @llvm.mips.fadd.w(<4 x float> %0, <4 x float> %0) %2 = bitcast <4 x float> %1 to <16 x i8> %3 = tail call <16 x i8> @llvm.mips.addv.b(<16 x i8> %2, <16 x i8> %2) @@ -714,7 +714,7 @@ entry: define void @v4f32_to_v8i16(<4 x float>* %src, <8 x i16>* %dst) nounwind { entry: - %0 = load volatile <4 x float>* %src + %0 = load volatile <4 x float>, <4 x float>* %src %1 = tail call <4 x float> @llvm.mips.fadd.w(<4 x float> %0, <4 x float> %0) %2 = bitcast <4 x float> %1 to <8 x i16> %3 = tail call <8 x i16> @llvm.mips.addv.h(<8 x i16> %2, <8 x i16> %2) @@ -741,7 +741,7 @@ entry: ; are no operations for v8f16 to put in the way. define void @v4f32_to_v8f16(<4 x float>* %src, <8 x half>* %dst) nounwind { entry: - %0 = load volatile <4 x float>* %src + %0 = load volatile <4 x float>, <4 x float>* %src %1 = tail call <4 x float> @llvm.mips.fadd.w(<4 x float> %0, <4 x float> %0) %2 = bitcast <4 x float> %1 to <8 x half> store <8 x half> %2, <8 x half>* %dst @@ -762,7 +762,7 @@ entry: define void @v4f32_to_v4i32(<4 x float>* %src, <4 x i32>* %dst) nounwind { entry: - %0 = load volatile <4 x float>* %src + %0 = load volatile <4 x float>, <4 x float>* %src %1 = tail call <4 x float> @llvm.mips.fadd.w(<4 x float> %0, <4 x float> %0) %2 = bitcast <4 x float> %1 to <4 x i32> %3 = tail call <4 x i32> @llvm.mips.addv.w(<4 x i32> %2, <4 x i32> %2) @@ -786,7 +786,7 @@ entry: define void @v4f32_to_v4f32(<4 x float>* %src, <4 x float>* %dst) nounwind { entry: - %0 = load volatile <4 x float>* %src + %0 = load volatile <4 x float>, <4 x float>* %src %1 = tail call <4 x float> @llvm.mips.fadd.w(<4 x float> %0, <4 x float> %0) %2 = bitcast <4 x float> %1 to <4 x float> %3 = tail call <4 x float> @llvm.mips.fadd.w(<4 x float> %2, <4 x float> %2) @@ -810,7 +810,7 @@ entry: define void @v4f32_to_v2i64(<4 x float>* %src, <2 x i64>* %dst) nounwind { entry: - %0 = load volatile <4 x float>* %src + %0 = load volatile <4 x float>, <4 x float>* %src %1 = tail call <4 x float> @llvm.mips.fadd.w(<4 x float> %0, <4 x float> %0) %2 = bitcast <4 x float> %1 to <2 x i64> %3 = tail call <2 x i64> @llvm.mips.addv.d(<2 x i64> %2, <2 x i64> %2) @@ -835,7 +835,7 @@ entry: define void @v4f32_to_v2f64(<4 x float>* %src, <2 x double>* %dst) nounwind { entry: - %0 = load volatile <4 x float>* %src + %0 = load volatile <4 x float>, <4 x float>* %src %1 = tail call <4 x float> @llvm.mips.fadd.w(<4 x float> %0, <4 x float> %0) %2 = bitcast <4 x float> %1 to <2 x double> %3 = tail call <2 x double> @llvm.mips.fadd.d(<2 x double> %2, <2 x double> %2) @@ -860,7 +860,7 @@ entry: define void @v2i64_to_v16i8(<2 x i64>* %src, <16 x i8>* %dst) nounwind { entry: - %0 = load volatile <2 x i64>* %src + %0 = load volatile <2 x i64>, <2 x i64>* %src %1 = tail call <2 x i64> @llvm.mips.addv.d(<2 x i64> %0, <2 x i64> %0) %2 = bitcast <2 x i64> %1 to <16 x i8> %3 = tail call <16 x i8> @llvm.mips.addv.b(<16 x i8> %2, <16 x i8> %2) @@ -886,7 +886,7 @@ entry: define void @v2i64_to_v8i16(<2 x i64>* %src, <8 x i16>* %dst) nounwind { entry: - %0 = load volatile <2 x i64>* %src + %0 = load volatile <2 x i64>, <2 x i64>* %src %1 = tail call <2 x i64> @llvm.mips.addv.d(<2 x i64> %0, <2 x i64> %0) %2 = bitcast <2 x i64> %1 to <8 x i16> %3 = tail call <8 x i16> @llvm.mips.addv.h(<8 x i16> %2, <8 x i16> %2) @@ -913,7 +913,7 @@ entry: ; are no operations for v8f16 to put in the way. define void @v2i64_to_v8f16(<2 x i64>* %src, <8 x half>* %dst) nounwind { entry: - %0 = load volatile <2 x i64>* %src + %0 = load volatile <2 x i64>, <2 x i64>* %src %1 = tail call <2 x i64> @llvm.mips.addv.d(<2 x i64> %0, <2 x i64> %0) %2 = bitcast <2 x i64> %1 to <8 x half> store <8 x half> %2, <8 x half>* %dst @@ -934,7 +934,7 @@ entry: define void @v2i64_to_v4i32(<2 x i64>* %src, <4 x i32>* %dst) nounwind { entry: - %0 = load volatile <2 x i64>* %src + %0 = load volatile <2 x i64>, <2 x i64>* %src %1 = tail call <2 x i64> @llvm.mips.addv.d(<2 x i64> %0, <2 x i64> %0) %2 = bitcast <2 x i64> %1 to <4 x i32> %3 = tail call <4 x i32> @llvm.mips.addv.w(<4 x i32> %2, <4 x i32> %2) @@ -959,7 +959,7 @@ entry: define void @v2i64_to_v4f32(<2 x i64>* %src, <4 x float>* %dst) nounwind { entry: - %0 = load volatile <2 x i64>* %src + %0 = load volatile <2 x i64>, <2 x i64>* %src %1 = tail call <2 x i64> @llvm.mips.addv.d(<2 x i64> %0, <2 x i64> %0) %2 = bitcast <2 x i64> %1 to <4 x float> %3 = tail call <4 x float> @llvm.mips.fadd.w(<4 x float> %2, <4 x float> %2) @@ -984,7 +984,7 @@ entry: define void @v2i64_to_v2i64(<2 x i64>* %src, <2 x i64>* %dst) nounwind { entry: - %0 = load volatile <2 x i64>* %src + %0 = load volatile <2 x i64>, <2 x i64>* %src %1 = tail call <2 x i64> @llvm.mips.addv.d(<2 x i64> %0, <2 x i64> %0) %2 = bitcast <2 x i64> %1 to <2 x i64> %3 = tail call <2 x i64> @llvm.mips.addv.d(<2 x i64> %2, <2 x i64> %2) @@ -1008,7 +1008,7 @@ entry: define void @v2i64_to_v2f64(<2 x i64>* %src, <2 x double>* %dst) nounwind { entry: - %0 = load volatile <2 x i64>* %src + %0 = load volatile <2 x i64>, <2 x i64>* %src %1 = tail call <2 x i64> @llvm.mips.addv.d(<2 x i64> %0, <2 x i64> %0) %2 = bitcast <2 x i64> %1 to <2 x double> %3 = tail call <2 x double> @llvm.mips.fadd.d(<2 x double> %2, <2 x double> %2) @@ -1032,7 +1032,7 @@ entry: define void @v2f64_to_v16i8(<2 x double>* %src, <16 x i8>* %dst) nounwind { entry: - %0 = load volatile <2 x double>* %src + %0 = load volatile <2 x double>, <2 x double>* %src %1 = tail call <2 x double> @llvm.mips.fadd.d(<2 x double> %0, <2 x double> %0) %2 = bitcast <2 x double> %1 to <16 x i8> %3 = tail call <16 x i8> @llvm.mips.addv.b(<16 x i8> %2, <16 x i8> %2) @@ -1058,7 +1058,7 @@ entry: define void @v2f64_to_v8i16(<2 x double>* %src, <8 x i16>* %dst) nounwind { entry: - %0 = load volatile <2 x double>* %src + %0 = load volatile <2 x double>, <2 x double>* %src %1 = tail call <2 x double> @llvm.mips.fadd.d(<2 x double> %0, <2 x double> %0) %2 = bitcast <2 x double> %1 to <8 x i16> %3 = tail call <8 x i16> @llvm.mips.addv.h(<8 x i16> %2, <8 x i16> %2) @@ -1085,7 +1085,7 @@ entry: ; are no operations for v8f16 to put in the way. define void @v2f64_to_v8f16(<2 x double>* %src, <8 x half>* %dst) nounwind { entry: - %0 = load volatile <2 x double>* %src + %0 = load volatile <2 x double>, <2 x double>* %src %1 = tail call <2 x double> @llvm.mips.fadd.d(<2 x double> %0, <2 x double> %0) %2 = bitcast <2 x double> %1 to <8 x half> store <8 x half> %2, <8 x half>* %dst @@ -1106,7 +1106,7 @@ entry: define void @v2f64_to_v4i32(<2 x double>* %src, <4 x i32>* %dst) nounwind { entry: - %0 = load volatile <2 x double>* %src + %0 = load volatile <2 x double>, <2 x double>* %src %1 = tail call <2 x double> @llvm.mips.fadd.d(<2 x double> %0, <2 x double> %0) %2 = bitcast <2 x double> %1 to <4 x i32> %3 = tail call <4 x i32> @llvm.mips.addv.w(<4 x i32> %2, <4 x i32> %2) @@ -1131,7 +1131,7 @@ entry: define void @v2f64_to_v4f32(<2 x double>* %src, <4 x float>* %dst) nounwind { entry: - %0 = load volatile <2 x double>* %src + %0 = load volatile <2 x double>, <2 x double>* %src %1 = tail call <2 x double> @llvm.mips.fadd.d(<2 x double> %0, <2 x double> %0) %2 = bitcast <2 x double> %1 to <4 x float> %3 = tail call <4 x float> @llvm.mips.fadd.w(<4 x float> %2, <4 x float> %2) @@ -1156,7 +1156,7 @@ entry: define void @v2f64_to_v2i64(<2 x double>* %src, <2 x i64>* %dst) nounwind { entry: - %0 = load volatile <2 x double>* %src + %0 = load volatile <2 x double>, <2 x double>* %src %1 = tail call <2 x double> @llvm.mips.fadd.d(<2 x double> %0, <2 x double> %0) %2 = bitcast <2 x double> %1 to <2 x i64> %3 = tail call <2 x i64> @llvm.mips.addv.d(<2 x i64> %2, <2 x i64> %2) @@ -1180,7 +1180,7 @@ entry: define void @v2f64_to_v2f64(<2 x double>* %src, <2 x double>* %dst) nounwind { entry: - %0 = load volatile <2 x double>* %src + %0 = load volatile <2 x double>, <2 x double>* %src %1 = tail call <2 x double> @llvm.mips.fadd.d(<2 x double> %0, <2 x double> %0) %2 = bitcast <2 x double> %1 to <2 x double> %3 = tail call <2 x double> @llvm.mips.fadd.d(<2 x double> %2, <2 x double> %2) diff --git a/test/CodeGen/Mips/msa/bitwise.ll b/test/CodeGen/Mips/msa/bitwise.ll index 5d57198a93558..2a260b2c57332 100644 --- a/test/CodeGen/Mips/msa/bitwise.ll +++ b/test/CodeGen/Mips/msa/bitwise.ll @@ -4,9 +4,9 @@ define void @and_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind { ; CHECK: and_v16i8: - %1 = load <16 x i8>* %a + %1 = load <16 x i8>, <16 x i8>* %a ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5) - %2 = load <16 x i8>* %b + %2 = load <16 x i8>, <16 x i8>* %b ; CHECK-DAG: ld.b [[R2:\$w[0-9]+]], 0($6) %3 = and <16 x i8> %1, %2 ; CHECK-DAG: and.v [[R3:\$w[0-9]+]], [[R1]], [[R2]] @@ -20,9 +20,9 @@ define void @and_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind { define void @and_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind { ; CHECK: and_v8i16: - %1 = load <8 x i16>* %a + %1 = load <8 x i16>, <8 x i16>* %a ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5) - %2 = load <8 x i16>* %b + %2 = load <8 x i16>, <8 x i16>* %b ; CHECK-DAG: ld.h [[R2:\$w[0-9]+]], 0($6) %3 = and <8 x i16> %1, %2 ; CHECK-DAG: and.v [[R3:\$w[0-9]+]], [[R1]], [[R2]] @@ -36,9 +36,9 @@ define void @and_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind { define void @and_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind { ; CHECK: and_v4i32: - %1 = load <4 x i32>* %a + %1 = load <4 x i32>, <4 x i32>* %a ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5) - %2 = load <4 x i32>* %b + %2 = load <4 x i32>, <4 x i32>* %b ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6) %3 = and <4 x i32> %1, %2 ; CHECK-DAG: and.v [[R3:\$w[0-9]+]], [[R1]], [[R2]] @@ -52,9 +52,9 @@ define void @and_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind { define void @and_v2i64(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind { ; CHECK: and_v2i64: - %1 = load <2 x i64>* %a + %1 = load <2 x i64>, <2 x i64>* %a ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5) - %2 = load <2 x i64>* %b + %2 = load <2 x i64>, <2 x i64>* %b ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6) %3 = and <2 x i64> %1, %2 ; CHECK-DAG: and.v [[R3:\$w[0-9]+]], [[R1]], [[R2]] @@ -68,7 +68,7 @@ define void @and_v2i64(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind { define void @and_v16i8_i(<16 x i8>* %c, <16 x i8>* %a) nounwind { ; CHECK: and_v16i8_i: - %1 = load <16 x i8>* %a + %1 = load <16 x i8>, <16 x i8>* %a ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5) %2 = and <16 x i8> %1, <i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1> ; CHECK-DAG: andi.b [[R4:\$w[0-9]+]], [[R1]], 1 @@ -82,7 +82,7 @@ define void @and_v16i8_i(<16 x i8>* %c, <16 x i8>* %a) nounwind { define void @and_v8i16_i(<8 x i16>* %c, <8 x i16>* %a) nounwind { ; CHECK: and_v8i16_i: - %1 = load <8 x i16>* %a + %1 = load <8 x i16>, <8 x i16>* %a ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5) %2 = and <8 x i16> %1, <i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1> ; CHECK-DAG: ldi.h [[R3:\$w[0-9]+]], 1 @@ -97,7 +97,7 @@ define void @and_v8i16_i(<8 x i16>* %c, <8 x i16>* %a) nounwind { define void @and_v4i32_i(<4 x i32>* %c, <4 x i32>* %a) nounwind { ; CHECK: and_v4i32_i: - %1 = load <4 x i32>* %a + %1 = load <4 x i32>, <4 x i32>* %a ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5) %2 = and <4 x i32> %1, <i32 1, i32 1, i32 1, i32 1> ; CHECK-DAG: ldi.w [[R3:\$w[0-9]+]], 1 @@ -112,7 +112,7 @@ define void @and_v4i32_i(<4 x i32>* %c, <4 x i32>* %a) nounwind { define void @and_v2i64_i(<2 x i64>* %c, <2 x i64>* %a) nounwind { ; CHECK: and_v2i64_i: - %1 = load <2 x i64>* %a + %1 = load <2 x i64>, <2 x i64>* %a ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5) %2 = and <2 x i64> %1, <i64 1, i64 1> ; CHECK-DAG: ldi.d [[R3:\$w[0-9]+]], 1 @@ -127,9 +127,9 @@ define void @and_v2i64_i(<2 x i64>* %c, <2 x i64>* %a) nounwind { define void @or_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind { ; CHECK: or_v16i8: - %1 = load <16 x i8>* %a + %1 = load <16 x i8>, <16 x i8>* %a ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5) - %2 = load <16 x i8>* %b + %2 = load <16 x i8>, <16 x i8>* %b ; CHECK-DAG: ld.b [[R2:\$w[0-9]+]], 0($6) %3 = or <16 x i8> %1, %2 ; CHECK-DAG: or.v [[R3:\$w[0-9]+]], [[R1]], [[R2]] @@ -143,9 +143,9 @@ define void @or_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind { define void @or_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind { ; CHECK: or_v8i16: - %1 = load <8 x i16>* %a + %1 = load <8 x i16>, <8 x i16>* %a ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5) - %2 = load <8 x i16>* %b + %2 = load <8 x i16>, <8 x i16>* %b ; CHECK-DAG: ld.h [[R2:\$w[0-9]+]], 0($6) %3 = or <8 x i16> %1, %2 ; CHECK-DAG: or.v [[R3:\$w[0-9]+]], [[R1]], [[R2]] @@ -159,9 +159,9 @@ define void @or_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind { define void @or_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind { ; CHECK: or_v4i32: - %1 = load <4 x i32>* %a + %1 = load <4 x i32>, <4 x i32>* %a ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5) - %2 = load <4 x i32>* %b + %2 = load <4 x i32>, <4 x i32>* %b ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6) %3 = or <4 x i32> %1, %2 ; CHECK-DAG: or.v [[R3:\$w[0-9]+]], [[R1]], [[R2]] @@ -175,9 +175,9 @@ define void @or_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind { define void @or_v2i64(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind { ; CHECK: or_v2i64: - %1 = load <2 x i64>* %a + %1 = load <2 x i64>, <2 x i64>* %a ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5) - %2 = load <2 x i64>* %b + %2 = load <2 x i64>, <2 x i64>* %b ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6) %3 = or <2 x i64> %1, %2 ; CHECK-DAG: or.v [[R3:\$w[0-9]+]], [[R1]], [[R2]] @@ -191,7 +191,7 @@ define void @or_v2i64(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind { define void @or_v16i8_i(<16 x i8>* %c, <16 x i8>* %a) nounwind { ; CHECK: or_v16i8_i: - %1 = load <16 x i8>* %a + %1 = load <16 x i8>, <16 x i8>* %a ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5) %2 = or <16 x i8> %1, <i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3> ; CHECK-DAG: ori.b [[R4:\$w[0-9]+]], [[R1]], 3 @@ -205,7 +205,7 @@ define void @or_v16i8_i(<16 x i8>* %c, <16 x i8>* %a) nounwind { define void @or_v8i16_i(<8 x i16>* %c, <8 x i16>* %a) nounwind { ; CHECK: or_v8i16_i: - %1 = load <8 x i16>* %a + %1 = load <8 x i16>, <8 x i16>* %a ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5) %2 = or <8 x i16> %1, <i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3> ; CHECK-DAG: ldi.h [[R3:\$w[0-9]+]], 3 @@ -220,7 +220,7 @@ define void @or_v8i16_i(<8 x i16>* %c, <8 x i16>* %a) nounwind { define void @or_v4i32_i(<4 x i32>* %c, <4 x i32>* %a) nounwind { ; CHECK: or_v4i32_i: - %1 = load <4 x i32>* %a + %1 = load <4 x i32>, <4 x i32>* %a ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5) %2 = or <4 x i32> %1, <i32 3, i32 3, i32 3, i32 3> ; CHECK-DAG: ldi.w [[R3:\$w[0-9]+]], 3 @@ -235,7 +235,7 @@ define void @or_v4i32_i(<4 x i32>* %c, <4 x i32>* %a) nounwind { define void @or_v2i64_i(<2 x i64>* %c, <2 x i64>* %a) nounwind { ; CHECK: or_v2i64_i: - %1 = load <2 x i64>* %a + %1 = load <2 x i64>, <2 x i64>* %a ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5) %2 = or <2 x i64> %1, <i64 3, i64 3> ; CHECK-DAG: ldi.d [[R3:\$w[0-9]+]], 3 @@ -250,9 +250,9 @@ define void @or_v2i64_i(<2 x i64>* %c, <2 x i64>* %a) nounwind { define void @nor_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind { ; CHECK: nor_v16i8: - %1 = load <16 x i8>* %a + %1 = load <16 x i8>, <16 x i8>* %a ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5) - %2 = load <16 x i8>* %b + %2 = load <16 x i8>, <16 x i8>* %b ; CHECK-DAG: ld.b [[R2:\$w[0-9]+]], 0($6) %3 = or <16 x i8> %1, %2 %4 = xor <16 x i8> %3, <i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1> @@ -267,9 +267,9 @@ define void @nor_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind { define void @nor_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind { ; CHECK: nor_v8i16: - %1 = load <8 x i16>* %a + %1 = load <8 x i16>, <8 x i16>* %a ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5) - %2 = load <8 x i16>* %b + %2 = load <8 x i16>, <8 x i16>* %b ; CHECK-DAG: ld.h [[R2:\$w[0-9]+]], 0($6) %3 = or <8 x i16> %1, %2 %4 = xor <8 x i16> %3, <i16 -1, i16 -1, i16 -1, i16 -1, i16 -1, i16 -1, i16 -1, i16 -1> @@ -284,9 +284,9 @@ define void @nor_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind { define void @nor_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind { ; CHECK: nor_v4i32: - %1 = load <4 x i32>* %a + %1 = load <4 x i32>, <4 x i32>* %a ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5) - %2 = load <4 x i32>* %b + %2 = load <4 x i32>, <4 x i32>* %b ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6) %3 = or <4 x i32> %1, %2 %4 = xor <4 x i32> %3, <i32 -1, i32 -1, i32 -1, i32 -1> @@ -301,9 +301,9 @@ define void @nor_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind { define void @nor_v2i64(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind { ; CHECK: nor_v2i64: - %1 = load <2 x i64>* %a + %1 = load <2 x i64>, <2 x i64>* %a ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5) - %2 = load <2 x i64>* %b + %2 = load <2 x i64>, <2 x i64>* %b ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6) %3 = or <2 x i64> %1, %2 %4 = xor <2 x i64> %3, <i64 -1, i64 -1> @@ -318,7 +318,7 @@ define void @nor_v2i64(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind { define void @nor_v16i8_i(<16 x i8>* %c, <16 x i8>* %a) nounwind { ; CHECK: nor_v16i8_i: - %1 = load <16 x i8>* %a + %1 = load <16 x i8>, <16 x i8>* %a ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5) %2 = or <16 x i8> %1, <i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1> %3 = xor <16 x i8> %2, <i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1> @@ -333,7 +333,7 @@ define void @nor_v16i8_i(<16 x i8>* %c, <16 x i8>* %a) nounwind { define void @nor_v8i16_i(<8 x i16>* %c, <8 x i16>* %a) nounwind { ; CHECK: nor_v8i16_i: - %1 = load <8 x i16>* %a + %1 = load <8 x i16>, <8 x i16>* %a ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5) %2 = or <8 x i16> %1, <i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1> %3 = xor <8 x i16> %2, <i16 -1, i16 -1, i16 -1, i16 -1, i16 -1, i16 -1, i16 -1, i16 -1> @@ -349,7 +349,7 @@ define void @nor_v8i16_i(<8 x i16>* %c, <8 x i16>* %a) nounwind { define void @nor_v4i32_i(<4 x i32>* %c, <4 x i32>* %a) nounwind { ; CHECK: nor_v4i32_i: - %1 = load <4 x i32>* %a + %1 = load <4 x i32>, <4 x i32>* %a ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5) %2 = or <4 x i32> %1, <i32 1, i32 1, i32 1, i32 1> %3 = xor <4 x i32> %2, <i32 -1, i32 -1, i32 -1, i32 -1> @@ -365,7 +365,7 @@ define void @nor_v4i32_i(<4 x i32>* %c, <4 x i32>* %a) nounwind { define void @nor_v2i64_i(<2 x i64>* %c, <2 x i64>* %a) nounwind { ; CHECK: nor_v2i64_i: - %1 = load <2 x i64>* %a + %1 = load <2 x i64>, <2 x i64>* %a ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5) %2 = or <2 x i64> %1, <i64 1, i64 1> %3 = xor <2 x i64> %2, <i64 -1, i64 -1> @@ -381,9 +381,9 @@ define void @nor_v2i64_i(<2 x i64>* %c, <2 x i64>* %a) nounwind { define void @xor_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind { ; CHECK: xor_v16i8: - %1 = load <16 x i8>* %a + %1 = load <16 x i8>, <16 x i8>* %a ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5) - %2 = load <16 x i8>* %b + %2 = load <16 x i8>, <16 x i8>* %b ; CHECK-DAG: ld.b [[R2:\$w[0-9]+]], 0($6) %3 = xor <16 x i8> %1, %2 ; CHECK-DAG: xor.v [[R3:\$w[0-9]+]], [[R1]], [[R2]] @@ -397,9 +397,9 @@ define void @xor_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind { define void @xor_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind { ; CHECK: xor_v8i16: - %1 = load <8 x i16>* %a + %1 = load <8 x i16>, <8 x i16>* %a ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5) - %2 = load <8 x i16>* %b + %2 = load <8 x i16>, <8 x i16>* %b ; CHECK-DAG: ld.h [[R2:\$w[0-9]+]], 0($6) %3 = xor <8 x i16> %1, %2 ; CHECK-DAG: xor.v [[R3:\$w[0-9]+]], [[R1]], [[R2]] @@ -413,9 +413,9 @@ define void @xor_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind { define void @xor_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind { ; CHECK: xor_v4i32: - %1 = load <4 x i32>* %a + %1 = load <4 x i32>, <4 x i32>* %a ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5) - %2 = load <4 x i32>* %b + %2 = load <4 x i32>, <4 x i32>* %b ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6) %3 = xor <4 x i32> %1, %2 ; CHECK-DAG: xor.v [[R3:\$w[0-9]+]], [[R1]], [[R2]] @@ -429,9 +429,9 @@ define void @xor_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind { define void @xor_v2i64(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind { ; CHECK: xor_v2i64: - %1 = load <2 x i64>* %a + %1 = load <2 x i64>, <2 x i64>* %a ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5) - %2 = load <2 x i64>* %b + %2 = load <2 x i64>, <2 x i64>* %b ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6) %3 = xor <2 x i64> %1, %2 ; CHECK-DAG: xor.v [[R3:\$w[0-9]+]], [[R1]], [[R2]] @@ -445,7 +445,7 @@ define void @xor_v2i64(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind { define void @xor_v16i8_i(<16 x i8>* %c, <16 x i8>* %a) nounwind { ; CHECK: xor_v16i8_i: - %1 = load <16 x i8>* %a + %1 = load <16 x i8>, <16 x i8>* %a ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5) %2 = xor <16 x i8> %1, <i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3> ; CHECK-DAG: xori.b [[R4:\$w[0-9]+]], [[R1]], 3 @@ -459,7 +459,7 @@ define void @xor_v16i8_i(<16 x i8>* %c, <16 x i8>* %a) nounwind { define void @xor_v8i16_i(<8 x i16>* %c, <8 x i16>* %a) nounwind { ; CHECK: xor_v8i16_i: - %1 = load <8 x i16>* %a + %1 = load <8 x i16>, <8 x i16>* %a ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5) %2 = xor <8 x i16> %1, <i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3> ; CHECK-DAG: ldi.h [[R3:\$w[0-9]+]], 3 @@ -474,7 +474,7 @@ define void @xor_v8i16_i(<8 x i16>* %c, <8 x i16>* %a) nounwind { define void @xor_v4i32_i(<4 x i32>* %c, <4 x i32>* %a) nounwind { ; CHECK: xor_v4i32_i: - %1 = load <4 x i32>* %a + %1 = load <4 x i32>, <4 x i32>* %a ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5) %2 = xor <4 x i32> %1, <i32 3, i32 3, i32 3, i32 3> ; CHECK-DAG: ldi.w [[R3:\$w[0-9]+]], 3 @@ -489,7 +489,7 @@ define void @xor_v4i32_i(<4 x i32>* %c, <4 x i32>* %a) nounwind { define void @xor_v2i64_i(<2 x i64>* %c, <2 x i64>* %a) nounwind { ; CHECK: xor_v2i64_i: - %1 = load <2 x i64>* %a + %1 = load <2 x i64>, <2 x i64>* %a ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5) %2 = xor <2 x i64> %1, <i64 3, i64 3> ; CHECK-DAG: ldi.d [[R3:\$w[0-9]+]], 3 @@ -504,9 +504,9 @@ define void @xor_v2i64_i(<2 x i64>* %c, <2 x i64>* %a) nounwind { define void @sll_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind { ; CHECK: sll_v16i8: - %1 = load <16 x i8>* %a + %1 = load <16 x i8>, <16 x i8>* %a ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5) - %2 = load <16 x i8>* %b + %2 = load <16 x i8>, <16 x i8>* %b ; CHECK-DAG: ld.b [[R2:\$w[0-9]+]], 0($6) %3 = shl <16 x i8> %1, %2 ; CHECK-DAG: sll.b [[R3:\$w[0-9]+]], [[R1]], [[R2]] @@ -520,9 +520,9 @@ define void @sll_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind { define void @sll_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind { ; CHECK: sll_v8i16: - %1 = load <8 x i16>* %a + %1 = load <8 x i16>, <8 x i16>* %a ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5) - %2 = load <8 x i16>* %b + %2 = load <8 x i16>, <8 x i16>* %b ; CHECK-DAG: ld.h [[R2:\$w[0-9]+]], 0($6) %3 = shl <8 x i16> %1, %2 ; CHECK-DAG: sll.h [[R3:\$w[0-9]+]], [[R1]], [[R2]] @@ -536,9 +536,9 @@ define void @sll_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind { define void @sll_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind { ; CHECK: sll_v4i32: - %1 = load <4 x i32>* %a + %1 = load <4 x i32>, <4 x i32>* %a ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5) - %2 = load <4 x i32>* %b + %2 = load <4 x i32>, <4 x i32>* %b ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6) %3 = shl <4 x i32> %1, %2 ; CHECK-DAG: sll.w [[R3:\$w[0-9]+]], [[R1]], [[R2]] @@ -552,9 +552,9 @@ define void @sll_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind { define void @sll_v2i64(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind { ; CHECK: sll_v2i64: - %1 = load <2 x i64>* %a + %1 = load <2 x i64>, <2 x i64>* %a ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5) - %2 = load <2 x i64>* %b + %2 = load <2 x i64>, <2 x i64>* %b ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6) %3 = shl <2 x i64> %1, %2 ; CHECK-DAG: sll.d [[R3:\$w[0-9]+]], [[R1]], [[R2]] @@ -568,7 +568,7 @@ define void @sll_v2i64(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind { define void @sll_v16i8_i(<16 x i8>* %c, <16 x i8>* %a) nounwind { ; CHECK: sll_v16i8_i: - %1 = load <16 x i8>* %a + %1 = load <16 x i8>, <16 x i8>* %a ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5) %2 = shl <16 x i8> %1, <i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1> ; CHECK-DAG: slli.b [[R4:\$w[0-9]+]], [[R1]], 1 @@ -582,7 +582,7 @@ define void @sll_v16i8_i(<16 x i8>* %c, <16 x i8>* %a) nounwind { define void @sll_v8i16_i(<8 x i16>* %c, <8 x i16>* %a) nounwind { ; CHECK: sll_v8i16_i: - %1 = load <8 x i16>* %a + %1 = load <8 x i16>, <8 x i16>* %a ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5) %2 = shl <8 x i16> %1, <i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1> ; CHECK-DAG: slli.h [[R4:\$w[0-9]+]], [[R1]], 1 @@ -596,7 +596,7 @@ define void @sll_v8i16_i(<8 x i16>* %c, <8 x i16>* %a) nounwind { define void @sll_v4i32_i(<4 x i32>* %c, <4 x i32>* %a) nounwind { ; CHECK: sll_v4i32_i: - %1 = load <4 x i32>* %a + %1 = load <4 x i32>, <4 x i32>* %a ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5) %2 = shl <4 x i32> %1, <i32 1, i32 1, i32 1, i32 1> ; CHECK-DAG: slli.w [[R4:\$w[0-9]+]], [[R1]], 1 @@ -610,7 +610,7 @@ define void @sll_v4i32_i(<4 x i32>* %c, <4 x i32>* %a) nounwind { define void @sll_v2i64_i(<2 x i64>* %c, <2 x i64>* %a) nounwind { ; CHECK: sll_v2i64_i: - %1 = load <2 x i64>* %a + %1 = load <2 x i64>, <2 x i64>* %a ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5) %2 = shl <2 x i64> %1, <i64 1, i64 1> ; CHECK-DAG: slli.d [[R4:\$w[0-9]+]], [[R1]], 1 @@ -624,9 +624,9 @@ define void @sll_v2i64_i(<2 x i64>* %c, <2 x i64>* %a) nounwind { define void @sra_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind { ; CHECK: sra_v16i8: - %1 = load <16 x i8>* %a + %1 = load <16 x i8>, <16 x i8>* %a ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5) - %2 = load <16 x i8>* %b + %2 = load <16 x i8>, <16 x i8>* %b ; CHECK-DAG: ld.b [[R2:\$w[0-9]+]], 0($6) %3 = ashr <16 x i8> %1, %2 ; CHECK-DAG: sra.b [[R3:\$w[0-9]+]], [[R1]], [[R2]] @@ -640,9 +640,9 @@ define void @sra_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind { define void @sra_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind { ; CHECK: sra_v8i16: - %1 = load <8 x i16>* %a + %1 = load <8 x i16>, <8 x i16>* %a ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5) - %2 = load <8 x i16>* %b + %2 = load <8 x i16>, <8 x i16>* %b ; CHECK-DAG: ld.h [[R2:\$w[0-9]+]], 0($6) %3 = ashr <8 x i16> %1, %2 ; CHECK-DAG: sra.h [[R3:\$w[0-9]+]], [[R1]], [[R2]] @@ -656,9 +656,9 @@ define void @sra_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind { define void @sra_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind { ; CHECK: sra_v4i32: - %1 = load <4 x i32>* %a + %1 = load <4 x i32>, <4 x i32>* %a ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5) - %2 = load <4 x i32>* %b + %2 = load <4 x i32>, <4 x i32>* %b ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6) %3 = ashr <4 x i32> %1, %2 ; CHECK-DAG: sra.w [[R3:\$w[0-9]+]], [[R1]], [[R2]] @@ -672,9 +672,9 @@ define void @sra_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind { define void @sra_v2i64(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind { ; CHECK: sra_v2i64: - %1 = load <2 x i64>* %a + %1 = load <2 x i64>, <2 x i64>* %a ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5) - %2 = load <2 x i64>* %b + %2 = load <2 x i64>, <2 x i64>* %b ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6) %3 = ashr <2 x i64> %1, %2 ; CHECK-DAG: sra.d [[R3:\$w[0-9]+]], [[R1]], [[R2]] @@ -688,7 +688,7 @@ define void @sra_v2i64(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind { define void @sra_v16i8_i(<16 x i8>* %c, <16 x i8>* %a) nounwind { ; CHECK: sra_v16i8_i: - %1 = load <16 x i8>* %a + %1 = load <16 x i8>, <16 x i8>* %a ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5) %2 = ashr <16 x i8> %1, <i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1> ; CHECK-DAG: srai.b [[R4:\$w[0-9]+]], [[R1]], 1 @@ -702,7 +702,7 @@ define void @sra_v16i8_i(<16 x i8>* %c, <16 x i8>* %a) nounwind { define void @sra_v8i16_i(<8 x i16>* %c, <8 x i16>* %a) nounwind { ; CHECK: sra_v8i16_i: - %1 = load <8 x i16>* %a + %1 = load <8 x i16>, <8 x i16>* %a ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5) %2 = ashr <8 x i16> %1, <i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1> ; CHECK-DAG: srai.h [[R4:\$w[0-9]+]], [[R1]], 1 @@ -716,7 +716,7 @@ define void @sra_v8i16_i(<8 x i16>* %c, <8 x i16>* %a) nounwind { define void @sra_v4i32_i(<4 x i32>* %c, <4 x i32>* %a) nounwind { ; CHECK: sra_v4i32_i: - %1 = load <4 x i32>* %a + %1 = load <4 x i32>, <4 x i32>* %a ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5) %2 = ashr <4 x i32> %1, <i32 1, i32 1, i32 1, i32 1> ; CHECK-DAG: srai.w [[R4:\$w[0-9]+]], [[R1]], 1 @@ -730,7 +730,7 @@ define void @sra_v4i32_i(<4 x i32>* %c, <4 x i32>* %a) nounwind { define void @sra_v2i64_i(<2 x i64>* %c, <2 x i64>* %a) nounwind { ; CHECK: sra_v2i64_i: - %1 = load <2 x i64>* %a + %1 = load <2 x i64>, <2 x i64>* %a ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5) %2 = ashr <2 x i64> %1, <i64 1, i64 1> ; CHECK-DAG: srai.d [[R4:\$w[0-9]+]], [[R1]], 1 @@ -744,9 +744,9 @@ define void @sra_v2i64_i(<2 x i64>* %c, <2 x i64>* %a) nounwind { define void @srl_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind { ; CHECK: srl_v16i8: - %1 = load <16 x i8>* %a + %1 = load <16 x i8>, <16 x i8>* %a ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5) - %2 = load <16 x i8>* %b + %2 = load <16 x i8>, <16 x i8>* %b ; CHECK-DAG: ld.b [[R2:\$w[0-9]+]], 0($6) %3 = lshr <16 x i8> %1, %2 ; CHECK-DAG: srl.b [[R3:\$w[0-9]+]], [[R1]], [[R2]] @@ -760,9 +760,9 @@ define void @srl_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind { define void @srl_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind { ; CHECK: srl_v8i16: - %1 = load <8 x i16>* %a + %1 = load <8 x i16>, <8 x i16>* %a ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5) - %2 = load <8 x i16>* %b + %2 = load <8 x i16>, <8 x i16>* %b ; CHECK-DAG: ld.h [[R2:\$w[0-9]+]], 0($6) %3 = lshr <8 x i16> %1, %2 ; CHECK-DAG: srl.h [[R3:\$w[0-9]+]], [[R1]], [[R2]] @@ -776,9 +776,9 @@ define void @srl_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind { define void @srl_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind { ; CHECK: srl_v4i32: - %1 = load <4 x i32>* %a + %1 = load <4 x i32>, <4 x i32>* %a ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5) - %2 = load <4 x i32>* %b + %2 = load <4 x i32>, <4 x i32>* %b ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6) %3 = lshr <4 x i32> %1, %2 ; CHECK-DAG: srl.w [[R3:\$w[0-9]+]], [[R1]], [[R2]] @@ -792,9 +792,9 @@ define void @srl_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind { define void @srl_v2i64(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind { ; CHECK: srl_v2i64: - %1 = load <2 x i64>* %a + %1 = load <2 x i64>, <2 x i64>* %a ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5) - %2 = load <2 x i64>* %b + %2 = load <2 x i64>, <2 x i64>* %b ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6) %3 = lshr <2 x i64> %1, %2 ; CHECK-DAG: srl.d [[R3:\$w[0-9]+]], [[R1]], [[R2]] @@ -808,7 +808,7 @@ define void @srl_v2i64(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind { define void @srl_v16i8_i(<16 x i8>* %c, <16 x i8>* %a) nounwind { ; CHECK: srl_v16i8_i: - %1 = load <16 x i8>* %a + %1 = load <16 x i8>, <16 x i8>* %a ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5) %2 = lshr <16 x i8> %1, <i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1> ; CHECK-DAG: srli.b [[R4:\$w[0-9]+]], [[R1]], 1 @@ -822,7 +822,7 @@ define void @srl_v16i8_i(<16 x i8>* %c, <16 x i8>* %a) nounwind { define void @srl_v8i16_i(<8 x i16>* %c, <8 x i16>* %a) nounwind { ; CHECK: srl_v8i16_i: - %1 = load <8 x i16>* %a + %1 = load <8 x i16>, <8 x i16>* %a ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5) %2 = lshr <8 x i16> %1, <i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1> ; CHECK-DAG: srli.h [[R4:\$w[0-9]+]], [[R1]], 1 @@ -836,7 +836,7 @@ define void @srl_v8i16_i(<8 x i16>* %c, <8 x i16>* %a) nounwind { define void @srl_v4i32_i(<4 x i32>* %c, <4 x i32>* %a) nounwind { ; CHECK: srl_v4i32_i: - %1 = load <4 x i32>* %a + %1 = load <4 x i32>, <4 x i32>* %a ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5) %2 = lshr <4 x i32> %1, <i32 1, i32 1, i32 1, i32 1> ; CHECK-DAG: srli.w [[R4:\$w[0-9]+]], [[R1]], 1 @@ -850,7 +850,7 @@ define void @srl_v4i32_i(<4 x i32>* %c, <4 x i32>* %a) nounwind { define void @srl_v2i64_i(<2 x i64>* %c, <2 x i64>* %a) nounwind { ; CHECK: srl_v2i64_i: - %1 = load <2 x i64>* %a + %1 = load <2 x i64>, <2 x i64>* %a ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5) %2 = lshr <2 x i64> %1, <i64 1, i64 1> ; CHECK-DAG: srli.d [[R4:\$w[0-9]+]], [[R1]], 1 @@ -864,7 +864,7 @@ define void @srl_v2i64_i(<2 x i64>* %c, <2 x i64>* %a) nounwind { define void @ctpop_v16i8(<16 x i8>* %c, <16 x i8>* %a) nounwind { ; CHECK: ctpop_v16i8: - %1 = load <16 x i8>* %a + %1 = load <16 x i8>, <16 x i8>* %a ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5) %2 = tail call <16 x i8> @llvm.ctpop.v16i8 (<16 x i8> %1) ; CHECK-DAG: pcnt.b [[R3:\$w[0-9]+]], [[R1]] @@ -878,7 +878,7 @@ define void @ctpop_v16i8(<16 x i8>* %c, <16 x i8>* %a) nounwind { define void @ctpop_v8i16(<8 x i16>* %c, <8 x i16>* %a) nounwind { ; CHECK: ctpop_v8i16: - %1 = load <8 x i16>* %a + %1 = load <8 x i16>, <8 x i16>* %a ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5) %2 = tail call <8 x i16> @llvm.ctpop.v8i16 (<8 x i16> %1) ; CHECK-DAG: pcnt.h [[R3:\$w[0-9]+]], [[R1]] @@ -892,7 +892,7 @@ define void @ctpop_v8i16(<8 x i16>* %c, <8 x i16>* %a) nounwind { define void @ctpop_v4i32(<4 x i32>* %c, <4 x i32>* %a) nounwind { ; CHECK: ctpop_v4i32: - %1 = load <4 x i32>* %a + %1 = load <4 x i32>, <4 x i32>* %a ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5) %2 = tail call <4 x i32> @llvm.ctpop.v4i32 (<4 x i32> %1) ; CHECK-DAG: pcnt.w [[R3:\$w[0-9]+]], [[R1]] @@ -906,7 +906,7 @@ define void @ctpop_v4i32(<4 x i32>* %c, <4 x i32>* %a) nounwind { define void @ctpop_v2i64(<2 x i64>* %c, <2 x i64>* %a) nounwind { ; CHECK: ctpop_v2i64: - %1 = load <2 x i64>* %a + %1 = load <2 x i64>, <2 x i64>* %a ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5) %2 = tail call <2 x i64> @llvm.ctpop.v2i64 (<2 x i64> %1) ; CHECK-DAG: pcnt.d [[R3:\$w[0-9]+]], [[R1]] @@ -920,7 +920,7 @@ define void @ctpop_v2i64(<2 x i64>* %c, <2 x i64>* %a) nounwind { define void @ctlz_v16i8(<16 x i8>* %c, <16 x i8>* %a) nounwind { ; CHECK: ctlz_v16i8: - %1 = load <16 x i8>* %a + %1 = load <16 x i8>, <16 x i8>* %a ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5) %2 = tail call <16 x i8> @llvm.ctlz.v16i8 (<16 x i8> %1) ; CHECK-DAG: nlzc.b [[R3:\$w[0-9]+]], [[R1]] @@ -934,7 +934,7 @@ define void @ctlz_v16i8(<16 x i8>* %c, <16 x i8>* %a) nounwind { define void @ctlz_v8i16(<8 x i16>* %c, <8 x i16>* %a) nounwind { ; CHECK: ctlz_v8i16: - %1 = load <8 x i16>* %a + %1 = load <8 x i16>, <8 x i16>* %a ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5) %2 = tail call <8 x i16> @llvm.ctlz.v8i16 (<8 x i16> %1) ; CHECK-DAG: nlzc.h [[R3:\$w[0-9]+]], [[R1]] @@ -948,7 +948,7 @@ define void @ctlz_v8i16(<8 x i16>* %c, <8 x i16>* %a) nounwind { define void @ctlz_v4i32(<4 x i32>* %c, <4 x i32>* %a) nounwind { ; CHECK: ctlz_v4i32: - %1 = load <4 x i32>* %a + %1 = load <4 x i32>, <4 x i32>* %a ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5) %2 = tail call <4 x i32> @llvm.ctlz.v4i32 (<4 x i32> %1) ; CHECK-DAG: nlzc.w [[R3:\$w[0-9]+]], [[R1]] @@ -962,7 +962,7 @@ define void @ctlz_v4i32(<4 x i32>* %c, <4 x i32>* %a) nounwind { define void @ctlz_v2i64(<2 x i64>* %c, <2 x i64>* %a) nounwind { ; CHECK: ctlz_v2i64: - %1 = load <2 x i64>* %a + %1 = load <2 x i64>, <2 x i64>* %a ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5) %2 = tail call <2 x i64> @llvm.ctlz.v2i64 (<2 x i64> %1) ; CHECK-DAG: nlzc.d [[R3:\$w[0-9]+]], [[R1]] @@ -976,11 +976,11 @@ define void @ctlz_v2i64(<2 x i64>* %c, <2 x i64>* %a) nounwind { define void @bsel_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b, <16 x i8>* %m) nounwind { ; CHECK: bsel_v16i8: - %1 = load <16 x i8>* %a + %1 = load <16 x i8>, <16 x i8>* %a ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5) - %2 = load <16 x i8>* %b + %2 = load <16 x i8>, <16 x i8>* %b ; CHECK-DAG: ld.b [[R2:\$w[0-9]+]], 0($6) - %3 = load <16 x i8>* %m + %3 = load <16 x i8>, <16 x i8>* %m ; CHECK-DAG: ld.b [[R3:\$w[0-9]+]], 0($7) %4 = xor <16 x i8> %3, <i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, @@ -1002,9 +1002,9 @@ define void @bsel_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b, <16 x i8>* define void @bsel_v16i8_i(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %m) nounwind { ; CHECK: bsel_v16i8_i: - %1 = load <16 x i8>* %a + %1 = load <16 x i8>, <16 x i8>* %a ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5) - %2 = load <16 x i8>* %m + %2 = load <16 x i8>, <16 x i8>* %m ; CHECK-DAG: ld.b [[R3:\$w[0-9]+]], 0($6) %3 = xor <16 x i8> %2, <i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, @@ -1027,9 +1027,9 @@ define void @bsel_v16i8_i(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %m) nounwind define void @bsel_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind { ; CHECK: bsel_v8i16: - %1 = load <8 x i16>* %a + %1 = load <8 x i16>, <8 x i16>* %a ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5) - %2 = load <8 x i16>* %b + %2 = load <8 x i16>, <8 x i16>* %b ; CHECK-DAG: ld.h [[R2:\$w[0-9]+]], 0($6) %3 = and <8 x i16> %1, <i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6> @@ -1048,9 +1048,9 @@ define void @bsel_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind { define void @bsel_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind { ; CHECK: bsel_v4i32: - %1 = load <4 x i32>* %a + %1 = load <4 x i32>, <4 x i32>* %a ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5) - %2 = load <4 x i32>* %b + %2 = load <4 x i32>, <4 x i32>* %b ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6) %3 = and <4 x i32> %1, <i32 6, i32 6, i32 6, i32 6> %4 = and <4 x i32> %2, <i32 4294967289, i32 4294967289, i32 4294967289, i32 4294967289> @@ -1067,9 +1067,9 @@ define void @bsel_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind { define void @bsel_v2i64(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind { ; CHECK: bsel_v2i64: - %1 = load <2 x i64>* %a + %1 = load <2 x i64>, <2 x i64>* %a ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5) - %2 = load <2 x i64>* %b + %2 = load <2 x i64>, <2 x i64>* %b ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6) %3 = and <2 x i64> %1, <i64 6, i64 6> %4 = and <2 x i64> %2, <i64 18446744073709551609, i64 18446744073709551609> @@ -1086,9 +1086,9 @@ define void @bsel_v2i64(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind { define void @binsl_v16i8_i(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind { ; CHECK: binsl_v16i8_i: - %1 = load <16 x i8>* %a + %1 = load <16 x i8>, <16 x i8>* %a ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5) - %2 = load <16 x i8>* %b + %2 = load <16 x i8>, <16 x i8>* %b ; CHECK-DAG: ld.b [[R2:\$w[0-9]+]], 0($6) %3 = and <16 x i8> %1, <i8 192, i8 192, i8 192, i8 192, i8 192, i8 192, i8 192, i8 192, @@ -1110,9 +1110,9 @@ define void @binsl_v16i8_i(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind define void @binsl_v8i16_i(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind { ; CHECK: binsl_v8i16_i: - %1 = load <8 x i16>* %a + %1 = load <8 x i16>, <8 x i16>* %a ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5) - %2 = load <8 x i16>* %b + %2 = load <8 x i16>, <8 x i16>* %b ; CHECK-DAG: ld.h [[R2:\$w[0-9]+]], 0($6) %3 = and <8 x i16> %1, <i16 49152, i16 49152, i16 49152, i16 49152, i16 49152, i16 49152, i16 49152, i16 49152> @@ -1130,9 +1130,9 @@ define void @binsl_v8i16_i(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind define void @binsl_v4i32_i(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind { ; CHECK: binsl_v4i32_i: - %1 = load <4 x i32>* %a + %1 = load <4 x i32>, <4 x i32>* %a ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5) - %2 = load <4 x i32>* %b + %2 = load <4 x i32>, <4 x i32>* %b ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6) %3 = and <4 x i32> %1, <i32 3221225472, i32 3221225472, i32 3221225472, i32 3221225472> %4 = and <4 x i32> %2, <i32 1073741823, i32 1073741823, i32 1073741823, i32 1073741823> @@ -1148,9 +1148,9 @@ define void @binsl_v4i32_i(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind define void @binsl_v2i64_i(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind { ; CHECK: binsl_v2i64_i: - %1 = load <2 x i64>* %a + %1 = load <2 x i64>, <2 x i64>* %a ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5) - %2 = load <2 x i64>* %b + %2 = load <2 x i64>, <2 x i64>* %b ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6) %3 = and <2 x i64> %1, <i64 18446744073709551608, i64 18446744073709551608> %4 = and <2 x i64> %2, <i64 7, i64 7> @@ -1170,9 +1170,9 @@ define void @binsl_v2i64_i(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind define void @binsr_v16i8_i(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind { ; CHECK: binsr_v16i8_i: - %1 = load <16 x i8>* %a + %1 = load <16 x i8>, <16 x i8>* %a ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5) - %2 = load <16 x i8>* %b + %2 = load <16 x i8>, <16 x i8>* %b ; CHECK-DAG: ld.b [[R2:\$w[0-9]+]], 0($6) %3 = and <16 x i8> %1, <i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3> @@ -1192,9 +1192,9 @@ define void @binsr_v16i8_i(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind define void @binsr_v8i16_i(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind { ; CHECK: binsr_v8i16_i: - %1 = load <8 x i16>* %a + %1 = load <8 x i16>, <8 x i16>* %a ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5) - %2 = load <8 x i16>* %b + %2 = load <8 x i16>, <8 x i16>* %b ; CHECK-DAG: ld.h [[R2:\$w[0-9]+]], 0($6) %3 = and <8 x i16> %1, <i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3> @@ -1212,9 +1212,9 @@ define void @binsr_v8i16_i(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind define void @binsr_v4i32_i(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind { ; CHECK: binsr_v4i32_i: - %1 = load <4 x i32>* %a + %1 = load <4 x i32>, <4 x i32>* %a ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5) - %2 = load <4 x i32>* %b + %2 = load <4 x i32>, <4 x i32>* %b ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6) %3 = and <4 x i32> %1, <i32 3, i32 3, i32 3, i32 3> %4 = and <4 x i32> %2, <i32 4294967292, i32 4294967292, i32 4294967292, i32 4294967292> @@ -1230,9 +1230,9 @@ define void @binsr_v4i32_i(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind define void @binsr_v2i64_i(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind { ; CHECK: binsr_v2i64_i: - %1 = load <2 x i64>* %a + %1 = load <2 x i64>, <2 x i64>* %a ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5) - %2 = load <2 x i64>* %b + %2 = load <2 x i64>, <2 x i64>* %b ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6) %3 = and <2 x i64> %1, <i64 3, i64 3> %4 = and <2 x i64> %2, <i64 18446744073709551612, i64 18446744073709551612> @@ -1248,9 +1248,9 @@ define void @binsr_v2i64_i(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind define void @bclr_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind { ; CHECK: bclr_v16i8: - %1 = load <16 x i8>* %a + %1 = load <16 x i8>, <16 x i8>* %a ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5) - %2 = load <16 x i8>* %b + %2 = load <16 x i8>, <16 x i8>* %b ; CHECK-DAG: ld.b [[R2:\$w[0-9]+]], 0($6) %3 = shl <16 x i8> <i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1>, %2 %4 = xor <16 x i8> %3, <i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1> @@ -1266,9 +1266,9 @@ define void @bclr_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind { define void @bclr_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind { ; CHECK: bclr_v8i16: - %1 = load <8 x i16>* %a + %1 = load <8 x i16>, <8 x i16>* %a ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5) - %2 = load <8 x i16>* %b + %2 = load <8 x i16>, <8 x i16>* %b ; CHECK-DAG: ld.h [[R2:\$w[0-9]+]], 0($6) %3 = shl <8 x i16> <i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1>, %2 %4 = xor <8 x i16> %3, <i16 -1, i16 -1, i16 -1, i16 -1, i16 -1, i16 -1, i16 -1, i16 -1> @@ -1284,9 +1284,9 @@ define void @bclr_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind { define void @bclr_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind { ; CHECK: bclr_v4i32: - %1 = load <4 x i32>* %a + %1 = load <4 x i32>, <4 x i32>* %a ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5) - %2 = load <4 x i32>* %b + %2 = load <4 x i32>, <4 x i32>* %b ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6) %3 = shl <4 x i32> <i32 1, i32 1, i32 1, i32 1>, %2 %4 = xor <4 x i32> %3, <i32 -1, i32 -1, i32 -1, i32 -1> @@ -1302,9 +1302,9 @@ define void @bclr_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind { define void @bclr_v2i64(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind { ; CHECK: bclr_v2i64: - %1 = load <2 x i64>* %a + %1 = load <2 x i64>, <2 x i64>* %a ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5) - %2 = load <2 x i64>* %b + %2 = load <2 x i64>, <2 x i64>* %b ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6) %3 = shl <2 x i64> <i64 1, i64 1>, %2 %4 = xor <2 x i64> %3, <i64 -1, i64 -1> @@ -1320,9 +1320,9 @@ define void @bclr_v2i64(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind { define void @bset_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind { ; CHECK: bset_v16i8: - %1 = load <16 x i8>* %a + %1 = load <16 x i8>, <16 x i8>* %a ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5) - %2 = load <16 x i8>* %b + %2 = load <16 x i8>, <16 x i8>* %b ; CHECK-DAG: ld.b [[R2:\$w[0-9]+]], 0($6) %3 = shl <16 x i8> <i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1>, %2 %4 = or <16 x i8> %1, %3 @@ -1337,9 +1337,9 @@ define void @bset_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind { define void @bset_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind { ; CHECK: bset_v8i16: - %1 = load <8 x i16>* %a + %1 = load <8 x i16>, <8 x i16>* %a ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5) - %2 = load <8 x i16>* %b + %2 = load <8 x i16>, <8 x i16>* %b ; CHECK-DAG: ld.h [[R2:\$w[0-9]+]], 0($6) %3 = shl <8 x i16> <i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1>, %2 %4 = or <8 x i16> %1, %3 @@ -1354,9 +1354,9 @@ define void @bset_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind { define void @bset_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind { ; CHECK: bset_v4i32: - %1 = load <4 x i32>* %a + %1 = load <4 x i32>, <4 x i32>* %a ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5) - %2 = load <4 x i32>* %b + %2 = load <4 x i32>, <4 x i32>* %b ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6) %3 = shl <4 x i32> <i32 1, i32 1, i32 1, i32 1>, %2 %4 = or <4 x i32> %1, %3 @@ -1371,9 +1371,9 @@ define void @bset_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind { define void @bset_v2i64(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind { ; CHECK: bset_v2i64: - %1 = load <2 x i64>* %a + %1 = load <2 x i64>, <2 x i64>* %a ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5) - %2 = load <2 x i64>* %b + %2 = load <2 x i64>, <2 x i64>* %b ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6) %3 = shl <2 x i64> <i64 1, i64 1>, %2 %4 = or <2 x i64> %1, %3 @@ -1388,9 +1388,9 @@ define void @bset_v2i64(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind { define void @bneg_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind { ; CHECK: bneg_v16i8: - %1 = load <16 x i8>* %a + %1 = load <16 x i8>, <16 x i8>* %a ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5) - %2 = load <16 x i8>* %b + %2 = load <16 x i8>, <16 x i8>* %b ; CHECK-DAG: ld.b [[R2:\$w[0-9]+]], 0($6) %3 = shl <16 x i8> <i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1>, %2 %4 = xor <16 x i8> %1, %3 @@ -1405,9 +1405,9 @@ define void @bneg_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind { define void @bneg_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind { ; CHECK: bneg_v8i16: - %1 = load <8 x i16>* %a + %1 = load <8 x i16>, <8 x i16>* %a ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5) - %2 = load <8 x i16>* %b + %2 = load <8 x i16>, <8 x i16>* %b ; CHECK-DAG: ld.h [[R2:\$w[0-9]+]], 0($6) %3 = shl <8 x i16> <i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1>, %2 %4 = xor <8 x i16> %1, %3 @@ -1422,9 +1422,9 @@ define void @bneg_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind { define void @bneg_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind { ; CHECK: bneg_v4i32: - %1 = load <4 x i32>* %a + %1 = load <4 x i32>, <4 x i32>* %a ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5) - %2 = load <4 x i32>* %b + %2 = load <4 x i32>, <4 x i32>* %b ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6) %3 = shl <4 x i32> <i32 1, i32 1, i32 1, i32 1>, %2 %4 = xor <4 x i32> %1, %3 @@ -1439,9 +1439,9 @@ define void @bneg_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind { define void @bneg_v2i64(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind { ; CHECK: bneg_v2i64: - %1 = load <2 x i64>* %a + %1 = load <2 x i64>, <2 x i64>* %a ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5) - %2 = load <2 x i64>* %b + %2 = load <2 x i64>, <2 x i64>* %b ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6) %3 = shl <2 x i64> <i64 1, i64 1>, %2 %4 = xor <2 x i64> %1, %3 @@ -1456,7 +1456,7 @@ define void @bneg_v2i64(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind { define void @bclri_v16i8(<16 x i8>* %c, <16 x i8>* %a) nounwind { ; CHECK: bclri_v16i8: - %1 = load <16 x i8>* %a + %1 = load <16 x i8>, <16 x i8>* %a ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5) %2 = xor <16 x i8> <i8 8, i8 8, i8 8, i8 8, i8 8, i8 8, i8 8, i8 8, i8 8, i8 8, i8 8, i8 8, i8 8, i8 8, i8 8, i8 8>, <i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1> @@ -1473,7 +1473,7 @@ define void @bclri_v16i8(<16 x i8>* %c, <16 x i8>* %a) nounwind { define void @bclri_v8i16(<8 x i16>* %c, <8 x i16>* %a) nounwind { ; CHECK: bclri_v8i16: - %1 = load <8 x i16>* %a + %1 = load <8 x i16>, <8 x i16>* %a ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5) %2 = xor <8 x i16> <i16 8, i16 8, i16 8, i16 8, i16 8, i16 8, i16 8, i16 8>, <i16 -1, i16 -1, i16 -1, i16 -1, i16 -1, i16 -1, i16 -1, i16 -1> @@ -1489,7 +1489,7 @@ define void @bclri_v8i16(<8 x i16>* %c, <8 x i16>* %a) nounwind { define void @bclri_v4i32(<4 x i32>* %c, <4 x i32>* %a) nounwind { ; CHECK: bclri_v4i32: - %1 = load <4 x i32>* %a + %1 = load <4 x i32>, <4 x i32>* %a ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5) %2 = xor <4 x i32> <i32 8, i32 8, i32 8, i32 8>, <i32 -1, i32 -1, i32 -1, i32 -1> @@ -1505,7 +1505,7 @@ define void @bclri_v4i32(<4 x i32>* %c, <4 x i32>* %a) nounwind { define void @bclri_v2i64(<2 x i64>* %c, <2 x i64>* %a) nounwind { ; CHECK: bclri_v2i64: - %1 = load <2 x i64>* %a + %1 = load <2 x i64>, <2 x i64>* %a ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5) %2 = xor <2 x i64> <i64 8, i64 8>, <i64 -1, i64 -1> @@ -1521,7 +1521,7 @@ define void @bclri_v2i64(<2 x i64>* %c, <2 x i64>* %a) nounwind { define void @bseti_v16i8(<16 x i8>* %c, <16 x i8>* %a) nounwind { ; CHECK: bseti_v16i8: - %1 = load <16 x i8>* %a + %1 = load <16 x i8>, <16 x i8>* %a ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5) %2 = or <16 x i8> %1, <i8 8, i8 8, i8 8, i8 8, i8 8, i8 8, i8 8, i8 8, i8 8, i8 8, i8 8, i8 8, i8 8, i8 8, i8 8, i8 8> ; CHECK-DAG: bseti.b [[R3:\$w[0-9]+]], [[R1]], 3 @@ -1535,7 +1535,7 @@ define void @bseti_v16i8(<16 x i8>* %c, <16 x i8>* %a) nounwind { define void @bseti_v8i16(<8 x i16>* %c, <8 x i16>* %a) nounwind { ; CHECK: bseti_v8i16: - %1 = load <8 x i16>* %a + %1 = load <8 x i16>, <8 x i16>* %a ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5) %2 = or <8 x i16> %1, <i16 8, i16 8, i16 8, i16 8, i16 8, i16 8, i16 8, i16 8> ; CHECK-DAG: bseti.h [[R3:\$w[0-9]+]], [[R1]], 3 @@ -1549,7 +1549,7 @@ define void @bseti_v8i16(<8 x i16>* %c, <8 x i16>* %a) nounwind { define void @bseti_v4i32(<4 x i32>* %c, <4 x i32>* %a) nounwind { ; CHECK: bseti_v4i32: - %1 = load <4 x i32>* %a + %1 = load <4 x i32>, <4 x i32>* %a ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5) %2 = or <4 x i32> %1, <i32 8, i32 8, i32 8, i32 8> ; CHECK-DAG: bseti.w [[R3:\$w[0-9]+]], [[R1]], 3 @@ -1563,7 +1563,7 @@ define void @bseti_v4i32(<4 x i32>* %c, <4 x i32>* %a) nounwind { define void @bseti_v2i64(<2 x i64>* %c, <2 x i64>* %a) nounwind { ; CHECK: bseti_v2i64: - %1 = load <2 x i64>* %a + %1 = load <2 x i64>, <2 x i64>* %a ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5) %2 = or <2 x i64> %1, <i64 8, i64 8> ; CHECK-DAG: bseti.d [[R3:\$w[0-9]+]], [[R1]], 3 @@ -1577,7 +1577,7 @@ define void @bseti_v2i64(<2 x i64>* %c, <2 x i64>* %a) nounwind { define void @bnegi_v16i8(<16 x i8>* %c, <16 x i8>* %a) nounwind { ; CHECK: bnegi_v16i8: - %1 = load <16 x i8>* %a + %1 = load <16 x i8>, <16 x i8>* %a ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5) %2 = xor <16 x i8> %1, <i8 8, i8 8, i8 8, i8 8, i8 8, i8 8, i8 8, i8 8, i8 8, i8 8, i8 8, i8 8, i8 8, i8 8, i8 8, i8 8> ; CHECK-DAG: bnegi.b [[R3:\$w[0-9]+]], [[R1]], 3 @@ -1591,7 +1591,7 @@ define void @bnegi_v16i8(<16 x i8>* %c, <16 x i8>* %a) nounwind { define void @bnegi_v8i16(<8 x i16>* %c, <8 x i16>* %a) nounwind { ; CHECK: bnegi_v8i16: - %1 = load <8 x i16>* %a + %1 = load <8 x i16>, <8 x i16>* %a ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5) %2 = xor <8 x i16> %1, <i16 8, i16 8, i16 8, i16 8, i16 8, i16 8, i16 8, i16 8> ; CHECK-DAG: bnegi.h [[R3:\$w[0-9]+]], [[R1]], 3 @@ -1605,7 +1605,7 @@ define void @bnegi_v8i16(<8 x i16>* %c, <8 x i16>* %a) nounwind { define void @bnegi_v4i32(<4 x i32>* %c, <4 x i32>* %a) nounwind { ; CHECK: bnegi_v4i32: - %1 = load <4 x i32>* %a + %1 = load <4 x i32>, <4 x i32>* %a ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5) %2 = xor <4 x i32> %1, <i32 8, i32 8, i32 8, i32 8> ; CHECK-DAG: bnegi.w [[R3:\$w[0-9]+]], [[R1]], 3 @@ -1619,7 +1619,7 @@ define void @bnegi_v4i32(<4 x i32>* %c, <4 x i32>* %a) nounwind { define void @bnegi_v2i64(<2 x i64>* %c, <2 x i64>* %a) nounwind { ; CHECK: bnegi_v2i64: - %1 = load <2 x i64>* %a + %1 = load <2 x i64>, <2 x i64>* %a ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5) %2 = xor <2 x i64> %1, <i64 8, i64 8> ; CHECK-DAG: bnegi.d [[R3:\$w[0-9]+]], [[R1]], 3 diff --git a/test/CodeGen/Mips/msa/compare.ll b/test/CodeGen/Mips/msa/compare.ll index 87ca1482da81f..bc4f6e7e39469 100644 --- a/test/CodeGen/Mips/msa/compare.ll +++ b/test/CodeGen/Mips/msa/compare.ll @@ -4,9 +4,9 @@ define void @ceq_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind { ; CHECK: ceq_v16i8: - %1 = load <16 x i8>* %a + %1 = load <16 x i8>, <16 x i8>* %a ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5) - %2 = load <16 x i8>* %b + %2 = load <16 x i8>, <16 x i8>* %b ; CHECK-DAG: ld.b [[R2:\$w[0-9]+]], 0($6) %3 = icmp eq <16 x i8> %1, %2 %4 = sext <16 x i1> %3 to <16 x i8> @@ -21,9 +21,9 @@ define void @ceq_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind { define void @ceq_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind { ; CHECK: ceq_v8i16: - %1 = load <8 x i16>* %a + %1 = load <8 x i16>, <8 x i16>* %a ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5) - %2 = load <8 x i16>* %b + %2 = load <8 x i16>, <8 x i16>* %b ; CHECK-DAG: ld.h [[R2:\$w[0-9]+]], 0($6) %3 = icmp eq <8 x i16> %1, %2 %4 = sext <8 x i1> %3 to <8 x i16> @@ -38,9 +38,9 @@ define void @ceq_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind { define void @ceq_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind { ; CHECK: ceq_v4i32: - %1 = load <4 x i32>* %a + %1 = load <4 x i32>, <4 x i32>* %a ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5) - %2 = load <4 x i32>* %b + %2 = load <4 x i32>, <4 x i32>* %b ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6) %3 = icmp eq <4 x i32> %1, %2 %4 = sext <4 x i1> %3 to <4 x i32> @@ -55,9 +55,9 @@ define void @ceq_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind { define void @ceq_v2i64(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind { ; CHECK: ceq_v2i64: - %1 = load <2 x i64>* %a + %1 = load <2 x i64>, <2 x i64>* %a ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5) - %2 = load <2 x i64>* %b + %2 = load <2 x i64>, <2 x i64>* %b ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6) %3 = icmp eq <2 x i64> %1, %2 %4 = sext <2 x i1> %3 to <2 x i64> @@ -72,9 +72,9 @@ define void @ceq_v2i64(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind { define void @cle_s_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind { ; CHECK: cle_s_v16i8: - %1 = load <16 x i8>* %a + %1 = load <16 x i8>, <16 x i8>* %a ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5) - %2 = load <16 x i8>* %b + %2 = load <16 x i8>, <16 x i8>* %b ; CHECK-DAG: ld.b [[R2:\$w[0-9]+]], 0($6) %3 = icmp sle <16 x i8> %1, %2 %4 = sext <16 x i1> %3 to <16 x i8> @@ -89,9 +89,9 @@ define void @cle_s_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind { define void @cle_s_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind { ; CHECK: cle_s_v8i16: - %1 = load <8 x i16>* %a + %1 = load <8 x i16>, <8 x i16>* %a ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5) - %2 = load <8 x i16>* %b + %2 = load <8 x i16>, <8 x i16>* %b ; CHECK-DAG: ld.h [[R2:\$w[0-9]+]], 0($6) %3 = icmp sle <8 x i16> %1, %2 %4 = sext <8 x i1> %3 to <8 x i16> @@ -106,9 +106,9 @@ define void @cle_s_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind { define void @cle_s_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind { ; CHECK: cle_s_v4i32: - %1 = load <4 x i32>* %a + %1 = load <4 x i32>, <4 x i32>* %a ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5) - %2 = load <4 x i32>* %b + %2 = load <4 x i32>, <4 x i32>* %b ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6) %3 = icmp sle <4 x i32> %1, %2 %4 = sext <4 x i1> %3 to <4 x i32> @@ -123,9 +123,9 @@ define void @cle_s_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind { define void @cle_s_v2i64(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind { ; CHECK: cle_s_v2i64: - %1 = load <2 x i64>* %a + %1 = load <2 x i64>, <2 x i64>* %a ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5) - %2 = load <2 x i64>* %b + %2 = load <2 x i64>, <2 x i64>* %b ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6) %3 = icmp sle <2 x i64> %1, %2 %4 = sext <2 x i1> %3 to <2 x i64> @@ -140,9 +140,9 @@ define void @cle_s_v2i64(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind { define void @cle_u_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind { ; CHECK: cle_u_v16i8: - %1 = load <16 x i8>* %a + %1 = load <16 x i8>, <16 x i8>* %a ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5) - %2 = load <16 x i8>* %b + %2 = load <16 x i8>, <16 x i8>* %b ; CHECK-DAG: ld.b [[R2:\$w[0-9]+]], 0($6) %3 = icmp ule <16 x i8> %1, %2 %4 = sext <16 x i1> %3 to <16 x i8> @@ -157,9 +157,9 @@ define void @cle_u_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind { define void @cle_u_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind { ; CHECK: cle_u_v8i16: - %1 = load <8 x i16>* %a + %1 = load <8 x i16>, <8 x i16>* %a ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5) - %2 = load <8 x i16>* %b + %2 = load <8 x i16>, <8 x i16>* %b ; CHECK-DAG: ld.h [[R2:\$w[0-9]+]], 0($6) %3 = icmp ule <8 x i16> %1, %2 %4 = sext <8 x i1> %3 to <8 x i16> @@ -174,9 +174,9 @@ define void @cle_u_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind { define void @cle_u_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind { ; CHECK: cle_u_v4i32: - %1 = load <4 x i32>* %a + %1 = load <4 x i32>, <4 x i32>* %a ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5) - %2 = load <4 x i32>* %b + %2 = load <4 x i32>, <4 x i32>* %b ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6) %3 = icmp ule <4 x i32> %1, %2 %4 = sext <4 x i1> %3 to <4 x i32> @@ -191,9 +191,9 @@ define void @cle_u_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind { define void @cle_u_v2i64(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind { ; CHECK: cle_u_v2i64: - %1 = load <2 x i64>* %a + %1 = load <2 x i64>, <2 x i64>* %a ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5) - %2 = load <2 x i64>* %b + %2 = load <2 x i64>, <2 x i64>* %b ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6) %3 = icmp ule <2 x i64> %1, %2 %4 = sext <2 x i1> %3 to <2 x i64> @@ -208,9 +208,9 @@ define void @cle_u_v2i64(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind { define void @clt_s_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind { ; CHECK: clt_s_v16i8: - %1 = load <16 x i8>* %a + %1 = load <16 x i8>, <16 x i8>* %a ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5) - %2 = load <16 x i8>* %b + %2 = load <16 x i8>, <16 x i8>* %b ; CHECK-DAG: ld.b [[R2:\$w[0-9]+]], 0($6) %3 = icmp slt <16 x i8> %1, %2 %4 = sext <16 x i1> %3 to <16 x i8> @@ -225,9 +225,9 @@ define void @clt_s_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind { define void @clt_s_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind { ; CHECK: clt_s_v8i16: - %1 = load <8 x i16>* %a + %1 = load <8 x i16>, <8 x i16>* %a ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5) - %2 = load <8 x i16>* %b + %2 = load <8 x i16>, <8 x i16>* %b ; CHECK-DAG: ld.h [[R2:\$w[0-9]+]], 0($6) %3 = icmp slt <8 x i16> %1, %2 %4 = sext <8 x i1> %3 to <8 x i16> @@ -242,9 +242,9 @@ define void @clt_s_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind { define void @clt_s_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind { ; CHECK: clt_s_v4i32: - %1 = load <4 x i32>* %a + %1 = load <4 x i32>, <4 x i32>* %a ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5) - %2 = load <4 x i32>* %b + %2 = load <4 x i32>, <4 x i32>* %b ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6) %3 = icmp slt <4 x i32> %1, %2 %4 = sext <4 x i1> %3 to <4 x i32> @@ -259,9 +259,9 @@ define void @clt_s_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind { define void @clt_s_v2i64(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind { ; CHECK: clt_s_v2i64: - %1 = load <2 x i64>* %a + %1 = load <2 x i64>, <2 x i64>* %a ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5) - %2 = load <2 x i64>* %b + %2 = load <2 x i64>, <2 x i64>* %b ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6) %3 = icmp slt <2 x i64> %1, %2 %4 = sext <2 x i1> %3 to <2 x i64> @@ -276,9 +276,9 @@ define void @clt_s_v2i64(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind { define void @clt_u_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind { ; CHECK: clt_u_v16i8: - %1 = load <16 x i8>* %a + %1 = load <16 x i8>, <16 x i8>* %a ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5) - %2 = load <16 x i8>* %b + %2 = load <16 x i8>, <16 x i8>* %b ; CHECK-DAG: ld.b [[R2:\$w[0-9]+]], 0($6) %3 = icmp ult <16 x i8> %1, %2 %4 = sext <16 x i1> %3 to <16 x i8> @@ -293,9 +293,9 @@ define void @clt_u_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind { define void @clt_u_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind { ; CHECK: clt_u_v8i16: - %1 = load <8 x i16>* %a + %1 = load <8 x i16>, <8 x i16>* %a ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5) - %2 = load <8 x i16>* %b + %2 = load <8 x i16>, <8 x i16>* %b ; CHECK-DAG: ld.h [[R2:\$w[0-9]+]], 0($6) %3 = icmp ult <8 x i16> %1, %2 %4 = sext <8 x i1> %3 to <8 x i16> @@ -310,9 +310,9 @@ define void @clt_u_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind { define void @clt_u_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind { ; CHECK: clt_u_v4i32: - %1 = load <4 x i32>* %a + %1 = load <4 x i32>, <4 x i32>* %a ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5) - %2 = load <4 x i32>* %b + %2 = load <4 x i32>, <4 x i32>* %b ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6) %3 = icmp ult <4 x i32> %1, %2 %4 = sext <4 x i1> %3 to <4 x i32> @@ -327,9 +327,9 @@ define void @clt_u_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind { define void @clt_u_v2i64(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind { ; CHECK: clt_u_v2i64: - %1 = load <2 x i64>* %a + %1 = load <2 x i64>, <2 x i64>* %a ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5) - %2 = load <2 x i64>* %b + %2 = load <2 x i64>, <2 x i64>* %b ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6) %3 = icmp ult <2 x i64> %1, %2 %4 = sext <2 x i1> %3 to <2 x i64> @@ -345,9 +345,9 @@ define void @clt_u_v2i64(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind { ; issues in this area. define void @cne_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind { ; CHECK: cne_v16i8: - %1 = load <16 x i8>* %a + %1 = load <16 x i8>, <16 x i8>* %a ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5) - %2 = load <16 x i8>* %b + %2 = load <16 x i8>, <16 x i8>* %b ; CHECK-DAG: ld.b [[R2:\$w[0-9]+]], 0($6) %3 = icmp ne <16 x i8> %1, %2 %4 = sext <16 x i1> %3 to <16 x i8> @@ -365,9 +365,9 @@ define void @cne_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind { define void @cne_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind { ; CHECK: cne_v8i16: - %1 = load <8 x i16>* %a + %1 = load <8 x i16>, <8 x i16>* %a ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5) - %2 = load <8 x i16>* %b + %2 = load <8 x i16>, <8 x i16>* %b ; CHECK-DAG: ld.h [[R2:\$w[0-9]+]], 0($6) %3 = icmp ne <8 x i16> %1, %2 %4 = sext <8 x i1> %3 to <8 x i16> @@ -387,9 +387,9 @@ define void @cne_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind { define void @cne_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind { ; CHECK: cne_v4i32: - %1 = load <4 x i32>* %a + %1 = load <4 x i32>, <4 x i32>* %a ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5) - %2 = load <4 x i32>* %b + %2 = load <4 x i32>, <4 x i32>* %b ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6) %3 = icmp ne <4 x i32> %1, %2 %4 = sext <4 x i1> %3 to <4 x i32> @@ -409,9 +409,9 @@ define void @cne_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind { define void @cne_v2i64(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind { ; CHECK: cne_v2i64: - %1 = load <2 x i64>* %a + %1 = load <2 x i64>, <2 x i64>* %a ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5) - %2 = load <2 x i64>* %b + %2 = load <2 x i64>, <2 x i64>* %b ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6) %3 = icmp ne <2 x i64> %1, %2 %4 = sext <2 x i1> %3 to <2 x i64> @@ -429,7 +429,7 @@ define void @cne_v2i64(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind { define void @ceqi_v16i8(<16 x i8>* %c, <16 x i8>* %a) nounwind { ; CHECK: ceqi_v16i8: - %1 = load <16 x i8>* %a + %1 = load <16 x i8>, <16 x i8>* %a ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5) %2 = icmp eq <16 x i8> %1, <i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1> %3 = sext <16 x i1> %2 to <16 x i8> @@ -444,7 +444,7 @@ define void @ceqi_v16i8(<16 x i8>* %c, <16 x i8>* %a) nounwind { define void @ceqi_v8i16(<8 x i16>* %c, <8 x i16>* %a) nounwind { ; CHECK: ceqi_v8i16: - %1 = load <8 x i16>* %a + %1 = load <8 x i16>, <8 x i16>* %a ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5) %2 = icmp eq <8 x i16> %1, <i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1> %3 = sext <8 x i1> %2 to <8 x i16> @@ -459,7 +459,7 @@ define void @ceqi_v8i16(<8 x i16>* %c, <8 x i16>* %a) nounwind { define void @ceqi_v4i32(<4 x i32>* %c, <4 x i32>* %a) nounwind { ; CHECK: ceqi_v4i32: - %1 = load <4 x i32>* %a + %1 = load <4 x i32>, <4 x i32>* %a ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5) %2 = icmp eq <4 x i32> %1, <i32 1, i32 1, i32 1, i32 1> %3 = sext <4 x i1> %2 to <4 x i32> @@ -474,7 +474,7 @@ define void @ceqi_v4i32(<4 x i32>* %c, <4 x i32>* %a) nounwind { define void @ceqi_v2i64(<2 x i64>* %c, <2 x i64>* %a) nounwind { ; CHECK: ceqi_v2i64: - %1 = load <2 x i64>* %a + %1 = load <2 x i64>, <2 x i64>* %a ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5) %2 = icmp eq <2 x i64> %1, <i64 1, i64 1> %3 = sext <2 x i1> %2 to <2 x i64> @@ -489,7 +489,7 @@ define void @ceqi_v2i64(<2 x i64>* %c, <2 x i64>* %a) nounwind { define void @clei_s_v16i8(<16 x i8>* %c, <16 x i8>* %a) nounwind { ; CHECK: clei_s_v16i8: - %1 = load <16 x i8>* %a + %1 = load <16 x i8>, <16 x i8>* %a ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5) %2 = icmp sle <16 x i8> %1, <i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1> %3 = sext <16 x i1> %2 to <16 x i8> @@ -504,7 +504,7 @@ define void @clei_s_v16i8(<16 x i8>* %c, <16 x i8>* %a) nounwind { define void @clei_s_v8i16(<8 x i16>* %c, <8 x i16>* %a) nounwind { ; CHECK: clei_s_v8i16: - %1 = load <8 x i16>* %a + %1 = load <8 x i16>, <8 x i16>* %a ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5) %2 = icmp sle <8 x i16> %1, <i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1> %3 = sext <8 x i1> %2 to <8 x i16> @@ -519,7 +519,7 @@ define void @clei_s_v8i16(<8 x i16>* %c, <8 x i16>* %a) nounwind { define void @clei_s_v4i32(<4 x i32>* %c, <4 x i32>* %a) nounwind { ; CHECK: clei_s_v4i32: - %1 = load <4 x i32>* %a + %1 = load <4 x i32>, <4 x i32>* %a ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5) %2 = icmp sle <4 x i32> %1, <i32 1, i32 1, i32 1, i32 1> %3 = sext <4 x i1> %2 to <4 x i32> @@ -534,7 +534,7 @@ define void @clei_s_v4i32(<4 x i32>* %c, <4 x i32>* %a) nounwind { define void @clei_s_v2i64(<2 x i64>* %c, <2 x i64>* %a) nounwind { ; CHECK: clei_s_v2i64: - %1 = load <2 x i64>* %a + %1 = load <2 x i64>, <2 x i64>* %a ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5) %2 = icmp sle <2 x i64> %1, <i64 1, i64 1> %3 = sext <2 x i1> %2 to <2 x i64> @@ -549,7 +549,7 @@ define void @clei_s_v2i64(<2 x i64>* %c, <2 x i64>* %a) nounwind { define void @clei_u_v16i8(<16 x i8>* %c, <16 x i8>* %a) nounwind { ; CHECK: clei_u_v16i8: - %1 = load <16 x i8>* %a + %1 = load <16 x i8>, <16 x i8>* %a ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5) %2 = icmp ule <16 x i8> %1, <i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1> %3 = sext <16 x i1> %2 to <16 x i8> @@ -564,7 +564,7 @@ define void @clei_u_v16i8(<16 x i8>* %c, <16 x i8>* %a) nounwind { define void @clei_u_v8i16(<8 x i16>* %c, <8 x i16>* %a) nounwind { ; CHECK: clei_u_v8i16: - %1 = load <8 x i16>* %a + %1 = load <8 x i16>, <8 x i16>* %a ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5) %2 = icmp ule <8 x i16> %1, <i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1> %3 = sext <8 x i1> %2 to <8 x i16> @@ -579,7 +579,7 @@ define void @clei_u_v8i16(<8 x i16>* %c, <8 x i16>* %a) nounwind { define void @clei_u_v4i32(<4 x i32>* %c, <4 x i32>* %a) nounwind { ; CHECK: clei_u_v4i32: - %1 = load <4 x i32>* %a + %1 = load <4 x i32>, <4 x i32>* %a ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5) %2 = icmp ule <4 x i32> %1, <i32 1, i32 1, i32 1, i32 1> %3 = sext <4 x i1> %2 to <4 x i32> @@ -594,7 +594,7 @@ define void @clei_u_v4i32(<4 x i32>* %c, <4 x i32>* %a) nounwind { define void @clei_u_v2i64(<2 x i64>* %c, <2 x i64>* %a) nounwind { ; CHECK: clei_u_v2i64: - %1 = load <2 x i64>* %a + %1 = load <2 x i64>, <2 x i64>* %a ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5) %2 = icmp ule <2 x i64> %1, <i64 1, i64 1> %3 = sext <2 x i1> %2 to <2 x i64> @@ -609,7 +609,7 @@ define void @clei_u_v2i64(<2 x i64>* %c, <2 x i64>* %a) nounwind { define void @clti_s_v16i8(<16 x i8>* %c, <16 x i8>* %a) nounwind { ; CHECK: clti_s_v16i8: - %1 = load <16 x i8>* %a + %1 = load <16 x i8>, <16 x i8>* %a ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5) %2 = icmp slt <16 x i8> %1, <i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1> %3 = sext <16 x i1> %2 to <16 x i8> @@ -624,7 +624,7 @@ define void @clti_s_v16i8(<16 x i8>* %c, <16 x i8>* %a) nounwind { define void @clti_s_v8i16(<8 x i16>* %c, <8 x i16>* %a) nounwind { ; CHECK: clti_s_v8i16: - %1 = load <8 x i16>* %a + %1 = load <8 x i16>, <8 x i16>* %a ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5) %2 = icmp slt <8 x i16> %1, <i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1> %3 = sext <8 x i1> %2 to <8 x i16> @@ -639,7 +639,7 @@ define void @clti_s_v8i16(<8 x i16>* %c, <8 x i16>* %a) nounwind { define void @clti_s_v4i32(<4 x i32>* %c, <4 x i32>* %a) nounwind { ; CHECK: clti_s_v4i32: - %1 = load <4 x i32>* %a + %1 = load <4 x i32>, <4 x i32>* %a ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5) %2 = icmp slt <4 x i32> %1, <i32 1, i32 1, i32 1, i32 1> %3 = sext <4 x i1> %2 to <4 x i32> @@ -654,7 +654,7 @@ define void @clti_s_v4i32(<4 x i32>* %c, <4 x i32>* %a) nounwind { define void @clti_s_v2i64(<2 x i64>* %c, <2 x i64>* %a) nounwind { ; CHECK: clti_s_v2i64: - %1 = load <2 x i64>* %a + %1 = load <2 x i64>, <2 x i64>* %a ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5) %2 = icmp slt <2 x i64> %1, <i64 1, i64 1> %3 = sext <2 x i1> %2 to <2 x i64> @@ -669,7 +669,7 @@ define void @clti_s_v2i64(<2 x i64>* %c, <2 x i64>* %a) nounwind { define void @clti_u_v16i8(<16 x i8>* %c, <16 x i8>* %a) nounwind { ; CHECK: clti_u_v16i8: - %1 = load <16 x i8>* %a + %1 = load <16 x i8>, <16 x i8>* %a ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5) %2 = icmp ult <16 x i8> %1, <i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1> %3 = sext <16 x i1> %2 to <16 x i8> @@ -684,7 +684,7 @@ define void @clti_u_v16i8(<16 x i8>* %c, <16 x i8>* %a) nounwind { define void @clti_u_v8i16(<8 x i16>* %c, <8 x i16>* %a) nounwind { ; CHECK: clti_u_v8i16: - %1 = load <8 x i16>* %a + %1 = load <8 x i16>, <8 x i16>* %a ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5) %2 = icmp ult <8 x i16> %1, <i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1> %3 = sext <8 x i1> %2 to <8 x i16> @@ -699,7 +699,7 @@ define void @clti_u_v8i16(<8 x i16>* %c, <8 x i16>* %a) nounwind { define void @clti_u_v4i32(<4 x i32>* %c, <4 x i32>* %a) nounwind { ; CHECK: clti_u_v4i32: - %1 = load <4 x i32>* %a + %1 = load <4 x i32>, <4 x i32>* %a ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5) %2 = icmp ult <4 x i32> %1, <i32 1, i32 1, i32 1, i32 1> %3 = sext <4 x i1> %2 to <4 x i32> @@ -714,7 +714,7 @@ define void @clti_u_v4i32(<4 x i32>* %c, <4 x i32>* %a) nounwind { define void @clti_u_v2i64(<2 x i64>* %c, <2 x i64>* %a) nounwind { ; CHECK: clti_u_v2i64: - %1 = load <2 x i64>* %a + %1 = load <2 x i64>, <2 x i64>* %a ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5) %2 = icmp ult <2 x i64> %1, <i64 1, i64 1> %3 = sext <2 x i1> %2 to <2 x i64> @@ -730,11 +730,11 @@ define void @bsel_s_v16i8(<16 x i8>* %d, <16 x i8>* %a, <16 x i8>* %b, <16 x i8>* %c) nounwind { ; CHECK: bsel_s_v16i8: - %1 = load <16 x i8>* %a + %1 = load <16 x i8>, <16 x i8>* %a ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5) - %2 = load <16 x i8>* %b + %2 = load <16 x i8>, <16 x i8>* %b ; CHECK-DAG: ld.b [[R2:\$w[0-9]+]], 0($6) - %3 = load <16 x i8>* %c + %3 = load <16 x i8>, <16 x i8>* %c ; CHECK-DAG: ld.b [[R3:\$w[0-9]+]], 0($7) %4 = icmp sgt <16 x i8> %1, %2 ; CHECK-DAG: clt_s.b [[R4:\$w[0-9]+]], [[R2]], [[R1]] @@ -752,11 +752,11 @@ define void @bsel_s_v8i16(<8 x i16>* %d, <8 x i16>* %a, <8 x i16>* %b, <8 x i16>* %c) nounwind { ; CHECK: bsel_s_v8i16: - %1 = load <8 x i16>* %a + %1 = load <8 x i16>, <8 x i16>* %a ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5) - %2 = load <8 x i16>* %b + %2 = load <8 x i16>, <8 x i16>* %b ; CHECK-DAG: ld.h [[R2:\$w[0-9]+]], 0($6) - %3 = load <8 x i16>* %c + %3 = load <8 x i16>, <8 x i16>* %c ; CHECK-DAG: ld.h [[R3:\$w[0-9]+]], 0($7) %4 = icmp sgt <8 x i16> %1, %2 ; CHECK-DAG: clt_s.h [[R4:\$w[0-9]+]], [[R2]], [[R1]] @@ -774,11 +774,11 @@ define void @bsel_s_v4i32(<4 x i32>* %d, <4 x i32>* %a, <4 x i32>* %b, <4 x i32>* %c) nounwind { ; CHECK: bsel_s_v4i32: - %1 = load <4 x i32>* %a + %1 = load <4 x i32>, <4 x i32>* %a ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5) - %2 = load <4 x i32>* %b + %2 = load <4 x i32>, <4 x i32>* %b ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6) - %3 = load <4 x i32>* %c + %3 = load <4 x i32>, <4 x i32>* %c ; CHECK-DAG: ld.w [[R3:\$w[0-9]+]], 0($7) %4 = icmp sgt <4 x i32> %1, %2 ; CHECK-DAG: clt_s.w [[R4:\$w[0-9]+]], [[R2]], [[R1]] @@ -796,11 +796,11 @@ define void @bsel_s_v2i64(<2 x i64>* %d, <2 x i64>* %a, <2 x i64>* %b, <2 x i64>* %c) nounwind { ; CHECK: bsel_s_v2i64: - %1 = load <2 x i64>* %a + %1 = load <2 x i64>, <2 x i64>* %a ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5) - %2 = load <2 x i64>* %b + %2 = load <2 x i64>, <2 x i64>* %b ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6) - %3 = load <2 x i64>* %c + %3 = load <2 x i64>, <2 x i64>* %c ; CHECK-DAG: ld.d [[R3:\$w[0-9]+]], 0($7) %4 = icmp sgt <2 x i64> %1, %2 ; CHECK-DAG: clt_s.d [[R4:\$w[0-9]+]], [[R2]], [[R1]] @@ -818,11 +818,11 @@ define void @bsel_u_v16i8(<16 x i8>* %d, <16 x i8>* %a, <16 x i8>* %b, <16 x i8>* %c) nounwind { ; CHECK: bsel_u_v16i8: - %1 = load <16 x i8>* %a + %1 = load <16 x i8>, <16 x i8>* %a ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5) - %2 = load <16 x i8>* %b + %2 = load <16 x i8>, <16 x i8>* %b ; CHECK-DAG: ld.b [[R2:\$w[0-9]+]], 0($6) - %3 = load <16 x i8>* %c + %3 = load <16 x i8>, <16 x i8>* %c ; CHECK-DAG: ld.b [[R3:\$w[0-9]+]], 0($7) %4 = icmp ugt <16 x i8> %1, %2 ; CHECK-DAG: clt_u.b [[R4:\$w[0-9]+]], [[R2]], [[R1]] @@ -840,11 +840,11 @@ define void @bsel_u_v8i16(<8 x i16>* %d, <8 x i16>* %a, <8 x i16>* %b, <8 x i16>* %c) nounwind { ; CHECK: bsel_u_v8i16: - %1 = load <8 x i16>* %a + %1 = load <8 x i16>, <8 x i16>* %a ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5) - %2 = load <8 x i16>* %b + %2 = load <8 x i16>, <8 x i16>* %b ; CHECK-DAG: ld.h [[R2:\$w[0-9]+]], 0($6) - %3 = load <8 x i16>* %c + %3 = load <8 x i16>, <8 x i16>* %c ; CHECK-DAG: ld.h [[R3:\$w[0-9]+]], 0($7) %4 = icmp ugt <8 x i16> %1, %2 ; CHECK-DAG: clt_u.h [[R4:\$w[0-9]+]], [[R2]], [[R1]] @@ -862,11 +862,11 @@ define void @bsel_u_v4i32(<4 x i32>* %d, <4 x i32>* %a, <4 x i32>* %b, <4 x i32>* %c) nounwind { ; CHECK: bsel_u_v4i32: - %1 = load <4 x i32>* %a + %1 = load <4 x i32>, <4 x i32>* %a ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5) - %2 = load <4 x i32>* %b + %2 = load <4 x i32>, <4 x i32>* %b ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6) - %3 = load <4 x i32>* %c + %3 = load <4 x i32>, <4 x i32>* %c ; CHECK-DAG: ld.w [[R3:\$w[0-9]+]], 0($7) %4 = icmp ugt <4 x i32> %1, %2 ; CHECK-DAG: clt_u.w [[R4:\$w[0-9]+]], [[R2]], [[R1]] @@ -884,11 +884,11 @@ define void @bsel_u_v2i64(<2 x i64>* %d, <2 x i64>* %a, <2 x i64>* %b, <2 x i64>* %c) nounwind { ; CHECK: bsel_u_v2i64: - %1 = load <2 x i64>* %a + %1 = load <2 x i64>, <2 x i64>* %a ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5) - %2 = load <2 x i64>* %b + %2 = load <2 x i64>, <2 x i64>* %b ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6) - %3 = load <2 x i64>* %c + %3 = load <2 x i64>, <2 x i64>* %c ; CHECK-DAG: ld.d [[R3:\$w[0-9]+]], 0($7) %4 = icmp ugt <2 x i64> %1, %2 ; CHECK-DAG: clt_u.d [[R4:\$w[0-9]+]], [[R2]], [[R1]] @@ -906,9 +906,9 @@ define void @bseli_s_v16i8(<16 x i8>* %d, <16 x i8>* %a, <16 x i8>* %b, <16 x i8>* %c) nounwind { ; CHECK: bseli_s_v16i8: - %1 = load <16 x i8>* %a + %1 = load <16 x i8>, <16 x i8>* %a ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5) - %2 = load <16 x i8>* %b + %2 = load <16 x i8>, <16 x i8>* %b ; CHECK-DAG: ld.b [[R2:\$w[0-9]+]], 0($6) %3 = icmp sgt <16 x i8> %1, %2 ; CHECK-DAG: clt_s.b [[R4:\$w[0-9]+]], [[R2]], [[R1]] @@ -925,9 +925,9 @@ define void @bseli_s_v8i16(<8 x i16>* %d, <8 x i16>* %a, <8 x i16>* %b, <8 x i16>* %c) nounwind { ; CHECK: bseli_s_v8i16: - %1 = load <8 x i16>* %a + %1 = load <8 x i16>, <8 x i16>* %a ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5) - %2 = load <8 x i16>* %b + %2 = load <8 x i16>, <8 x i16>* %b ; CHECK-DAG: ld.h [[R2:\$w[0-9]+]], 0($6) %3 = icmp sgt <8 x i16> %1, %2 ; CHECK-DAG: clt_s.h [[R4:\$w[0-9]+]], [[R2]], [[R1]] @@ -945,9 +945,9 @@ define void @bseli_s_v4i32(<4 x i32>* %d, <4 x i32>* %a, <4 x i32>* %b, <4 x i32>* %c) nounwind { ; CHECK: bseli_s_v4i32: - %1 = load <4 x i32>* %a + %1 = load <4 x i32>, <4 x i32>* %a ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5) - %2 = load <4 x i32>* %b + %2 = load <4 x i32>, <4 x i32>* %b ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6) %3 = icmp sgt <4 x i32> %1, %2 ; CHECK-DAG: clt_s.w [[R4:\$w[0-9]+]], [[R2]], [[R1]] @@ -965,9 +965,9 @@ define void @bseli_s_v2i64(<2 x i64>* %d, <2 x i64>* %a, <2 x i64>* %b, <2 x i64>* %c) nounwind { ; CHECK: bseli_s_v2i64: - %1 = load <2 x i64>* %a + %1 = load <2 x i64>, <2 x i64>* %a ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5) - %2 = load <2 x i64>* %b + %2 = load <2 x i64>, <2 x i64>* %b ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6) %3 = icmp sgt <2 x i64> %1, %2 ; CHECK-DAG: clt_s.d [[R4:\$w[0-9]+]], [[R2]], [[R1]] @@ -985,9 +985,9 @@ define void @bseli_u_v16i8(<16 x i8>* %d, <16 x i8>* %a, <16 x i8>* %b, <16 x i8>* %c) nounwind { ; CHECK: bseli_u_v16i8: - %1 = load <16 x i8>* %a + %1 = load <16 x i8>, <16 x i8>* %a ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5) - %2 = load <16 x i8>* %b + %2 = load <16 x i8>, <16 x i8>* %b ; CHECK-DAG: ld.b [[R2:\$w[0-9]+]], 0($6) %3 = icmp ugt <16 x i8> %1, %2 ; CHECK-DAG: clt_u.b [[R4:\$w[0-9]+]], [[R2]], [[R1]] @@ -1004,9 +1004,9 @@ define void @bseli_u_v8i16(<8 x i16>* %d, <8 x i16>* %a, <8 x i16>* %b, <8 x i16>* %c) nounwind { ; CHECK: bseli_u_v8i16: - %1 = load <8 x i16>* %a + %1 = load <8 x i16>, <8 x i16>* %a ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5) - %2 = load <8 x i16>* %b + %2 = load <8 x i16>, <8 x i16>* %b ; CHECK-DAG: ld.h [[R2:\$w[0-9]+]], 0($6) %3 = icmp ugt <8 x i16> %1, %2 ; CHECK-DAG: clt_u.h [[R4:\$w[0-9]+]], [[R2]], [[R1]] @@ -1024,9 +1024,9 @@ define void @bseli_u_v4i32(<4 x i32>* %d, <4 x i32>* %a, <4 x i32>* %b, <4 x i32>* %c) nounwind { ; CHECK: bseli_u_v4i32: - %1 = load <4 x i32>* %a + %1 = load <4 x i32>, <4 x i32>* %a ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5) - %2 = load <4 x i32>* %b + %2 = load <4 x i32>, <4 x i32>* %b ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6) %3 = icmp ugt <4 x i32> %1, %2 ; CHECK-DAG: clt_u.w [[R4:\$w[0-9]+]], [[R2]], [[R1]] @@ -1044,9 +1044,9 @@ define void @bseli_u_v2i64(<2 x i64>* %d, <2 x i64>* %a, <2 x i64>* %b, <2 x i64>* %c) nounwind { ; CHECK: bseli_u_v2i64: - %1 = load <2 x i64>* %a + %1 = load <2 x i64>, <2 x i64>* %a ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5) - %2 = load <2 x i64>* %b + %2 = load <2 x i64>, <2 x i64>* %b ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6) %3 = icmp ugt <2 x i64> %1, %2 ; CHECK-DAG: clt_u.d [[R4:\$w[0-9]+]], [[R2]], [[R1]] @@ -1063,9 +1063,9 @@ define void @bseli_u_v2i64(<2 x i64>* %d, <2 x i64>* %a, <2 x i64>* %b, define void @max_s_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind { ; CHECK: max_s_v16i8: - %1 = load <16 x i8>* %a + %1 = load <16 x i8>, <16 x i8>* %a ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5) - %2 = load <16 x i8>* %b + %2 = load <16 x i8>, <16 x i8>* %b ; CHECK-DAG: ld.b [[R2:\$w[0-9]+]], 0($6) %3 = icmp sgt <16 x i8> %1, %2 %4 = select <16 x i1> %3, <16 x i8> %1, <16 x i8> %2 @@ -1080,9 +1080,9 @@ define void @max_s_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind { define void @max_s_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind { ; CHECK: max_s_v8i16: - %1 = load <8 x i16>* %a + %1 = load <8 x i16>, <8 x i16>* %a ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5) - %2 = load <8 x i16>* %b + %2 = load <8 x i16>, <8 x i16>* %b ; CHECK-DAG: ld.h [[R2:\$w[0-9]+]], 0($6) %3 = icmp sgt <8 x i16> %1, %2 %4 = select <8 x i1> %3, <8 x i16> %1, <8 x i16> %2 @@ -1097,9 +1097,9 @@ define void @max_s_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind { define void @max_s_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind { ; CHECK: max_s_v4i32: - %1 = load <4 x i32>* %a + %1 = load <4 x i32>, <4 x i32>* %a ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5) - %2 = load <4 x i32>* %b + %2 = load <4 x i32>, <4 x i32>* %b ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6) %3 = icmp sgt <4 x i32> %1, %2 %4 = select <4 x i1> %3, <4 x i32> %1, <4 x i32> %2 @@ -1114,9 +1114,9 @@ define void @max_s_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind { define void @max_s_v2i64(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind { ; CHECK: max_s_v2i64: - %1 = load <2 x i64>* %a + %1 = load <2 x i64>, <2 x i64>* %a ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5) - %2 = load <2 x i64>* %b + %2 = load <2 x i64>, <2 x i64>* %b ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6) %3 = icmp sgt <2 x i64> %1, %2 %4 = select <2 x i1> %3, <2 x i64> %1, <2 x i64> %2 @@ -1131,9 +1131,9 @@ define void @max_s_v2i64(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind { define void @max_u_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind { ; CHECK: max_u_v16i8: - %1 = load <16 x i8>* %a + %1 = load <16 x i8>, <16 x i8>* %a ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5) - %2 = load <16 x i8>* %b + %2 = load <16 x i8>, <16 x i8>* %b ; CHECK-DAG: ld.b [[R2:\$w[0-9]+]], 0($6) %3 = icmp ugt <16 x i8> %1, %2 %4 = select <16 x i1> %3, <16 x i8> %1, <16 x i8> %2 @@ -1148,9 +1148,9 @@ define void @max_u_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind { define void @max_u_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind { ; CHECK: max_u_v8i16: - %1 = load <8 x i16>* %a + %1 = load <8 x i16>, <8 x i16>* %a ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5) - %2 = load <8 x i16>* %b + %2 = load <8 x i16>, <8 x i16>* %b ; CHECK-DAG: ld.h [[R2:\$w[0-9]+]], 0($6) %3 = icmp ugt <8 x i16> %1, %2 %4 = select <8 x i1> %3, <8 x i16> %1, <8 x i16> %2 @@ -1165,9 +1165,9 @@ define void @max_u_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind { define void @max_u_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind { ; CHECK: max_u_v4i32: - %1 = load <4 x i32>* %a + %1 = load <4 x i32>, <4 x i32>* %a ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5) - %2 = load <4 x i32>* %b + %2 = load <4 x i32>, <4 x i32>* %b ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6) %3 = icmp ugt <4 x i32> %1, %2 %4 = select <4 x i1> %3, <4 x i32> %1, <4 x i32> %2 @@ -1182,9 +1182,9 @@ define void @max_u_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind { define void @max_u_v2i64(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind { ; CHECK: max_u_v2i64: - %1 = load <2 x i64>* %a + %1 = load <2 x i64>, <2 x i64>* %a ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5) - %2 = load <2 x i64>* %b + %2 = load <2 x i64>, <2 x i64>* %b ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6) %3 = icmp ugt <2 x i64> %1, %2 %4 = select <2 x i1> %3, <2 x i64> %1, <2 x i64> %2 @@ -1199,9 +1199,9 @@ define void @max_u_v2i64(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind { define void @max_s_eq_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind { ; CHECK: max_s_eq_v16i8: - %1 = load <16 x i8>* %a + %1 = load <16 x i8>, <16 x i8>* %a ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5) - %2 = load <16 x i8>* %b + %2 = load <16 x i8>, <16 x i8>* %b ; CHECK-DAG: ld.b [[R2:\$w[0-9]+]], 0($6) %3 = icmp sge <16 x i8> %1, %2 %4 = select <16 x i1> %3, <16 x i8> %1, <16 x i8> %2 @@ -1216,9 +1216,9 @@ define void @max_s_eq_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwin define void @max_s_eq_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind { ; CHECK: max_s_eq_v8i16: - %1 = load <8 x i16>* %a + %1 = load <8 x i16>, <8 x i16>* %a ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5) - %2 = load <8 x i16>* %b + %2 = load <8 x i16>, <8 x i16>* %b ; CHECK-DAG: ld.h [[R2:\$w[0-9]+]], 0($6) %3 = icmp sge <8 x i16> %1, %2 %4 = select <8 x i1> %3, <8 x i16> %1, <8 x i16> %2 @@ -1233,9 +1233,9 @@ define void @max_s_eq_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwin define void @max_s_eq_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind { ; CHECK: max_s_eq_v4i32: - %1 = load <4 x i32>* %a + %1 = load <4 x i32>, <4 x i32>* %a ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5) - %2 = load <4 x i32>* %b + %2 = load <4 x i32>, <4 x i32>* %b ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6) %3 = icmp sge <4 x i32> %1, %2 %4 = select <4 x i1> %3, <4 x i32> %1, <4 x i32> %2 @@ -1250,9 +1250,9 @@ define void @max_s_eq_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwin define void @max_s_eq_v2i64(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind { ; CHECK: max_s_eq_v2i64: - %1 = load <2 x i64>* %a + %1 = load <2 x i64>, <2 x i64>* %a ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5) - %2 = load <2 x i64>* %b + %2 = load <2 x i64>, <2 x i64>* %b ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6) %3 = icmp sge <2 x i64> %1, %2 %4 = select <2 x i1> %3, <2 x i64> %1, <2 x i64> %2 @@ -1267,9 +1267,9 @@ define void @max_s_eq_v2i64(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwin define void @max_u_eq_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind { ; CHECK: max_u_eq_v16i8: - %1 = load <16 x i8>* %a + %1 = load <16 x i8>, <16 x i8>* %a ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5) - %2 = load <16 x i8>* %b + %2 = load <16 x i8>, <16 x i8>* %b ; CHECK-DAG: ld.b [[R2:\$w[0-9]+]], 0($6) %3 = icmp uge <16 x i8> %1, %2 %4 = select <16 x i1> %3, <16 x i8> %1, <16 x i8> %2 @@ -1284,9 +1284,9 @@ define void @max_u_eq_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwin define void @max_u_eq_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind { ; CHECK: max_u_eq_v8i16: - %1 = load <8 x i16>* %a + %1 = load <8 x i16>, <8 x i16>* %a ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5) - %2 = load <8 x i16>* %b + %2 = load <8 x i16>, <8 x i16>* %b ; CHECK-DAG: ld.h [[R2:\$w[0-9]+]], 0($6) %3 = icmp uge <8 x i16> %1, %2 %4 = select <8 x i1> %3, <8 x i16> %1, <8 x i16> %2 @@ -1301,9 +1301,9 @@ define void @max_u_eq_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwin define void @max_u_eq_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind { ; CHECK: max_u_eq_v4i32: - %1 = load <4 x i32>* %a + %1 = load <4 x i32>, <4 x i32>* %a ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5) - %2 = load <4 x i32>* %b + %2 = load <4 x i32>, <4 x i32>* %b ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6) %3 = icmp uge <4 x i32> %1, %2 %4 = select <4 x i1> %3, <4 x i32> %1, <4 x i32> %2 @@ -1318,9 +1318,9 @@ define void @max_u_eq_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwin define void @max_u_eq_v2i64(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind { ; CHECK: max_u_eq_v2i64: - %1 = load <2 x i64>* %a + %1 = load <2 x i64>, <2 x i64>* %a ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5) - %2 = load <2 x i64>* %b + %2 = load <2 x i64>, <2 x i64>* %b ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6) %3 = icmp uge <2 x i64> %1, %2 %4 = select <2 x i1> %3, <2 x i64> %1, <2 x i64> %2 @@ -1335,7 +1335,7 @@ define void @max_u_eq_v2i64(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwin define void @maxi_s_v16i8(<16 x i8>* %c, <16 x i8>* %a) nounwind { ; CHECK: maxi_s_v16i8: - %1 = load <16 x i8>* %a + %1 = load <16 x i8>, <16 x i8>* %a ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5) %2 = icmp sgt <16 x i8> %1, <i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1> %3 = select <16 x i1> %2, <16 x i8> %1, <16 x i8> <i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1> @@ -1350,7 +1350,7 @@ define void @maxi_s_v16i8(<16 x i8>* %c, <16 x i8>* %a) nounwind { define void @maxi_s_v8i16(<8 x i16>* %c, <8 x i16>* %a) nounwind { ; CHECK: maxi_s_v8i16: - %1 = load <8 x i16>* %a + %1 = load <8 x i16>, <8 x i16>* %a ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5) %2 = icmp sgt <8 x i16> %1, <i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1> %3 = select <8 x i1> %2, <8 x i16> %1, <8 x i16> <i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1> @@ -1365,7 +1365,7 @@ define void @maxi_s_v8i16(<8 x i16>* %c, <8 x i16>* %a) nounwind { define void @maxi_s_v4i32(<4 x i32>* %c, <4 x i32>* %a) nounwind { ; CHECK: maxi_s_v4i32: - %1 = load <4 x i32>* %a + %1 = load <4 x i32>, <4 x i32>* %a ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5) %2 = icmp sgt <4 x i32> %1, <i32 1, i32 1, i32 1, i32 1> %3 = select <4 x i1> %2, <4 x i32> %1, <4 x i32> <i32 1, i32 1, i32 1, i32 1> @@ -1380,7 +1380,7 @@ define void @maxi_s_v4i32(<4 x i32>* %c, <4 x i32>* %a) nounwind { define void @maxi_s_v2i64(<2 x i64>* %c, <2 x i64>* %a) nounwind { ; CHECK: maxi_s_v2i64: - %1 = load <2 x i64>* %a + %1 = load <2 x i64>, <2 x i64>* %a ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5) %2 = icmp sgt <2 x i64> %1, <i64 1, i64 1> %3 = select <2 x i1> %2, <2 x i64> %1, <2 x i64> <i64 1, i64 1> @@ -1395,7 +1395,7 @@ define void @maxi_s_v2i64(<2 x i64>* %c, <2 x i64>* %a) nounwind { define void @maxi_u_v16i8(<16 x i8>* %c, <16 x i8>* %a) nounwind { ; CHECK: maxi_u_v16i8: - %1 = load <16 x i8>* %a + %1 = load <16 x i8>, <16 x i8>* %a ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5) %2 = icmp ugt <16 x i8> %1, <i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1> %3 = select <16 x i1> %2, <16 x i8> %1, <16 x i8> <i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1> @@ -1410,7 +1410,7 @@ define void @maxi_u_v16i8(<16 x i8>* %c, <16 x i8>* %a) nounwind { define void @maxi_u_v8i16(<8 x i16>* %c, <8 x i16>* %a) nounwind { ; CHECK: maxi_u_v8i16: - %1 = load <8 x i16>* %a + %1 = load <8 x i16>, <8 x i16>* %a ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5) %2 = icmp ugt <8 x i16> %1, <i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1> %3 = select <8 x i1> %2, <8 x i16> %1, <8 x i16> <i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1> @@ -1425,7 +1425,7 @@ define void @maxi_u_v8i16(<8 x i16>* %c, <8 x i16>* %a) nounwind { define void @maxi_u_v4i32(<4 x i32>* %c, <4 x i32>* %a) nounwind { ; CHECK: maxi_u_v4i32: - %1 = load <4 x i32>* %a + %1 = load <4 x i32>, <4 x i32>* %a ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5) %2 = icmp ugt <4 x i32> %1, <i32 1, i32 1, i32 1, i32 1> %3 = select <4 x i1> %2, <4 x i32> %1, <4 x i32> <i32 1, i32 1, i32 1, i32 1> @@ -1440,7 +1440,7 @@ define void @maxi_u_v4i32(<4 x i32>* %c, <4 x i32>* %a) nounwind { define void @maxi_u_v2i64(<2 x i64>* %c, <2 x i64>* %a) nounwind { ; CHECK: maxi_u_v2i64: - %1 = load <2 x i64>* %a + %1 = load <2 x i64>, <2 x i64>* %a ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5) %2 = icmp ugt <2 x i64> %1, <i64 1, i64 1> %3 = select <2 x i1> %2, <2 x i64> %1, <2 x i64> <i64 1, i64 1> @@ -1455,7 +1455,7 @@ define void @maxi_u_v2i64(<2 x i64>* %c, <2 x i64>* %a) nounwind { define void @maxi_s_eq_v16i8(<16 x i8>* %c, <16 x i8>* %a) nounwind { ; CHECK: maxi_s_eq_v16i8: - %1 = load <16 x i8>* %a + %1 = load <16 x i8>, <16 x i8>* %a ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5) %2 = icmp sge <16 x i8> %1, <i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1> %3 = select <16 x i1> %2, <16 x i8> %1, <16 x i8> <i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1> @@ -1470,7 +1470,7 @@ define void @maxi_s_eq_v16i8(<16 x i8>* %c, <16 x i8>* %a) nounwind { define void @maxi_s_eq_v8i16(<8 x i16>* %c, <8 x i16>* %a) nounwind { ; CHECK: maxi_s_eq_v8i16: - %1 = load <8 x i16>* %a + %1 = load <8 x i16>, <8 x i16>* %a ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5) %2 = icmp sge <8 x i16> %1, <i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1> %3 = select <8 x i1> %2, <8 x i16> %1, <8 x i16> <i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1> @@ -1485,7 +1485,7 @@ define void @maxi_s_eq_v8i16(<8 x i16>* %c, <8 x i16>* %a) nounwind { define void @maxi_s_eq_v4i32(<4 x i32>* %c, <4 x i32>* %a) nounwind { ; CHECK: maxi_s_eq_v4i32: - %1 = load <4 x i32>* %a + %1 = load <4 x i32>, <4 x i32>* %a ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5) %2 = icmp sge <4 x i32> %1, <i32 1, i32 1, i32 1, i32 1> %3 = select <4 x i1> %2, <4 x i32> %1, <4 x i32> <i32 1, i32 1, i32 1, i32 1> @@ -1500,7 +1500,7 @@ define void @maxi_s_eq_v4i32(<4 x i32>* %c, <4 x i32>* %a) nounwind { define void @maxi_s_eq_v2i64(<2 x i64>* %c, <2 x i64>* %a) nounwind { ; CHECK: maxi_s_eq_v2i64: - %1 = load <2 x i64>* %a + %1 = load <2 x i64>, <2 x i64>* %a ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5) %2 = icmp sge <2 x i64> %1, <i64 1, i64 1> %3 = select <2 x i1> %2, <2 x i64> %1, <2 x i64> <i64 1, i64 1> @@ -1515,7 +1515,7 @@ define void @maxi_s_eq_v2i64(<2 x i64>* %c, <2 x i64>* %a) nounwind { define void @maxi_u_eq_v16i8(<16 x i8>* %c, <16 x i8>* %a) nounwind { ; CHECK: maxi_u_eq_v16i8: - %1 = load <16 x i8>* %a + %1 = load <16 x i8>, <16 x i8>* %a ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5) %2 = icmp uge <16 x i8> %1, <i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1> %3 = select <16 x i1> %2, <16 x i8> %1, <16 x i8> <i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1> @@ -1530,7 +1530,7 @@ define void @maxi_u_eq_v16i8(<16 x i8>* %c, <16 x i8>* %a) nounwind { define void @maxi_u_eq_v8i16(<8 x i16>* %c, <8 x i16>* %a) nounwind { ; CHECK: maxi_u_eq_v8i16: - %1 = load <8 x i16>* %a + %1 = load <8 x i16>, <8 x i16>* %a ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5) %2 = icmp uge <8 x i16> %1, <i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1> %3 = select <8 x i1> %2, <8 x i16> %1, <8 x i16> <i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1> @@ -1545,7 +1545,7 @@ define void @maxi_u_eq_v8i16(<8 x i16>* %c, <8 x i16>* %a) nounwind { define void @maxi_u_eq_v4i32(<4 x i32>* %c, <4 x i32>* %a) nounwind { ; CHECK: maxi_u_eq_v4i32: - %1 = load <4 x i32>* %a + %1 = load <4 x i32>, <4 x i32>* %a ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5) %2 = icmp uge <4 x i32> %1, <i32 1, i32 1, i32 1, i32 1> %3 = select <4 x i1> %2, <4 x i32> %1, <4 x i32> <i32 1, i32 1, i32 1, i32 1> @@ -1560,7 +1560,7 @@ define void @maxi_u_eq_v4i32(<4 x i32>* %c, <4 x i32>* %a) nounwind { define void @maxi_u_eq_v2i64(<2 x i64>* %c, <2 x i64>* %a) nounwind { ; CHECK: maxi_u_eq_v2i64: - %1 = load <2 x i64>* %a + %1 = load <2 x i64>, <2 x i64>* %a ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5) %2 = icmp uge <2 x i64> %1, <i64 1, i64 1> %3 = select <2 x i1> %2, <2 x i64> %1, <2 x i64> <i64 1, i64 1> @@ -1575,9 +1575,9 @@ define void @maxi_u_eq_v2i64(<2 x i64>* %c, <2 x i64>* %a) nounwind { define void @min_s_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind { ; CHECK: min_s_v16i8: - %1 = load <16 x i8>* %a + %1 = load <16 x i8>, <16 x i8>* %a ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5) - %2 = load <16 x i8>* %b + %2 = load <16 x i8>, <16 x i8>* %b ; CHECK-DAG: ld.b [[R2:\$w[0-9]+]], 0($6) %3 = icmp sle <16 x i8> %1, %2 %4 = select <16 x i1> %3, <16 x i8> %1, <16 x i8> %2 @@ -1592,9 +1592,9 @@ define void @min_s_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind { define void @min_s_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind { ; CHECK: min_s_v8i16: - %1 = load <8 x i16>* %a + %1 = load <8 x i16>, <8 x i16>* %a ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5) - %2 = load <8 x i16>* %b + %2 = load <8 x i16>, <8 x i16>* %b ; CHECK-DAG: ld.h [[R2:\$w[0-9]+]], 0($6) %3 = icmp slt <8 x i16> %1, %2 %4 = select <8 x i1> %3, <8 x i16> %1, <8 x i16> %2 @@ -1609,9 +1609,9 @@ define void @min_s_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind { define void @min_s_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind { ; CHECK: min_s_v4i32: - %1 = load <4 x i32>* %a + %1 = load <4 x i32>, <4 x i32>* %a ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5) - %2 = load <4 x i32>* %b + %2 = load <4 x i32>, <4 x i32>* %b ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6) %3 = icmp slt <4 x i32> %1, %2 %4 = select <4 x i1> %3, <4 x i32> %1, <4 x i32> %2 @@ -1626,9 +1626,9 @@ define void @min_s_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind { define void @min_s_v2i64(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind { ; CHECK: min_s_v2i64: - %1 = load <2 x i64>* %a + %1 = load <2 x i64>, <2 x i64>* %a ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5) - %2 = load <2 x i64>* %b + %2 = load <2 x i64>, <2 x i64>* %b ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6) %3 = icmp slt <2 x i64> %1, %2 %4 = select <2 x i1> %3, <2 x i64> %1, <2 x i64> %2 @@ -1643,9 +1643,9 @@ define void @min_s_v2i64(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind { define void @min_u_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind { ; CHECK: min_u_v16i8: - %1 = load <16 x i8>* %a + %1 = load <16 x i8>, <16 x i8>* %a ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5) - %2 = load <16 x i8>* %b + %2 = load <16 x i8>, <16 x i8>* %b ; CHECK-DAG: ld.b [[R2:\$w[0-9]+]], 0($6) %3 = icmp ult <16 x i8> %1, %2 %4 = select <16 x i1> %3, <16 x i8> %1, <16 x i8> %2 @@ -1660,9 +1660,9 @@ define void @min_u_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind { define void @min_u_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind { ; CHECK: min_u_v8i16: - %1 = load <8 x i16>* %a + %1 = load <8 x i16>, <8 x i16>* %a ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5) - %2 = load <8 x i16>* %b + %2 = load <8 x i16>, <8 x i16>* %b ; CHECK-DAG: ld.h [[R2:\$w[0-9]+]], 0($6) %3 = icmp ult <8 x i16> %1, %2 %4 = select <8 x i1> %3, <8 x i16> %1, <8 x i16> %2 @@ -1677,9 +1677,9 @@ define void @min_u_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind { define void @min_u_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind { ; CHECK: min_u_v4i32: - %1 = load <4 x i32>* %a + %1 = load <4 x i32>, <4 x i32>* %a ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5) - %2 = load <4 x i32>* %b + %2 = load <4 x i32>, <4 x i32>* %b ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6) %3 = icmp ult <4 x i32> %1, %2 %4 = select <4 x i1> %3, <4 x i32> %1, <4 x i32> %2 @@ -1694,9 +1694,9 @@ define void @min_u_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind { define void @min_u_v2i64(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind { ; CHECK: min_u_v2i64: - %1 = load <2 x i64>* %a + %1 = load <2 x i64>, <2 x i64>* %a ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5) - %2 = load <2 x i64>* %b + %2 = load <2 x i64>, <2 x i64>* %b ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6) %3 = icmp ult <2 x i64> %1, %2 %4 = select <2 x i1> %3, <2 x i64> %1, <2 x i64> %2 @@ -1711,9 +1711,9 @@ define void @min_u_v2i64(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind { define void @min_s_eq_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind { ; CHECK: min_s_eq_v16i8: - %1 = load <16 x i8>* %a + %1 = load <16 x i8>, <16 x i8>* %a ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5) - %2 = load <16 x i8>* %b + %2 = load <16 x i8>, <16 x i8>* %b ; CHECK-DAG: ld.b [[R2:\$w[0-9]+]], 0($6) %3 = icmp sle <16 x i8> %1, %2 %4 = select <16 x i1> %3, <16 x i8> %1, <16 x i8> %2 @@ -1728,9 +1728,9 @@ define void @min_s_eq_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwin define void @min_s_eq_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind { ; CHECK: min_s_eq_v8i16: - %1 = load <8 x i16>* %a + %1 = load <8 x i16>, <8 x i16>* %a ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5) - %2 = load <8 x i16>* %b + %2 = load <8 x i16>, <8 x i16>* %b ; CHECK-DAG: ld.h [[R2:\$w[0-9]+]], 0($6) %3 = icmp sle <8 x i16> %1, %2 %4 = select <8 x i1> %3, <8 x i16> %1, <8 x i16> %2 @@ -1745,9 +1745,9 @@ define void @min_s_eq_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwin define void @min_s_eq_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind { ; CHECK: min_s_eq_v4i32: - %1 = load <4 x i32>* %a + %1 = load <4 x i32>, <4 x i32>* %a ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5) - %2 = load <4 x i32>* %b + %2 = load <4 x i32>, <4 x i32>* %b ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6) %3 = icmp sle <4 x i32> %1, %2 %4 = select <4 x i1> %3, <4 x i32> %1, <4 x i32> %2 @@ -1762,9 +1762,9 @@ define void @min_s_eq_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwin define void @min_s_eq_v2i64(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind { ; CHECK: min_s_eq_v2i64: - %1 = load <2 x i64>* %a + %1 = load <2 x i64>, <2 x i64>* %a ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5) - %2 = load <2 x i64>* %b + %2 = load <2 x i64>, <2 x i64>* %b ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6) %3 = icmp sle <2 x i64> %1, %2 %4 = select <2 x i1> %3, <2 x i64> %1, <2 x i64> %2 @@ -1779,9 +1779,9 @@ define void @min_s_eq_v2i64(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwin define void @min_u_eq_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind { ; CHECK: min_u_eq_v16i8: - %1 = load <16 x i8>* %a + %1 = load <16 x i8>, <16 x i8>* %a ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5) - %2 = load <16 x i8>* %b + %2 = load <16 x i8>, <16 x i8>* %b ; CHECK-DAG: ld.b [[R2:\$w[0-9]+]], 0($6) %3 = icmp ule <16 x i8> %1, %2 %4 = select <16 x i1> %3, <16 x i8> %1, <16 x i8> %2 @@ -1796,9 +1796,9 @@ define void @min_u_eq_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwin define void @min_u_eq_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind { ; CHECK: min_u_eq_v8i16: - %1 = load <8 x i16>* %a + %1 = load <8 x i16>, <8 x i16>* %a ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5) - %2 = load <8 x i16>* %b + %2 = load <8 x i16>, <8 x i16>* %b ; CHECK-DAG: ld.h [[R2:\$w[0-9]+]], 0($6) %3 = icmp ule <8 x i16> %1, %2 %4 = select <8 x i1> %3, <8 x i16> %1, <8 x i16> %2 @@ -1813,9 +1813,9 @@ define void @min_u_eq_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwin define void @min_u_eq_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind { ; CHECK: min_u_eq_v4i32: - %1 = load <4 x i32>* %a + %1 = load <4 x i32>, <4 x i32>* %a ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5) - %2 = load <4 x i32>* %b + %2 = load <4 x i32>, <4 x i32>* %b ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6) %3 = icmp ule <4 x i32> %1, %2 %4 = select <4 x i1> %3, <4 x i32> %1, <4 x i32> %2 @@ -1830,9 +1830,9 @@ define void @min_u_eq_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwin define void @min_u_eq_v2i64(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind { ; CHECK: min_u_eq_v2i64: - %1 = load <2 x i64>* %a + %1 = load <2 x i64>, <2 x i64>* %a ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5) - %2 = load <2 x i64>* %b + %2 = load <2 x i64>, <2 x i64>* %b ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6) %3 = icmp ule <2 x i64> %1, %2 %4 = select <2 x i1> %3, <2 x i64> %1, <2 x i64> %2 @@ -1847,7 +1847,7 @@ define void @min_u_eq_v2i64(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwin define void @mini_s_v16i8(<16 x i8>* %c, <16 x i8>* %a) nounwind { ; CHECK: mini_s_v16i8: - %1 = load <16 x i8>* %a + %1 = load <16 x i8>, <16 x i8>* %a ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5) %2 = icmp slt <16 x i8> %1, <i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1> %3 = select <16 x i1> %2, <16 x i8> %1, <16 x i8> <i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1> @@ -1862,7 +1862,7 @@ define void @mini_s_v16i8(<16 x i8>* %c, <16 x i8>* %a) nounwind { define void @mini_s_v8i16(<8 x i16>* %c, <8 x i16>* %a) nounwind { ; CHECK: mini_s_v8i16: - %1 = load <8 x i16>* %a + %1 = load <8 x i16>, <8 x i16>* %a ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5) %2 = icmp slt <8 x i16> %1, <i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1> %3 = select <8 x i1> %2, <8 x i16> %1, <8 x i16> <i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1> @@ -1877,7 +1877,7 @@ define void @mini_s_v8i16(<8 x i16>* %c, <8 x i16>* %a) nounwind { define void @mini_s_v4i32(<4 x i32>* %c, <4 x i32>* %a) nounwind { ; CHECK: mini_s_v4i32: - %1 = load <4 x i32>* %a + %1 = load <4 x i32>, <4 x i32>* %a ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5) %2 = icmp slt <4 x i32> %1, <i32 1, i32 1, i32 1, i32 1> %3 = select <4 x i1> %2, <4 x i32> %1, <4 x i32> <i32 1, i32 1, i32 1, i32 1> @@ -1892,7 +1892,7 @@ define void @mini_s_v4i32(<4 x i32>* %c, <4 x i32>* %a) nounwind { define void @mini_s_v2i64(<2 x i64>* %c, <2 x i64>* %a) nounwind { ; CHECK: mini_s_v2i64: - %1 = load <2 x i64>* %a + %1 = load <2 x i64>, <2 x i64>* %a ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5) %2 = icmp slt <2 x i64> %1, <i64 1, i64 1> %3 = select <2 x i1> %2, <2 x i64> %1, <2 x i64> <i64 1, i64 1> @@ -1907,7 +1907,7 @@ define void @mini_s_v2i64(<2 x i64>* %c, <2 x i64>* %a) nounwind { define void @mini_u_v16i8(<16 x i8>* %c, <16 x i8>* %a) nounwind { ; CHECK: mini_u_v16i8: - %1 = load <16 x i8>* %a + %1 = load <16 x i8>, <16 x i8>* %a ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5) %2 = icmp ult <16 x i8> %1, <i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1> %3 = select <16 x i1> %2, <16 x i8> %1, <16 x i8> <i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1> @@ -1922,7 +1922,7 @@ define void @mini_u_v16i8(<16 x i8>* %c, <16 x i8>* %a) nounwind { define void @mini_u_v8i16(<8 x i16>* %c, <8 x i16>* %a) nounwind { ; CHECK: mini_u_v8i16: - %1 = load <8 x i16>* %a + %1 = load <8 x i16>, <8 x i16>* %a ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5) %2 = icmp ult <8 x i16> %1, <i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1> %3 = select <8 x i1> %2, <8 x i16> %1, <8 x i16> <i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1> @@ -1937,7 +1937,7 @@ define void @mini_u_v8i16(<8 x i16>* %c, <8 x i16>* %a) nounwind { define void @mini_u_v4i32(<4 x i32>* %c, <4 x i32>* %a) nounwind { ; CHECK: mini_u_v4i32: - %1 = load <4 x i32>* %a + %1 = load <4 x i32>, <4 x i32>* %a ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5) %2 = icmp ult <4 x i32> %1, <i32 1, i32 1, i32 1, i32 1> %3 = select <4 x i1> %2, <4 x i32> %1, <4 x i32> <i32 1, i32 1, i32 1, i32 1> @@ -1952,7 +1952,7 @@ define void @mini_u_v4i32(<4 x i32>* %c, <4 x i32>* %a) nounwind { define void @mini_u_v2i64(<2 x i64>* %c, <2 x i64>* %a) nounwind { ; CHECK: mini_u_v2i64: - %1 = load <2 x i64>* %a + %1 = load <2 x i64>, <2 x i64>* %a ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5) %2 = icmp ult <2 x i64> %1, <i64 1, i64 1> %3 = select <2 x i1> %2, <2 x i64> %1, <2 x i64> <i64 1, i64 1> @@ -1967,7 +1967,7 @@ define void @mini_u_v2i64(<2 x i64>* %c, <2 x i64>* %a) nounwind { define void @mini_s_eq_v16i8(<16 x i8>* %c, <16 x i8>* %a) nounwind { ; CHECK: mini_s_eq_v16i8: - %1 = load <16 x i8>* %a + %1 = load <16 x i8>, <16 x i8>* %a ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5) %2 = icmp sle <16 x i8> %1, <i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1> %3 = select <16 x i1> %2, <16 x i8> %1, <16 x i8> <i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1> @@ -1982,7 +1982,7 @@ define void @mini_s_eq_v16i8(<16 x i8>* %c, <16 x i8>* %a) nounwind { define void @mini_s_eq_v8i16(<8 x i16>* %c, <8 x i16>* %a) nounwind { ; CHECK: mini_s_eq_v8i16: - %1 = load <8 x i16>* %a + %1 = load <8 x i16>, <8 x i16>* %a ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5) %2 = icmp sle <8 x i16> %1, <i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1> %3 = select <8 x i1> %2, <8 x i16> %1, <8 x i16> <i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1> @@ -1997,7 +1997,7 @@ define void @mini_s_eq_v8i16(<8 x i16>* %c, <8 x i16>* %a) nounwind { define void @mini_s_eq_v4i32(<4 x i32>* %c, <4 x i32>* %a) nounwind { ; CHECK: mini_s_eq_v4i32: - %1 = load <4 x i32>* %a + %1 = load <4 x i32>, <4 x i32>* %a ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5) %2 = icmp sle <4 x i32> %1, <i32 1, i32 1, i32 1, i32 1> %3 = select <4 x i1> %2, <4 x i32> %1, <4 x i32> <i32 1, i32 1, i32 1, i32 1> @@ -2012,7 +2012,7 @@ define void @mini_s_eq_v4i32(<4 x i32>* %c, <4 x i32>* %a) nounwind { define void @mini_s_eq_v2i64(<2 x i64>* %c, <2 x i64>* %a) nounwind { ; CHECK: mini_s_eq_v2i64: - %1 = load <2 x i64>* %a + %1 = load <2 x i64>, <2 x i64>* %a ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5) %2 = icmp sle <2 x i64> %1, <i64 1, i64 1> %3 = select <2 x i1> %2, <2 x i64> %1, <2 x i64> <i64 1, i64 1> @@ -2027,7 +2027,7 @@ define void @mini_s_eq_v2i64(<2 x i64>* %c, <2 x i64>* %a) nounwind { define void @mini_u_eq_v16i8(<16 x i8>* %c, <16 x i8>* %a) nounwind { ; CHECK: mini_u_eq_v16i8: - %1 = load <16 x i8>* %a + %1 = load <16 x i8>, <16 x i8>* %a ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5) %2 = icmp ule <16 x i8> %1, <i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1> %3 = select <16 x i1> %2, <16 x i8> %1, <16 x i8> <i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1> @@ -2042,7 +2042,7 @@ define void @mini_u_eq_v16i8(<16 x i8>* %c, <16 x i8>* %a) nounwind { define void @mini_u_eq_v8i16(<8 x i16>* %c, <8 x i16>* %a) nounwind { ; CHECK: mini_u_eq_v8i16: - %1 = load <8 x i16>* %a + %1 = load <8 x i16>, <8 x i16>* %a ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5) %2 = icmp ule <8 x i16> %1, <i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1> %3 = select <8 x i1> %2, <8 x i16> %1, <8 x i16> <i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1> @@ -2057,7 +2057,7 @@ define void @mini_u_eq_v8i16(<8 x i16>* %c, <8 x i16>* %a) nounwind { define void @mini_u_eq_v4i32(<4 x i32>* %c, <4 x i32>* %a) nounwind { ; CHECK: mini_u_eq_v4i32: - %1 = load <4 x i32>* %a + %1 = load <4 x i32>, <4 x i32>* %a ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5) %2 = icmp ule <4 x i32> %1, <i32 1, i32 1, i32 1, i32 1> %3 = select <4 x i1> %2, <4 x i32> %1, <4 x i32> <i32 1, i32 1, i32 1, i32 1> @@ -2072,7 +2072,7 @@ define void @mini_u_eq_v4i32(<4 x i32>* %c, <4 x i32>* %a) nounwind { define void @mini_u_eq_v2i64(<2 x i64>* %c, <2 x i64>* %a) nounwind { ; CHECK: mini_u_eq_v2i64: - %1 = load <2 x i64>* %a + %1 = load <2 x i64>, <2 x i64>* %a ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5) %2 = icmp ule <2 x i64> %1, <i64 1, i64 1> %3 = select <2 x i1> %2, <2 x i64> %1, <2 x i64> <i64 1, i64 1> diff --git a/test/CodeGen/Mips/msa/compare_float.ll b/test/CodeGen/Mips/msa/compare_float.ll index e93221b93612f..3229d027d95d0 100644 --- a/test/CodeGen/Mips/msa/compare_float.ll +++ b/test/CodeGen/Mips/msa/compare_float.ll @@ -9,8 +9,8 @@ declare <2 x double> @llvm.mips.fmin.d(<2 x double>, <2 x double>) nounwind define void @false_v4f32(<4 x i32>* %c, <4 x float>* %a, <4 x float>* %b) nounwind { ; CHECK: false_v4f32: - %1 = load <4 x float>* %a - %2 = load <4 x float>* %b + %1 = load <4 x float>, <4 x float>* %a + %2 = load <4 x float>, <4 x float>* %b %3 = fcmp false <4 x float> %1, %2 %4 = sext <4 x i1> %3 to <4 x i32> store <4 x i32> %4, <4 x i32>* %c @@ -25,8 +25,8 @@ define void @false_v4f32(<4 x i32>* %c, <4 x float>* %a, <4 x float>* %b) nounwi define void @false_v2f64(<2 x i64>* %c, <2 x double>* %a, <2 x double>* %b) nounwind { ; CHECK: false_v2f64: - %1 = load <2 x double>* %a - %2 = load <2 x double>* %b + %1 = load <2 x double>, <2 x double>* %a + %2 = load <2 x double>, <2 x double>* %b %3 = fcmp false <2 x double> %1, %2 %4 = sext <2 x i1> %3 to <2 x i64> store <2 x i64> %4, <2 x i64>* %c @@ -41,9 +41,9 @@ define void @false_v2f64(<2 x i64>* %c, <2 x double>* %a, <2 x double>* %b) noun define void @oeq_v4f32(<4 x i32>* %c, <4 x float>* %a, <4 x float>* %b) nounwind { ; CHECK: oeq_v4f32: - %1 = load <4 x float>* %a + %1 = load <4 x float>, <4 x float>* %a ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5) - %2 = load <4 x float>* %b + %2 = load <4 x float>, <4 x float>* %b ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6) %3 = fcmp oeq <4 x float> %1, %2 %4 = sext <4 x i1> %3 to <4 x i32> @@ -58,9 +58,9 @@ define void @oeq_v4f32(<4 x i32>* %c, <4 x float>* %a, <4 x float>* %b) nounwind define void @oeq_v2f64(<2 x i64>* %c, <2 x double>* %a, <2 x double>* %b) nounwind { ; CHECK: oeq_v2f64: - %1 = load <2 x double>* %a + %1 = load <2 x double>, <2 x double>* %a ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5) - %2 = load <2 x double>* %b + %2 = load <2 x double>, <2 x double>* %b ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6) %3 = fcmp oeq <2 x double> %1, %2 %4 = sext <2 x i1> %3 to <2 x i64> @@ -75,9 +75,9 @@ define void @oeq_v2f64(<2 x i64>* %c, <2 x double>* %a, <2 x double>* %b) nounwi define void @oge_v4f32(<4 x i32>* %c, <4 x float>* %a, <4 x float>* %b) nounwind { ; CHECK: oge_v4f32: - %1 = load <4 x float>* %a + %1 = load <4 x float>, <4 x float>* %a ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5) - %2 = load <4 x float>* %b + %2 = load <4 x float>, <4 x float>* %b ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6) %3 = fcmp oge <4 x float> %1, %2 %4 = sext <4 x i1> %3 to <4 x i32> @@ -92,9 +92,9 @@ define void @oge_v4f32(<4 x i32>* %c, <4 x float>* %a, <4 x float>* %b) nounwind define void @oge_v2f64(<2 x i64>* %c, <2 x double>* %a, <2 x double>* %b) nounwind { ; CHECK: oge_v2f64: - %1 = load <2 x double>* %a + %1 = load <2 x double>, <2 x double>* %a ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5) - %2 = load <2 x double>* %b + %2 = load <2 x double>, <2 x double>* %b ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6) %3 = fcmp oge <2 x double> %1, %2 %4 = sext <2 x i1> %3 to <2 x i64> @@ -109,9 +109,9 @@ define void @oge_v2f64(<2 x i64>* %c, <2 x double>* %a, <2 x double>* %b) nounwi define void @ogt_v4f32(<4 x i32>* %c, <4 x float>* %a, <4 x float>* %b) nounwind { ; CHECK: ogt_v4f32: - %1 = load <4 x float>* %a + %1 = load <4 x float>, <4 x float>* %a ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5) - %2 = load <4 x float>* %b + %2 = load <4 x float>, <4 x float>* %b ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6) %3 = fcmp ogt <4 x float> %1, %2 %4 = sext <4 x i1> %3 to <4 x i32> @@ -126,9 +126,9 @@ define void @ogt_v4f32(<4 x i32>* %c, <4 x float>* %a, <4 x float>* %b) nounwind define void @ogt_v2f64(<2 x i64>* %c, <2 x double>* %a, <2 x double>* %b) nounwind { ; CHECK: ogt_v2f64: - %1 = load <2 x double>* %a + %1 = load <2 x double>, <2 x double>* %a ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5) - %2 = load <2 x double>* %b + %2 = load <2 x double>, <2 x double>* %b ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6) %3 = fcmp ogt <2 x double> %1, %2 %4 = sext <2 x i1> %3 to <2 x i64> @@ -143,9 +143,9 @@ define void @ogt_v2f64(<2 x i64>* %c, <2 x double>* %a, <2 x double>* %b) nounwi define void @ole_v4f32(<4 x i32>* %c, <4 x float>* %a, <4 x float>* %b) nounwind { ; CHECK: ole_v4f32: - %1 = load <4 x float>* %a + %1 = load <4 x float>, <4 x float>* %a ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5) - %2 = load <4 x float>* %b + %2 = load <4 x float>, <4 x float>* %b ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6) %3 = fcmp ole <4 x float> %1, %2 %4 = sext <4 x i1> %3 to <4 x i32> @@ -160,9 +160,9 @@ define void @ole_v4f32(<4 x i32>* %c, <4 x float>* %a, <4 x float>* %b) nounwind define void @ole_v2f64(<2 x i64>* %c, <2 x double>* %a, <2 x double>* %b) nounwind { ; CHECK: ole_v2f64: - %1 = load <2 x double>* %a + %1 = load <2 x double>, <2 x double>* %a ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5) - %2 = load <2 x double>* %b + %2 = load <2 x double>, <2 x double>* %b ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6) %3 = fcmp ole <2 x double> %1, %2 %4 = sext <2 x i1> %3 to <2 x i64> @@ -177,9 +177,9 @@ define void @ole_v2f64(<2 x i64>* %c, <2 x double>* %a, <2 x double>* %b) nounwi define void @olt_v4f32(<4 x i32>* %c, <4 x float>* %a, <4 x float>* %b) nounwind { ; CHECK: olt_v4f32: - %1 = load <4 x float>* %a + %1 = load <4 x float>, <4 x float>* %a ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5) - %2 = load <4 x float>* %b + %2 = load <4 x float>, <4 x float>* %b ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6) %3 = fcmp olt <4 x float> %1, %2 %4 = sext <4 x i1> %3 to <4 x i32> @@ -194,9 +194,9 @@ define void @olt_v4f32(<4 x i32>* %c, <4 x float>* %a, <4 x float>* %b) nounwind define void @olt_v2f64(<2 x i64>* %c, <2 x double>* %a, <2 x double>* %b) nounwind { ; CHECK: olt_v2f64: - %1 = load <2 x double>* %a + %1 = load <2 x double>, <2 x double>* %a ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5) - %2 = load <2 x double>* %b + %2 = load <2 x double>, <2 x double>* %b ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6) %3 = fcmp olt <2 x double> %1, %2 %4 = sext <2 x i1> %3 to <2 x i64> @@ -211,9 +211,9 @@ define void @olt_v2f64(<2 x i64>* %c, <2 x double>* %a, <2 x double>* %b) nounwi define void @one_v4f32(<4 x i32>* %c, <4 x float>* %a, <4 x float>* %b) nounwind { ; CHECK: one_v4f32: - %1 = load <4 x float>* %a + %1 = load <4 x float>, <4 x float>* %a ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5) - %2 = load <4 x float>* %b + %2 = load <4 x float>, <4 x float>* %b ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6) %3 = fcmp one <4 x float> %1, %2 %4 = sext <4 x i1> %3 to <4 x i32> @@ -228,9 +228,9 @@ define void @one_v4f32(<4 x i32>* %c, <4 x float>* %a, <4 x float>* %b) nounwind define void @one_v2f64(<2 x i64>* %c, <2 x double>* %a, <2 x double>* %b) nounwind { ; CHECK: one_v2f64: - %1 = load <2 x double>* %a + %1 = load <2 x double>, <2 x double>* %a ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5) - %2 = load <2 x double>* %b + %2 = load <2 x double>, <2 x double>* %b ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6) %3 = fcmp one <2 x double> %1, %2 %4 = sext <2 x i1> %3 to <2 x i64> @@ -245,9 +245,9 @@ define void @one_v2f64(<2 x i64>* %c, <2 x double>* %a, <2 x double>* %b) nounwi define void @ord_v4f32(<4 x i32>* %c, <4 x float>* %a, <4 x float>* %b) nounwind { ; CHECK: ord_v4f32: - %1 = load <4 x float>* %a + %1 = load <4 x float>, <4 x float>* %a ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5) - %2 = load <4 x float>* %b + %2 = load <4 x float>, <4 x float>* %b ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6) %3 = fcmp ord <4 x float> %1, %2 %4 = sext <4 x i1> %3 to <4 x i32> @@ -262,9 +262,9 @@ define void @ord_v4f32(<4 x i32>* %c, <4 x float>* %a, <4 x float>* %b) nounwind define void @ord_v2f64(<2 x i64>* %c, <2 x double>* %a, <2 x double>* %b) nounwind { ; CHECK: ord_v2f64: - %1 = load <2 x double>* %a + %1 = load <2 x double>, <2 x double>* %a ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5) - %2 = load <2 x double>* %b + %2 = load <2 x double>, <2 x double>* %b ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6) %3 = fcmp ord <2 x double> %1, %2 %4 = sext <2 x i1> %3 to <2 x i64> @@ -279,9 +279,9 @@ define void @ord_v2f64(<2 x i64>* %c, <2 x double>* %a, <2 x double>* %b) nounwi define void @ueq_v4f32(<4 x i32>* %c, <4 x float>* %a, <4 x float>* %b) nounwind { ; CHECK: ueq_v4f32: - %1 = load <4 x float>* %a + %1 = load <4 x float>, <4 x float>* %a ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5) - %2 = load <4 x float>* %b + %2 = load <4 x float>, <4 x float>* %b ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6) %3 = fcmp ueq <4 x float> %1, %2 %4 = sext <4 x i1> %3 to <4 x i32> @@ -296,9 +296,9 @@ define void @ueq_v4f32(<4 x i32>* %c, <4 x float>* %a, <4 x float>* %b) nounwind define void @ueq_v2f64(<2 x i64>* %c, <2 x double>* %a, <2 x double>* %b) nounwind { ; CHECK: ueq_v2f64: - %1 = load <2 x double>* %a + %1 = load <2 x double>, <2 x double>* %a ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5) - %2 = load <2 x double>* %b + %2 = load <2 x double>, <2 x double>* %b ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6) %3 = fcmp ueq <2 x double> %1, %2 %4 = sext <2 x i1> %3 to <2 x i64> @@ -313,9 +313,9 @@ define void @ueq_v2f64(<2 x i64>* %c, <2 x double>* %a, <2 x double>* %b) nounwi define void @uge_v4f32(<4 x i32>* %c, <4 x float>* %a, <4 x float>* %b) nounwind { ; CHECK: uge_v4f32: - %1 = load <4 x float>* %a + %1 = load <4 x float>, <4 x float>* %a ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5) - %2 = load <4 x float>* %b + %2 = load <4 x float>, <4 x float>* %b ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6) %3 = fcmp uge <4 x float> %1, %2 %4 = sext <4 x i1> %3 to <4 x i32> @@ -330,9 +330,9 @@ define void @uge_v4f32(<4 x i32>* %c, <4 x float>* %a, <4 x float>* %b) nounwind define void @uge_v2f64(<2 x i64>* %c, <2 x double>* %a, <2 x double>* %b) nounwind { ; CHECK: uge_v2f64: - %1 = load <2 x double>* %a + %1 = load <2 x double>, <2 x double>* %a ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5) - %2 = load <2 x double>* %b + %2 = load <2 x double>, <2 x double>* %b ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6) %3 = fcmp uge <2 x double> %1, %2 %4 = sext <2 x i1> %3 to <2 x i64> @@ -347,9 +347,9 @@ define void @uge_v2f64(<2 x i64>* %c, <2 x double>* %a, <2 x double>* %b) nounwi define void @ugt_v4f32(<4 x i32>* %c, <4 x float>* %a, <4 x float>* %b) nounwind { ; CHECK: ugt_v4f32: - %1 = load <4 x float>* %a + %1 = load <4 x float>, <4 x float>* %a ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5) - %2 = load <4 x float>* %b + %2 = load <4 x float>, <4 x float>* %b ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6) %3 = fcmp ugt <4 x float> %1, %2 %4 = sext <4 x i1> %3 to <4 x i32> @@ -364,9 +364,9 @@ define void @ugt_v4f32(<4 x i32>* %c, <4 x float>* %a, <4 x float>* %b) nounwind define void @ugt_v2f64(<2 x i64>* %c, <2 x double>* %a, <2 x double>* %b) nounwind { ; CHECK: ugt_v2f64: - %1 = load <2 x double>* %a + %1 = load <2 x double>, <2 x double>* %a ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5) - %2 = load <2 x double>* %b + %2 = load <2 x double>, <2 x double>* %b ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6) %3 = fcmp ugt <2 x double> %1, %2 %4 = sext <2 x i1> %3 to <2 x i64> @@ -381,9 +381,9 @@ define void @ugt_v2f64(<2 x i64>* %c, <2 x double>* %a, <2 x double>* %b) nounwi define void @ule_v4f32(<4 x i32>* %c, <4 x float>* %a, <4 x float>* %b) nounwind { ; CHECK: ule_v4f32: - %1 = load <4 x float>* %a + %1 = load <4 x float>, <4 x float>* %a ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5) - %2 = load <4 x float>* %b + %2 = load <4 x float>, <4 x float>* %b ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6) %3 = fcmp ule <4 x float> %1, %2 %4 = sext <4 x i1> %3 to <4 x i32> @@ -398,9 +398,9 @@ define void @ule_v4f32(<4 x i32>* %c, <4 x float>* %a, <4 x float>* %b) nounwind define void @ule_v2f64(<2 x i64>* %c, <2 x double>* %a, <2 x double>* %b) nounwind { ; CHECK: ule_v2f64: - %1 = load <2 x double>* %a + %1 = load <2 x double>, <2 x double>* %a ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5) - %2 = load <2 x double>* %b + %2 = load <2 x double>, <2 x double>* %b ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6) %3 = fcmp ule <2 x double> %1, %2 %4 = sext <2 x i1> %3 to <2 x i64> @@ -415,9 +415,9 @@ define void @ule_v2f64(<2 x i64>* %c, <2 x double>* %a, <2 x double>* %b) nounwi define void @ult_v4f32(<4 x i32>* %c, <4 x float>* %a, <4 x float>* %b) nounwind { ; CHECK: ult_v4f32: - %1 = load <4 x float>* %a + %1 = load <4 x float>, <4 x float>* %a ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5) - %2 = load <4 x float>* %b + %2 = load <4 x float>, <4 x float>* %b ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6) %3 = fcmp ult <4 x float> %1, %2 %4 = sext <4 x i1> %3 to <4 x i32> @@ -432,9 +432,9 @@ define void @ult_v4f32(<4 x i32>* %c, <4 x float>* %a, <4 x float>* %b) nounwind define void @ult_v2f64(<2 x i64>* %c, <2 x double>* %a, <2 x double>* %b) nounwind { ; CHECK: ult_v2f64: - %1 = load <2 x double>* %a + %1 = load <2 x double>, <2 x double>* %a ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5) - %2 = load <2 x double>* %b + %2 = load <2 x double>, <2 x double>* %b ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6) %3 = fcmp ult <2 x double> %1, %2 %4 = sext <2 x i1> %3 to <2 x i64> @@ -449,9 +449,9 @@ define void @ult_v2f64(<2 x i64>* %c, <2 x double>* %a, <2 x double>* %b) nounwi define void @uno_v4f32(<4 x i32>* %c, <4 x float>* %a, <4 x float>* %b) nounwind { ; CHECK: uno_v4f32: - %1 = load <4 x float>* %a + %1 = load <4 x float>, <4 x float>* %a ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5) - %2 = load <4 x float>* %b + %2 = load <4 x float>, <4 x float>* %b ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6) %3 = fcmp uno <4 x float> %1, %2 %4 = sext <4 x i1> %3 to <4 x i32> @@ -466,9 +466,9 @@ define void @uno_v4f32(<4 x i32>* %c, <4 x float>* %a, <4 x float>* %b) nounwind define void @uno_v2f64(<2 x i64>* %c, <2 x double>* %a, <2 x double>* %b) nounwind { ; CHECK: uno_v2f64: - %1 = load <2 x double>* %a + %1 = load <2 x double>, <2 x double>* %a ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5) - %2 = load <2 x double>* %b + %2 = load <2 x double>, <2 x double>* %b ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6) %3 = fcmp uno <2 x double> %1, %2 %4 = sext <2 x i1> %3 to <2 x i64> @@ -483,8 +483,8 @@ define void @uno_v2f64(<2 x i64>* %c, <2 x double>* %a, <2 x double>* %b) nounwi define void @true_v4f32(<4 x i32>* %c, <4 x float>* %a, <4 x float>* %b) nounwind { ; CHECK: true_v4f32: - %1 = load <4 x float>* %a - %2 = load <4 x float>* %b + %1 = load <4 x float>, <4 x float>* %a + %2 = load <4 x float>, <4 x float>* %b %3 = fcmp true <4 x float> %1, %2 %4 = sext <4 x i1> %3 to <4 x i32> store <4 x i32> %4, <4 x i32>* %c @@ -499,8 +499,8 @@ define void @true_v4f32(<4 x i32>* %c, <4 x float>* %a, <4 x float>* %b) nounwin define void @true_v2f64(<2 x i64>* %c, <2 x double>* %a, <2 x double>* %b) nounwind { ; CHECK: true_v2f64: - %1 = load <2 x double>* %a - %2 = load <2 x double>* %b + %1 = load <2 x double>, <2 x double>* %a + %2 = load <2 x double>, <2 x double>* %b %3 = fcmp true <2 x double> %1, %2 %4 = sext <2 x i1> %3 to <2 x i64> store <2 x i64> %4, <2 x i64>* %c @@ -516,11 +516,11 @@ define void @bsel_v4f32(<4 x float>* %d, <4 x float>* %a, <4 x float>* %b, <4 x float>* %c) nounwind { ; CHECK: bsel_v4f32: - %1 = load <4 x float>* %a + %1 = load <4 x float>, <4 x float>* %a ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5) - %2 = load <4 x float>* %b + %2 = load <4 x float>, <4 x float>* %b ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6) - %3 = load <4 x float>* %c + %3 = load <4 x float>, <4 x float>* %c ; CHECK-DAG: ld.w [[R3:\$w[0-9]+]], 0($7) %4 = fcmp ogt <4 x float> %1, %2 ; CHECK-DAG: fclt.w [[R4:\$w[0-9]+]], [[R2]], [[R1]] @@ -538,11 +538,11 @@ define void @bsel_v2f64(<2 x double>* %d, <2 x double>* %a, <2 x double>* %b, <2 x double>* %c) nounwind { ; CHECK: bsel_v2f64: - %1 = load <2 x double>* %a + %1 = load <2 x double>, <2 x double>* %a ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5) - %2 = load <2 x double>* %b + %2 = load <2 x double>, <2 x double>* %b ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6) - %3 = load <2 x double>* %c + %3 = load <2 x double>, <2 x double>* %c ; CHECK-DAG: ld.d [[R3:\$w[0-9]+]], 0($7) %4 = fcmp ogt <2 x double> %1, %2 ; CHECK-DAG: fclt.d [[R4:\$w[0-9]+]], [[R2]], [[R1]] @@ -560,9 +560,9 @@ define void @bseli_v4f32(<4 x float>* %d, <4 x float>* %a, <4 x float>* %b, <4 x float>* %c) nounwind { ; CHECK: bseli_v4f32: - %1 = load <4 x float>* %a + %1 = load <4 x float>, <4 x float>* %a ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5) - %2 = load <4 x float>* %b + %2 = load <4 x float>, <4 x float>* %b ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6) %3 = fcmp ogt <4 x float> %1, %2 ; CHECK-DAG: fclt.w [[R4:\$w[0-9]+]], [[R2]], [[R1]] @@ -580,9 +580,9 @@ define void @bseli_v2f64(<2 x double>* %d, <2 x double>* %a, <2 x double>* %b, <2 x double>* %c) nounwind { ; CHECK: bseli_v2f64: - %1 = load <2 x double>* %a + %1 = load <2 x double>, <2 x double>* %a ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5) - %2 = load <2 x double>* %b + %2 = load <2 x double>, <2 x double>* %b ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6) %3 = fcmp ogt <2 x double> %1, %2 ; CHECK-DAG: fclt.d [[R4:\$w[0-9]+]], [[R2]], [[R1]] @@ -599,9 +599,9 @@ define void @bseli_v2f64(<2 x double>* %d, <2 x double>* %a, <2 x double>* %b, define void @max_v4f32(<4 x float>* %c, <4 x float>* %a, <4 x float>* %b) nounwind { ; CHECK: max_v4f32: - %1 = load <4 x float>* %a + %1 = load <4 x float>, <4 x float>* %a ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5) - %2 = load <4 x float>* %b + %2 = load <4 x float>, <4 x float>* %b ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6) %3 = tail call <4 x float> @llvm.mips.fmax.w(<4 x float> %1, <4 x float> %2) ; CHECK-DAG: fmax.w [[R3:\$w[0-9]+]], [[R1]], [[R2]] @@ -615,9 +615,9 @@ define void @max_v4f32(<4 x float>* %c, <4 x float>* %a, <4 x float>* %b) nounwi define void @max_v2f64(<2 x double>* %c, <2 x double>* %a, <2 x double>* %b) nounwind { ; CHECK: max_v2f64: - %1 = load <2 x double>* %a + %1 = load <2 x double>, <2 x double>* %a ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5) - %2 = load <2 x double>* %b + %2 = load <2 x double>, <2 x double>* %b ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6) %3 = tail call <2 x double> @llvm.mips.fmax.d(<2 x double> %1, <2 x double> %2) ; CHECK-DAG: fmax.d [[R3:\$w[0-9]+]], [[R1]], [[R2]] @@ -631,9 +631,9 @@ define void @max_v2f64(<2 x double>* %c, <2 x double>* %a, <2 x double>* %b) nou define void @min_v4f32(<4 x float>* %c, <4 x float>* %a, <4 x float>* %b) nounwind { ; CHECK: min_v4f32: - %1 = load <4 x float>* %a + %1 = load <4 x float>, <4 x float>* %a ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5) - %2 = load <4 x float>* %b + %2 = load <4 x float>, <4 x float>* %b ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6) %3 = tail call <4 x float> @llvm.mips.fmin.w(<4 x float> %1, <4 x float> %2) ; CHECK-DAG: fmin.w [[R3:\$w[0-9]+]], [[R1]], [[R2]] @@ -647,9 +647,9 @@ define void @min_v4f32(<4 x float>* %c, <4 x float>* %a, <4 x float>* %b) nounwi define void @min_v2f64(<2 x double>* %c, <2 x double>* %a, <2 x double>* %b) nounwind { ; CHECK: min_v2f64: - %1 = load <2 x double>* %a + %1 = load <2 x double>, <2 x double>* %a ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5) - %2 = load <2 x double>* %b + %2 = load <2 x double>, <2 x double>* %b ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6) %3 = tail call <2 x double> @llvm.mips.fmin.d(<2 x double> %1, <2 x double> %2) ; CHECK-DAG: fmin.d [[R3:\$w[0-9]+]], [[R1]], [[R2]] diff --git a/test/CodeGen/Mips/msa/elm_copy.ll b/test/CodeGen/Mips/msa/elm_copy.ll index 0dd75fa3db12a..2a0d74f445246 100644 --- a/test/CodeGen/Mips/msa/elm_copy.ll +++ b/test/CodeGen/Mips/msa/elm_copy.ll @@ -15,7 +15,7 @@ define void @llvm_mips_copy_s_b_test() nounwind { entry: - %0 = load <16 x i8>* @llvm_mips_copy_s_b_ARG1 + %0 = load <16 x i8>, <16 x i8>* @llvm_mips_copy_s_b_ARG1 %1 = tail call i32 @llvm.mips.copy.s.b(<16 x i8> %0, i32 1) store i32 %1, i32* @llvm_mips_copy_s_b_RES ret void @@ -38,7 +38,7 @@ declare i32 @llvm.mips.copy.s.b(<16 x i8>, i32) nounwind define void @llvm_mips_copy_s_h_test() nounwind { entry: - %0 = load <8 x i16>* @llvm_mips_copy_s_h_ARG1 + %0 = load <8 x i16>, <8 x i16>* @llvm_mips_copy_s_h_ARG1 %1 = tail call i32 @llvm.mips.copy.s.h(<8 x i16> %0, i32 1) store i32 %1, i32* @llvm_mips_copy_s_h_RES ret void @@ -61,7 +61,7 @@ declare i32 @llvm.mips.copy.s.h(<8 x i16>, i32) nounwind define void @llvm_mips_copy_s_w_test() nounwind { entry: - %0 = load <4 x i32>* @llvm_mips_copy_s_w_ARG1 + %0 = load <4 x i32>, <4 x i32>* @llvm_mips_copy_s_w_ARG1 %1 = tail call i32 @llvm.mips.copy.s.w(<4 x i32> %0, i32 1) store i32 %1, i32* @llvm_mips_copy_s_w_RES ret void @@ -84,7 +84,7 @@ declare i32 @llvm.mips.copy.s.w(<4 x i32>, i32) nounwind define void @llvm_mips_copy_s_d_test() nounwind { entry: - %0 = load <2 x i64>* @llvm_mips_copy_s_d_ARG1 + %0 = load <2 x i64>, <2 x i64>* @llvm_mips_copy_s_d_ARG1 %1 = tail call i64 @llvm.mips.copy.s.d(<2 x i64> %0, i32 1) store i64 %1, i64* @llvm_mips_copy_s_d_RES ret void @@ -112,7 +112,7 @@ declare i64 @llvm.mips.copy.s.d(<2 x i64>, i32) nounwind define void @llvm_mips_copy_u_b_test() nounwind { entry: - %0 = load <16 x i8>* @llvm_mips_copy_u_b_ARG1 + %0 = load <16 x i8>, <16 x i8>* @llvm_mips_copy_u_b_ARG1 %1 = tail call i32 @llvm.mips.copy.u.b(<16 x i8> %0, i32 1) store i32 %1, i32* @llvm_mips_copy_u_b_RES ret void @@ -135,7 +135,7 @@ declare i32 @llvm.mips.copy.u.b(<16 x i8>, i32) nounwind define void @llvm_mips_copy_u_h_test() nounwind { entry: - %0 = load <8 x i16>* @llvm_mips_copy_u_h_ARG1 + %0 = load <8 x i16>, <8 x i16>* @llvm_mips_copy_u_h_ARG1 %1 = tail call i32 @llvm.mips.copy.u.h(<8 x i16> %0, i32 1) store i32 %1, i32* @llvm_mips_copy_u_h_RES ret void @@ -158,7 +158,7 @@ declare i32 @llvm.mips.copy.u.h(<8 x i16>, i32) nounwind define void @llvm_mips_copy_u_w_test() nounwind { entry: - %0 = load <4 x i32>* @llvm_mips_copy_u_w_ARG1 + %0 = load <4 x i32>, <4 x i32>* @llvm_mips_copy_u_w_ARG1 %1 = tail call i32 @llvm.mips.copy.u.w(<4 x i32> %0, i32 1) store i32 %1, i32* @llvm_mips_copy_u_w_RES ret void @@ -181,7 +181,7 @@ declare i32 @llvm.mips.copy.u.w(<4 x i32>, i32) nounwind define void @llvm_mips_copy_u_d_test() nounwind { entry: - %0 = load <2 x i64>* @llvm_mips_copy_u_d_ARG1 + %0 = load <2 x i64>, <2 x i64>* @llvm_mips_copy_u_d_ARG1 %1 = tail call i64 @llvm.mips.copy.u.d(<2 x i64> %0, i32 1) store i64 %1, i64* @llvm_mips_copy_u_d_RES ret void diff --git a/test/CodeGen/Mips/msa/elm_insv.ll b/test/CodeGen/Mips/msa/elm_insv.ll index c746e523def6d..46e6289189df0 100644 --- a/test/CodeGen/Mips/msa/elm_insv.ll +++ b/test/CodeGen/Mips/msa/elm_insv.ll @@ -16,8 +16,8 @@ define void @llvm_mips_insert_b_test() nounwind { entry: - %0 = load <16 x i8>* @llvm_mips_insert_b_ARG1 - %1 = load i32* @llvm_mips_insert_b_ARG3 + %0 = load <16 x i8>, <16 x i8>* @llvm_mips_insert_b_ARG1 + %1 = load i32, i32* @llvm_mips_insert_b_ARG3 %2 = tail call <16 x i8> @llvm.mips.insert.b(<16 x i8> %0, i32 1, i32 %1) store <16 x i8> %2, <16 x i8>* @llvm_mips_insert_b_RES ret void @@ -38,8 +38,8 @@ declare <16 x i8> @llvm.mips.insert.b(<16 x i8>, i32, i32) nounwind define void @llvm_mips_insert_h_test() nounwind { entry: - %0 = load <8 x i16>* @llvm_mips_insert_h_ARG1 - %1 = load i32* @llvm_mips_insert_h_ARG3 + %0 = load <8 x i16>, <8 x i16>* @llvm_mips_insert_h_ARG1 + %1 = load i32, i32* @llvm_mips_insert_h_ARG3 %2 = tail call <8 x i16> @llvm.mips.insert.h(<8 x i16> %0, i32 1, i32 %1) store <8 x i16> %2, <8 x i16>* @llvm_mips_insert_h_RES ret void @@ -60,8 +60,8 @@ declare <8 x i16> @llvm.mips.insert.h(<8 x i16>, i32, i32) nounwind define void @llvm_mips_insert_w_test() nounwind { entry: - %0 = load <4 x i32>* @llvm_mips_insert_w_ARG1 - %1 = load i32* @llvm_mips_insert_w_ARG3 + %0 = load <4 x i32>, <4 x i32>* @llvm_mips_insert_w_ARG1 + %1 = load i32, i32* @llvm_mips_insert_w_ARG3 %2 = tail call <4 x i32> @llvm.mips.insert.w(<4 x i32> %0, i32 1, i32 %1) store <4 x i32> %2, <4 x i32>* @llvm_mips_insert_w_RES ret void @@ -82,8 +82,8 @@ declare <4 x i32> @llvm.mips.insert.w(<4 x i32>, i32, i32) nounwind define void @llvm_mips_insert_d_test() nounwind { entry: - %0 = load <2 x i64>* @llvm_mips_insert_d_ARG1 - %1 = load i64* @llvm_mips_insert_d_ARG3 + %0 = load <2 x i64>, <2 x i64>* @llvm_mips_insert_d_ARG1 + %1 = load i64, i64* @llvm_mips_insert_d_ARG3 %2 = tail call <2 x i64> @llvm.mips.insert.d(<2 x i64> %0, i32 1, i64 %1) store <2 x i64> %2, <2 x i64>* @llvm_mips_insert_d_RES ret void @@ -110,8 +110,8 @@ declare <2 x i64> @llvm.mips.insert.d(<2 x i64>, i32, i64) nounwind define void @llvm_mips_insve_b_test() nounwind { entry: - %0 = load <16 x i8>* @llvm_mips_insve_b_ARG1 - %1 = load <16 x i8>* @llvm_mips_insve_b_ARG3 + %0 = load <16 x i8>, <16 x i8>* @llvm_mips_insve_b_ARG1 + %1 = load <16 x i8>, <16 x i8>* @llvm_mips_insve_b_ARG3 %2 = tail call <16 x i8> @llvm.mips.insve.b(<16 x i8> %0, i32 1, <16 x i8> %1) store <16 x i8> %2, <16 x i8>* @llvm_mips_insve_b_RES ret void @@ -136,8 +136,8 @@ declare <16 x i8> @llvm.mips.insve.b(<16 x i8>, i32, <16 x i8>) nounwind define void @llvm_mips_insve_h_test() nounwind { entry: - %0 = load <8 x i16>* @llvm_mips_insve_h_ARG1 - %1 = load <8 x i16>* @llvm_mips_insve_h_ARG3 + %0 = load <8 x i16>, <8 x i16>* @llvm_mips_insve_h_ARG1 + %1 = load <8 x i16>, <8 x i16>* @llvm_mips_insve_h_ARG3 %2 = tail call <8 x i16> @llvm.mips.insve.h(<8 x i16> %0, i32 1, <8 x i16> %1) store <8 x i16> %2, <8 x i16>* @llvm_mips_insve_h_RES ret void @@ -162,8 +162,8 @@ declare <8 x i16> @llvm.mips.insve.h(<8 x i16>, i32, <8 x i16>) nounwind define void @llvm_mips_insve_w_test() nounwind { entry: - %0 = load <4 x i32>* @llvm_mips_insve_w_ARG1 - %1 = load <4 x i32>* @llvm_mips_insve_w_ARG3 + %0 = load <4 x i32>, <4 x i32>* @llvm_mips_insve_w_ARG1 + %1 = load <4 x i32>, <4 x i32>* @llvm_mips_insve_w_ARG3 %2 = tail call <4 x i32> @llvm.mips.insve.w(<4 x i32> %0, i32 1, <4 x i32> %1) store <4 x i32> %2, <4 x i32>* @llvm_mips_insve_w_RES ret void @@ -188,8 +188,8 @@ declare <4 x i32> @llvm.mips.insve.w(<4 x i32>, i32, <4 x i32>) nounwind define void @llvm_mips_insve_d_test() nounwind { entry: - %0 = load <2 x i64>* @llvm_mips_insve_d_ARG1 - %1 = load <2 x i64>* @llvm_mips_insve_d_ARG3 + %0 = load <2 x i64>, <2 x i64>* @llvm_mips_insve_d_ARG1 + %1 = load <2 x i64>, <2 x i64>* @llvm_mips_insve_d_ARG3 %2 = tail call <2 x i64> @llvm.mips.insve.d(<2 x i64> %0, i32 1, <2 x i64> %1) store <2 x i64> %2, <2 x i64>* @llvm_mips_insve_d_RES ret void diff --git a/test/CodeGen/Mips/msa/elm_move.ll b/test/CodeGen/Mips/msa/elm_move.ll index 98c06c732c36c..9665b6d688f39 100644 --- a/test/CodeGen/Mips/msa/elm_move.ll +++ b/test/CodeGen/Mips/msa/elm_move.ll @@ -9,7 +9,7 @@ define void @llvm_mips_move_vb_test() nounwind { entry: - %0 = load <16 x i8>* @llvm_mips_move_vb_ARG1 + %0 = load <16 x i8>, <16 x i8>* @llvm_mips_move_vb_ARG1 %1 = tail call <16 x i8> @llvm.mips.move.v(<16 x i8> %0) store <16 x i8> %1, <16 x i8>* @llvm_mips_move_vb_RES ret void diff --git a/test/CodeGen/Mips/msa/elm_shift_slide.ll b/test/CodeGen/Mips/msa/elm_shift_slide.ll index 00a6544b12070..87f15f1a8c927 100644 --- a/test/CodeGen/Mips/msa/elm_shift_slide.ll +++ b/test/CodeGen/Mips/msa/elm_shift_slide.ll @@ -10,8 +10,8 @@ define void @llvm_mips_sldi_b_test() nounwind { entry: - %0 = load <16 x i8>* @llvm_mips_sldi_b_ARG1 - %1 = load <16 x i8>* @llvm_mips_sldi_b_ARG2 + %0 = load <16 x i8>, <16 x i8>* @llvm_mips_sldi_b_ARG1 + %1 = load <16 x i8>, <16 x i8>* @llvm_mips_sldi_b_ARG2 %2 = tail call <16 x i8> @llvm.mips.sldi.b(<16 x i8> %0, <16 x i8> %1, i32 1) store <16 x i8> %2, <16 x i8>* @llvm_mips_sldi_b_RES ret void @@ -31,8 +31,8 @@ declare <16 x i8> @llvm.mips.sldi.b(<16 x i8>, <16 x i8>, i32) nounwind define void @llvm_mips_sldi_h_test() nounwind { entry: - %0 = load <8 x i16>* @llvm_mips_sldi_h_ARG1 - %1 = load <8 x i16>* @llvm_mips_sldi_h_ARG2 + %0 = load <8 x i16>, <8 x i16>* @llvm_mips_sldi_h_ARG1 + %1 = load <8 x i16>, <8 x i16>* @llvm_mips_sldi_h_ARG2 %2 = tail call <8 x i16> @llvm.mips.sldi.h(<8 x i16> %0, <8 x i16> %1, i32 1) store <8 x i16> %2, <8 x i16>* @llvm_mips_sldi_h_RES ret void @@ -52,8 +52,8 @@ declare <8 x i16> @llvm.mips.sldi.h(<8 x i16>, <8 x i16>, i32) nounwind define void @llvm_mips_sldi_w_test() nounwind { entry: - %0 = load <4 x i32>* @llvm_mips_sldi_w_ARG1 - %1 = load <4 x i32>* @llvm_mips_sldi_w_ARG2 + %0 = load <4 x i32>, <4 x i32>* @llvm_mips_sldi_w_ARG1 + %1 = load <4 x i32>, <4 x i32>* @llvm_mips_sldi_w_ARG2 %2 = tail call <4 x i32> @llvm.mips.sldi.w(<4 x i32> %0, <4 x i32> %1, i32 1) store <4 x i32> %2, <4 x i32>* @llvm_mips_sldi_w_RES ret void @@ -73,8 +73,8 @@ declare <4 x i32> @llvm.mips.sldi.w(<4 x i32>, <4 x i32>, i32) nounwind define void @llvm_mips_sldi_d_test() nounwind { entry: - %0 = load <2 x i64>* @llvm_mips_sldi_d_ARG1 - %1 = load <2 x i64>* @llvm_mips_sldi_d_ARG2 + %0 = load <2 x i64>, <2 x i64>* @llvm_mips_sldi_d_ARG1 + %1 = load <2 x i64>, <2 x i64>* @llvm_mips_sldi_d_ARG2 %2 = tail call <2 x i64> @llvm.mips.sldi.d(<2 x i64> %0, <2 x i64> %1, i32 1) store <2 x i64> %2, <2 x i64>* @llvm_mips_sldi_d_RES ret void @@ -93,7 +93,7 @@ declare <2 x i64> @llvm.mips.sldi.d(<2 x i64>, <2 x i64>, i32) nounwind define void @llvm_mips_splati_b_test() nounwind { entry: - %0 = load <16 x i8>* @llvm_mips_splati_b_ARG1 + %0 = load <16 x i8>, <16 x i8>* @llvm_mips_splati_b_ARG1 %1 = tail call <16 x i8> @llvm.mips.splati.b(<16 x i8> %0, i32 1) store <16 x i8> %1, <16 x i8>* @llvm_mips_splati_b_RES ret void @@ -112,7 +112,7 @@ declare <16 x i8> @llvm.mips.splati.b(<16 x i8>, i32) nounwind define void @llvm_mips_splati_h_test() nounwind { entry: - %0 = load <8 x i16>* @llvm_mips_splati_h_ARG1 + %0 = load <8 x i16>, <8 x i16>* @llvm_mips_splati_h_ARG1 %1 = tail call <8 x i16> @llvm.mips.splati.h(<8 x i16> %0, i32 1) store <8 x i16> %1, <8 x i16>* @llvm_mips_splati_h_RES ret void @@ -131,7 +131,7 @@ declare <8 x i16> @llvm.mips.splati.h(<8 x i16>, i32) nounwind define void @llvm_mips_splati_w_test() nounwind { entry: - %0 = load <4 x i32>* @llvm_mips_splati_w_ARG1 + %0 = load <4 x i32>, <4 x i32>* @llvm_mips_splati_w_ARG1 %1 = tail call <4 x i32> @llvm.mips.splati.w(<4 x i32> %0, i32 1) store <4 x i32> %1, <4 x i32>* @llvm_mips_splati_w_RES ret void @@ -150,7 +150,7 @@ declare <4 x i32> @llvm.mips.splati.w(<4 x i32>, i32) nounwind define void @llvm_mips_splati_d_test() nounwind { entry: - %0 = load <2 x i64>* @llvm_mips_splati_d_ARG1 + %0 = load <2 x i64>, <2 x i64>* @llvm_mips_splati_d_ARG1 %1 = tail call <2 x i64> @llvm.mips.splati.d(<2 x i64> %0, i32 1) store <2 x i64> %1, <2 x i64>* @llvm_mips_splati_d_RES ret void diff --git a/test/CodeGen/Mips/msa/frameindex.ll b/test/CodeGen/Mips/msa/frameindex.ll index ebec465a3e331..afd28ae184dd7 100644 --- a/test/CodeGen/Mips/msa/frameindex.ll +++ b/test/CodeGen/Mips/msa/frameindex.ll @@ -5,7 +5,7 @@ define void @loadstore_v16i8_near() nounwind { ; MIPS32-AE: loadstore_v16i8_near: %1 = alloca <16 x i8> - %2 = load volatile <16 x i8>* %1 + %2 = load volatile <16 x i8>, <16 x i8>* %1 ; MIPS32-AE: ld.b [[R1:\$w[0-9]+]], 0($sp) store volatile <16 x i8> %2, <16 x i8>* %1 ; MIPS32-AE: st.b [[R1]], 0($sp) @@ -20,7 +20,7 @@ define void @loadstore_v16i8_just_under_simm10() nounwind { %1 = alloca <16 x i8> %2 = alloca [496 x i8] ; Push the frame right up to 512 bytes - %3 = load volatile <16 x i8>* %1 + %3 = load volatile <16 x i8>, <16 x i8>* %1 ; MIPS32-AE: ld.b [[R1:\$w[0-9]+]], 496($sp) store volatile <16 x i8> %3, <16 x i8>* %1 ; MIPS32-AE: st.b [[R1]], 496($sp) @@ -35,7 +35,7 @@ define void @loadstore_v16i8_just_over_simm10() nounwind { %1 = alloca <16 x i8> %2 = alloca [497 x i8] ; Push the frame just over 512 bytes - %3 = load volatile <16 x i8>* %1 + %3 = load volatile <16 x i8>, <16 x i8>* %1 ; MIPS32-AE: addiu [[BASE:\$([0-9]+|gp)]], $sp, 512 ; MIPS32-AE: ld.b [[R1:\$w[0-9]+]], 0([[BASE]]) store volatile <16 x i8> %3, <16 x i8>* %1 @@ -52,7 +52,7 @@ define void @loadstore_v16i8_just_under_simm16() nounwind { %1 = alloca <16 x i8> %2 = alloca [32752 x i8] ; Push the frame right up to 32768 bytes - %3 = load volatile <16 x i8>* %1 + %3 = load volatile <16 x i8>, <16 x i8>* %1 ; MIPS32-AE: ori [[R2:\$([0-9]+|gp)]], $zero, 32768 ; MIPS32-AE: addu [[BASE:\$([0-9]+|gp)]], $sp, [[R2]] ; MIPS32-AE: ld.b [[R1:\$w[0-9]+]], 0([[BASE]]) @@ -71,7 +71,7 @@ define void @loadstore_v16i8_just_over_simm16() nounwind { %1 = alloca <16 x i8> %2 = alloca [32753 x i8] ; Push the frame just over 32768 bytes - %3 = load volatile <16 x i8>* %1 + %3 = load volatile <16 x i8>, <16 x i8>* %1 ; MIPS32-AE: ori [[R2:\$([0-9]+|gp)]], $zero, 32768 ; MIPS32-AE: addu [[BASE:\$([0-9]+|gp)]], $sp, [[R2]] ; MIPS32-AE: ld.b [[R1:\$w[0-9]+]], 0([[BASE]]) @@ -88,7 +88,7 @@ define void @loadstore_v8i16_near() nounwind { ; MIPS32-AE: loadstore_v8i16_near: %1 = alloca <8 x i16> - %2 = load volatile <8 x i16>* %1 + %2 = load volatile <8 x i16>, <8 x i16>* %1 ; MIPS32-AE: ld.h [[R1:\$w[0-9]+]], 0($sp) store volatile <8 x i16> %2, <8 x i16>* %1 ; MIPS32-AE: st.h [[R1]], 0($sp) @@ -102,11 +102,11 @@ define void @loadstore_v8i16_unaligned() nounwind { %1 = alloca [2 x <8 x i16>] %2 = bitcast [2 x <8 x i16>]* %1 to i8* - %3 = getelementptr i8* %2, i32 1 + %3 = getelementptr i8, i8* %2, i32 1 %4 = bitcast i8* %3 to [2 x <8 x i16>]* - %5 = getelementptr [2 x <8 x i16>]* %4, i32 0, i32 0 + %5 = getelementptr [2 x <8 x i16>], [2 x <8 x i16>]* %4, i32 0, i32 0 - %6 = load volatile <8 x i16>* %5 + %6 = load volatile <8 x i16>, <8 x i16>* %5 ; MIPS32-AE: addiu [[BASE:\$([0-9]+|gp)]], $sp, 1 ; MIPS32-AE: ld.h [[R1:\$w[0-9]+]], 0([[BASE]]) store volatile <8 x i16> %6, <8 x i16>* %5 @@ -123,7 +123,7 @@ define void @loadstore_v8i16_just_under_simm10() nounwind { %1 = alloca <8 x i16> %2 = alloca [1008 x i8] ; Push the frame right up to 1024 bytes - %3 = load volatile <8 x i16>* %1 + %3 = load volatile <8 x i16>, <8 x i16>* %1 ; MIPS32-AE: ld.h [[R1:\$w[0-9]+]], 1008($sp) store volatile <8 x i16> %3, <8 x i16>* %1 ; MIPS32-AE: st.h [[R1]], 1008($sp) @@ -138,7 +138,7 @@ define void @loadstore_v8i16_just_over_simm10() nounwind { %1 = alloca <8 x i16> %2 = alloca [1009 x i8] ; Push the frame just over 1024 bytes - %3 = load volatile <8 x i16>* %1 + %3 = load volatile <8 x i16>, <8 x i16>* %1 ; MIPS32-AE: addiu [[BASE:\$([0-9]+|gp)]], $sp, 1024 ; MIPS32-AE: ld.h [[R1:\$w[0-9]+]], 0([[BASE]]) store volatile <8 x i16> %3, <8 x i16>* %1 @@ -155,7 +155,7 @@ define void @loadstore_v8i16_just_under_simm16() nounwind { %1 = alloca <8 x i16> %2 = alloca [32752 x i8] ; Push the frame right up to 32768 bytes - %3 = load volatile <8 x i16>* %1 + %3 = load volatile <8 x i16>, <8 x i16>* %1 ; MIPS32-AE: ori [[R2:\$([0-9]+|gp)]], $zero, 32768 ; MIPS32-AE: addu [[BASE:\$([0-9]+|gp)]], $sp, [[R2]] ; MIPS32-AE: ld.h [[R1:\$w[0-9]+]], 0([[BASE]]) @@ -174,7 +174,7 @@ define void @loadstore_v8i16_just_over_simm16() nounwind { %1 = alloca <8 x i16> %2 = alloca [32753 x i8] ; Push the frame just over 32768 bytes - %3 = load volatile <8 x i16>* %1 + %3 = load volatile <8 x i16>, <8 x i16>* %1 ; MIPS32-AE: ori [[R2:\$([0-9]+|gp)]], $zero, 32768 ; MIPS32-AE: addu [[BASE:\$([0-9]+|gp)]], $sp, [[R2]] ; MIPS32-AE: ld.h [[R1:\$w[0-9]+]], 0([[BASE]]) @@ -191,7 +191,7 @@ define void @loadstore_v4i32_near() nounwind { ; MIPS32-AE: loadstore_v4i32_near: %1 = alloca <4 x i32> - %2 = load volatile <4 x i32>* %1 + %2 = load volatile <4 x i32>, <4 x i32>* %1 ; MIPS32-AE: ld.w [[R1:\$w[0-9]+]], 0($sp) store volatile <4 x i32> %2, <4 x i32>* %1 ; MIPS32-AE: st.w [[R1]], 0($sp) @@ -205,11 +205,11 @@ define void @loadstore_v4i32_unaligned() nounwind { %1 = alloca [2 x <4 x i32>] %2 = bitcast [2 x <4 x i32>]* %1 to i8* - %3 = getelementptr i8* %2, i32 1 + %3 = getelementptr i8, i8* %2, i32 1 %4 = bitcast i8* %3 to [2 x <4 x i32>]* - %5 = getelementptr [2 x <4 x i32>]* %4, i32 0, i32 0 + %5 = getelementptr [2 x <4 x i32>], [2 x <4 x i32>]* %4, i32 0, i32 0 - %6 = load volatile <4 x i32>* %5 + %6 = load volatile <4 x i32>, <4 x i32>* %5 ; MIPS32-AE: addiu [[BASE:\$([0-9]+|gp)]], $sp, 1 ; MIPS32-AE: ld.w [[R1:\$w[0-9]+]], 0([[BASE]]) store volatile <4 x i32> %6, <4 x i32>* %5 @@ -226,7 +226,7 @@ define void @loadstore_v4i32_just_under_simm10() nounwind { %1 = alloca <4 x i32> %2 = alloca [2032 x i8] ; Push the frame right up to 2048 bytes - %3 = load volatile <4 x i32>* %1 + %3 = load volatile <4 x i32>, <4 x i32>* %1 ; MIPS32-AE: ld.w [[R1:\$w[0-9]+]], 2032($sp) store volatile <4 x i32> %3, <4 x i32>* %1 ; MIPS32-AE: st.w [[R1]], 2032($sp) @@ -241,7 +241,7 @@ define void @loadstore_v4i32_just_over_simm10() nounwind { %1 = alloca <4 x i32> %2 = alloca [2033 x i8] ; Push the frame just over 2048 bytes - %3 = load volatile <4 x i32>* %1 + %3 = load volatile <4 x i32>, <4 x i32>* %1 ; MIPS32-AE: addiu [[BASE:\$([0-9]+|gp)]], $sp, 2048 ; MIPS32-AE: ld.w [[R1:\$w[0-9]+]], 0([[BASE]]) store volatile <4 x i32> %3, <4 x i32>* %1 @@ -258,7 +258,7 @@ define void @loadstore_v4i32_just_under_simm16() nounwind { %1 = alloca <4 x i32> %2 = alloca [32752 x i8] ; Push the frame right up to 32768 bytes - %3 = load volatile <4 x i32>* %1 + %3 = load volatile <4 x i32>, <4 x i32>* %1 ; MIPS32-AE: ori [[R2:\$([0-9]+|gp)]], $zero, 32768 ; MIPS32-AE: addu [[BASE:\$([0-9]+|gp)]], $sp, [[R2]] ; MIPS32-AE: ld.w [[R1:\$w[0-9]+]], 0([[BASE]]) @@ -277,7 +277,7 @@ define void @loadstore_v4i32_just_over_simm16() nounwind { %1 = alloca <4 x i32> %2 = alloca [32753 x i8] ; Push the frame just over 32768 bytes - %3 = load volatile <4 x i32>* %1 + %3 = load volatile <4 x i32>, <4 x i32>* %1 ; MIPS32-AE: ori [[R2:\$([0-9]+|gp)]], $zero, 32768 ; MIPS32-AE: addu [[BASE:\$([0-9]+|gp)]], $sp, [[R2]] ; MIPS32-AE: ld.w [[R1:\$w[0-9]+]], 0([[BASE]]) @@ -294,7 +294,7 @@ define void @loadstore_v2i64_near() nounwind { ; MIPS32-AE: loadstore_v2i64_near: %1 = alloca <2 x i64> - %2 = load volatile <2 x i64>* %1 + %2 = load volatile <2 x i64>, <2 x i64>* %1 ; MIPS32-AE: ld.d [[R1:\$w[0-9]+]], 0($sp) store volatile <2 x i64> %2, <2 x i64>* %1 ; MIPS32-AE: st.d [[R1]], 0($sp) @@ -308,11 +308,11 @@ define void @loadstore_v2i64_unaligned() nounwind { %1 = alloca [2 x <2 x i64>] %2 = bitcast [2 x <2 x i64>]* %1 to i8* - %3 = getelementptr i8* %2, i32 1 + %3 = getelementptr i8, i8* %2, i32 1 %4 = bitcast i8* %3 to [2 x <2 x i64>]* - %5 = getelementptr [2 x <2 x i64>]* %4, i32 0, i32 0 + %5 = getelementptr [2 x <2 x i64>], [2 x <2 x i64>]* %4, i32 0, i32 0 - %6 = load volatile <2 x i64>* %5 + %6 = load volatile <2 x i64>, <2 x i64>* %5 ; MIPS32-AE: addiu [[BASE:\$([0-9]+|gp)]], $sp, 1 ; MIPS32-AE: ld.d [[R1:\$w[0-9]+]], 0([[BASE]]) store volatile <2 x i64> %6, <2 x i64>* %5 @@ -329,7 +329,7 @@ define void @loadstore_v2i64_just_under_simm10() nounwind { %1 = alloca <2 x i64> %2 = alloca [4080 x i8] ; Push the frame right up to 4096 bytes - %3 = load volatile <2 x i64>* %1 + %3 = load volatile <2 x i64>, <2 x i64>* %1 ; MIPS32-AE: ld.d [[R1:\$w[0-9]+]], 4080($sp) store volatile <2 x i64> %3, <2 x i64>* %1 ; MIPS32-AE: st.d [[R1]], 4080($sp) @@ -344,7 +344,7 @@ define void @loadstore_v2i64_just_over_simm10() nounwind { %1 = alloca <2 x i64> %2 = alloca [4081 x i8] ; Push the frame just over 4096 bytes - %3 = load volatile <2 x i64>* %1 + %3 = load volatile <2 x i64>, <2 x i64>* %1 ; MIPS32-AE: addiu [[BASE:\$([0-9]+|gp)]], $sp, 4096 ; MIPS32-AE: ld.d [[R1:\$w[0-9]+]], 0([[BASE]]) store volatile <2 x i64> %3, <2 x i64>* %1 @@ -361,7 +361,7 @@ define void @loadstore_v2i64_just_under_simm16() nounwind { %1 = alloca <2 x i64> %2 = alloca [32752 x i8] ; Push the frame right up to 32768 bytes - %3 = load volatile <2 x i64>* %1 + %3 = load volatile <2 x i64>, <2 x i64>* %1 ; MIPS32-AE: ori [[R2:\$([0-9]+|gp)]], $zero, 32768 ; MIPS32-AE: addu [[BASE:\$([0-9]+|gp)]], $sp, [[R2]] ; MIPS32-AE: ld.d [[R1:\$w[0-9]+]], 0([[BASE]]) @@ -380,7 +380,7 @@ define void @loadstore_v2i64_just_over_simm16() nounwind { %1 = alloca <2 x i64> %2 = alloca [32753 x i8] ; Push the frame just over 32768 bytes - %3 = load volatile <2 x i64>* %1 + %3 = load volatile <2 x i64>, <2 x i64>* %1 ; MIPS32-AE: ori [[R2:\$([0-9]+|gp)]], $zero, 32768 ; MIPS32-AE: addu [[BASE:\$([0-9]+|gp)]], $sp, [[R2]] ; MIPS32-AE: ld.d [[R1:\$w[0-9]+]], 0([[BASE]]) diff --git a/test/CodeGen/Mips/msa/i10.ll b/test/CodeGen/Mips/msa/i10.ll index c5a96174a7349..204884bbf0252 100644 --- a/test/CodeGen/Mips/msa/i10.ll +++ b/test/CodeGen/Mips/msa/i10.ll @@ -7,7 +7,7 @@ define i32 @llvm_mips_bnz_b_test() nounwind { entry: - %0 = load <16 x i8>* @llvm_mips_bnz_b_ARG1 + %0 = load <16 x i8>, <16 x i8>* @llvm_mips_bnz_b_ARG1 %1 = tail call i32 @llvm.mips.bnz.b(<16 x i8> %0) %2 = icmp eq i32 %1, 0 br i1 %2, label %true, label %false @@ -28,7 +28,7 @@ declare i32 @llvm.mips.bnz.b(<16 x i8>) nounwind define i32 @llvm_mips_bnz_h_test() nounwind { entry: - %0 = load <8 x i16>* @llvm_mips_bnz_h_ARG1 + %0 = load <8 x i16>, <8 x i16>* @llvm_mips_bnz_h_ARG1 %1 = tail call i32 @llvm.mips.bnz.h(<8 x i16> %0) %2 = icmp eq i32 %1, 0 br i1 %2, label %true, label %false @@ -49,7 +49,7 @@ declare i32 @llvm.mips.bnz.h(<8 x i16>) nounwind define i32 @llvm_mips_bnz_w_test() nounwind { entry: - %0 = load <4 x i32>* @llvm_mips_bnz_w_ARG1 + %0 = load <4 x i32>, <4 x i32>* @llvm_mips_bnz_w_ARG1 %1 = tail call i32 @llvm.mips.bnz.w(<4 x i32> %0) %2 = icmp eq i32 %1, 0 br i1 %2, label %true, label %false @@ -70,7 +70,7 @@ declare i32 @llvm.mips.bnz.w(<4 x i32>) nounwind define i32 @llvm_mips_bnz_d_test() nounwind { entry: - %0 = load <2 x i64>* @llvm_mips_bnz_d_ARG1 + %0 = load <2 x i64>, <2 x i64>* @llvm_mips_bnz_d_ARG1 %1 = tail call i32 @llvm.mips.bnz.d(<2 x i64> %0) %2 = icmp eq i32 %1, 0 br i1 %2, label %true, label %false diff --git a/test/CodeGen/Mips/msa/i5-a.ll b/test/CodeGen/Mips/msa/i5-a.ll index 0b507208f4298..f9486b17e0ad1 100644 --- a/test/CodeGen/Mips/msa/i5-a.ll +++ b/test/CodeGen/Mips/msa/i5-a.ll @@ -9,7 +9,7 @@ define void @llvm_mips_addvi_b_test() nounwind { entry: - %0 = load <16 x i8>* @llvm_mips_addvi_b_ARG1 + %0 = load <16 x i8>, <16 x i8>* @llvm_mips_addvi_b_ARG1 %1 = tail call <16 x i8> @llvm.mips.addvi.b(<16 x i8> %0, i32 14) store <16 x i8> %1, <16 x i8>* @llvm_mips_addvi_b_RES ret void @@ -28,7 +28,7 @@ declare <16 x i8> @llvm.mips.addvi.b(<16 x i8>, i32) nounwind define void @llvm_mips_addvi_h_test() nounwind { entry: - %0 = load <8 x i16>* @llvm_mips_addvi_h_ARG1 + %0 = load <8 x i16>, <8 x i16>* @llvm_mips_addvi_h_ARG1 %1 = tail call <8 x i16> @llvm.mips.addvi.h(<8 x i16> %0, i32 14) store <8 x i16> %1, <8 x i16>* @llvm_mips_addvi_h_RES ret void @@ -47,7 +47,7 @@ declare <8 x i16> @llvm.mips.addvi.h(<8 x i16>, i32) nounwind define void @llvm_mips_addvi_w_test() nounwind { entry: - %0 = load <4 x i32>* @llvm_mips_addvi_w_ARG1 + %0 = load <4 x i32>, <4 x i32>* @llvm_mips_addvi_w_ARG1 %1 = tail call <4 x i32> @llvm.mips.addvi.w(<4 x i32> %0, i32 14) store <4 x i32> %1, <4 x i32>* @llvm_mips_addvi_w_RES ret void @@ -66,7 +66,7 @@ declare <4 x i32> @llvm.mips.addvi.w(<4 x i32>, i32) nounwind define void @llvm_mips_addvi_d_test() nounwind { entry: - %0 = load <2 x i64>* @llvm_mips_addvi_d_ARG1 + %0 = load <2 x i64>, <2 x i64>* @llvm_mips_addvi_d_ARG1 %1 = tail call <2 x i64> @llvm.mips.addvi.d(<2 x i64> %0, i32 14) store <2 x i64> %1, <2 x i64>* @llvm_mips_addvi_d_RES ret void diff --git a/test/CodeGen/Mips/msa/i5-b.ll b/test/CodeGen/Mips/msa/i5-b.ll index da6be669f0dda..40ab095f68097 100644 --- a/test/CodeGen/Mips/msa/i5-b.ll +++ b/test/CodeGen/Mips/msa/i5-b.ll @@ -9,7 +9,7 @@ define void @llvm_mips_bclri_b_test() nounwind { entry: - %0 = load <16 x i8>* @llvm_mips_bclri_b_ARG1 + %0 = load <16 x i8>, <16 x i8>* @llvm_mips_bclri_b_ARG1 %1 = tail call <16 x i8> @llvm.mips.bclri.b(<16 x i8> %0, i32 7) store <16 x i8> %1, <16 x i8>* @llvm_mips_bclri_b_RES ret void @@ -29,7 +29,7 @@ declare <16 x i8> @llvm.mips.bclri.b(<16 x i8>, i32) nounwind define void @llvm_mips_bclri_h_test() nounwind { entry: - %0 = load <8 x i16>* @llvm_mips_bclri_h_ARG1 + %0 = load <8 x i16>, <8 x i16>* @llvm_mips_bclri_h_ARG1 %1 = tail call <8 x i16> @llvm.mips.bclri.h(<8 x i16> %0, i32 7) store <8 x i16> %1, <8 x i16>* @llvm_mips_bclri_h_RES ret void @@ -48,7 +48,7 @@ declare <8 x i16> @llvm.mips.bclri.h(<8 x i16>, i32) nounwind define void @llvm_mips_bclri_w_test() nounwind { entry: - %0 = load <4 x i32>* @llvm_mips_bclri_w_ARG1 + %0 = load <4 x i32>, <4 x i32>* @llvm_mips_bclri_w_ARG1 %1 = tail call <4 x i32> @llvm.mips.bclri.w(<4 x i32> %0, i32 7) store <4 x i32> %1, <4 x i32>* @llvm_mips_bclri_w_RES ret void @@ -67,7 +67,7 @@ declare <4 x i32> @llvm.mips.bclri.w(<4 x i32>, i32) nounwind define void @llvm_mips_bclri_d_test() nounwind { entry: - %0 = load <2 x i64>* @llvm_mips_bclri_d_ARG1 + %0 = load <2 x i64>, <2 x i64>* @llvm_mips_bclri_d_ARG1 %1 = tail call <2 x i64> @llvm.mips.bclri.d(<2 x i64> %0, i32 7) store <2 x i64> %1, <2 x i64>* @llvm_mips_bclri_d_RES ret void @@ -87,8 +87,8 @@ declare <2 x i64> @llvm.mips.bclri.d(<2 x i64>, i32) nounwind define void @llvm_mips_binsli_b_test() nounwind { entry: - %0 = load <16 x i8>* @llvm_mips_binsli_b_ARG1 - %1 = load <16 x i8>* @llvm_mips_binsli_b_ARG2 + %0 = load <16 x i8>, <16 x i8>* @llvm_mips_binsli_b_ARG1 + %1 = load <16 x i8>, <16 x i8>* @llvm_mips_binsli_b_ARG2 %2 = tail call <16 x i8> @llvm.mips.binsli.b(<16 x i8> %0, <16 x i8> %1, i32 7) store <16 x i8> %2, <16 x i8>* @llvm_mips_binsli_b_RES ret void @@ -112,8 +112,8 @@ declare <16 x i8> @llvm.mips.binsli.b(<16 x i8>, <16 x i8>, i32) nounwind define void @llvm_mips_binsli_h_test() nounwind { entry: - %0 = load <8 x i16>* @llvm_mips_binsli_h_ARG1 - %1 = load <8 x i16>* @llvm_mips_binsli_h_ARG2 + %0 = load <8 x i16>, <8 x i16>* @llvm_mips_binsli_h_ARG1 + %1 = load <8 x i16>, <8 x i16>* @llvm_mips_binsli_h_ARG2 %2 = tail call <8 x i16> @llvm.mips.binsli.h(<8 x i16> %0, <8 x i16> %1, i32 7) store <8 x i16> %2, <8 x i16>* @llvm_mips_binsli_h_RES ret void @@ -137,8 +137,8 @@ declare <8 x i16> @llvm.mips.binsli.h(<8 x i16>, <8 x i16>, i32) nounwind define void @llvm_mips_binsli_w_test() nounwind { entry: - %0 = load <4 x i32>* @llvm_mips_binsli_w_ARG1 - %1 = load <4 x i32>* @llvm_mips_binsli_w_ARG2 + %0 = load <4 x i32>, <4 x i32>* @llvm_mips_binsli_w_ARG1 + %1 = load <4 x i32>, <4 x i32>* @llvm_mips_binsli_w_ARG2 %2 = tail call <4 x i32> @llvm.mips.binsli.w(<4 x i32> %0, <4 x i32> %1, i32 7) store <4 x i32> %2, <4 x i32>* @llvm_mips_binsli_w_RES ret void @@ -162,8 +162,8 @@ declare <4 x i32> @llvm.mips.binsli.w(<4 x i32>, <4 x i32>, i32) nounwind define void @llvm_mips_binsli_d_test() nounwind { entry: - %0 = load <2 x i64>* @llvm_mips_binsli_d_ARG1 - %1 = load <2 x i64>* @llvm_mips_binsli_d_ARG2 + %0 = load <2 x i64>, <2 x i64>* @llvm_mips_binsli_d_ARG1 + %1 = load <2 x i64>, <2 x i64>* @llvm_mips_binsli_d_ARG2 ; TODO: We use a particularly wide mask here to work around a legalization ; issue. If the mask doesn't fit within a 10-bit immediate, it gets ; legalized into a constant pool. We should add a test to cover the @@ -191,8 +191,8 @@ declare <2 x i64> @llvm.mips.binsli.d(<2 x i64>, <2 x i64>, i32) nounwind define void @llvm_mips_binsri_b_test() nounwind { entry: - %0 = load <16 x i8>* @llvm_mips_binsri_b_ARG1 - %1 = load <16 x i8>* @llvm_mips_binsri_b_ARG2 + %0 = load <16 x i8>, <16 x i8>* @llvm_mips_binsri_b_ARG1 + %1 = load <16 x i8>, <16 x i8>* @llvm_mips_binsri_b_ARG2 %2 = tail call <16 x i8> @llvm.mips.binsri.b(<16 x i8> %0, <16 x i8> %1, i32 7) store <16 x i8> %2, <16 x i8>* @llvm_mips_binsri_b_RES ret void @@ -216,8 +216,8 @@ declare <16 x i8> @llvm.mips.binsri.b(<16 x i8>, <16 x i8>, i32) nounwind define void @llvm_mips_binsri_h_test() nounwind { entry: - %0 = load <8 x i16>* @llvm_mips_binsri_h_ARG1 - %1 = load <8 x i16>* @llvm_mips_binsri_h_ARG2 + %0 = load <8 x i16>, <8 x i16>* @llvm_mips_binsri_h_ARG1 + %1 = load <8 x i16>, <8 x i16>* @llvm_mips_binsri_h_ARG2 %2 = tail call <8 x i16> @llvm.mips.binsri.h(<8 x i16> %0, <8 x i16> %1, i32 7) store <8 x i16> %2, <8 x i16>* @llvm_mips_binsri_h_RES ret void @@ -241,8 +241,8 @@ declare <8 x i16> @llvm.mips.binsri.h(<8 x i16>, <8 x i16>, i32) nounwind define void @llvm_mips_binsri_w_test() nounwind { entry: - %0 = load <4 x i32>* @llvm_mips_binsri_w_ARG1 - %1 = load <4 x i32>* @llvm_mips_binsri_w_ARG2 + %0 = load <4 x i32>, <4 x i32>* @llvm_mips_binsri_w_ARG1 + %1 = load <4 x i32>, <4 x i32>* @llvm_mips_binsri_w_ARG2 %2 = tail call <4 x i32> @llvm.mips.binsri.w(<4 x i32> %0, <4 x i32> %1, i32 7) store <4 x i32> %2, <4 x i32>* @llvm_mips_binsri_w_RES ret void @@ -266,8 +266,8 @@ declare <4 x i32> @llvm.mips.binsri.w(<4 x i32>, <4 x i32>, i32) nounwind define void @llvm_mips_binsri_d_test() nounwind { entry: - %0 = load <2 x i64>* @llvm_mips_binsri_d_ARG1 - %1 = load <2 x i64>* @llvm_mips_binsri_d_ARG2 + %0 = load <2 x i64>, <2 x i64>* @llvm_mips_binsri_d_ARG1 + %1 = load <2 x i64>, <2 x i64>* @llvm_mips_binsri_d_ARG2 %2 = tail call <2 x i64> @llvm.mips.binsri.d(<2 x i64> %0, <2 x i64> %1, i32 7) store <2 x i64> %2, <2 x i64>* @llvm_mips_binsri_d_RES ret void @@ -290,7 +290,7 @@ declare <2 x i64> @llvm.mips.binsri.d(<2 x i64>, <2 x i64>, i32) nounwind define void @llvm_mips_bnegi_b_test() nounwind { entry: - %0 = load <16 x i8>* @llvm_mips_bnegi_b_ARG1 + %0 = load <16 x i8>, <16 x i8>* @llvm_mips_bnegi_b_ARG1 %1 = tail call <16 x i8> @llvm.mips.bnegi.b(<16 x i8> %0, i32 7) store <16 x i8> %1, <16 x i8>* @llvm_mips_bnegi_b_RES ret void @@ -309,7 +309,7 @@ declare <16 x i8> @llvm.mips.bnegi.b(<16 x i8>, i32) nounwind define void @llvm_mips_bnegi_h_test() nounwind { entry: - %0 = load <8 x i16>* @llvm_mips_bnegi_h_ARG1 + %0 = load <8 x i16>, <8 x i16>* @llvm_mips_bnegi_h_ARG1 %1 = tail call <8 x i16> @llvm.mips.bnegi.h(<8 x i16> %0, i32 7) store <8 x i16> %1, <8 x i16>* @llvm_mips_bnegi_h_RES ret void @@ -328,7 +328,7 @@ declare <8 x i16> @llvm.mips.bnegi.h(<8 x i16>, i32) nounwind define void @llvm_mips_bnegi_w_test() nounwind { entry: - %0 = load <4 x i32>* @llvm_mips_bnegi_w_ARG1 + %0 = load <4 x i32>, <4 x i32>* @llvm_mips_bnegi_w_ARG1 %1 = tail call <4 x i32> @llvm.mips.bnegi.w(<4 x i32> %0, i32 7) store <4 x i32> %1, <4 x i32>* @llvm_mips_bnegi_w_RES ret void @@ -347,7 +347,7 @@ declare <4 x i32> @llvm.mips.bnegi.w(<4 x i32>, i32) nounwind define void @llvm_mips_bnegi_d_test() nounwind { entry: - %0 = load <2 x i64>* @llvm_mips_bnegi_d_ARG1 + %0 = load <2 x i64>, <2 x i64>* @llvm_mips_bnegi_d_ARG1 %1 = tail call <2 x i64> @llvm.mips.bnegi.d(<2 x i64> %0, i32 7) store <2 x i64> %1, <2 x i64>* @llvm_mips_bnegi_d_RES ret void @@ -366,7 +366,7 @@ declare <2 x i64> @llvm.mips.bnegi.d(<2 x i64>, i32) nounwind define void @llvm_mips_bseti_b_test() nounwind { entry: - %0 = load <16 x i8>* @llvm_mips_bseti_b_ARG1 + %0 = load <16 x i8>, <16 x i8>* @llvm_mips_bseti_b_ARG1 %1 = tail call <16 x i8> @llvm.mips.bseti.b(<16 x i8> %0, i32 7) store <16 x i8> %1, <16 x i8>* @llvm_mips_bseti_b_RES ret void @@ -385,7 +385,7 @@ declare <16 x i8> @llvm.mips.bseti.b(<16 x i8>, i32) nounwind define void @llvm_mips_bseti_h_test() nounwind { entry: - %0 = load <8 x i16>* @llvm_mips_bseti_h_ARG1 + %0 = load <8 x i16>, <8 x i16>* @llvm_mips_bseti_h_ARG1 %1 = tail call <8 x i16> @llvm.mips.bseti.h(<8 x i16> %0, i32 7) store <8 x i16> %1, <8 x i16>* @llvm_mips_bseti_h_RES ret void @@ -404,7 +404,7 @@ declare <8 x i16> @llvm.mips.bseti.h(<8 x i16>, i32) nounwind define void @llvm_mips_bseti_w_test() nounwind { entry: - %0 = load <4 x i32>* @llvm_mips_bseti_w_ARG1 + %0 = load <4 x i32>, <4 x i32>* @llvm_mips_bseti_w_ARG1 %1 = tail call <4 x i32> @llvm.mips.bseti.w(<4 x i32> %0, i32 7) store <4 x i32> %1, <4 x i32>* @llvm_mips_bseti_w_RES ret void @@ -423,7 +423,7 @@ declare <4 x i32> @llvm.mips.bseti.w(<4 x i32>, i32) nounwind define void @llvm_mips_bseti_d_test() nounwind { entry: - %0 = load <2 x i64>* @llvm_mips_bseti_d_ARG1 + %0 = load <2 x i64>, <2 x i64>* @llvm_mips_bseti_d_ARG1 %1 = tail call <2 x i64> @llvm.mips.bseti.d(<2 x i64> %0, i32 7) store <2 x i64> %1, <2 x i64>* @llvm_mips_bseti_d_RES ret void diff --git a/test/CodeGen/Mips/msa/i5-c.ll b/test/CodeGen/Mips/msa/i5-c.ll index bf1578f30f326..815825013ea19 100644 --- a/test/CodeGen/Mips/msa/i5-c.ll +++ b/test/CodeGen/Mips/msa/i5-c.ll @@ -9,7 +9,7 @@ define void @llvm_mips_ceqi_b_test() nounwind { entry: - %0 = load <16 x i8>* @llvm_mips_ceqi_b_ARG1 + %0 = load <16 x i8>, <16 x i8>* @llvm_mips_ceqi_b_ARG1 %1 = tail call <16 x i8> @llvm.mips.ceqi.b(<16 x i8> %0, i32 14) store <16 x i8> %1, <16 x i8>* @llvm_mips_ceqi_b_RES ret void @@ -28,7 +28,7 @@ declare <16 x i8> @llvm.mips.ceqi.b(<16 x i8>, i32) nounwind define void @llvm_mips_ceqi_h_test() nounwind { entry: - %0 = load <8 x i16>* @llvm_mips_ceqi_h_ARG1 + %0 = load <8 x i16>, <8 x i16>* @llvm_mips_ceqi_h_ARG1 %1 = tail call <8 x i16> @llvm.mips.ceqi.h(<8 x i16> %0, i32 14) store <8 x i16> %1, <8 x i16>* @llvm_mips_ceqi_h_RES ret void @@ -47,7 +47,7 @@ declare <8 x i16> @llvm.mips.ceqi.h(<8 x i16>, i32) nounwind define void @llvm_mips_ceqi_w_test() nounwind { entry: - %0 = load <4 x i32>* @llvm_mips_ceqi_w_ARG1 + %0 = load <4 x i32>, <4 x i32>* @llvm_mips_ceqi_w_ARG1 %1 = tail call <4 x i32> @llvm.mips.ceqi.w(<4 x i32> %0, i32 14) store <4 x i32> %1, <4 x i32>* @llvm_mips_ceqi_w_RES ret void @@ -66,7 +66,7 @@ declare <4 x i32> @llvm.mips.ceqi.w(<4 x i32>, i32) nounwind define void @llvm_mips_ceqi_d_test() nounwind { entry: - %0 = load <2 x i64>* @llvm_mips_ceqi_d_ARG1 + %0 = load <2 x i64>, <2 x i64>* @llvm_mips_ceqi_d_ARG1 %1 = tail call <2 x i64> @llvm.mips.ceqi.d(<2 x i64> %0, i32 14) store <2 x i64> %1, <2 x i64>* @llvm_mips_ceqi_d_RES ret void @@ -85,7 +85,7 @@ declare <2 x i64> @llvm.mips.ceqi.d(<2 x i64>, i32) nounwind define void @llvm_mips_clei_s_b_test() nounwind { entry: - %0 = load <16 x i8>* @llvm_mips_clei_s_b_ARG1 + %0 = load <16 x i8>, <16 x i8>* @llvm_mips_clei_s_b_ARG1 %1 = tail call <16 x i8> @llvm.mips.clei.s.b(<16 x i8> %0, i32 14) store <16 x i8> %1, <16 x i8>* @llvm_mips_clei_s_b_RES ret void @@ -104,7 +104,7 @@ declare <16 x i8> @llvm.mips.clei.s.b(<16 x i8>, i32) nounwind define void @llvm_mips_clei_s_h_test() nounwind { entry: - %0 = load <8 x i16>* @llvm_mips_clei_s_h_ARG1 + %0 = load <8 x i16>, <8 x i16>* @llvm_mips_clei_s_h_ARG1 %1 = tail call <8 x i16> @llvm.mips.clei.s.h(<8 x i16> %0, i32 14) store <8 x i16> %1, <8 x i16>* @llvm_mips_clei_s_h_RES ret void @@ -123,7 +123,7 @@ declare <8 x i16> @llvm.mips.clei.s.h(<8 x i16>, i32) nounwind define void @llvm_mips_clei_s_w_test() nounwind { entry: - %0 = load <4 x i32>* @llvm_mips_clei_s_w_ARG1 + %0 = load <4 x i32>, <4 x i32>* @llvm_mips_clei_s_w_ARG1 %1 = tail call <4 x i32> @llvm.mips.clei.s.w(<4 x i32> %0, i32 14) store <4 x i32> %1, <4 x i32>* @llvm_mips_clei_s_w_RES ret void @@ -142,7 +142,7 @@ declare <4 x i32> @llvm.mips.clei.s.w(<4 x i32>, i32) nounwind define void @llvm_mips_clei_s_d_test() nounwind { entry: - %0 = load <2 x i64>* @llvm_mips_clei_s_d_ARG1 + %0 = load <2 x i64>, <2 x i64>* @llvm_mips_clei_s_d_ARG1 %1 = tail call <2 x i64> @llvm.mips.clei.s.d(<2 x i64> %0, i32 14) store <2 x i64> %1, <2 x i64>* @llvm_mips_clei_s_d_RES ret void @@ -161,7 +161,7 @@ declare <2 x i64> @llvm.mips.clei.s.d(<2 x i64>, i32) nounwind define void @llvm_mips_clei_u_b_test() nounwind { entry: - %0 = load <16 x i8>* @llvm_mips_clei_u_b_ARG1 + %0 = load <16 x i8>, <16 x i8>* @llvm_mips_clei_u_b_ARG1 %1 = tail call <16 x i8> @llvm.mips.clei.u.b(<16 x i8> %0, i32 14) store <16 x i8> %1, <16 x i8>* @llvm_mips_clei_u_b_RES ret void @@ -180,7 +180,7 @@ declare <16 x i8> @llvm.mips.clei.u.b(<16 x i8>, i32) nounwind define void @llvm_mips_clei_u_h_test() nounwind { entry: - %0 = load <8 x i16>* @llvm_mips_clei_u_h_ARG1 + %0 = load <8 x i16>, <8 x i16>* @llvm_mips_clei_u_h_ARG1 %1 = tail call <8 x i16> @llvm.mips.clei.u.h(<8 x i16> %0, i32 14) store <8 x i16> %1, <8 x i16>* @llvm_mips_clei_u_h_RES ret void @@ -199,7 +199,7 @@ declare <8 x i16> @llvm.mips.clei.u.h(<8 x i16>, i32) nounwind define void @llvm_mips_clei_u_w_test() nounwind { entry: - %0 = load <4 x i32>* @llvm_mips_clei_u_w_ARG1 + %0 = load <4 x i32>, <4 x i32>* @llvm_mips_clei_u_w_ARG1 %1 = tail call <4 x i32> @llvm.mips.clei.u.w(<4 x i32> %0, i32 14) store <4 x i32> %1, <4 x i32>* @llvm_mips_clei_u_w_RES ret void @@ -218,7 +218,7 @@ declare <4 x i32> @llvm.mips.clei.u.w(<4 x i32>, i32) nounwind define void @llvm_mips_clei_u_d_test() nounwind { entry: - %0 = load <2 x i64>* @llvm_mips_clei_u_d_ARG1 + %0 = load <2 x i64>, <2 x i64>* @llvm_mips_clei_u_d_ARG1 %1 = tail call <2 x i64> @llvm.mips.clei.u.d(<2 x i64> %0, i32 14) store <2 x i64> %1, <2 x i64>* @llvm_mips_clei_u_d_RES ret void @@ -237,7 +237,7 @@ declare <2 x i64> @llvm.mips.clei.u.d(<2 x i64>, i32) nounwind define void @llvm_mips_clti_s_b_test() nounwind { entry: - %0 = load <16 x i8>* @llvm_mips_clti_s_b_ARG1 + %0 = load <16 x i8>, <16 x i8>* @llvm_mips_clti_s_b_ARG1 %1 = tail call <16 x i8> @llvm.mips.clti.s.b(<16 x i8> %0, i32 14) store <16 x i8> %1, <16 x i8>* @llvm_mips_clti_s_b_RES ret void @@ -256,7 +256,7 @@ declare <16 x i8> @llvm.mips.clti.s.b(<16 x i8>, i32) nounwind define void @llvm_mips_clti_s_h_test() nounwind { entry: - %0 = load <8 x i16>* @llvm_mips_clti_s_h_ARG1 + %0 = load <8 x i16>, <8 x i16>* @llvm_mips_clti_s_h_ARG1 %1 = tail call <8 x i16> @llvm.mips.clti.s.h(<8 x i16> %0, i32 14) store <8 x i16> %1, <8 x i16>* @llvm_mips_clti_s_h_RES ret void @@ -275,7 +275,7 @@ declare <8 x i16> @llvm.mips.clti.s.h(<8 x i16>, i32) nounwind define void @llvm_mips_clti_s_w_test() nounwind { entry: - %0 = load <4 x i32>* @llvm_mips_clti_s_w_ARG1 + %0 = load <4 x i32>, <4 x i32>* @llvm_mips_clti_s_w_ARG1 %1 = tail call <4 x i32> @llvm.mips.clti.s.w(<4 x i32> %0, i32 14) store <4 x i32> %1, <4 x i32>* @llvm_mips_clti_s_w_RES ret void @@ -294,7 +294,7 @@ declare <4 x i32> @llvm.mips.clti.s.w(<4 x i32>, i32) nounwind define void @llvm_mips_clti_s_d_test() nounwind { entry: - %0 = load <2 x i64>* @llvm_mips_clti_s_d_ARG1 + %0 = load <2 x i64>, <2 x i64>* @llvm_mips_clti_s_d_ARG1 %1 = tail call <2 x i64> @llvm.mips.clti.s.d(<2 x i64> %0, i32 14) store <2 x i64> %1, <2 x i64>* @llvm_mips_clti_s_d_RES ret void @@ -313,7 +313,7 @@ declare <2 x i64> @llvm.mips.clti.s.d(<2 x i64>, i32) nounwind define void @llvm_mips_clti_u_b_test() nounwind { entry: - %0 = load <16 x i8>* @llvm_mips_clti_u_b_ARG1 + %0 = load <16 x i8>, <16 x i8>* @llvm_mips_clti_u_b_ARG1 %1 = tail call <16 x i8> @llvm.mips.clti.u.b(<16 x i8> %0, i32 14) store <16 x i8> %1, <16 x i8>* @llvm_mips_clti_u_b_RES ret void @@ -332,7 +332,7 @@ declare <16 x i8> @llvm.mips.clti.u.b(<16 x i8>, i32) nounwind define void @llvm_mips_clti_u_h_test() nounwind { entry: - %0 = load <8 x i16>* @llvm_mips_clti_u_h_ARG1 + %0 = load <8 x i16>, <8 x i16>* @llvm_mips_clti_u_h_ARG1 %1 = tail call <8 x i16> @llvm.mips.clti.u.h(<8 x i16> %0, i32 14) store <8 x i16> %1, <8 x i16>* @llvm_mips_clti_u_h_RES ret void @@ -351,7 +351,7 @@ declare <8 x i16> @llvm.mips.clti.u.h(<8 x i16>, i32) nounwind define void @llvm_mips_clti_u_w_test() nounwind { entry: - %0 = load <4 x i32>* @llvm_mips_clti_u_w_ARG1 + %0 = load <4 x i32>, <4 x i32>* @llvm_mips_clti_u_w_ARG1 %1 = tail call <4 x i32> @llvm.mips.clti.u.w(<4 x i32> %0, i32 14) store <4 x i32> %1, <4 x i32>* @llvm_mips_clti_u_w_RES ret void @@ -370,7 +370,7 @@ declare <4 x i32> @llvm.mips.clti.u.w(<4 x i32>, i32) nounwind define void @llvm_mips_clti_u_d_test() nounwind { entry: - %0 = load <2 x i64>* @llvm_mips_clti_u_d_ARG1 + %0 = load <2 x i64>, <2 x i64>* @llvm_mips_clti_u_d_ARG1 %1 = tail call <2 x i64> @llvm.mips.clti.u.d(<2 x i64> %0, i32 14) store <2 x i64> %1, <2 x i64>* @llvm_mips_clti_u_d_RES ret void diff --git a/test/CodeGen/Mips/msa/i5-m.ll b/test/CodeGen/Mips/msa/i5-m.ll index 27663494324d9..ba6e9d2384a77 100644 --- a/test/CodeGen/Mips/msa/i5-m.ll +++ b/test/CodeGen/Mips/msa/i5-m.ll @@ -9,7 +9,7 @@ define void @llvm_mips_maxi_s_b_test() nounwind { entry: - %0 = load <16 x i8>* @llvm_mips_maxi_s_b_ARG1 + %0 = load <16 x i8>, <16 x i8>* @llvm_mips_maxi_s_b_ARG1 %1 = tail call <16 x i8> @llvm.mips.maxi.s.b(<16 x i8> %0, i32 14) store <16 x i8> %1, <16 x i8>* @llvm_mips_maxi_s_b_RES ret void @@ -28,7 +28,7 @@ declare <16 x i8> @llvm.mips.maxi.s.b(<16 x i8>, i32) nounwind define void @llvm_mips_maxi_s_h_test() nounwind { entry: - %0 = load <8 x i16>* @llvm_mips_maxi_s_h_ARG1 + %0 = load <8 x i16>, <8 x i16>* @llvm_mips_maxi_s_h_ARG1 %1 = tail call <8 x i16> @llvm.mips.maxi.s.h(<8 x i16> %0, i32 14) store <8 x i16> %1, <8 x i16>* @llvm_mips_maxi_s_h_RES ret void @@ -47,7 +47,7 @@ declare <8 x i16> @llvm.mips.maxi.s.h(<8 x i16>, i32) nounwind define void @llvm_mips_maxi_s_w_test() nounwind { entry: - %0 = load <4 x i32>* @llvm_mips_maxi_s_w_ARG1 + %0 = load <4 x i32>, <4 x i32>* @llvm_mips_maxi_s_w_ARG1 %1 = tail call <4 x i32> @llvm.mips.maxi.s.w(<4 x i32> %0, i32 14) store <4 x i32> %1, <4 x i32>* @llvm_mips_maxi_s_w_RES ret void @@ -66,7 +66,7 @@ declare <4 x i32> @llvm.mips.maxi.s.w(<4 x i32>, i32) nounwind define void @llvm_mips_maxi_s_d_test() nounwind { entry: - %0 = load <2 x i64>* @llvm_mips_maxi_s_d_ARG1 + %0 = load <2 x i64>, <2 x i64>* @llvm_mips_maxi_s_d_ARG1 %1 = tail call <2 x i64> @llvm.mips.maxi.s.d(<2 x i64> %0, i32 14) store <2 x i64> %1, <2 x i64>* @llvm_mips_maxi_s_d_RES ret void @@ -85,7 +85,7 @@ declare <2 x i64> @llvm.mips.maxi.s.d(<2 x i64>, i32) nounwind define void @llvm_mips_maxi_u_b_test() nounwind { entry: - %0 = load <16 x i8>* @llvm_mips_maxi_u_b_ARG1 + %0 = load <16 x i8>, <16 x i8>* @llvm_mips_maxi_u_b_ARG1 %1 = tail call <16 x i8> @llvm.mips.maxi.u.b(<16 x i8> %0, i32 14) store <16 x i8> %1, <16 x i8>* @llvm_mips_maxi_u_b_RES ret void @@ -104,7 +104,7 @@ declare <16 x i8> @llvm.mips.maxi.u.b(<16 x i8>, i32) nounwind define void @llvm_mips_maxi_u_h_test() nounwind { entry: - %0 = load <8 x i16>* @llvm_mips_maxi_u_h_ARG1 + %0 = load <8 x i16>, <8 x i16>* @llvm_mips_maxi_u_h_ARG1 %1 = tail call <8 x i16> @llvm.mips.maxi.u.h(<8 x i16> %0, i32 14) store <8 x i16> %1, <8 x i16>* @llvm_mips_maxi_u_h_RES ret void @@ -123,7 +123,7 @@ declare <8 x i16> @llvm.mips.maxi.u.h(<8 x i16>, i32) nounwind define void @llvm_mips_maxi_u_w_test() nounwind { entry: - %0 = load <4 x i32>* @llvm_mips_maxi_u_w_ARG1 + %0 = load <4 x i32>, <4 x i32>* @llvm_mips_maxi_u_w_ARG1 %1 = tail call <4 x i32> @llvm.mips.maxi.u.w(<4 x i32> %0, i32 14) store <4 x i32> %1, <4 x i32>* @llvm_mips_maxi_u_w_RES ret void @@ -142,7 +142,7 @@ declare <4 x i32> @llvm.mips.maxi.u.w(<4 x i32>, i32) nounwind define void @llvm_mips_maxi_u_d_test() nounwind { entry: - %0 = load <2 x i64>* @llvm_mips_maxi_u_d_ARG1 + %0 = load <2 x i64>, <2 x i64>* @llvm_mips_maxi_u_d_ARG1 %1 = tail call <2 x i64> @llvm.mips.maxi.u.d(<2 x i64> %0, i32 14) store <2 x i64> %1, <2 x i64>* @llvm_mips_maxi_u_d_RES ret void @@ -161,7 +161,7 @@ declare <2 x i64> @llvm.mips.maxi.u.d(<2 x i64>, i32) nounwind define void @llvm_mips_mini_s_b_test() nounwind { entry: - %0 = load <16 x i8>* @llvm_mips_mini_s_b_ARG1 + %0 = load <16 x i8>, <16 x i8>* @llvm_mips_mini_s_b_ARG1 %1 = tail call <16 x i8> @llvm.mips.mini.s.b(<16 x i8> %0, i32 14) store <16 x i8> %1, <16 x i8>* @llvm_mips_mini_s_b_RES ret void @@ -180,7 +180,7 @@ declare <16 x i8> @llvm.mips.mini.s.b(<16 x i8>, i32) nounwind define void @llvm_mips_mini_s_h_test() nounwind { entry: - %0 = load <8 x i16>* @llvm_mips_mini_s_h_ARG1 + %0 = load <8 x i16>, <8 x i16>* @llvm_mips_mini_s_h_ARG1 %1 = tail call <8 x i16> @llvm.mips.mini.s.h(<8 x i16> %0, i32 14) store <8 x i16> %1, <8 x i16>* @llvm_mips_mini_s_h_RES ret void @@ -199,7 +199,7 @@ declare <8 x i16> @llvm.mips.mini.s.h(<8 x i16>, i32) nounwind define void @llvm_mips_mini_s_w_test() nounwind { entry: - %0 = load <4 x i32>* @llvm_mips_mini_s_w_ARG1 + %0 = load <4 x i32>, <4 x i32>* @llvm_mips_mini_s_w_ARG1 %1 = tail call <4 x i32> @llvm.mips.mini.s.w(<4 x i32> %0, i32 14) store <4 x i32> %1, <4 x i32>* @llvm_mips_mini_s_w_RES ret void @@ -218,7 +218,7 @@ declare <4 x i32> @llvm.mips.mini.s.w(<4 x i32>, i32) nounwind define void @llvm_mips_mini_s_d_test() nounwind { entry: - %0 = load <2 x i64>* @llvm_mips_mini_s_d_ARG1 + %0 = load <2 x i64>, <2 x i64>* @llvm_mips_mini_s_d_ARG1 %1 = tail call <2 x i64> @llvm.mips.mini.s.d(<2 x i64> %0, i32 14) store <2 x i64> %1, <2 x i64>* @llvm_mips_mini_s_d_RES ret void @@ -237,7 +237,7 @@ declare <2 x i64> @llvm.mips.mini.s.d(<2 x i64>, i32) nounwind define void @llvm_mips_mini_u_b_test() nounwind { entry: - %0 = load <16 x i8>* @llvm_mips_mini_u_b_ARG1 + %0 = load <16 x i8>, <16 x i8>* @llvm_mips_mini_u_b_ARG1 %1 = tail call <16 x i8> @llvm.mips.mini.u.b(<16 x i8> %0, i32 14) store <16 x i8> %1, <16 x i8>* @llvm_mips_mini_u_b_RES ret void @@ -256,7 +256,7 @@ declare <16 x i8> @llvm.mips.mini.u.b(<16 x i8>, i32) nounwind define void @llvm_mips_mini_u_h_test() nounwind { entry: - %0 = load <8 x i16>* @llvm_mips_mini_u_h_ARG1 + %0 = load <8 x i16>, <8 x i16>* @llvm_mips_mini_u_h_ARG1 %1 = tail call <8 x i16> @llvm.mips.mini.u.h(<8 x i16> %0, i32 14) store <8 x i16> %1, <8 x i16>* @llvm_mips_mini_u_h_RES ret void @@ -275,7 +275,7 @@ declare <8 x i16> @llvm.mips.mini.u.h(<8 x i16>, i32) nounwind define void @llvm_mips_mini_u_w_test() nounwind { entry: - %0 = load <4 x i32>* @llvm_mips_mini_u_w_ARG1 + %0 = load <4 x i32>, <4 x i32>* @llvm_mips_mini_u_w_ARG1 %1 = tail call <4 x i32> @llvm.mips.mini.u.w(<4 x i32> %0, i32 14) store <4 x i32> %1, <4 x i32>* @llvm_mips_mini_u_w_RES ret void @@ -294,7 +294,7 @@ declare <4 x i32> @llvm.mips.mini.u.w(<4 x i32>, i32) nounwind define void @llvm_mips_mini_u_d_test() nounwind { entry: - %0 = load <2 x i64>* @llvm_mips_mini_u_d_ARG1 + %0 = load <2 x i64>, <2 x i64>* @llvm_mips_mini_u_d_ARG1 %1 = tail call <2 x i64> @llvm.mips.mini.u.d(<2 x i64> %0, i32 14) store <2 x i64> %1, <2 x i64>* @llvm_mips_mini_u_d_RES ret void diff --git a/test/CodeGen/Mips/msa/i5-s.ll b/test/CodeGen/Mips/msa/i5-s.ll index 184172f63b852..db331b1476ce5 100644 --- a/test/CodeGen/Mips/msa/i5-s.ll +++ b/test/CodeGen/Mips/msa/i5-s.ll @@ -9,7 +9,7 @@ define void @llvm_mips_subvi_b_test() nounwind { entry: - %0 = load <16 x i8>* @llvm_mips_subvi_b_ARG1 + %0 = load <16 x i8>, <16 x i8>* @llvm_mips_subvi_b_ARG1 %1 = tail call <16 x i8> @llvm.mips.subvi.b(<16 x i8> %0, i32 14) store <16 x i8> %1, <16 x i8>* @llvm_mips_subvi_b_RES ret void @@ -28,7 +28,7 @@ declare <16 x i8> @llvm.mips.subvi.b(<16 x i8>, i32) nounwind define void @llvm_mips_subvi_h_test() nounwind { entry: - %0 = load <8 x i16>* @llvm_mips_subvi_h_ARG1 + %0 = load <8 x i16>, <8 x i16>* @llvm_mips_subvi_h_ARG1 %1 = tail call <8 x i16> @llvm.mips.subvi.h(<8 x i16> %0, i32 14) store <8 x i16> %1, <8 x i16>* @llvm_mips_subvi_h_RES ret void @@ -47,7 +47,7 @@ declare <8 x i16> @llvm.mips.subvi.h(<8 x i16>, i32) nounwind define void @llvm_mips_subvi_w_test() nounwind { entry: - %0 = load <4 x i32>* @llvm_mips_subvi_w_ARG1 + %0 = load <4 x i32>, <4 x i32>* @llvm_mips_subvi_w_ARG1 %1 = tail call <4 x i32> @llvm.mips.subvi.w(<4 x i32> %0, i32 14) store <4 x i32> %1, <4 x i32>* @llvm_mips_subvi_w_RES ret void @@ -66,7 +66,7 @@ declare <4 x i32> @llvm.mips.subvi.w(<4 x i32>, i32) nounwind define void @llvm_mips_subvi_d_test() nounwind { entry: - %0 = load <2 x i64>* @llvm_mips_subvi_d_ARG1 + %0 = load <2 x i64>, <2 x i64>* @llvm_mips_subvi_d_ARG1 %1 = tail call <2 x i64> @llvm.mips.subvi.d(<2 x i64> %0, i32 14) store <2 x i64> %1, <2 x i64>* @llvm_mips_subvi_d_RES ret void diff --git a/test/CodeGen/Mips/msa/i5_ld_st.ll b/test/CodeGen/Mips/msa/i5_ld_st.ll index 7cc55f2904be7..991bb8436b339 100644 --- a/test/CodeGen/Mips/msa/i5_ld_st.ll +++ b/test/CodeGen/Mips/msa/i5_ld_st.ll @@ -81,7 +81,7 @@ declare <2 x i64> @llvm.mips.ld.d(i8*, i32) nounwind define void @llvm_mips_st_b_test() nounwind { entry: - %0 = load <16 x i8>* @llvm_mips_st_b_ARG + %0 = load <16 x i8>, <16 x i8>* @llvm_mips_st_b_ARG %1 = bitcast <16 x i8>* @llvm_mips_st_b_RES to i8* tail call void @llvm.mips.st.b(<16 x i8> %0, i8* %1, i32 16) ret void @@ -99,7 +99,7 @@ declare void @llvm.mips.st.b(<16 x i8>, i8*, i32) nounwind define void @llvm_mips_st_h_test() nounwind { entry: - %0 = load <8 x i16>* @llvm_mips_st_h_ARG + %0 = load <8 x i16>, <8 x i16>* @llvm_mips_st_h_ARG %1 = bitcast <8 x i16>* @llvm_mips_st_h_RES to i8* tail call void @llvm.mips.st.h(<8 x i16> %0, i8* %1, i32 16) ret void @@ -117,7 +117,7 @@ declare void @llvm.mips.st.h(<8 x i16>, i8*, i32) nounwind define void @llvm_mips_st_w_test() nounwind { entry: - %0 = load <4 x i32>* @llvm_mips_st_w_ARG + %0 = load <4 x i32>, <4 x i32>* @llvm_mips_st_w_ARG %1 = bitcast <4 x i32>* @llvm_mips_st_w_RES to i8* tail call void @llvm.mips.st.w(<4 x i32> %0, i8* %1, i32 16) ret void @@ -135,7 +135,7 @@ declare void @llvm.mips.st.w(<4 x i32>, i8*, i32) nounwind define void @llvm_mips_st_d_test() nounwind { entry: - %0 = load <2 x i64>* @llvm_mips_st_d_ARG + %0 = load <2 x i64>, <2 x i64>* @llvm_mips_st_d_ARG %1 = bitcast <2 x i64>* @llvm_mips_st_d_RES to i8* tail call void @llvm.mips.st.d(<2 x i64> %0, i8* %1, i32 16) ret void diff --git a/test/CodeGen/Mips/msa/i8.ll b/test/CodeGen/Mips/msa/i8.ll index d2931a72feaab..4af9c588fdef8 100644 --- a/test/CodeGen/Mips/msa/i8.ll +++ b/test/CodeGen/Mips/msa/i8.ll @@ -8,7 +8,7 @@ define void @llvm_mips_andi_b_test() nounwind { entry: - %0 = load <16 x i8>* @llvm_mips_andi_b_ARG1 + %0 = load <16 x i8>, <16 x i8>* @llvm_mips_andi_b_ARG1 %1 = tail call <16 x i8> @llvm.mips.andi.b(<16 x i8> %0, i32 25) store <16 x i8> %1, <16 x i8>* @llvm_mips_andi_b_RES ret void @@ -28,8 +28,8 @@ declare <16 x i8> @llvm.mips.andi.b(<16 x i8>, i32) nounwind define void @llvm_mips_bmnzi_b_test() nounwind { entry: - %0 = load <16 x i8>* @llvm_mips_bmnzi_b_ARG1 - %1 = load <16 x i8>* @llvm_mips_bmnzi_b_ARG2 + %0 = load <16 x i8>, <16 x i8>* @llvm_mips_bmnzi_b_ARG1 + %1 = load <16 x i8>, <16 x i8>* @llvm_mips_bmnzi_b_ARG2 %2 = tail call <16 x i8> @llvm.mips.bmnzi.b(<16 x i8> %0, <16 x i8> %1, i32 25) store <16 x i8> %2, <16 x i8>* @llvm_mips_bmnzi_b_RES ret void @@ -52,8 +52,8 @@ declare <16 x i8> @llvm.mips.bmnzi.b(<16 x i8>, <16 x i8>, i32) nounwind define void @llvm_mips_bmzi_b_test() nounwind { entry: - %0 = load <16 x i8>* @llvm_mips_bmzi_b_ARG1 - %1 = load <16 x i8>* @llvm_mips_bmzi_b_ARG2 + %0 = load <16 x i8>, <16 x i8>* @llvm_mips_bmzi_b_ARG1 + %1 = load <16 x i8>, <16 x i8>* @llvm_mips_bmzi_b_ARG2 %2 = tail call <16 x i8> @llvm.mips.bmzi.b(<16 x i8> %0, <16 x i8> %1, i32 25) store <16 x i8> %2, <16 x i8>* @llvm_mips_bmzi_b_RES ret void @@ -77,8 +77,8 @@ declare <16 x i8> @llvm.mips.bmzi.b(<16 x i8>, <16 x i8>, i32) nounwind define void @llvm_mips_bseli_b_test() nounwind { entry: - %0 = load <16 x i8>* @llvm_mips_bseli_b_ARG1 - %1 = load <16 x i8>* @llvm_mips_bseli_b_ARG2 + %0 = load <16 x i8>, <16 x i8>* @llvm_mips_bseli_b_ARG1 + %1 = load <16 x i8>, <16 x i8>* @llvm_mips_bseli_b_ARG2 %2 = tail call <16 x i8> @llvm.mips.bseli.b(<16 x i8> %0, <16 x i8> %1, i32 25) store <16 x i8> %2, <16 x i8>* @llvm_mips_bseli_b_RES ret void @@ -100,7 +100,7 @@ declare <16 x i8> @llvm.mips.bseli.b(<16 x i8>, <16 x i8>, i32) nounwind define void @llvm_mips_nori_b_test() nounwind { entry: - %0 = load <16 x i8>* @llvm_mips_nori_b_ARG1 + %0 = load <16 x i8>, <16 x i8>* @llvm_mips_nori_b_ARG1 %1 = tail call <16 x i8> @llvm.mips.nori.b(<16 x i8> %0, i32 25) store <16 x i8> %1, <16 x i8>* @llvm_mips_nori_b_RES ret void @@ -119,7 +119,7 @@ declare <16 x i8> @llvm.mips.nori.b(<16 x i8>, i32) nounwind define void @llvm_mips_ori_b_test() nounwind { entry: - %0 = load <16 x i8>* @llvm_mips_ori_b_ARG1 + %0 = load <16 x i8>, <16 x i8>* @llvm_mips_ori_b_ARG1 %1 = tail call <16 x i8> @llvm.mips.ori.b(<16 x i8> %0, i32 25) store <16 x i8> %1, <16 x i8>* @llvm_mips_ori_b_RES ret void @@ -138,7 +138,7 @@ declare <16 x i8> @llvm.mips.ori.b(<16 x i8>, i32) nounwind define void @llvm_mips_shf_b_test() nounwind { entry: - %0 = load <16 x i8>* @llvm_mips_shf_b_ARG1 + %0 = load <16 x i8>, <16 x i8>* @llvm_mips_shf_b_ARG1 %1 = tail call <16 x i8> @llvm.mips.shf.b(<16 x i8> %0, i32 25) store <16 x i8> %1, <16 x i8>* @llvm_mips_shf_b_RES ret void @@ -157,7 +157,7 @@ declare <16 x i8> @llvm.mips.shf.b(<16 x i8>, i32) nounwind define void @llvm_mips_shf_h_test() nounwind { entry: - %0 = load <8 x i16>* @llvm_mips_shf_h_ARG1 + %0 = load <8 x i16>, <8 x i16>* @llvm_mips_shf_h_ARG1 %1 = tail call <8 x i16> @llvm.mips.shf.h(<8 x i16> %0, i32 25) store <8 x i16> %1, <8 x i16>* @llvm_mips_shf_h_RES ret void @@ -176,7 +176,7 @@ declare <8 x i16> @llvm.mips.shf.h(<8 x i16>, i32) nounwind define void @llvm_mips_shf_w_test() nounwind { entry: - %0 = load <4 x i32>* @llvm_mips_shf_w_ARG1 + %0 = load <4 x i32>, <4 x i32>* @llvm_mips_shf_w_ARG1 %1 = tail call <4 x i32> @llvm.mips.shf.w(<4 x i32> %0, i32 25) store <4 x i32> %1, <4 x i32>* @llvm_mips_shf_w_RES ret void @@ -195,7 +195,7 @@ declare <4 x i32> @llvm.mips.shf.w(<4 x i32>, i32) nounwind define void @llvm_mips_xori_b_test() nounwind { entry: - %0 = load <16 x i8>* @llvm_mips_xori_b_ARG1 + %0 = load <16 x i8>, <16 x i8>* @llvm_mips_xori_b_ARG1 %1 = tail call <16 x i8> @llvm.mips.xori.b(<16 x i8> %0, i32 25) store <16 x i8> %1, <16 x i8>* @llvm_mips_xori_b_RES ret void diff --git a/test/CodeGen/Mips/msa/inline-asm.ll b/test/CodeGen/Mips/msa/inline-asm.ll index 4a34273f3c00c..85da87b6f8a72 100644 --- a/test/CodeGen/Mips/msa/inline-asm.ll +++ b/test/CodeGen/Mips/msa/inline-asm.ll @@ -16,7 +16,7 @@ entry: define void @test2() nounwind { entry: ; CHECK-LABEL: test2: - %0 = load <4 x i32>* @v4i32_r + %0 = load <4 x i32>, <4 x i32>* @v4i32_r %1 = call <4 x i32> asm "addvi.w ${0:w}, ${1:w}, 1", "=f,f"(<4 x i32> %0) ; CHECK: addvi.w $w{{[1-3]?[0-9]}}, $w{{[1-3]?[0-9]}}, 1 store <4 x i32> %1, <4 x i32>* @v4i32_r @@ -26,7 +26,7 @@ entry: define void @test3() nounwind { entry: ; CHECK-LABEL: test3: - %0 = load <4 x i32>* @v4i32_r + %0 = load <4 x i32>, <4 x i32>* @v4i32_r %1 = call <4 x i32> asm sideeffect "addvi.w ${0:w}, ${1:w}, 1", "=f,f,~{$w0}"(<4 x i32> %0) ; CHECK: addvi.w $w{{([1-9]|[1-3][0-9])}}, $w{{([1-9]|[1-3][0-9])}}, 1 store <4 x i32> %1, <4 x i32>* @v4i32_r diff --git a/test/CodeGen/Mips/msa/llvm-stress-s1704963983.ll b/test/CodeGen/Mips/msa/llvm-stress-s1704963983.ll index 4beaaa9c18415..beb361bc9f3d8 100644 --- a/test/CodeGen/Mips/msa/llvm-stress-s1704963983.ll +++ b/test/CodeGen/Mips/msa/llvm-stress-s1704963983.ll @@ -14,7 +14,7 @@ BB: %A2 = alloca <1 x double> %A1 = alloca double %A = alloca i32 - %L = load i8* %0 + %L = load i8, i8* %0 store i8 77, i8* %0 %E = extractelement <8 x i64> zeroinitializer, i32 2 %Shuff = shufflevector <8 x i64> zeroinitializer, <8 x i64> zeroinitializer, <8 x i32> <i32 5, i32 7, i32 undef, i32 undef, i32 13, i32 15, i32 1, i32 3> @@ -24,7 +24,7 @@ BB: br label %CF CF: ; preds = %CF, %CF78, %BB - %L5 = load i8* %Sl + %L5 = load i8, i8* %Sl store i8 %L, i8* %Sl %E6 = extractelement <8 x i32> zeroinitializer, i32 2 %Shuff7 = shufflevector <8 x i64> zeroinitializer, <8 x i64> %Shuff, <8 x i32> <i32 13, i32 15, i32 1, i32 3, i32 5, i32 7, i32 9, i32 undef> @@ -33,7 +33,7 @@ CF: ; preds = %CF, %CF78, %BB %FC = sitofp <8 x i64> zeroinitializer to <8 x float> %Sl9 = select i1 %Cmp, i8 77, i8 77 %Cmp10 = icmp uge <8 x i64> %Shuff, zeroinitializer - %L11 = load i8* %0 + %L11 = load i8, i8* %0 store i8 %Sl9, i8* %0 %E12 = extractelement <1 x i16> zeroinitializer, i32 0 %Shuff13 = shufflevector <8 x i64> zeroinitializer, <8 x i64> %Shuff, <8 x i32> <i32 9, i32 11, i32 13, i32 15, i32 undef, i32 3, i32 5, i32 7> @@ -42,7 +42,7 @@ CF: ; preds = %CF, %CF78, %BB %Tr = trunc <8 x i64> %Shuff to <8 x i32> %Sl16 = select i1 %Cmp, i8 77, i8 %5 %Cmp17 = icmp ult <8 x i1> %Cmp10, %Cmp10 - %L18 = load i8* %Sl + %L18 = load i8, i8* %Sl store i8 -1, i8* %Sl %E19 = extractelement <8 x i32> zeroinitializer, i32 3 %Shuff20 = shufflevector <8 x float> %FC, <8 x float> %FC, <8 x i32> <i32 6, i32 8, i32 undef, i32 12, i32 14, i32 0, i32 2, i32 undef> @@ -54,7 +54,7 @@ CF: ; preds = %CF, %CF78, %BB br i1 %Cmp25, label %CF, label %CF78 CF78: ; preds = %CF - %L26 = load i8* %Sl + %L26 = load i8, i8* %Sl store i32 50347, i32* %A %E27 = extractelement <8 x i1> %Cmp10, i32 2 br i1 %E27, label %CF, label %CF77 @@ -65,7 +65,7 @@ CF77: ; preds = %CF77, %CF81, %CF78 %B30 = urem <8 x i32> %Tr, zeroinitializer %Tr31 = trunc i32 0 to i16 %Sl32 = select i1 %Cmp, <2 x i1> zeroinitializer, <2 x i1> zeroinitializer - %L33 = load i8* %Sl + %L33 = load i8, i8* %Sl store i8 %L26, i8* %Sl %E34 = extractelement <4 x i32> zeroinitializer, i32 0 %Shuff35 = shufflevector <1 x i16> zeroinitializer, <1 x i16> %B, <1 x i32> undef @@ -73,7 +73,7 @@ CF77: ; preds = %CF77, %CF81, %CF78 %B37 = srem <1 x i16> %I29, zeroinitializer %FC38 = sitofp <8 x i32> %B30 to <8 x double> %Sl39 = select i1 %Cmp, double 0.000000e+00, double %Sl24 - %L40 = load i8* %Sl + %L40 = load i8, i8* %Sl store i8 %Sl16, i8* %Sl %E41 = extractelement <1 x i16> zeroinitializer, i32 0 %Shuff42 = shufflevector <8 x i1> %Cmp17, <8 x i1> %Cmp10, <8 x i32> <i32 14, i32 undef, i32 2, i32 4, i32 undef, i32 8, i32 10, i32 12> @@ -85,7 +85,7 @@ CF77: ; preds = %CF77, %CF81, %CF78 br i1 %Cmp46, label %CF77, label %CF80 CF80: ; preds = %CF80, %CF77 - %L47 = load i64* %PC + %L47 = load i64, i64* %PC store i8 77, i8* %Sl %E48 = extractelement <8 x i64> zeroinitializer, i32 2 %Shuff49 = shufflevector <8 x i64> zeroinitializer, <8 x i64> %Shuff7, <8 x i32> <i32 5, i32 7, i32 9, i32 undef, i32 undef, i32 undef, i32 undef, i32 3> @@ -97,7 +97,7 @@ CF80: ; preds = %CF80, %CF77 br i1 %Cmp54, label %CF80, label %CF81 CF81: ; preds = %CF80 - %L55 = load i8* %Sl + %L55 = load i8, i8* %Sl store i8 %Sl16, i8* %Sl %E56 = extractelement <1 x i16> %B, i32 0 %Shuff57 = shufflevector <1 x i16> zeroinitializer, <1 x i16> zeroinitializer, <1 x i32> <i32 1> @@ -105,7 +105,7 @@ CF81: ; preds = %CF80 %B59 = srem i32 %E19, %E19 %Sl60 = select i1 %Cmp, i8 77, i8 77 %Cmp61 = icmp ult <1 x i16> zeroinitializer, %B - %L62 = load i8* %Sl + %L62 = load i8, i8* %Sl store i64 %L47, i64* %PC52 %E63 = extractelement <4 x i32> %I43, i32 2 %Shuff64 = shufflevector <4 x i1> zeroinitializer, <4 x i1> zeroinitializer, <4 x i32> <i32 undef, i32 undef, i32 1, i32 3> @@ -117,7 +117,7 @@ CF81: ; preds = %CF80 br i1 %Cmp69, label %CF77, label %CF79 CF79: ; preds = %CF81 - %L70 = load i32* %A + %L70 = load i32, i32* %A store i64 %4, i64* %PC %E71 = extractelement <4 x i32> zeroinitializer, i32 0 %Shuff72 = shufflevector <8 x i32> zeroinitializer, <8 x i32> %B44, <8 x i32> <i32 11, i32 undef, i32 15, i32 1, i32 3, i32 undef, i32 7, i32 9> diff --git a/test/CodeGen/Mips/msa/llvm-stress-s1935737938.ll b/test/CodeGen/Mips/msa/llvm-stress-s1935737938.ll index f9cab037e7ccb..bdf6eafdf4efe 100644 --- a/test/CodeGen/Mips/msa/llvm-stress-s1935737938.ll +++ b/test/CodeGen/Mips/msa/llvm-stress-s1935737938.ll @@ -14,7 +14,7 @@ BB: %A2 = alloca i64 %A1 = alloca i32 %A = alloca <2 x i64> - %L = load i8* %0 + %L = load i8, i8* %0 store i8 -1, i8* %0 %E = extractelement <2 x i32> zeroinitializer, i32 0 %Shuff = shufflevector <2 x i32> zeroinitializer, <2 x i32> zeroinitializer, <2 x i32> <i32 1, i32 3> @@ -22,7 +22,7 @@ BB: %B = lshr i8 %L, -69 %ZE = fpext float 0xBF2AA5FE80000000 to double %Sl = select i1 true, <1 x i64> <i64 -1>, <1 x i64> <i64 -1> - %L5 = load i8* %0 + %L5 = load i8, i8* %0 store i8 -69, i8* %0 %E6 = extractelement <16 x i64> <i64 -1, i64 -1, i64 -1, i64 -1, i64 -1, i64 -1, i64 -1, i64 -1, i64 -1, i64 -1, i64 -1, i64 -1, i64 -1, i64 -1, i64 -1, i64 -1>, i32 14 %Shuff7 = shufflevector <2 x i32> zeroinitializer, <2 x i32> zeroinitializer, <2 x i32> <i32 1, i32 3> @@ -31,7 +31,7 @@ BB: %FC = uitofp i32 %3 to double %Sl10 = select i1 true, <1 x i1> zeroinitializer, <1 x i1> zeroinitializer %Cmp = icmp ne <1 x i64> %I, <i64 -1> - %L11 = load i8* %0 + %L11 = load i8, i8* %0 store i8 %L11, i8* %0 %E12 = extractelement <1 x i64> <i64 -1>, i32 0 %Shuff13 = shufflevector <1 x i64> %Sl, <1 x i64> <i64 -1>, <1 x i32> <i32 1> @@ -42,7 +42,7 @@ BB: br label %CF74 CF74: ; preds = %CF74, %CF80, %CF76, %BB - %L18 = load i8* %0 + %L18 = load i8, i8* %0 store i8 -69, i8* %0 %E19 = extractelement <1 x i64> %Sl, i32 0 %Shuff20 = shufflevector <8 x i8> <i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1>, <8 x i8> <i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1>, <8 x i32> <i32 12, i32 14, i32 0, i32 2, i32 4, i32 6, i32 8, i32 10> @@ -50,7 +50,7 @@ CF74: ; preds = %CF74, %CF80, %CF76, %B22 = urem i32 135673, %3 %FC23 = sitofp i8 %L to float %Sl24 = select i1 true, i8 %B, i8 %L18 - %L25 = load i8* %0 + %L25 = load i8, i8* %0 store i8 %L, i8* %0 %E26 = extractelement <2 x i32> %Shuff, i32 1 %Shuff27 = shufflevector <2 x i32> zeroinitializer, <2 x i32> zeroinitializer, <2 x i32> <i32 2, i32 0> @@ -62,7 +62,7 @@ CF74: ; preds = %CF74, %CF80, %CF76, br i1 %Cmp31, label %CF74, label %CF80 CF80: ; preds = %CF74 - %L32 = load i8* %0 + %L32 = load i8, i8* %0 store i8 -1, i8* %0 %E33 = extractelement <2 x i32> zeroinitializer, i32 1 %Shuff34 = shufflevector <1 x i64> %Shuff13, <1 x i64> <i64 -1>, <1 x i32> zeroinitializer @@ -70,7 +70,7 @@ CF80: ; preds = %CF74 %FC36 = sitofp <1 x i1> %Cmp to <1 x float> %Sl37 = select i1 true, <8 x i8> %Shuff20, <8 x i8> <i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1> %Cmp38 = icmp sgt <2 x i32> %I21, %Shuff27 - %L39 = load i8* %0 + %L39 = load i8, i8* %0 store i8 %Sl24, i8* %0 %E40 = extractelement <8 x i64> zeroinitializer, i32 1 %Shuff41 = shufflevector <2 x i1> zeroinitializer, <2 x i1> %Cmp38, <2 x i32> <i32 0, i32 2> @@ -81,7 +81,7 @@ CF80: ; preds = %CF74 br i1 %Cmp45, label %CF74, label %CF76 CF76: ; preds = %CF80 - %L46 = load i8* %0 + %L46 = load i8, i8* %0 store i8 %L39, i8* %0 %E47 = extractelement <2 x i32> %Shuff27, i32 0 %Shuff48 = shufflevector <1 x i1> %Sl10, <1 x i1> %Sl10, <1 x i32> <i32 1> @@ -92,7 +92,7 @@ CF76: ; preds = %CF80 br i1 %Cmp52, label %CF74, label %CF75 CF75: ; preds = %CF75, %CF76 - %L53 = load i8* %0 + %L53 = load i8, i8* %0 store i8 %L18, i8* %0 %E54 = extractelement <8 x i8> %Shuff20, i32 5 %Shuff55 = shufflevector <2 x i32> %Shuff, <2 x i32> zeroinitializer, <2 x i32> <i32 0, i32 2> @@ -103,7 +103,7 @@ CF75: ; preds = %CF75, %CF76 br i1 %Cmp59, label %CF75, label %CF78 CF78: ; preds = %CF75 - %L60 = load i8* %0 + %L60 = load i8, i8* %0 store i8 -69, i8* %0 %E61 = extractelement <2 x i32> zeroinitializer, i32 0 %Shuff62 = shufflevector <2 x i32> %Shuff7, <2 x i32> %I21, <2 x i32> <i32 1, i32 3> @@ -115,7 +115,7 @@ CF78: ; preds = %CF75 br label %CF CF: ; preds = %CF, %CF78 - %L68 = load i8* %0 + %L68 = load i8, i8* %0 store i64 %B57, i64* %2 %E69 = extractelement <2 x i1> %Shuff41, i32 1 br i1 %E69, label %CF, label %CF77 diff --git a/test/CodeGen/Mips/msa/llvm-stress-s2704903805.ll b/test/CodeGen/Mips/msa/llvm-stress-s2704903805.ll index e14f405320cb8..8f23a8ca5177b 100644 --- a/test/CodeGen/Mips/msa/llvm-stress-s2704903805.ll +++ b/test/CodeGen/Mips/msa/llvm-stress-s2704903805.ll @@ -13,7 +13,7 @@ BB: %A2 = alloca i8 %A1 = alloca i32 %A = alloca i8 - %L = load i8* %0 + %L = load i8, i8* %0 store i8 %5, i8* %0 %E = extractelement <2 x i16> zeroinitializer, i32 0 %Shuff = shufflevector <1 x i8> <i8 -1>, <1 x i8> <i8 -1>, <1 x i32> undef @@ -25,7 +25,7 @@ BB: br label %CF83 CF83: ; preds = %BB - %L5 = load i8* %0 + %L5 = load i8, i8* %0 store i8 85, i8* %0 %E6 = extractelement <1 x i8> <i8 -1>, i32 0 %Shuff7 = shufflevector <2 x i16> zeroinitializer, <2 x i16> zeroinitializer, <2 x i32> <i32 1, i32 3> @@ -37,7 +37,7 @@ CF83: ; preds = %BB br label %CF CF: ; preds = %CF, %CF81, %CF83 - %L13 = load i8* %0 + %L13 = load i8, i8* %0 store i8 0, i8* %0 %E14 = extractelement <2 x i64> zeroinitializer, i32 0 %Shuff15 = shufflevector <4 x i64> <i64 -1, i64 -1, i64 -1, i64 -1>, <4 x i64> <i64 -1, i64 -1, i64 -1, i64 -1>, <4 x i32> <i32 3, i32 5, i32 7, i32 undef> @@ -52,7 +52,7 @@ CF80: ; preds = %CF80, %CF br i1 %Cmp19, label %CF80, label %CF81 CF81: ; preds = %CF80 - %L20 = load i8* %0 + %L20 = load i8, i8* %0 store i8 85, i8* %0 %E21 = extractelement <1 x i8> <i8 -1>, i32 0 %Shuff22 = shufflevector <1 x i8> <i8 -1>, <1 x i8> %Shuff, <1 x i32> zeroinitializer @@ -60,7 +60,7 @@ CF81: ; preds = %CF80 %FC24 = fptoui <4 x float> %FC to <4 x i16> %Sl25 = select i1 %Cmp, <2 x i32> zeroinitializer, <2 x i32> <i32 -1, i32 -1> %Cmp26 = icmp ult <4 x i64> %I16, %Shuff15 - %L27 = load i8* %0 + %L27 = load i8, i8* %0 store i8 %L, i8* %0 %E28 = extractelement <1 x i8> <i8 -1>, i32 0 %Shuff29 = shufflevector <8 x i16> zeroinitializer, <8 x i16> zeroinitializer, <8 x i32> <i32 11, i32 undef, i32 15, i32 1, i32 3, i32 5, i32 undef, i32 9> @@ -68,7 +68,7 @@ CF81: ; preds = %CF80 %B31 = mul i8 %E28, 85 %PC = bitcast i32* %A3 to i32* %Sl32 = select i1 %Cmp12, float %FC10, float 0x4712BFE680000000 - %L33 = load i32* %PC + %L33 = load i32, i32* %PC store i32 %L33, i32* %PC %E34 = extractelement <2 x i16> zeroinitializer, i32 1 %Shuff35 = shufflevector <1 x i8> %Shuff, <1 x i8> <i8 -1>, <1 x i32> zeroinitializer @@ -79,7 +79,7 @@ CF81: ; preds = %CF80 br i1 %Cmp39, label %CF, label %CF77 CF77: ; preds = %CF77, %CF81 - %L40 = load i32* %PC + %L40 = load i32, i32* %PC store i32 %3, i32* %PC %E41 = extractelement <2 x i32> zeroinitializer, i32 0 %Shuff42 = shufflevector <2 x i32> <i32 -1, i32 -1>, <2 x i32> zeroinitializer, <2 x i32> <i32 1, i32 3> @@ -88,7 +88,7 @@ CF77: ; preds = %CF77, %CF81 %Se = sext i32 %3 to i64 %Sl45 = select i1 true, <1 x i8> %Shuff, <1 x i8> %I43 %Cmp46 = icmp sge <1 x i8> %I36, %Shuff - %L47 = load i32* %PC + %L47 = load i32, i32* %PC store i32 %L33, i32* %PC %E48 = extractelement <2 x i16> zeroinitializer, i32 0 %Shuff49 = shufflevector <1 x i8> <i8 -1>, <1 x i8> <i8 -1>, <1 x i32> <i32 1> @@ -100,7 +100,7 @@ CF77: ; preds = %CF77, %CF81 br i1 %Cmp54, label %CF77, label %CF78 CF78: ; preds = %CF78, %CF77 - %L55 = load i32* %PC + %L55 = load i32, i32* %PC store i32 %L33, i32* %PC %E56 = extractelement <8 x i16> %Shuff29, i32 4 %Shuff57 = shufflevector <1 x i8> <i8 -1>, <1 x i8> <i8 -1>, <1 x i32> <i32 1> @@ -111,7 +111,7 @@ CF78: ; preds = %CF78, %CF77 br i1 %Cmp60, label %CF78, label %CF79 CF79: ; preds = %CF79, %CF78 - %L61 = load i32* %PC + %L61 = load i32, i32* %PC store i32 %L33, i32* %A3 %E62 = extractelement <4 x i64> %Shuff15, i32 1 %Shuff63 = shufflevector <8 x i16> %Shuff29, <8 x i16> %Shuff29, <8 x i32> <i32 undef, i32 10, i32 12, i32 undef, i32 undef, i32 undef, i32 4, i32 6> @@ -123,7 +123,7 @@ CF79: ; preds = %CF79, %CF78 br i1 %Cmp68, label %CF79, label %CF82 CF82: ; preds = %CF79 - %L69 = load i32* %PC + %L69 = load i32, i32* %PC store i32 %L33, i32* %PC %E70 = extractelement <8 x i16> zeroinitializer, i32 3 %Shuff71 = shufflevector <4 x i64> %Shuff15, <4 x i64> <i64 -1, i64 -1, i64 -1, i64 -1>, <4 x i32> <i32 6, i32 undef, i32 2, i32 4> diff --git a/test/CodeGen/Mips/msa/llvm-stress-s3861334421.ll b/test/CodeGen/Mips/msa/llvm-stress-s3861334421.ll index 1a03e55d9d540..e3cf7964497f6 100644 --- a/test/CodeGen/Mips/msa/llvm-stress-s3861334421.ll +++ b/test/CodeGen/Mips/msa/llvm-stress-s3861334421.ll @@ -14,7 +14,7 @@ BB: %A2 = alloca i64 %A1 = alloca i64 %A = alloca double - %L = load i8* %0 + %L = load i8, i8* %0 store i8 -101, i8* %0 %E = extractelement <4 x i32> <i32 -1, i32 -1, i32 -1, i32 -1>, i32 0 %Shuff = shufflevector <8 x i64> zeroinitializer, <8 x i64> zeroinitializer, <8 x i32> <i32 3, i32 5, i32 7, i32 9, i32 11, i32 13, i32 undef, i32 1> @@ -22,7 +22,7 @@ BB: %B = and i64 116376, 57247 %FC = uitofp i8 7 to double %Sl = select i1 false, <8 x i8> <i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1>, <8 x i8> <i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1> - %L5 = load i8* %0 + %L5 = load i8, i8* %0 store i8 %L, i8* %0 %E6 = extractelement <4 x i32> <i32 -1, i32 -1, i32 -1, i32 -1>, i32 3 %Shuff7 = shufflevector <4 x i64> zeroinitializer, <4 x i64> zeroinitializer, <4 x i32> <i32 2, i32 4, i32 6, i32 0> @@ -33,7 +33,7 @@ BB: br label %CF CF: ; preds = %CF, %BB - %L11 = load i8* %0 + %L11 = load i8, i8* %0 store i8 -87, i8* %0 %E12 = extractelement <4 x i64> zeroinitializer, i32 0 %Shuff13 = shufflevector <8 x i64> zeroinitializer, <8 x i64> zeroinitializer, <8 x i32> <i32 7, i32 9, i32 11, i32 13, i32 undef, i32 1, i32 3, i32 5> @@ -45,7 +45,7 @@ CF: ; preds = %CF, %BB br i1 %Cmp18, label %CF, label %CF80 CF80: ; preds = %CF80, %CF88, %CF - %L19 = load i8* %0 + %L19 = load i8, i8* %0 store i8 -101, i8* %0 %E20 = extractelement <4 x i64> zeroinitializer, i32 0 %Shuff21 = shufflevector <4 x i64> zeroinitializer, <4 x i64> %Shuff7, <4 x i32> <i32 7, i32 1, i32 3, i32 5> @@ -56,7 +56,7 @@ CF80: ; preds = %CF80, %CF88, %CF br i1 %Cmp25, label %CF80, label %CF83 CF83: ; preds = %CF83, %CF80 - %L26 = load i8* %0 + %L26 = load i8, i8* %0 store i8 -87, i8* %0 %E27 = extractelement <4 x i32> <i32 -1, i32 -1, i32 -1, i32 -1>, i32 0 %Shuff28 = shufflevector <4 x i32> <i32 -1, i32 -1, i32 -1, i32 -1>, <4 x i32> <i32 -1, i32 -1, i32 -1, i32 -1>, <4 x i32> <i32 7, i32 1, i32 3, i32 5> @@ -68,7 +68,7 @@ CF83: ; preds = %CF83, %CF80 br i1 %Cmp33, label %CF83, label %CF88 CF88: ; preds = %CF83 - %L34 = load i8* %0 + %L34 = load i8, i8* %0 store i8 -87, i8* %0 %E35 = extractelement <8 x i64> %Shuff, i32 7 %Shuff36 = shufflevector <4 x i32> <i32 -1, i32 -1, i32 -1, i32 -1>, <4 x i32> %Shuff28, <4 x i32> <i32 2, i32 undef, i32 undef, i32 0> @@ -80,7 +80,7 @@ CF88: ; preds = %CF83 br i1 %Cmp40, label %CF80, label %CF81 CF81: ; preds = %CF81, %CF85, %CF87, %CF88 - %L41 = load i8* %0 + %L41 = load i8, i8* %0 store i8 %L34, i8* %0 %E42 = extractelement <8 x i64> %Shuff13, i32 6 %Shuff43 = shufflevector <4 x i64> zeroinitializer, <4 x i64> zeroinitializer, <4 x i32> <i32 undef, i32 undef, i32 undef, i32 7> @@ -92,7 +92,7 @@ CF81: ; preds = %CF81, %CF85, %CF87, br i1 %Cmp47, label %CF81, label %CF85 CF85: ; preds = %CF81 - %L48 = load i8* %0 + %L48 = load i8, i8* %0 store i8 -101, i8* %0 %E49 = extractelement <8 x i8> <i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1>, i32 2 %Shuff50 = shufflevector <4 x i64> zeroinitializer, <4 x i64> zeroinitializer, <4 x i32> <i32 5, i32 7, i32 1, i32 3> @@ -101,7 +101,7 @@ CF85: ; preds = %CF81 %FC53 = uitofp i8 %L48 to double %Sl54 = select i1 %Cmp47, i32 %3, i32 %Sl24 %Cmp55 = icmp ne <8 x i64> %Shuff13, zeroinitializer - %L56 = load i8* %0 + %L56 = load i8, i8* %0 store i8 %L11, i8* %0 %E57 = extractelement <4 x i64> %Shuff21, i32 1 %Shuff58 = shufflevector <8 x i64> %Shuff, <8 x i64> zeroinitializer, <8 x i32> <i32 4, i32 6, i32 undef, i32 10, i32 12, i32 undef, i32 0, i32 2> @@ -113,7 +113,7 @@ CF85: ; preds = %CF81 CF84: ; preds = %CF84, %CF85 %Sl62 = select i1 false, i8 %L, i8 %L48 %Cmp63 = icmp ne <8 x i64> %I, zeroinitializer - %L64 = load i8* %0 + %L64 = load i8, i8* %0 store i8 %5, i8* %0 %E65 = extractelement <8 x i1> %Cmp55, i32 0 br i1 %E65, label %CF84, label %CF87 @@ -125,7 +125,7 @@ CF87: ; preds = %CF84 %ZE69 = zext <8 x i8> %Sl32 to <8 x i64> %Sl70 = select i1 %Tr61, i64 %E20, i64 %E12 %Cmp71 = icmp slt <8 x i64> %I, %Shuff - %L72 = load i8* %0 + %L72 = load i8, i8* %0 store i8 %L72, i8* %0 %E73 = extractelement <8 x i1> %Cmp55, i32 6 br i1 %E73, label %CF81, label %CF82 diff --git a/test/CodeGen/Mips/msa/llvm-stress-s3926023935.ll b/test/CodeGen/Mips/msa/llvm-stress-s3926023935.ll index 96547d90cb401..6f338107825e7 100644 --- a/test/CodeGen/Mips/msa/llvm-stress-s3926023935.ll +++ b/test/CodeGen/Mips/msa/llvm-stress-s3926023935.ll @@ -14,7 +14,7 @@ BB: %A2 = alloca double %A1 = alloca float %A = alloca double - %L = load i8* %0 + %L = load i8, i8* %0 store i8 -123, i8* %0 %E = extractelement <4 x i64> zeroinitializer, i32 1 %Shuff = shufflevector <4 x i32> <i32 -1, i32 -1, i32 -1, i32 -1>, <4 x i32> zeroinitializer, <4 x i32> <i32 0, i32 2, i32 4, i32 6> @@ -22,7 +22,7 @@ BB: %BC = bitcast i64 181325 to double %Sl = select i1 false, <2 x i32> zeroinitializer, <2 x i32> zeroinitializer %Cmp = icmp ne <4 x i64> zeroinitializer, zeroinitializer - %L5 = load i8* %0 + %L5 = load i8, i8* %0 store i8 %L, i8* %0 %E6 = extractelement <4 x i64> zeroinitializer, i32 3 %Shuff7 = shufflevector <2 x i16> zeroinitializer, <2 x i16> zeroinitializer, <2 x i32> <i32 2, i32 0> @@ -33,7 +33,7 @@ BB: br label %CF80 CF80: ; preds = %BB - %L11 = load i8* %0 + %L11 = load i8, i8* %0 store i8 -123, i8* %0 %E12 = extractelement <2 x i16> zeroinitializer, i32 1 %Shuff13 = shufflevector <4 x i64> zeroinitializer, <4 x i64> zeroinitializer, <4 x i32> <i32 0, i32 2, i32 4, i32 6> @@ -42,7 +42,7 @@ CF80: ; preds = %BB %PC = bitcast i1* %A4 to i64* %Sl16 = select i1 %Cmp10, <4 x i32> zeroinitializer, <4 x i32> <i32 -1, i32 -1, i32 -1, i32 -1> %Cmp17 = icmp ule <4 x i32> <i32 -1, i32 -1, i32 -1, i32 -1>, %Sl16 - %L18 = load double* %A2 + %L18 = load double, double* %A2 store i64 498254, i64* %PC %E19 = extractelement <4 x i64> zeroinitializer, i32 0 %Shuff20 = shufflevector <2 x i1> zeroinitializer, <2 x i1> %I, <2 x i32> <i32 3, i32 1> @@ -51,7 +51,7 @@ CF80: ; preds = %BB %ZE = zext <2 x i1> %Shuff20 to <2 x i32> %Sl23 = select i1 %Cmp10, <2 x i1> %Shuff20, <2 x i1> zeroinitializer %Cmp24 = icmp ult <2 x i32> zeroinitializer, zeroinitializer - %L25 = load i8* %0 + %L25 = load i8, i8* %0 store i8 %L25, i8* %0 %E26 = extractelement <4 x i8> <i8 -1, i8 -1, i8 -1, i8 -1>, i32 3 %Shuff27 = shufflevector <4 x i32> %Shuff, <4 x i32> %I14, <4 x i32> <i32 6, i32 0, i32 undef, i32 4> @@ -63,7 +63,7 @@ CF80: ; preds = %BB CF79: ; preds = %CF80 %Sl30 = select i1 false, i8 %B29, i8 -123 %Cmp31 = icmp sge <2 x i1> %I, %I - %L32 = load i64* %PC + %L32 = load i64, i64* %PC store i8 -123, i8* %0 %E33 = extractelement <8 x i64> <i64 -1, i64 -1, i64 -1, i64 -1, i64 -1, i64 -1, i64 -1, i64 -1>, i32 2 %Shuff34 = shufflevector <4 x i64> zeroinitializer, <4 x i64> %Shuff13, <4 x i32> <i32 5, i32 7, i32 1, i32 3> @@ -75,7 +75,7 @@ CF79: ; preds = %CF80 br label %CF CF: ; preds = %CF, %CF79 - %L40 = load double* %A + %L40 = load double, double* %A store i1 %Cmp39, i1* %PC37 %E41 = extractelement <4 x i64> zeroinitializer, i32 3 %Shuff42 = shufflevector <2 x i32> zeroinitializer, <2 x i32> %ZE, <2 x i32> <i32 2, i32 undef> @@ -90,7 +90,7 @@ CF77: ; preds = %CF77, %CF br i1 %Cmp46, label %CF77, label %CF78 CF78: ; preds = %CF78, %CF83, %CF82, %CF77 - %L47 = load i64* %PC + %L47 = load i64, i64* %PC store i8 -123, i8* %0 %E48 = extractelement <4 x i64> zeroinitializer, i32 3 %Shuff49 = shufflevector <4 x i32> <i32 -1, i32 -1, i32 -1, i32 -1>, <4 x i32> zeroinitializer, <4 x i32> <i32 2, i32 4, i32 6, i32 undef> @@ -105,7 +105,7 @@ CF83: ; preds = %CF78 br i1 %Cmp54, label %CF78, label %CF82 CF82: ; preds = %CF83 - %L55 = load i64* %PC + %L55 = load i64, i64* %PC store i64 %L32, i64* %PC %E56 = extractelement <2 x i16> %Shuff7, i32 1 %Shuff57 = shufflevector <4 x i64> zeroinitializer, <4 x i64> zeroinitializer, <4 x i32> <i32 2, i32 4, i32 6, i32 0> @@ -114,7 +114,7 @@ CF82: ; preds = %CF83 %FC = sitofp i64 498254 to double %Sl60 = select i1 false, i64 %E6, i64 -1 %Cmp61 = icmp sgt <4 x i32> %Shuff27, %I43 - %L62 = load i64* %PC + %L62 = load i64, i64* %PC store i64 %Sl9, i64* %PC %E63 = extractelement <2 x i32> %ZE, i32 0 %Shuff64 = shufflevector <4 x i64> zeroinitializer, <4 x i64> %Shuff13, <4 x i32> <i32 1, i32 3, i32 undef, i32 7> @@ -126,7 +126,7 @@ CF82: ; preds = %CF83 CF81: ; preds = %CF82 %Cmp69 = icmp ne <8 x i64> <i64 -1, i64 -1, i64 -1, i64 -1, i64 -1, i64 -1, i64 -1, i64 -1>, %B36 - %L70 = load i8* %0 + %L70 = load i8, i8* %0 store i64 %L55, i64* %PC %E71 = extractelement <4 x i32> %Shuff49, i32 1 %Shuff72 = shufflevector <4 x i64> zeroinitializer, <4 x i64> %Shuff34, <4 x i32> <i32 0, i32 2, i32 4, i32 6> diff --git a/test/CodeGen/Mips/msa/llvm-stress-s3997499501.ll b/test/CodeGen/Mips/msa/llvm-stress-s3997499501.ll index bef75f3645c8a..181f72abd3786 100644 --- a/test/CodeGen/Mips/msa/llvm-stress-s3997499501.ll +++ b/test/CodeGen/Mips/msa/llvm-stress-s3997499501.ll @@ -14,7 +14,7 @@ BB: %A2 = alloca float %A1 = alloca double %A = alloca double - %L = load i8* %0 + %L = load i8, i8* %0 store i8 97, i8* %0 %E = extractelement <16 x i64> <i64 -1, i64 -1, i64 -1, i64 -1, i64 -1, i64 -1, i64 -1, i64 -1, i64 -1, i64 -1, i64 -1, i64 -1, i64 -1, i64 -1, i64 -1, i64 -1>, i32 14 %Shuff = shufflevector <2 x i1> zeroinitializer, <2 x i1> zeroinitializer, <2 x i32> <i32 1, i32 3> @@ -22,7 +22,7 @@ BB: %Tr = trunc <1 x i64> zeroinitializer to <1 x i8> %Sl = select i1 false, double* %A1, double* %A %Cmp = icmp ne <2 x i64> zeroinitializer, zeroinitializer - %L5 = load double* %Sl + %L5 = load double, double* %Sl store float -4.374162e+06, float* %A2 %E6 = extractelement <4 x i64> zeroinitializer, i32 3 %Shuff7 = shufflevector <4 x i64> zeroinitializer, <4 x i64> %I, <4 x i32> <i32 2, i32 4, i32 6, i32 undef> @@ -34,7 +34,7 @@ BB: br label %CF72 CF72: ; preds = %CF72, %CF80, %CF78, %BB - %L11 = load double* %Sl + %L11 = load double, double* %Sl store double 0.000000e+00, double* %Sl %E12 = extractelement <2 x i1> zeroinitializer, i32 0 br i1 %E12, label %CF72, label %CF80 @@ -49,7 +49,7 @@ CF80: ; preds = %CF72 br i1 %Cmp17, label %CF72, label %CF77 CF77: ; preds = %CF77, %CF80 - %L18 = load double* %Sl + %L18 = load double, double* %Sl store double 0.000000e+00, double* %Sl %E19 = extractelement <2 x i1> zeroinitializer, i32 0 br i1 %E19, label %CF77, label %CF78 @@ -60,7 +60,7 @@ CF78: ; preds = %CF77 %B22 = sdiv <4 x i64> %Shuff7, zeroinitializer %FC = uitofp i8 97 to double %Sl23 = select i1 %Cmp10, <2 x i1> zeroinitializer, <2 x i1> zeroinitializer - %L24 = load double* %Sl + %L24 = load double, double* %Sl store float %Sl16, float* %PC %E25 = extractelement <2 x i1> %Shuff, i32 1 br i1 %E25, label %CF72, label %CF76 @@ -71,7 +71,7 @@ CF76: ; preds = %CF78 %B28 = mul <4 x i64> %I27, zeroinitializer %ZE = zext <8 x i1> zeroinitializer to <8 x i64> %Sl29 = select i1 %Cmp17, float -4.374162e+06, float -4.374162e+06 - %L30 = load i8* %0 + %L30 = load i8, i8* %0 store double %L5, double* %Sl %E31 = extractelement <8 x i1> zeroinitializer, i32 5 br label %CF @@ -85,7 +85,7 @@ CF: ; preds = %CF, %CF81, %CF76 br i1 %Cmp36, label %CF, label %CF74 CF74: ; preds = %CF74, %CF - %L37 = load float* %PC + %L37 = load float, float* %PC store double 0.000000e+00, double* %Sl %E38 = extractelement <2 x i1> %Sl23, i32 1 br i1 %E38, label %CF74, label %CF75 @@ -95,7 +95,7 @@ CF75: ; preds = %CF75, %CF82, %CF74 %I40 = insertelement <4 x i64> zeroinitializer, i64 %4, i32 2 %Sl41 = select i1 %Cmp10, i32 0, i32 %3 %Cmp42 = icmp ne <1 x i64> zeroinitializer, zeroinitializer - %L43 = load double* %Sl + %L43 = load double, double* %Sl store i64 %4, i64* %2 %E44 = extractelement <2 x i1> %Shuff20, i32 1 br i1 %E44, label %CF75, label %CF82 @@ -109,7 +109,7 @@ CF82: ; preds = %CF75 br i1 %Cmp49, label %CF75, label %CF81 CF81: ; preds = %CF82 - %L50 = load i8* %0 + %L50 = load i8, i8* %0 store double %L43, double* %Sl %E51 = extractelement <4 x i64> %Shuff7, i32 3 %Shuff52 = shufflevector <4 x float> %BC34, <4 x float> %BC34, <4 x i32> <i32 2, i32 4, i32 6, i32 0> @@ -117,7 +117,7 @@ CF81: ; preds = %CF82 %B54 = fdiv double %L24, %L43 %BC55 = bitcast <4 x i64> zeroinitializer to <4 x double> %Sl56 = select i1 false, i8 %5, i8 97 - %L57 = load i8* %0 + %L57 = load i8, i8* %0 store i8 %L50, i8* %0 %E58 = extractelement <2 x i1> %Shuff20, i32 1 br i1 %E58, label %CF, label %CF73 @@ -129,7 +129,7 @@ CF73: ; preds = %CF73, %CF81 %PC62 = bitcast double* %A3 to float* %Sl63 = select i1 %Cmp10, <1 x i64> zeroinitializer, <1 x i64> zeroinitializer %Cmp64 = icmp ne <2 x i1> %Cmp, %Shuff - %L65 = load double* %A1 + %L65 = load double, double* %A1 store float -4.374162e+06, float* %PC62 %E66 = extractelement <8 x i1> %I21, i32 3 br i1 %E66, label %CF73, label %CF79 diff --git a/test/CodeGen/Mips/msa/llvm-stress-s525530439.ll b/test/CodeGen/Mips/msa/llvm-stress-s525530439.ll index 697871df797df..c0bc905634847 100644 --- a/test/CodeGen/Mips/msa/llvm-stress-s525530439.ll +++ b/test/CodeGen/Mips/msa/llvm-stress-s525530439.ll @@ -14,7 +14,7 @@ BB: %A2 = alloca <1 x double> %A1 = alloca <8 x double> %A = alloca i64 - %L = load i8* %0 + %L = load i8, i8* %0 store i64 33695, i64* %A %E = extractelement <4 x i32> zeroinitializer, i32 3 %Shuff = shufflevector <2 x i32> <i32 -1, i32 -1>, <2 x i32> <i32 -1, i32 -1>, <2 x i32> <i32 2, i32 0> @@ -22,7 +22,7 @@ BB: %B = lshr <8 x i32> <i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1>, <i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1> %ZE = fpext float 0x3B64A2B880000000 to double %Sl = select i1 true, i16 -1, i16 -11642 - %L5 = load i8* %0 + %L5 = load i8, i8* %0 store i8 0, i8* %0 %E6 = extractelement <4 x i32> zeroinitializer, i32 2 %Shuff7 = shufflevector <8 x i1> zeroinitializer, <8 x i1> zeroinitializer, <8 x i32> <i32 undef, i32 7, i32 9, i32 11, i32 13, i32 15, i32 1, i32 undef> @@ -31,7 +31,7 @@ BB: %BC = bitcast <2 x i32> <i32 -1, i32 -1> to <2 x float> %Sl10 = select i1 true, i32* %1, i32* %1 %Cmp = icmp sge <8 x i64> zeroinitializer, zeroinitializer - %L11 = load i32* %Sl10 + %L11 = load i32, i32* %Sl10 store <1 x double> zeroinitializer, <1 x double>* %A2 %E12 = extractelement <4 x i16> zeroinitializer, i32 0 %Shuff13 = shufflevector <1 x i64> zeroinitializer, <1 x i64> zeroinitializer, <1 x i32> undef @@ -43,7 +43,7 @@ BB: br label %CF75 CF75: ; preds = %CF75, %BB - %L19 = load i32* %Sl10 + %L19 = load i32, i32* %Sl10 store i32 %L11, i32* %Sl10 %E20 = extractelement <4 x i32> zeroinitializer, i32 1 %Shuff21 = shufflevector <4 x i32> zeroinitializer, <4 x i32> %I8, <4 x i32> <i32 undef, i32 2, i32 4, i32 6> @@ -55,7 +55,7 @@ CF75: ; preds = %CF75, %BB br i1 %Cmp26, label %CF75, label %CF76 CF76: ; preds = %CF75 - %L27 = load i32* %Sl10 + %L27 = load i32, i32* %Sl10 store i32 439732, i32* %Sl10 %E28 = extractelement <4 x i32> %Shuff21, i32 3 %Shuff29 = shufflevector <8 x i32> <i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1>, <8 x i32> <i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1>, <8 x i32> <i32 2, i32 4, i32 6, i32 8, i32 10, i32 12, i32 14, i32 0> @@ -65,7 +65,7 @@ CF76: ; preds = %CF75 br label %CF74 CF74: ; preds = %CF74, %CF80, %CF78, %CF76 - %L33 = load i64* %2 + %L33 = load i64, i64* %2 store i32 71140, i32* %Sl10 %E34 = extractelement <4 x i32> zeroinitializer, i32 1 %Shuff35 = shufflevector <1 x i16> zeroinitializer, <1 x i16> zeroinitializer, <1 x i32> undef @@ -76,7 +76,7 @@ CF74: ; preds = %CF74, %CF80, %CF78, br i1 %Cmp39, label %CF74, label %CF80 CF80: ; preds = %CF74 - %L40 = load i8* %0 + %L40 = load i8, i8* %0 store i32 0, i32* %Sl10 %E41 = extractelement <8 x i64> zeroinitializer, i32 1 %Shuff42 = shufflevector <1 x i16> %I14, <1 x i16> %I14, <1 x i32> undef @@ -86,7 +86,7 @@ CF80: ; preds = %CF74 br i1 %Sl44, label %CF74, label %CF78 CF78: ; preds = %CF80 - %L45 = load i32* %Sl10 + %L45 = load i32, i32* %Sl10 store i8 %L5, i8* %0 %E46 = extractelement <8 x i1> %Shuff7, i32 2 br i1 %E46, label %CF74, label %CF77 @@ -101,7 +101,7 @@ CF77: ; preds = %CF77, %CF78 br i1 %Cmp52, label %CF77, label %CF79 CF79: ; preds = %CF77 - %L53 = load i32* %Sl10 + %L53 = load i32, i32* %Sl10 store i8 %L40, i8* %0 %E54 = extractelement <4 x i32> zeroinitializer, i32 1 %Shuff55 = shufflevector <4 x i32> %Shuff21, <4 x i32> %I8, <4 x i32> <i32 4, i32 6, i32 undef, i32 2> @@ -109,7 +109,7 @@ CF79: ; preds = %CF77 %Tr = trunc <1 x i64> %Shuff13 to <1 x i16> %Sl57 = select i1 %Cmp18, <2 x i32> <i32 -1, i32 -1>, <2 x i32> <i32 -1, i32 -1> %Cmp58 = icmp uge <4 x i32> <i32 -1, i32 -1, i32 -1, i32 -1>, %I56 - %L59 = load i8* %0 + %L59 = load i8, i8* %0 store <1 x double> zeroinitializer, <1 x double>* %A2 %E60 = extractelement <4 x i32> zeroinitializer, i32 0 %Shuff61 = shufflevector <4 x i32> %I8, <4 x i32> %I8, <4 x i32> <i32 undef, i32 1, i32 undef, i32 undef> @@ -121,7 +121,7 @@ CF79: ; preds = %CF77 br label %CF CF: ; preds = %CF79 - %L66 = load i32* %Sl10 + %L66 = load i32, i32* %Sl10 store i32 %E6, i32* %PC %E67 = extractelement <4 x i32> <i32 -1, i32 -1, i32 -1, i32 -1>, i32 2 %Shuff68 = shufflevector <4 x i32> %Sl64, <4 x i32> %I8, <4 x i32> <i32 5, i32 undef, i32 1, i32 undef> diff --git a/test/CodeGen/Mips/msa/llvm-stress-s997348632.ll b/test/CodeGen/Mips/msa/llvm-stress-s997348632.ll index dc4200ad42851..a3150e9a67d59 100644 --- a/test/CodeGen/Mips/msa/llvm-stress-s997348632.ll +++ b/test/CodeGen/Mips/msa/llvm-stress-s997348632.ll @@ -14,14 +14,14 @@ BB: %A2 = alloca <4 x i1> %A1 = alloca <4 x i16> %A = alloca <2 x i32> - %L = load i8* %0 + %L = load i8, i8* %0 store i8 %L, i8* %0 %E = extractelement <4 x i32> zeroinitializer, i32 0 %Shuff = shufflevector <4 x i64> zeroinitializer, <4 x i64> zeroinitializer, <4 x i32> <i32 undef, i32 1, i32 3, i32 5> %I = insertelement <2 x i1> zeroinitializer, i1 false, i32 1 %FC = sitofp <4 x i32> zeroinitializer to <4 x double> %Sl = select i1 false, <4 x i64> %Shuff, <4 x i64> %Shuff - %L5 = load i8* %0 + %L5 = load i8, i8* %0 store i8 %5, i8* %0 %E6 = extractelement <1 x i16> zeroinitializer, i32 0 %Shuff7 = shufflevector <2 x i1> %I, <2 x i1> %I, <2 x i32> <i32 1, i32 undef> @@ -30,7 +30,7 @@ BB: %FC9 = fptoui float 0x406DB70180000000 to i64 %Sl10 = select i1 false, <8 x i32> <i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1>, <8 x i32> <i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1> %Cmp = icmp ult <4 x i64> zeroinitializer, zeroinitializer - %L11 = load i8* %0 + %L11 = load i8, i8* %0 store i8 %L, i8* %0 %E12 = extractelement <4 x i64> zeroinitializer, i32 2 %Shuff13 = shufflevector <4 x i32> zeroinitializer, <4 x i32> zeroinitializer, <4 x i32> <i32 5, i32 7, i32 undef, i32 3> @@ -42,7 +42,7 @@ BB: br label %CF CF: ; preds = %CF, %CF79, %CF84, %BB - %L18 = load i8* %0 + %L18 = load i8, i8* %0 store i8 %L, i8* %0 %E19 = extractelement <4 x i64> %Sl, i32 3 %Shuff20 = shufflevector <2 x i1> %Shuff7, <2 x i1> %I, <2 x i32> <i32 2, i32 0> @@ -54,7 +54,7 @@ CF: ; preds = %CF, %CF79, %CF84, % br i1 %Cmp25, label %CF, label %CF79 CF79: ; preds = %CF - %L26 = load i8* %0 + %L26 = load i8, i8* %0 store i8 %L26, i8* %0 %E27 = extractelement <1 x i16> zeroinitializer, i32 0 %Shuff28 = shufflevector <16 x i32> <i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1>, <16 x i32> <i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1>, <16 x i32> <i32 13, i32 15, i32 17, i32 19, i32 21, i32 23, i32 25, i32 27, i32 29, i32 31, i32 1, i32 3, i32 5, i32 7, i32 9, i32 11> @@ -65,7 +65,7 @@ CF79: ; preds = %CF br i1 %Cmp32, label %CF, label %CF78 CF78: ; preds = %CF78, %CF79 - %L33 = load i8* %0 + %L33 = load i8, i8* %0 store i8 %L, i8* %0 %E34 = extractelement <16 x i32> %Shuff28, i32 1 %Shuff35 = shufflevector <4 x i64> zeroinitializer, <4 x i64> %I21, <4 x i32> <i32 undef, i32 6, i32 0, i32 2> @@ -76,7 +76,7 @@ CF78: ; preds = %CF78, %CF79 br i1 %Cmp38, label %CF78, label %CF80 CF80: ; preds = %CF80, %CF82, %CF78 - %L39 = load i8* %0 + %L39 = load i8, i8* %0 store i8 %L, i8* %0 %E40 = extractelement <2 x i1> %Shuff20, i32 1 br i1 %E40, label %CF80, label %CF82 @@ -87,7 +87,7 @@ CF82: ; preds = %CF80 %B43 = sub i32 %E, 0 %Sl44 = select i1 %Cmp32, <16 x i32> %Shuff28, <16 x i32> %Shuff28 %Cmp45 = icmp sgt <4 x i64> zeroinitializer, %I21 - %L46 = load i8* %0 + %L46 = load i8, i8* %0 store i8 %L11, i8* %0 %E47 = extractelement <8 x i32> %Sl16, i32 4 %Shuff48 = shufflevector <2 x i1> zeroinitializer, <2 x i1> %Shuff7, <2 x i32> <i32 undef, i32 1> @@ -99,7 +99,7 @@ CF82: ; preds = %CF80 CF81: ; preds = %CF81, %CF82 %Sl52 = select i1 false, float -6.749110e+06, float 0x406DB70180000000 %Cmp53 = icmp uge <2 x i32> <i32 -1, i32 -1>, <i32 -1, i32 -1> - %L54 = load i8* %0 + %L54 = load i8, i8* %0 store i8 %L5, i8* %0 %E55 = extractelement <8 x i32> zeroinitializer, i32 7 %Shuff56 = shufflevector <4 x i64> zeroinitializer, <4 x i64> zeroinitializer, <4 x i32> <i32 undef, i32 4, i32 6, i32 0> @@ -108,7 +108,7 @@ CF81: ; preds = %CF81, %CF82 %FC59 = fptoui <4 x double> %I36 to <4 x i16> %Sl60 = select i1 %Cmp17, <2 x i1> %I, <2 x i1> %I57 %Cmp61 = icmp ule <8 x i32> %B50, <i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1> - %L62 = load i8* %0 + %L62 = load i8, i8* %0 store i8 %L33, i8* %0 %E63 = extractelement <4 x i64> %Shuff, i32 2 %Shuff64 = shufflevector <4 x i64> %Shuff56, <4 x i64> %Shuff56, <4 x i32> <i32 5, i32 7, i32 1, i32 undef> @@ -126,7 +126,7 @@ CF84: ; preds = %CF83 br i1 %Cmp69, label %CF, label %CF77 CF77: ; preds = %CF84 - %L70 = load i8* %0 + %L70 = load i8, i8* %0 store i8 %L, i8* %0 %E71 = extractelement <4 x i64> %Shuff, i32 0 %Shuff72 = shufflevector <2 x i1> zeroinitializer, <2 x i1> %I, <2 x i32> <i32 3, i32 1> diff --git a/test/CodeGen/Mips/msa/shuffle.ll b/test/CodeGen/Mips/msa/shuffle.ll index faeec5d58dd49..aadff7d5a5c83 100644 --- a/test/CodeGen/Mips/msa/shuffle.ll +++ b/test/CodeGen/Mips/msa/shuffle.ll @@ -2,9 +2,9 @@ ; RUN: llc -march=mipsel -mattr=+msa,+fp64 < %s | FileCheck %s define void @vshf_v16i8_0(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind { - ; CHECK: vshf_v16i8_0: + ; CHECK-LABEL: vshf_v16i8_0: - %1 = load <16 x i8>* %a + %1 = load <16 x i8>, <16 x i8>* %a ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5) %2 = shufflevector <16 x i8> %1, <16 x i8> undef, <16 x i32> <i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0> ; CHECK-DAG: addiu [[PTR_A:\$[0-9]+]], {{.*}}, %lo($ @@ -14,13 +14,12 @@ define void @vshf_v16i8_0(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind ; CHECK-DAG: st.b [[R3]], 0($4) ret void - ; CHECK: .size vshf_v16i8_0 } define void @vshf_v16i8_1(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind { - ; CHECK: vshf_v16i8_1: + ; CHECK-LABEL: vshf_v16i8_1: - %1 = load <16 x i8>* %a + %1 = load <16 x i8>, <16 x i8>* %a ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5) %2 = shufflevector <16 x i8> %1, <16 x i8> undef, <16 x i32> <i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1> ; CHECK-DAG: splati.b [[R3:\$w[0-9]+]], [[R1]][1] @@ -28,14 +27,13 @@ define void @vshf_v16i8_1(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind ; CHECK-DAG: st.b [[R3]], 0($4) ret void - ; CHECK: .size vshf_v16i8_1 } define void @vshf_v16i8_2(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind { - ; CHECK: vshf_v16i8_2: + ; CHECK-LABEL: vshf_v16i8_2: - %1 = load <16 x i8>* %a - %2 = load <16 x i8>* %b + %1 = load <16 x i8>, <16 x i8>* %a + %2 = load <16 x i8>, <16 x i8>* %b ; CHECK-DAG: ld.b [[R2:\$w[0-9]+]], 0($6) %3 = shufflevector <16 x i8> %1, <16 x i8> %2, <16 x i32> <i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 16> ; CHECK-DAG: addiu [[PTR_A:\$[0-9]+]], {{.*}}, %lo($ @@ -45,15 +43,14 @@ define void @vshf_v16i8_2(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind ; CHECK-DAG: st.b [[R3]], 0($4) ret void - ; CHECK: .size vshf_v16i8_2 } define void @vshf_v16i8_3(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind { - ; CHECK: vshf_v16i8_3: + ; CHECK-LABEL: vshf_v16i8_3: - %1 = load <16 x i8>* %a + %1 = load <16 x i8>, <16 x i8>* %a ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5) - %2 = load <16 x i8>* %b + %2 = load <16 x i8>, <16 x i8>* %b ; CHECK-DAG: ld.b [[R2:\$w[0-9]+]], 0($6) %3 = shufflevector <16 x i8> %1, <16 x i8> %2, <16 x i32> <i32 17, i32 24, i32 25, i32 18, i32 19, i32 20, i32 28, i32 19, i32 1, i32 8, i32 9, i32 2, i32 3, i32 4, i32 12, i32 3> ; CHECK-DAG: addiu [[PTR_A:\$[0-9]+]], {{.*}}, %lo($ @@ -65,13 +62,12 @@ define void @vshf_v16i8_3(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind ; CHECK-DAG: st.b [[R3]], 0($4) ret void - ; CHECK: .size vshf_v16i8_3 } define void @vshf_v16i8_4(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind { - ; CHECK: vshf_v16i8_4: + ; CHECK-LABEL: vshf_v16i8_4: - %1 = load <16 x i8>* %a + %1 = load <16 x i8>, <16 x i8>* %a ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5) %2 = shufflevector <16 x i8> %1, <16 x i8> %1, <16 x i32> <i32 1, i32 17, i32 1, i32 17, i32 1, i32 17, i32 1, i32 17, i32 1, i32 17, i32 1, i32 17, i32 1, i32 17, i32 1, i32 17> ; CHECK-DAG: splati.b [[R3:\$w[0-9]+]], [[R1]][1] @@ -79,13 +75,12 @@ define void @vshf_v16i8_4(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind ; CHECK-DAG: st.b [[R3]], 0($4) ret void - ; CHECK: .size vshf_v16i8_4 } define void @vshf_v8i16_0(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind { - ; CHECK: vshf_v8i16_0: + ; CHECK-LABEL: vshf_v8i16_0: - %1 = load <8 x i16>* %a + %1 = load <8 x i16>, <8 x i16>* %a ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5) %2 = shufflevector <8 x i16> %1, <8 x i16> undef, <8 x i32> <i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0> ; CHECK-DAG: addiu [[PTR_A:\$[0-9]+]], {{.*}}, %lo($ @@ -95,13 +90,12 @@ define void @vshf_v8i16_0(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind ; CHECK-DAG: st.h [[R3]], 0($4) ret void - ; CHECK: .size vshf_v8i16_0 } define void @vshf_v8i16_1(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind { - ; CHECK: vshf_v8i16_1: + ; CHECK-LABEL: vshf_v8i16_1: - %1 = load <8 x i16>* %a + %1 = load <8 x i16>, <8 x i16>* %a ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5) %2 = shufflevector <8 x i16> %1, <8 x i16> undef, <8 x i32> <i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1> ; CHECK-DAG: splati.h [[R3:\$w[0-9]+]], [[R1]][1] @@ -109,14 +103,13 @@ define void @vshf_v8i16_1(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind ; CHECK-DAG: st.h [[R3]], 0($4) ret void - ; CHECK: .size vshf_v8i16_1 } define void @vshf_v8i16_2(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind { - ; CHECK: vshf_v8i16_2: + ; CHECK-LABEL: vshf_v8i16_2: - %1 = load <8 x i16>* %a - %2 = load <8 x i16>* %b + %1 = load <8 x i16>, <8 x i16>* %a + %2 = load <8 x i16>, <8 x i16>* %b ; CHECK-DAG: ld.h [[R2:\$w[0-9]+]], 0($6) %3 = shufflevector <8 x i16> %1, <8 x i16> %2, <8 x i32> <i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 8> ; CHECK-DAG: addiu [[PTR_A:\$[0-9]+]], {{.*}}, %lo($ @@ -126,15 +119,14 @@ define void @vshf_v8i16_2(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind ; CHECK-DAG: st.h [[R3]], 0($4) ret void - ; CHECK: .size vshf_v8i16_2 } define void @vshf_v8i16_3(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind { - ; CHECK: vshf_v8i16_3: + ; CHECK-LABEL: vshf_v8i16_3: - %1 = load <8 x i16>* %a + %1 = load <8 x i16>, <8 x i16>* %a ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5) - %2 = load <8 x i16>* %b + %2 = load <8 x i16>, <8 x i16>* %b ; CHECK-DAG: ld.h [[R2:\$w[0-9]+]], 0($6) %3 = shufflevector <8 x i16> %1, <8 x i16> %2, <8 x i32> <i32 1, i32 8, i32 9, i32 2, i32 3, i32 4, i32 12, i32 3> ; CHECK-DAG: addiu [[PTR_A:\$[0-9]+]], {{.*}}, %lo($ @@ -146,13 +138,12 @@ define void @vshf_v8i16_3(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind ; CHECK-DAG: st.h [[R3]], 0($4) ret void - ; CHECK: .size vshf_v8i16_3 } define void @vshf_v8i16_4(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind { - ; CHECK: vshf_v8i16_4: + ; CHECK-LABEL: vshf_v8i16_4: - %1 = load <8 x i16>* %a + %1 = load <8 x i16>, <8 x i16>* %a ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5) %2 = shufflevector <8 x i16> %1, <8 x i16> %1, <8 x i32> <i32 1, i32 9, i32 1, i32 9, i32 1, i32 9, i32 1, i32 9> ; CHECK-DAG: splati.h [[R3:\$w[0-9]+]], [[R1]][1] @@ -160,16 +151,15 @@ define void @vshf_v8i16_4(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind ; CHECK-DAG: st.h [[R3]], 0($4) ret void - ; CHECK: .size vshf_v8i16_4 } ; Note: v4i32 only has one 4-element set so it's impossible to get a vshf.w ; instruction when using a single vector. define void @vshf_v4i32_0(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind { - ; CHECK: vshf_v4i32_0: + ; CHECK-LABEL: vshf_v4i32_0: - %1 = load <4 x i32>* %a + %1 = load <4 x i32>, <4 x i32>* %a ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5) %2 = shufflevector <4 x i32> %1, <4 x i32> undef, <4 x i32> <i32 3, i32 2, i32 1, i32 0> ; CHECK-DAG: shf.w [[R3:\$w[0-9]+]], [[R1]], 27 @@ -177,28 +167,26 @@ define void @vshf_v4i32_0(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind ; CHECK-DAG: st.w [[R3]], 0($4) ret void - ; CHECK: .size vshf_v4i32_0 } define void @vshf_v4i32_1(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind { - ; CHECK: vshf_v4i32_1: + ; CHECK-LABEL: vshf_v4i32_1: - %1 = load <4 x i32>* %a + %1 = load <4 x i32>, <4 x i32>* %a ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5) %2 = shufflevector <4 x i32> %1, <4 x i32> undef, <4 x i32> <i32 1, i32 1, i32 1, i32 1> - ; CHECK-DAG: shf.w [[R3:\$w[0-9]+]], [[R1]], 85 + ; CHECK-DAG: splati.w [[R3:\$w[0-9]+]], [[R1]][1] store <4 x i32> %2, <4 x i32>* %c ; CHECK-DAG: st.w [[R3]], 0($4) ret void - ; CHECK: .size vshf_v4i32_1 } define void @vshf_v4i32_2(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind { - ; CHECK: vshf_v4i32_2: + ; CHECK-LABEL: vshf_v4i32_2: - %1 = load <4 x i32>* %a - %2 = load <4 x i32>* %b + %1 = load <4 x i32>, <4 x i32>* %a + %2 = load <4 x i32>, <4 x i32>* %b ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6) %3 = shufflevector <4 x i32> %1, <4 x i32> %2, <4 x i32> <i32 4, i32 5, i32 6, i32 4> ; CHECK-DAG: shf.w [[R3:\$w[0-9]+]], [[R2]], 36 @@ -206,15 +194,14 @@ define void @vshf_v4i32_2(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind ; CHECK-DAG: st.w [[R3]], 0($4) ret void - ; CHECK: .size vshf_v4i32_2 } define void @vshf_v4i32_3(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind { - ; CHECK: vshf_v4i32_3: + ; CHECK-LABEL: vshf_v4i32_3: - %1 = load <4 x i32>* %a + %1 = load <4 x i32>, <4 x i32>* %a ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5) - %2 = load <4 x i32>* %b + %2 = load <4 x i32>, <4 x i32>* %b ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6) %3 = shufflevector <4 x i32> %1, <4 x i32> %2, <4 x i32> <i32 1, i32 5, i32 6, i32 4> ; CHECK-DAG: addiu [[PTR_A:\$[0-9]+]], {{.*}}, %lo($ @@ -226,27 +213,26 @@ define void @vshf_v4i32_3(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind ; CHECK-DAG: st.w [[R3]], 0($4) ret void - ; CHECK: .size vshf_v4i32_3 } define void @vshf_v4i32_4(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind { - ; CHECK: vshf_v4i32_4: + ; CHECK-LABEL: vshf_v4i32_4: - %1 = load <4 x i32>* %a + %1 = load <4 x i32>, <4 x i32>* %a ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5) %2 = shufflevector <4 x i32> %1, <4 x i32> %1, <4 x i32> <i32 1, i32 5, i32 5, i32 1> - ; CHECK-DAG: shf.w [[R3:\$w[0-9]+]], [[R1]], 85 + ; The two operand vectors are the same so element 1 and 5 are equivalent. + ; CHECK-DAG: splati.w [[R3:\$w[0-9]+]], [[R1]][1] store <4 x i32> %2, <4 x i32>* %c ; CHECK-DAG: st.w [[R3]], 0($4) ret void - ; CHECK: .size vshf_v4i32_4 } define void @vshf_v2i64_0(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind { - ; CHECK: vshf_v2i64_0: + ; CHECK-LABEL: vshf_v2i64_0: - %1 = load <2 x i64>* %a + %1 = load <2 x i64>, <2 x i64>* %a ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5) %2 = shufflevector <2 x i64> %1, <2 x i64> undef, <2 x i32> <i32 1, i32 0> ; CHECK-DAG: addiu [[PTR_A:\$[0-9]+]], {{.*}}, %lo($ @@ -256,13 +242,12 @@ define void @vshf_v2i64_0(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind ; CHECK-DAG: st.d [[R3]], 0($4) ret void - ; CHECK: .size vshf_v2i64_0 } define void @vshf_v2i64_1(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind { - ; CHECK: vshf_v2i64_1: + ; CHECK-LABEL: vshf_v2i64_1: - %1 = load <2 x i64>* %a + %1 = load <2 x i64>, <2 x i64>* %a ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5) %2 = shufflevector <2 x i64> %1, <2 x i64> undef, <2 x i32> <i32 1, i32 1> ; CHECK-DAG: splati.d [[R3:\$w[0-9]+]], [[R1]][1] @@ -270,14 +255,13 @@ define void @vshf_v2i64_1(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind ; CHECK-DAG: st.d [[R3]], 0($4) ret void - ; CHECK: .size vshf_v2i64_1 } define void @vshf_v2i64_2(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind { - ; CHECK: vshf_v2i64_2: + ; CHECK-LABEL: vshf_v2i64_2: - %1 = load <2 x i64>* %a - %2 = load <2 x i64>* %b + %1 = load <2 x i64>, <2 x i64>* %a + %2 = load <2 x i64>, <2 x i64>* %b ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6) %3 = shufflevector <2 x i64> %1, <2 x i64> %2, <2 x i32> <i32 3, i32 2> ; CHECK-DAG: addiu [[PTR_A:\$[0-9]+]], {{.*}}, %lo($ @@ -287,15 +271,14 @@ define void @vshf_v2i64_2(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind ; CHECK-DAG: st.d [[R3]], 0($4) ret void - ; CHECK: .size vshf_v2i64_2 } define void @vshf_v2i64_3(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind { - ; CHECK: vshf_v2i64_3: + ; CHECK-LABEL: vshf_v2i64_3: - %1 = load <2 x i64>* %a + %1 = load <2 x i64>, <2 x i64>* %a ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5) - %2 = load <2 x i64>* %b + %2 = load <2 x i64>, <2 x i64>* %b ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6) %3 = shufflevector <2 x i64> %1, <2 x i64> %2, <2 x i32> <i32 1, i32 2> ; CHECK-DAG: addiu [[PTR_A:\$[0-9]+]], {{.*}}, %lo($ @@ -307,13 +290,12 @@ define void @vshf_v2i64_3(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind ; CHECK-DAG: st.d [[R3]], 0($4) ret void - ; CHECK: .size vshf_v2i64_3 } define void @vshf_v2i64_4(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind { - ; CHECK: vshf_v2i64_4: + ; CHECK-LABEL: vshf_v2i64_4: - %1 = load <2 x i64>* %a + %1 = load <2 x i64>, <2 x i64>* %a ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5) %2 = shufflevector <2 x i64> %1, <2 x i64> %1, <2 x i32> <i32 1, i32 3> ; CHECK-DAG: splati.d [[R3:\$w[0-9]+]], [[R1]][1] @@ -321,13 +303,12 @@ define void @vshf_v2i64_4(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind ; CHECK-DAG: st.d [[R3]], 0($4) ret void - ; CHECK: .size vshf_v2i64_4 } define void @shf_v16i8_0(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind { - ; CHECK: shf_v16i8_0: + ; CHECK-LABEL: shf_v16i8_0: - %1 = load <16 x i8>* %a + %1 = load <16 x i8>, <16 x i8>* %a ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5) %2 = shufflevector <16 x i8> %1, <16 x i8> undef, <16 x i32> <i32 1, i32 3, i32 2, i32 0, i32 5, i32 7, i32 6, i32 4, i32 9, i32 11, i32 10, i32 8, i32 13, i32 15, i32 14, i32 12> ; CHECK-DAG: shf.b [[R3:\$w[0-9]+]], [[R1]], 45 @@ -335,13 +316,12 @@ define void @shf_v16i8_0(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind { ; CHECK-DAG: st.b [[R3]], 0($4) ret void - ; CHECK: .size shf_v16i8_0 } define void @shf_v8i16_0(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind { - ; CHECK: shf_v8i16_0: + ; CHECK-LABEL: shf_v8i16_0: - %1 = load <8 x i16>* %a + %1 = load <8 x i16>, <8 x i16>* %a ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5) %2 = shufflevector <8 x i16> %1, <8 x i16> undef, <8 x i32> <i32 3, i32 2, i32 1, i32 0, i32 7, i32 6, i32 5, i32 4> ; CHECK-DAG: shf.h [[R3:\$w[0-9]+]], [[R1]], 27 @@ -349,13 +329,12 @@ define void @shf_v8i16_0(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind { ; CHECK-DAG: st.h [[R3]], 0($4) ret void - ; CHECK: .size shf_v8i16_0 } define void @shf_v4i32_0(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind { - ; CHECK: shf_v4i32_0: + ; CHECK-LABEL: shf_v4i32_0: - %1 = load <4 x i32>* %a + %1 = load <4 x i32>, <4 x i32>* %a ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5) %2 = shufflevector <4 x i32> %1, <4 x i32> undef, <4 x i32> <i32 3, i32 2, i32 1, i32 0> ; CHECK-DAG: shf.w [[R3:\$w[0-9]+]], [[R1]], 27 @@ -363,82 +342,194 @@ define void @shf_v4i32_0(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind { ; CHECK-DAG: st.w [[R3]], 0($4) ret void - ; CHECK: .size shf_v4i32_0 } ; shf.d does not exist define void @ilvev_v16i8_0(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind { - ; CHECK: ilvev_v16i8_0: + ; CHECK-LABEL: ilvev_v16i8_0: - %1 = load <16 x i8>* %a + %1 = load <16 x i8>, <16 x i8>* %a ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5) - %2 = load <16 x i8>* %b + %2 = load <16 x i8>, <16 x i8>* %b ; CHECK-DAG: ld.b [[R2:\$w[0-9]+]], 0($6) %3 = shufflevector <16 x i8> %1, <16 x i8> %2, <16 x i32> <i32 0, i32 16, i32 2, i32 18, i32 4, i32 20, i32 6, i32 22, i32 8, i32 24, i32 10, i32 26, i32 12, i32 28, i32 14, i32 30> - ; CHECK-DAG: ilvev.b [[R3:\$w[0-9]+]], [[R1]], [[R2]] + ; CHECK-DAG: ilvev.b [[R3:\$w[0-9]+]], [[R2]], [[R1]] store <16 x i8> %3, <16 x i8>* %c ; CHECK-DAG: st.b [[R3]], 0($4) ret void - ; CHECK: .size ilvev_v16i8_0 } define void @ilvev_v8i16_0(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind { - ; CHECK: ilvev_v8i16_0: + ; CHECK-LABEL: ilvev_v8i16_0: - %1 = load <8 x i16>* %a + %1 = load <8 x i16>, <8 x i16>* %a ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5) - %2 = load <8 x i16>* %b + %2 = load <8 x i16>, <8 x i16>* %b ; CHECK-DAG: ld.h [[R2:\$w[0-9]+]], 0($6) %3 = shufflevector <8 x i16> %1, <8 x i16> %2, <8 x i32> <i32 0, i32 8, i32 2, i32 10, i32 4, i32 12, i32 6, i32 14> - ; CHECK-DAG: ilvev.h [[R3:\$w[0-9]+]], [[R1]], [[R2]] + ; CHECK-DAG: ilvev.h [[R3:\$w[0-9]+]], [[R2]], [[R1]] store <8 x i16> %3, <8 x i16>* %c ; CHECK-DAG: st.h [[R3]], 0($4) ret void - ; CHECK: .size ilvev_v8i16_0 } define void @ilvev_v4i32_0(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind { - ; CHECK: ilvev_v4i32_0: + ; CHECK-LABEL: ilvev_v4i32_0: - %1 = load <4 x i32>* %a + %1 = load <4 x i32>, <4 x i32>* %a ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5) - %2 = load <4 x i32>* %b + %2 = load <4 x i32>, <4 x i32>* %b ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6) %3 = shufflevector <4 x i32> %1, <4 x i32> %2, <4 x i32> <i32 0, i32 4, i32 2, i32 6> - ; CHECK-DAG: ilvev.w [[R3:\$w[0-9]+]], [[R1]], [[R2]] + ; CHECK-DAG: ilvev.w [[R3:\$w[0-9]+]], [[R2]], [[R1]] store <4 x i32> %3, <4 x i32>* %c ; CHECK-DAG: st.w [[R3]], 0($4) ret void - ; CHECK: .size ilvev_v4i32_0 } define void @ilvev_v2i64_0(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind { - ; CHECK: ilvev_v2i64_0: + ; CHECK-LABEL: ilvev_v2i64_0: - %1 = load <2 x i64>* %a + %1 = load <2 x i64>, <2 x i64>* %a ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5) - %2 = load <2 x i64>* %b + %2 = load <2 x i64>, <2 x i64>* %b ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6) %3 = shufflevector <2 x i64> %1, <2 x i64> %2, <2 x i32> <i32 0, i32 2> - ; CHECK-DAG: ilvev.d [[R3:\$w[0-9]+]], [[R1]], [[R2]] + ; CHECK-DAG: ilvev.d [[R3:\$w[0-9]+]], [[R2]], [[R1]] + store <2 x i64> %3, <2 x i64>* %c + ; CHECK-DAG: st.d [[R3]], 0($4) + + ret void +} + +; Interleaving one operand with itself. +define void @ilvev_v16i8_1(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind { + ; CHECK-LABEL: ilvev_v16i8_1: + + %1 = load <16 x i8>, <16 x i8>* %a + ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5) + %2 = load <16 x i8>, <16 x i8>* %b + %3 = shufflevector <16 x i8> %1, <16 x i8> %2, + <16 x i32> <i32 0, i32 0, i32 2, i32 2, i32 4, i32 4, i32 6, i32 6, i32 8, i32 8, i32 10, i32 10, i32 12, i32 12, i32 14, i32 14> + ; CHECK-DAG: ilvev.b [[R3:\$w[0-9]+]], [[R1]], [[R1]] + store <16 x i8> %3, <16 x i8>* %c + ; CHECK-DAG: st.b [[R3]], 0($4) + + ret void +} + +define void @ilvev_v8i16_1(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind { + ; CHECK-LABEL: ilvev_v8i16_1: + + %1 = load <8 x i16>, <8 x i16>* %a + ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5) + %2 = load <8 x i16>, <8 x i16>* %b + %3 = shufflevector <8 x i16> %1, <8 x i16> %2, <8 x i32> <i32 0, i32 0, i32 2, i32 2, i32 4, i32 4, i32 6, i32 6> + ; CHECK-DAG: ilvev.h [[R3:\$w[0-9]+]], [[R1]], [[R1]] + store <8 x i16> %3, <8 x i16>* %c + ; CHECK-DAG: st.h [[R3]], 0($4) + + ret void +} + +define void @ilvev_v4i32_1(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind { + ; CHECK-LABEL: ilvev_v4i32_1: + + %1 = load <4 x i32>, <4 x i32>* %a + ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5) + %2 = load <4 x i32>, <4 x i32>* %b + %3 = shufflevector <4 x i32> %1, <4 x i32> %2, <4 x i32> <i32 0, i32 0, i32 2, i32 2> + ; CHECK-DAG: ilvev.w [[R3:\$w[0-9]+]], [[R1]], [[R1]] + store <4 x i32> %3, <4 x i32>* %c + ; CHECK-DAG: st.w [[R3]], 0($4) + + ret void +} + +define void @ilvev_v2i64_1(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind { + ; CHECK-LABEL: ilvev_v2i64_1: + + %1 = load <2 x i64>, <2 x i64>* %a + ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5) + %2 = load <2 x i64>, <2 x i64>* %b + %3 = shufflevector <2 x i64> %1, <2 x i64> %2, <2 x i32> <i32 0, i32 0> + ; ilvev.d with two identical operands is equivalent to splati.d + ; CHECK-DAG: splati.d [[R3:\$w[0-9]+]], [[R1]][0] + store <2 x i64> %3, <2 x i64>* %c + ; CHECK-DAG: st.d [[R3]], 0($4) + + ret void +} + +define void @ilvev_v16i8_2(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind { + ; CHECK-LABEL: ilvev_v16i8_2: + + %1 = load <16 x i8>, <16 x i8>* %a + %2 = load <16 x i8>, <16 x i8>* %b + ; CHECK-DAG: ld.b [[R2:\$w[0-9]+]], 0($6) + %3 = shufflevector <16 x i8> %1, <16 x i8> %2, + <16 x i32> <i32 16, i32 16, i32 18, i32 18, i32 20, i32 20, i32 22, i32 22, i32 24, i32 24, i32 26, i32 26, i32 28, i32 28, i32 30, i32 30> + ; CHECK-DAG: ilvev.b [[R3:\$w[0-9]+]], [[R2]], [[R2]] + store <16 x i8> %3, <16 x i8>* %c + ; CHECK-DAG: st.b [[R3]], 0($4) + + ret void +} + +define void @ilvev_v8i16_2(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind { + ; CHECK-LABEL: ilvev_v8i16_2: + + %1 = load <8 x i16>, <8 x i16>* %a + %2 = load <8 x i16>, <8 x i16>* %b + ; CHECK-DAG: ld.h [[R2:\$w[0-9]+]], 0($6) + %3 = shufflevector <8 x i16> %1, <8 x i16> %2, <8 x i32> <i32 8, i32 8, i32 10, i32 10, i32 12, i32 12, i32 14, i32 14> + ; CHECK-DAG: ilvev.h [[R3:\$w[0-9]+]], [[R2]], [[R2]] + store <8 x i16> %3, <8 x i16>* %c + ; CHECK-DAG: st.h [[R3]], 0($4) + + ret void +} + +define void @ilvev_v4i32_2(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind { + ; CHECK-LABEL: ilvev_v4i32_2: + + %1 = load <4 x i32>, <4 x i32>* %a + %2 = load <4 x i32>, <4 x i32>* %b + ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6) + %3 = shufflevector <4 x i32> %1, <4 x i32> %2, <4 x i32> <i32 4, i32 4, i32 6, i32 6> + ; CHECK-DAG: ilvev.w [[R3:\$w[0-9]+]], [[R2]], [[R2]] + store <4 x i32> %3, <4 x i32>* %c + ; CHECK-DAG: st.w [[R3]], 0($4) + + ret void +} + +define void @ilvev_v2i64_2(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind { + ; CHECK-LABEL: ilvev_v2i64_2: + + %1 = load <2 x i64>, <2 x i64>* %a + %2 = load <2 x i64>, <2 x i64>* %b + ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6) + %3 = shufflevector <2 x i64> %1, <2 x i64> %2, <2 x i32> <i32 2, i32 2> + ; ilvev.d with two identical operands is equivalent to splati.d + ; CHECK-DAG: splati.d [[R3:\$w[0-9]+]], [[R2]][0] store <2 x i64> %3, <2 x i64>* %c ; CHECK-DAG: st.d [[R3]], 0($4) ret void - ; CHECK: .size ilvev_v2i64_0 } define void @ilvod_v16i8_0(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind { - ; CHECK: ilvod_v16i8_0: + ; CHECK-LABEL: ilvod_v16i8_0: - %1 = load <16 x i8>* %a + %1 = load <16 x i8>, <16 x i8>* %a ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5) - %2 = load <16 x i8>* %b + %2 = load <16 x i8>, <16 x i8>* %b ; CHECK-DAG: ld.b [[R2:\$w[0-9]+]], 0($6) %3 = shufflevector <16 x i8> %1, <16 x i8> %2, <16 x i32> <i32 1, i32 17, i32 3, i32 19, i32 5, i32 21, i32 7, i32 23, i32 9, i32 25, i32 11, i32 27, i32 13, i32 29, i32 15, i32 31> @@ -447,15 +538,14 @@ define void @ilvod_v16i8_0(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind ; CHECK-DAG: st.b [[R3]], 0($4) ret void - ; CHECK: .size ilvod_v16i8_0 } define void @ilvod_v8i16_0(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind { - ; CHECK: ilvod_v8i16_0: + ; CHECK-LABEL: ilvod_v8i16_0: - %1 = load <8 x i16>* %a + %1 = load <8 x i16>, <8 x i16>* %a ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5) - %2 = load <8 x i16>* %b + %2 = load <8 x i16>, <8 x i16>* %b ; CHECK-DAG: ld.h [[R2:\$w[0-9]+]], 0($6) %3 = shufflevector <8 x i16> %1, <8 x i16> %2, <8 x i32> <i32 1, i32 9, i32 3, i32 11, i32 5, i32 13, i32 7, i32 15> ; CHECK-DAG: ilvod.h [[R3:\$w[0-9]+]], [[R1]], [[R2]] @@ -463,15 +553,14 @@ define void @ilvod_v8i16_0(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind ; CHECK-DAG: st.h [[R3]], 0($4) ret void - ; CHECK: .size ilvod_v8i16_0 } define void @ilvod_v4i32_0(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind { - ; CHECK: ilvod_v4i32_0: + ; CHECK-LABEL: ilvod_v4i32_0: - %1 = load <4 x i32>* %a + %1 = load <4 x i32>, <4 x i32>* %a ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5) - %2 = load <4 x i32>* %b + %2 = load <4 x i32>, <4 x i32>* %b ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6) %3 = shufflevector <4 x i32> %1, <4 x i32> %2, <4 x i32> <i32 1, i32 5, i32 3, i32 7> ; CHECK-DAG: ilvod.w [[R3:\$w[0-9]+]], [[R1]], [[R2]] @@ -479,15 +568,14 @@ define void @ilvod_v4i32_0(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind ; CHECK-DAG: st.w [[R3]], 0($4) ret void - ; CHECK: .size ilvod_v4i32_0 } define void @ilvod_v2i64_0(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind { - ; CHECK: ilvod_v2i64_0: + ; CHECK-LABEL: ilvod_v2i64_0: - %1 = load <2 x i64>* %a + %1 = load <2 x i64>, <2 x i64>* %a ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5) - %2 = load <2 x i64>* %b + %2 = load <2 x i64>, <2 x i64>* %b ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6) %3 = shufflevector <2 x i64> %1, <2 x i64> %2, <2 x i32> <i32 1, i32 3> ; CHECK-DAG: ilvod.d [[R3:\$w[0-9]+]], [[R1]], [[R2]] @@ -495,262 +583,710 @@ define void @ilvod_v2i64_0(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind ; CHECK-DAG: st.d [[R3]], 0($4) ret void - ; CHECK: .size ilvod_v2i64_0 } -define void @ilvl_v16i8_0(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind { - ; CHECK: ilvl_v16i8_0: +define void @ilvod_v16i8_1(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind { + ; CHECK-LABEL: ilvod_v16i8_1: + + %1 = load <16 x i8>, <16 x i8>* %a + ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5) + %2 = load <16 x i8>, <16 x i8>* %b + %3 = shufflevector <16 x i8> %1, <16 x i8> %2, + <16 x i32> <i32 1, i32 1, i32 3, i32 3, i32 5, i32 5, i32 7, i32 7, i32 9, i32 9, i32 11, i32 11, i32 13, i32 13, i32 15, i32 15> + ; CHECK-DAG: ilvod.b [[R3:\$w[0-9]+]], [[R1]], [[R1]] + store <16 x i8> %3, <16 x i8>* %c + ; CHECK-DAG: st.b [[R3]], 0($4) + + ret void +} + +define void @ilvod_v8i16_1(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind { + ; CHECK-LABEL: ilvod_v8i16_1: + + %1 = load <8 x i16>, <8 x i16>* %a + ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5) + %2 = load <8 x i16>, <8 x i16>* %b + %3 = shufflevector <8 x i16> %1, <8 x i16> %2, <8 x i32> <i32 1, i32 1, i32 3, i32 3, i32 5, i32 5, i32 7, i32 7> + ; CHECK-DAG: ilvod.h [[R3:\$w[0-9]+]], [[R1]], [[R1]] + store <8 x i16> %3, <8 x i16>* %c + ; CHECK-DAG: st.h [[R3]], 0($4) + + ret void +} + +define void @ilvod_v4i32_1(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind { + ; CHECK-LABEL: ilvod_v4i32_1: + + %1 = load <4 x i32>, <4 x i32>* %a + ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5) + %2 = load <4 x i32>, <4 x i32>* %b + %3 = shufflevector <4 x i32> %1, <4 x i32> %2, <4 x i32> <i32 1, i32 1, i32 3, i32 3> + ; CHECK-DAG: ilvod.w [[R3:\$w[0-9]+]], [[R1]], [[R1]] + store <4 x i32> %3, <4 x i32>* %c + ; CHECK-DAG: st.w [[R3]], 0($4) + + ret void +} + +define void @ilvod_v2i64_1(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind { + ; CHECK-LABEL: ilvod_v2i64_1: + + %1 = load <2 x i64>, <2 x i64>* %a + ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5) + %2 = load <2 x i64>, <2 x i64>* %b + %3 = shufflevector <2 x i64> %1, <2 x i64> %2, <2 x i32> <i32 1, i32 1> + ; ilvod.d with two identical operands is equivalent to splati.d + ; CHECK-DAG: splati.d [[R3:\$w[0-9]+]], [[R1]][1] + store <2 x i64> %3, <2 x i64>* %c + ; CHECK-DAG: st.d [[R3]], 0($4) + + ret void +} + +define void @ilvod_v16i8_2(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind { + ; CHECK-LABEL: ilvod_v16i8_2: + + %1 = load <16 x i8>, <16 x i8>* %a + %2 = load <16 x i8>, <16 x i8>* %b + ; CHECK-DAG: ld.b [[R2:\$w[0-9]+]], 0($6) + %3 = shufflevector <16 x i8> %1, <16 x i8> %2, + <16 x i32> <i32 17, i32 17, i32 19, i32 19, i32 21, i32 21, i32 23, i32 23, i32 25, i32 25, i32 27, i32 27, i32 29, i32 29, i32 31, i32 31> + ; CHECK-DAG: ilvod.b [[R3:\$w[0-9]+]], [[R2]], [[R2]] + store <16 x i8> %3, <16 x i8>* %c + ; CHECK-DAG: st.b [[R3]], 0($4) + + ret void +} + +define void @ilvod_v8i16_2(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind { + ; CHECK-LABEL: ilvod_v8i16_2: + + %1 = load <8 x i16>, <8 x i16>* %a + %2 = load <8 x i16>, <8 x i16>* %b + ; CHECK-DAG: ld.h [[R2:\$w[0-9]+]], 0($6) + %3 = shufflevector <8 x i16> %1, <8 x i16> %2, <8 x i32> <i32 9, i32 9, i32 11, i32 11, i32 13, i32 13, i32 15, i32 15> + ; CHECK-DAG: ilvod.h [[R3:\$w[0-9]+]], [[R2]], [[R2]] + store <8 x i16> %3, <8 x i16>* %c + ; CHECK-DAG: st.h [[R3]], 0($4) + + ret void +} + +define void @ilvod_v4i32_2(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind { + ; CHECK-LABEL: ilvod_v4i32_2: + + %1 = load <4 x i32>, <4 x i32>* %a + %2 = load <4 x i32>, <4 x i32>* %b + ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6) + %3 = shufflevector <4 x i32> %1, <4 x i32> %2, <4 x i32> <i32 5, i32 5, i32 7, i32 7> + ; CHECK-DAG: ilvod.w [[R3:\$w[0-9]+]], [[R2]], [[R2]] + store <4 x i32> %3, <4 x i32>* %c + ; CHECK-DAG: st.w [[R3]], 0($4) + + ret void +} + +define void @ilvod_v2i64_2(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind { + ; CHECK-LABEL: ilvod_v2i64_2: + + %1 = load <2 x i64>, <2 x i64>* %a + %2 = load <2 x i64>, <2 x i64>* %b + ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6) + %3 = shufflevector <2 x i64> %1, <2 x i64> %2, <2 x i32> <i32 3, i32 3> + ; ilvod.d with two identical operands is equivalent to splati.d + ; CHECK-DAG: splati.d [[R3:\$w[0-9]+]], [[R2]][1] + store <2 x i64> %3, <2 x i64>* %c + ; CHECK-DAG: st.d [[R3]], 0($4) - %1 = load <16 x i8>* %a + ret void +} + +define void @ilvr_v16i8_0(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind { + ; CHECK-LABEL: ilvr_v16i8_0: + + %1 = load <16 x i8>, <16 x i8>* %a ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5) - %2 = load <16 x i8>* %b + %2 = load <16 x i8>, <16 x i8>* %b ; CHECK-DAG: ld.b [[R2:\$w[0-9]+]], 0($6) %3 = shufflevector <16 x i8> %1, <16 x i8> %2, <16 x i32> <i32 0, i32 16, i32 1, i32 17, i32 2, i32 18, i32 3, i32 19, i32 4, i32 20, i32 5, i32 21, i32 6, i32 22, i32 7, i32 23> - ; CHECK-DAG: ilvl.b [[R3:\$w[0-9]+]], [[R1]], [[R2]] + ; CHECK-DAG: ilvr.b [[R3:\$w[0-9]+]], [[R2]], [[R1]] store <16 x i8> %3, <16 x i8>* %c ; CHECK-DAG: st.b [[R3]], 0($4) ret void - ; CHECK: .size ilvl_v16i8_0 } -define void @ilvl_v8i16_0(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind { - ; CHECK: ilvl_v8i16_0: +define void @ilvr_v8i16_0(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind { + ; CHECK-LABEL: ilvr_v8i16_0: - %1 = load <8 x i16>* %a + %1 = load <8 x i16>, <8 x i16>* %a ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5) - %2 = load <8 x i16>* %b + %2 = load <8 x i16>, <8 x i16>* %b ; CHECK-DAG: ld.h [[R2:\$w[0-9]+]], 0($6) %3 = shufflevector <8 x i16> %1, <8 x i16> %2, <8 x i32> <i32 0, i32 8, i32 1, i32 9, i32 2, i32 10, i32 3, i32 11> - ; CHECK-DAG: ilvl.h [[R3:\$w[0-9]+]], [[R1]], [[R2]] + ; CHECK-DAG: ilvr.h [[R3:\$w[0-9]+]], [[R2]], [[R1]] store <8 x i16> %3, <8 x i16>* %c ; CHECK-DAG: st.h [[R3]], 0($4) ret void - ; CHECK: .size ilvl_v8i16_0 } -define void @ilvl_v4i32_0(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind { - ; CHECK: ilvl_v4i32_0: +define void @ilvr_v4i32_0(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind { + ; CHECK-LABEL: ilvr_v4i32_0: - %1 = load <4 x i32>* %a + %1 = load <4 x i32>, <4 x i32>* %a ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5) - %2 = load <4 x i32>* %b + %2 = load <4 x i32>, <4 x i32>* %b ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6) %3 = shufflevector <4 x i32> %1, <4 x i32> %2, <4 x i32> <i32 0, i32 4, i32 1, i32 5> - ; CHECK-DAG: ilvl.w [[R3:\$w[0-9]+]], [[R1]], [[R2]] + ; CHECK-DAG: ilvr.w [[R3:\$w[0-9]+]], [[R2]], [[R1]] store <4 x i32> %3, <4 x i32>* %c ; CHECK-DAG: st.w [[R3]], 0($4) ret void - ; CHECK: .size ilvl_v4i32_0 } -define void @ilvl_v2i64_0(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind { - ; CHECK: ilvl_v2i64_0: +define void @ilvr_v2i64_0(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind { + ; CHECK-LABEL: ilvr_v2i64_0: - %1 = load <2 x i64>* %a + %1 = load <2 x i64>, <2 x i64>* %a ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5) - %2 = load <2 x i64>* %b + %2 = load <2 x i64>, <2 x i64>* %b ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6) %3 = shufflevector <2 x i64> %1, <2 x i64> %2, <2 x i32> <i32 0, i32 2> - ; ilvl.d and ilvev.d are equivalent for v2i64 - ; CHECK-DAG: ilvev.d [[R3:\$w[0-9]+]], [[R1]], [[R2]] + ; ilvr.d and ilvev.d are equivalent for v2i64 + ; CHECK-DAG: ilvev.d [[R3:\$w[0-9]+]], [[R2]], [[R1]] store <2 x i64> %3, <2 x i64>* %c ; CHECK-DAG: st.d [[R3]], 0($4) ret void - ; CHECK: .size ilvl_v2i64_0 } -define void @ilvr_v16i8_0(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind { - ; CHECK: ilvr_v16i8_0: +define void @ilvr_v16i8_1(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind { + ; CHECK-LABEL: ilvr_v16i8_1: + + %1 = load <16 x i8>, <16 x i8>* %a + %2 = load <16 x i8>, <16 x i8>* %b + ; CHECK-DAG: ld.b [[R2:\$w[0-9]+]], 0($6) + %3 = shufflevector <16 x i8> %1, <16 x i8> %2, + <16 x i32> <i32 16, i32 16, i32 17, i32 17, i32 18, i32 18, i32 19, i32 19, i32 20, i32 20, i32 21, i32 21, i32 22, i32 22, i32 23, i32 23> + ; CHECK-DAG: ilvr.b [[R3:\$w[0-9]+]], [[R2]], [[R2]] + store <16 x i8> %3, <16 x i8>* %c + ; CHECK-DAG: st.b [[R3]], 0($4) + + ret void +} + +define void @ilvr_v8i16_1(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind { + ; CHECK-LABEL: ilvr_v8i16_1: + + %1 = load <8 x i16>, <8 x i16>* %a + %2 = load <8 x i16>, <8 x i16>* %b + ; CHECK-DAG: ld.h [[R2:\$w[0-9]+]], 0($6) + %3 = shufflevector <8 x i16> %1, <8 x i16> %2, <8 x i32> <i32 8, i32 8, i32 9, i32 9, i32 10, i32 10, i32 11, i32 11> + ; CHECK-DAG: ilvr.h [[R3:\$w[0-9]+]], [[R2]], [[R2]] + store <8 x i16> %3, <8 x i16>* %c + ; CHECK-DAG: st.h [[R3]], 0($4) + + ret void +} + +define void @ilvr_v4i32_1(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind { + ; CHECK-LABEL: ilvr_v4i32_1: + + %1 = load <4 x i32>, <4 x i32>* %a + %2 = load <4 x i32>, <4 x i32>* %b + ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6) + %3 = shufflevector <4 x i32> %1, <4 x i32> %2, <4 x i32> <i32 4, i32 4, i32 5, i32 5> + ; CHECK-DAG: ilvr.w [[R3:\$w[0-9]+]], [[R2]], [[R2]] + store <4 x i32> %3, <4 x i32>* %c + ; CHECK-DAG: st.w [[R3]], 0($4) + + ret void +} + +define void @ilvr_v2i64_1(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind { + ; CHECK-LABEL: ilvr_v2i64_1: + + %1 = load <2 x i64>, <2 x i64>* %a + %2 = load <2 x i64>, <2 x i64>* %b + ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6) + %3 = shufflevector <2 x i64> %1, <2 x i64> %2, <2 x i32> <i32 2, i32 2> + ; ilvr.d and splati.d are equivalent for v2i64 + ; CHECK-DAG: splati.d [[R3:\$w[0-9]+]], [[R2]][0] + store <2 x i64> %3, <2 x i64>* %c + ; CHECK-DAG: st.d [[R3]], 0($4) - %1 = load <16 x i8>* %a + ret void +} + +define void @ilvr_v16i8_2(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind { + ; CHECK-LABEL: ilvr_v16i8_2: + + %1 = load <16 x i8>, <16 x i8>* %a ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5) - %2 = load <16 x i8>* %b + %2 = load <16 x i8>, <16 x i8>* %b + %3 = shufflevector <16 x i8> %1, <16 x i8> %2, + <16 x i32> <i32 0, i32 0, i32 1, i32 1, i32 2, i32 2, i32 3, i32 3, i32 4, i32 4, i32 5, i32 5, i32 6, i32 6, i32 7, i32 7> + ; CHECK-DAG: ilvr.b [[R3:\$w[0-9]+]], [[R1]], [[R1]] + store <16 x i8> %3, <16 x i8>* %c + ; CHECK-DAG: st.b [[R3]], 0($4) + + ret void +} + +define void @ilvr_v8i16_2(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind { + ; CHECK-LABEL: ilvr_v8i16_2: + + %1 = load <8 x i16>, <8 x i16>* %a + ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5) + %2 = load <8 x i16>, <8 x i16>* %b + %3 = shufflevector <8 x i16> %1, <8 x i16> %2, <8 x i32> <i32 0, i32 0, i32 1, i32 1, i32 2, i32 2, i32 3, i32 3> + ; CHECK-DAG: ilvr.h [[R3:\$w[0-9]+]], [[R1]], [[R1]] + store <8 x i16> %3, <8 x i16>* %c + ; CHECK-DAG: st.h [[R3]], 0($4) + + ret void +} + +define void @ilvr_v4i32_2(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind { + ; CHECK-LABEL: ilvr_v4i32_2: + + %1 = load <4 x i32>, <4 x i32>* %a + ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5) + %2 = load <4 x i32>, <4 x i32>* %b + %3 = shufflevector <4 x i32> %1, <4 x i32> %2, <4 x i32> <i32 0, i32 0, i32 1, i32 1> + ; CHECK-DAG: ilvr.w [[R3:\$w[0-9]+]], [[R1]], [[R1]] + store <4 x i32> %3, <4 x i32>* %c + ; CHECK-DAG: st.w [[R3]], 0($4) + + ret void +} + +define void @ilvr_v2i64_2(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind { + ; CHECK-LABEL: ilvr_v2i64_2: + + %1 = load <2 x i64>, <2 x i64>* %a + ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5) + %2 = load <2 x i64>, <2 x i64>* %b + %3 = shufflevector <2 x i64> %1, <2 x i64> %2, <2 x i32> <i32 0, i32 0> + ; ilvr.d and splati.d are equivalent for v2i64 + ; CHECK-DAG: splati.d [[R3:\$w[0-9]+]], [[R1]][0] + store <2 x i64> %3, <2 x i64>* %c + ; CHECK-DAG: st.d [[R3]], 0($4) + + ret void +} + +define void @ilvl_v16i8_0(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind { + ; CHECK-LABEL: ilvl_v16i8_0: + + %1 = load <16 x i8>, <16 x i8>* %a + ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5) + %2 = load <16 x i8>, <16 x i8>* %b ; CHECK-DAG: ld.b [[R2:\$w[0-9]+]], 0($6) %3 = shufflevector <16 x i8> %1, <16 x i8> %2, <16 x i32> <i32 8, i32 24, i32 9, i32 25, i32 10, i32 26, i32 11, i32 27, i32 12, i32 28, i32 13, i32 29, i32 14, i32 30, i32 15, i32 31> - ; CHECK-DAG: ilvr.b [[R3:\$w[0-9]+]], [[R1]], [[R2]] + ; CHECK-DAG: ilvl.b [[R3:\$w[0-9]+]], [[R2]], [[R1]] store <16 x i8> %3, <16 x i8>* %c ; CHECK-DAG: st.b [[R3]], 0($4) ret void - ; CHECK: .size ilvr_v16i8_0 } -define void @ilvr_v8i16_0(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind { - ; CHECK: ilvr_v8i16_0: +define void @ilvl_v8i16_0(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind { + ; CHECK-LABEL: ilvl_v8i16_0: - %1 = load <8 x i16>* %a + %1 = load <8 x i16>, <8 x i16>* %a ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5) - %2 = load <8 x i16>* %b + %2 = load <8 x i16>, <8 x i16>* %b ; CHECK-DAG: ld.h [[R2:\$w[0-9]+]], 0($6) %3 = shufflevector <8 x i16> %1, <8 x i16> %2, <8 x i32> <i32 4, i32 12, i32 5, i32 13, i32 6, i32 14, i32 7, i32 15> - ; CHECK-DAG: ilvr.h [[R3:\$w[0-9]+]], [[R1]], [[R2]] + ; CHECK-DAG: ilvl.h [[R3:\$w[0-9]+]], [[R2]], [[R1]] store <8 x i16> %3, <8 x i16>* %c ; CHECK-DAG: st.h [[R3]], 0($4) ret void - ; CHECK: .size ilvr_v8i16_0 } -define void @ilvr_v4i32_0(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind { - ; CHECK: ilvr_v4i32_0: +define void @ilvl_v4i32_0(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind { + ; CHECK-LABEL: ilvl_v4i32_0: - %1 = load <4 x i32>* %a + %1 = load <4 x i32>, <4 x i32>* %a ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5) - %2 = load <4 x i32>* %b + %2 = load <4 x i32>, <4 x i32>* %b ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6) %3 = shufflevector <4 x i32> %1, <4 x i32> %2, <4 x i32> <i32 2, i32 6, i32 3, i32 7> - ; CHECK-DAG: ilvr.w [[R3:\$w[0-9]+]], [[R1]], [[R2]] + ; CHECK-DAG: ilvl.w [[R3:\$w[0-9]+]], [[R2]], [[R1]] store <4 x i32> %3, <4 x i32>* %c ; CHECK-DAG: st.w [[R3]], 0($4) ret void - ; CHECK: .size ilvr_v4i32_0 } -define void @ilvr_v2i64_0(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind { - ; CHECK: ilvr_v2i64_0: +define void @ilvl_v2i64_0(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind { + ; CHECK-LABEL: ilvl_v2i64_0: - %1 = load <2 x i64>* %a + %1 = load <2 x i64>, <2 x i64>* %a ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5) - %2 = load <2 x i64>* %b + %2 = load <2 x i64>, <2 x i64>* %b ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6) %3 = shufflevector <2 x i64> %1, <2 x i64> %2, <2 x i32> <i32 1, i32 3> - ; ilvr.d and ilvod.d are equivalent for v2i64 + ; ilvl.d and ilvod.d are equivalent for v2i64 ; CHECK-DAG: ilvod.d [[R3:\$w[0-9]+]], [[R1]], [[R2]] store <2 x i64> %3, <2 x i64>* %c ; CHECK-DAG: st.d [[R3]], 0($4) ret void - ; CHECK: .size ilvr_v2i64_0 +} + +define void @ilvl_v16i8_1(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind { + ; CHECK-LABEL: ilvl_v16i8_1: + + %1 = load <16 x i8>, <16 x i8>* %a + %2 = load <16 x i8>, <16 x i8>* %b + ; CHECK-DAG: ld.b [[R2:\$w[0-9]+]], 0($6) + %3 = shufflevector <16 x i8> %1, <16 x i8> %2, + <16 x i32> <i32 24, i32 24, i32 25, i32 25, i32 26, i32 26, i32 27, i32 27, i32 28, i32 28, i32 29, i32 29, i32 30, i32 30, i32 31, i32 31> + ; CHECK-DAG: ilvl.b [[R3:\$w[0-9]+]], [[R2]], [[R2]] + store <16 x i8> %3, <16 x i8>* %c + ; CHECK-DAG: st.b [[R3]], 0($4) + + ret void +} + +define void @ilvl_v8i16_1(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind { + ; CHECK-LABEL: ilvl_v8i16_1: + + %1 = load <8 x i16>, <8 x i16>* %a + %2 = load <8 x i16>, <8 x i16>* %b + ; CHECK-DAG: ld.h [[R2:\$w[0-9]+]], 0($6) + %3 = shufflevector <8 x i16> %1, <8 x i16> %2, <8 x i32> <i32 12, i32 12, i32 13, i32 13, i32 14, i32 14, i32 15, i32 15> + ; CHECK-DAG: ilvl.h [[R3:\$w[0-9]+]], [[R2]], [[R2]] + store <8 x i16> %3, <8 x i16>* %c + ; CHECK-DAG: st.h [[R3]], 0($4) + + ret void +} + +define void @ilvl_v4i32_1(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind { + ; CHECK-LABEL: ilvl_v4i32_1: + + %1 = load <4 x i32>, <4 x i32>* %a + %2 = load <4 x i32>, <4 x i32>* %b + ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6) + %3 = shufflevector <4 x i32> %1, <4 x i32> %2, <4 x i32> <i32 6, i32 6, i32 7, i32 7> + ; CHECK-DAG: ilvl.w [[R3:\$w[0-9]+]], [[R2]], [[R2]] + store <4 x i32> %3, <4 x i32>* %c + ; CHECK-DAG: st.w [[R3]], 0($4) + + ret void +} + +define void @ilvl_v2i64_1(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind { + ; CHECK-LABEL: ilvl_v2i64_1: + + %1 = load <2 x i64>, <2 x i64>* %a + %2 = load <2 x i64>, <2 x i64>* %b + ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6) + %3 = shufflevector <2 x i64> %1, <2 x i64> %2, <2 x i32> <i32 3, i32 3> + ; ilvl.d and splati.d are equivalent for v2i64 + ; CHECK-DAG: splati.d [[R3:\$w[0-9]+]], [[R2]][1] + store <2 x i64> %3, <2 x i64>* %c + ; CHECK-DAG: st.d [[R3]], 0($4) + + ret void +} + +define void @ilvl_v16i8_2(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind { + ; CHECK-LABEL: ilvl_v16i8_2: + + %1 = load <16 x i8>, <16 x i8>* %a + ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5) + %2 = load <16 x i8>, <16 x i8>* %b + %3 = shufflevector <16 x i8> %1, <16 x i8> %2, + <16 x i32> <i32 8, i32 8, i32 9, i32 9, i32 10, i32 10, i32 11, i32 11, i32 12, i32 12, i32 13, i32 13, i32 14, i32 14, i32 15, i32 15> + ; CHECK-DAG: ilvl.b [[R3:\$w[0-9]+]], [[R1]], [[R1]] + store <16 x i8> %3, <16 x i8>* %c + ; CHECK-DAG: st.b [[R3]], 0($4) + + ret void +} + +define void @ilvl_v8i16_2(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind { + ; CHECK-LABEL: ilvl_v8i16_2: + + %1 = load <8 x i16>, <8 x i16>* %a + ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5) + %2 = load <8 x i16>, <8 x i16>* %b + %3 = shufflevector <8 x i16> %1, <8 x i16> %2, <8 x i32> <i32 4, i32 4, i32 5, i32 5, i32 6, i32 6, i32 7, i32 7> + ; CHECK-DAG: ilvl.h [[R3:\$w[0-9]+]], [[R1]], [[R1]] + store <8 x i16> %3, <8 x i16>* %c + ; CHECK-DAG: st.h [[R3]], 0($4) + + ret void +} + +define void @ilvl_v4i32_2(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind { + ; CHECK-LABEL: ilvl_v4i32_2: + + %1 = load <4 x i32>, <4 x i32>* %a + ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5) + %2 = load <4 x i32>, <4 x i32>* %b + %3 = shufflevector <4 x i32> %1, <4 x i32> %2, <4 x i32> <i32 2, i32 2, i32 3, i32 3> + ; CHECK-DAG: ilvl.w [[R3:\$w[0-9]+]], [[R1]], [[R1]] + store <4 x i32> %3, <4 x i32>* %c + ; CHECK-DAG: st.w [[R3]], 0($4) + + ret void +} + +define void @ilvl_v2i64_2(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind { + ; CHECK-LABEL: ilvl_v2i64_2: + + %1 = load <2 x i64>, <2 x i64>* %a + ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5) + %2 = load <2 x i64>, <2 x i64>* %b + %3 = shufflevector <2 x i64> %1, <2 x i64> %2, <2 x i32> <i32 1, i32 1> + ; ilvl.d and splati.d are equivalent for v2i64 + ; CHECK-DAG: splati.d [[R3:\$w[0-9]+]], [[R1]][1] + store <2 x i64> %3, <2 x i64>* %c + ; CHECK-DAG: st.d [[R3]], 0($4) + + ret void } define void @pckev_v16i8_0(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind { - ; CHECK: pckev_v16i8_0: + ; CHECK-LABEL: pckev_v16i8_0: - %1 = load <16 x i8>* %a + %1 = load <16 x i8>, <16 x i8>* %a ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5) - %2 = load <16 x i8>* %b + %2 = load <16 x i8>, <16 x i8>* %b ; CHECK-DAG: ld.b [[R2:\$w[0-9]+]], 0($6) %3 = shufflevector <16 x i8> %1, <16 x i8> %2, <16 x i32> <i32 0, i32 2, i32 4, i32 6, i32 8, i32 10, i32 12, i32 14, i32 16, i32 18, i32 20, i32 22, i32 24, i32 26, i32 28, i32 30> - ; CHECK-DAG: pckev.b [[R3:\$w[0-9]+]], [[R1]], [[R2]] + ; CHECK-DAG: pckev.b [[R3:\$w[0-9]+]], [[R2]], [[R1]] store <16 x i8> %3, <16 x i8>* %c ; CHECK-DAG: st.b [[R3]], 0($4) ret void - ; CHECK: .size pckev_v16i8_0 } define void @pckev_v8i16_0(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind { - ; CHECK: pckev_v8i16_0: + ; CHECK-LABEL: pckev_v8i16_0: - %1 = load <8 x i16>* %a + %1 = load <8 x i16>, <8 x i16>* %a ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5) - %2 = load <8 x i16>* %b + %2 = load <8 x i16>, <8 x i16>* %b ; CHECK-DAG: ld.h [[R2:\$w[0-9]+]], 0($6) %3 = shufflevector <8 x i16> %1, <8 x i16> %2, <8 x i32> <i32 0, i32 2, i32 4, i32 6, i32 8, i32 10, i32 12, i32 14> - ; CHECK-DAG: pckev.h [[R3:\$w[0-9]+]], [[R1]], [[R2]] + ; CHECK-DAG: pckev.h [[R3:\$w[0-9]+]], [[R2]], [[R1]] store <8 x i16> %3, <8 x i16>* %c ; CHECK-DAG: st.h [[R3]], 0($4) ret void - ; CHECK: .size pckev_v8i16_0 } define void @pckev_v4i32_0(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind { - ; CHECK: pckev_v4i32_0: + ; CHECK-LABEL: pckev_v4i32_0: - %1 = load <4 x i32>* %a + %1 = load <4 x i32>, <4 x i32>* %a ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5) - %2 = load <4 x i32>* %b + %2 = load <4 x i32>, <4 x i32>* %b ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6) %3 = shufflevector <4 x i32> %1, <4 x i32> %2, <4 x i32> <i32 0, i32 2, i32 4, i32 6> - ; CHECK-DAG: pckev.w [[R3:\$w[0-9]+]], [[R1]], [[R2]] + ; CHECK-DAG: pckev.w [[R3:\$w[0-9]+]], [[R2]], [[R1]] store <4 x i32> %3, <4 x i32>* %c ; CHECK-DAG: st.w [[R3]], 0($4) ret void - ; CHECK: .size pckev_v4i32_0 } define void @pckev_v2i64_0(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind { - ; CHECK: pckev_v2i64_0: + ; CHECK-LABEL: pckev_v2i64_0: - %1 = load <2 x i64>* %a + %1 = load <2 x i64>, <2 x i64>* %a ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5) - %2 = load <2 x i64>* %b + %2 = load <2 x i64>, <2 x i64>* %b ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6) %3 = shufflevector <2 x i64> %1, <2 x i64> %2, <2 x i32> <i32 0, i32 2> ; pckev.d and ilvev.d are equivalent for v2i64 - ; CHECK-DAG: ilvev.d [[R3:\$w[0-9]+]], [[R1]], [[R2]] + ; CHECK-DAG: ilvev.d [[R3:\$w[0-9]+]], [[R2]], [[R1]] + store <2 x i64> %3, <2 x i64>* %c + ; CHECK-DAG: st.d [[R3]], 0($4) + + ret void +} + +define void @pckev_v16i8_1(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind { + ; CHECK-LABEL: pckev_v16i8_1: + + %1 = load <16 x i8>, <16 x i8>* %a + %2 = load <16 x i8>, <16 x i8>* %b + ; CHECK-DAG: ld.b [[R2:\$w[0-9]+]], 0($6) + %3 = shufflevector <16 x i8> %1, <16 x i8> %2, + <16 x i32> <i32 16, i32 18, i32 20, i32 22, i32 24, i32 26, i32 28, i32 30, i32 16, i32 18, i32 20, i32 22, i32 24, i32 26, i32 28, i32 30> + ; CHECK-DAG: pckev.b [[R3:\$w[0-9]+]], [[R2]], [[R2]] + store <16 x i8> %3, <16 x i8>* %c + ; CHECK-DAG: st.b [[R3]], 0($4) + + ret void +} + +define void @pckev_v8i16_1(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind { + ; CHECK-LABEL: pckev_v8i16_1: + + %1 = load <8 x i16>, <8 x i16>* %a + %2 = load <8 x i16>, <8 x i16>* %b + ; CHECK-DAG: ld.h [[R2:\$w[0-9]+]], 0($6) + %3 = shufflevector <8 x i16> %1, <8 x i16> %2, <8 x i32> <i32 8, i32 10, i32 12, i32 14, i32 8, i32 10, i32 12, i32 14> + ; CHECK-DAG: pckev.h [[R3:\$w[0-9]+]], [[R2]], [[R2]] + store <8 x i16> %3, <8 x i16>* %c + ; CHECK-DAG: st.h [[R3]], 0($4) + + ret void +} + +define void @pckev_v4i32_1(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind { + ; CHECK-LABEL: pckev_v4i32_1: + + %1 = load <4 x i32>, <4 x i32>* %a + %2 = load <4 x i32>, <4 x i32>* %b + ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6) + %3 = shufflevector <4 x i32> %1, <4 x i32> %2, <4 x i32> <i32 4, i32 6, i32 4, i32 6> + ; CHECK-DAG: pckev.w [[R3:\$w[0-9]+]], [[R2]], [[R2]] + store <4 x i32> %3, <4 x i32>* %c + ; CHECK-DAG: st.w [[R3]], 0($4) + + ret void +} + +define void @pckev_v2i64_1(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind { + ; CHECK-LABEL: pckev_v2i64_1: + + %1 = load <2 x i64>, <2 x i64>* %a + %2 = load <2 x i64>, <2 x i64>* %b + ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6) + %3 = shufflevector <2 x i64> %1, <2 x i64> %2, <2 x i32> <i32 2, i32 2> + ; pckev.d and splati.d are equivalent for v2i64 + ; CHECK-DAG: splati.d [[R3:\$w[0-9]+]], [[R2]][0] + store <2 x i64> %3, <2 x i64>* %c + ; CHECK-DAG: st.d [[R3]], 0($4) + + ret void +} + +define void @pckev_v16i8_2(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind { + ; CHECK-LABEL: pckev_v16i8_2: + + %1 = load <16 x i8>, <16 x i8>* %a + ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5) + %2 = load <16 x i8>, <16 x i8>* %b + %3 = shufflevector <16 x i8> %1, <16 x i8> %2, + <16 x i32> <i32 0, i32 2, i32 4, i32 6, i32 8, i32 10, i32 12, i32 14, i32 0, i32 2, i32 4, i32 6, i32 8, i32 10, i32 12, i32 14> + ; CHECK-DAG: pckev.b [[R3:\$w[0-9]+]], [[R1]], [[R1]] + store <16 x i8> %3, <16 x i8>* %c + ; CHECK-DAG: st.b [[R3]], 0($4) + + ret void +} + +define void @pckev_v8i16_2(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind { + ; CHECK-LABEL: pckev_v8i16_2: + + %1 = load <8 x i16>, <8 x i16>* %a + ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5) + %2 = load <8 x i16>, <8 x i16>* %b + %3 = shufflevector <8 x i16> %1, <8 x i16> %2, <8 x i32> <i32 0, i32 2, i32 4, i32 6, i32 0, i32 2, i32 4, i32 6> + ; CHECK-DAG: pckev.h [[R3:\$w[0-9]+]], [[R1]], [[R1]] + store <8 x i16> %3, <8 x i16>* %c + ; CHECK-DAG: st.h [[R3]], 0($4) + + ret void +} + +define void @pckev_v4i32_2(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind { + ; CHECK-LABEL: pckev_v4i32_2: + + %1 = load <4 x i32>, <4 x i32>* %a + ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5) + %2 = load <4 x i32>, <4 x i32>* %b + %3 = shufflevector <4 x i32> %1, <4 x i32> %2, <4 x i32> <i32 0, i32 2, i32 0, i32 2> + ; CHECK-DAG: pckev.w [[R3:\$w[0-9]+]], [[R1]], [[R1]] + store <4 x i32> %3, <4 x i32>* %c + ; CHECK-DAG: st.w [[R3]], 0($4) + + ret void +} + +define void @pckev_v2i64_2(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind { + ; CHECK-LABEL: pckev_v2i64_2: + + %1 = load <2 x i64>, <2 x i64>* %a + ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5) + %2 = load <2 x i64>, <2 x i64>* %b + %3 = shufflevector <2 x i64> %1, <2 x i64> %2, <2 x i32> <i32 0, i32 0> + ; pckev.d and splati.d are equivalent for v2i64 + ; CHECK-DAG: splati.d [[R3:\$w[0-9]+]], [[R1]][0] store <2 x i64> %3, <2 x i64>* %c ; CHECK-DAG: st.d [[R3]], 0($4) ret void - ; CHECK: .size pckev_v2i64_0 } define void @pckod_v16i8_0(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind { - ; CHECK: pckod_v16i8_0: + ; CHECK-LABEL: pckod_v16i8_0: - %1 = load <16 x i8>* %a + %1 = load <16 x i8>, <16 x i8>* %a ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5) - %2 = load <16 x i8>* %b + %2 = load <16 x i8>, <16 x i8>* %b ; CHECK-DAG: ld.b [[R2:\$w[0-9]+]], 0($6) %3 = shufflevector <16 x i8> %1, <16 x i8> %2, <16 x i32> <i32 1, i32 3, i32 5, i32 7, i32 9, i32 11, i32 13, i32 15, i32 17, i32 19, i32 21, i32 23, i32 25, i32 27, i32 29, i32 31> - ; CHECK-DAG: pckod.b [[R3:\$w[0-9]+]], [[R1]], [[R2]] + ; CHECK-DAG: pckod.b [[R3:\$w[0-9]+]], [[R2]], [[R1]] store <16 x i8> %3, <16 x i8>* %c ; CHECK-DAG: st.b [[R3]], 0($4) ret void - ; CHECK: .size pckod_v16i8_0 } define void @pckod_v8i16_0(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind { - ; CHECK: pckod_v8i16_0: + ; CHECK-LABEL: pckod_v8i16_0: - %1 = load <8 x i16>* %a + %1 = load <8 x i16>, <8 x i16>* %a ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5) - %2 = load <8 x i16>* %b + %2 = load <8 x i16>, <8 x i16>* %b ; CHECK-DAG: ld.h [[R2:\$w[0-9]+]], 0($6) %3 = shufflevector <8 x i16> %1, <8 x i16> %2, <8 x i32> <i32 1, i32 3, i32 5, i32 7, i32 9, i32 11, i32 13, i32 15> - ; CHECK-DAG: pckod.h [[R3:\$w[0-9]+]], [[R1]], [[R2]] + ; CHECK-DAG: pckod.h [[R3:\$w[0-9]+]], [[R2]], [[R1]] store <8 x i16> %3, <8 x i16>* %c ; CHECK-DAG: st.h [[R3]], 0($4) ret void - ; CHECK: .size pckod_v8i16_0 } define void @pckod_v4i32_0(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind { - ; CHECK: pckod_v4i32_0: + ; CHECK-LABEL: pckod_v4i32_0: - %1 = load <4 x i32>* %a + %1 = load <4 x i32>, <4 x i32>* %a ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5) - %2 = load <4 x i32>* %b + %2 = load <4 x i32>, <4 x i32>* %b ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6) %3 = shufflevector <4 x i32> %1, <4 x i32> %2, <4 x i32> <i32 1, i32 3, i32 5, i32 7> - ; CHECK-DAG: pckod.w [[R3:\$w[0-9]+]], [[R1]], [[R2]] + ; CHECK-DAG: pckod.w [[R3:\$w[0-9]+]], [[R2]], [[R1]] store <4 x i32> %3, <4 x i32>* %c ; CHECK-DAG: st.w [[R3]], 0($4) ret void - ; CHECK: .size pckod_v4i32_0 } define void @pckod_v2i64_0(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind { - ; CHECK: pckod_v2i64_0: + ; CHECK-LABEL: pckod_v2i64_0: - %1 = load <2 x i64>* %a + %1 = load <2 x i64>, <2 x i64>* %a ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5) - %2 = load <2 x i64>* %b + %2 = load <2 x i64>, <2 x i64>* %b ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6) %3 = shufflevector <2 x i64> %1, <2 x i64> %2, <2 x i32> <i32 1, i32 3> ; pckod.d and ilvod.d are equivalent for v2i64 @@ -759,13 +1295,128 @@ define void @pckod_v2i64_0(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind ; CHECK-DAG: st.d [[R3]], 0($4) ret void - ; CHECK: .size pckod_v2i64_0 +} + +define void @pckod_v16i8_1(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind { + ; CHECK-LABEL: pckod_v16i8_1: + + %1 = load <16 x i8>, <16 x i8>* %a + %2 = load <16 x i8>, <16 x i8>* %b + ; CHECK-DAG: ld.b [[R2:\$w[0-9]+]], 0($6) + %3 = shufflevector <16 x i8> %1, <16 x i8> %2, + <16 x i32> <i32 17, i32 19, i32 21, i32 23, i32 25, i32 27, i32 29, i32 31, i32 17, i32 19, i32 21, i32 23, i32 25, i32 27, i32 29, i32 31> + ; CHECK-DAG: pckod.b [[R3:\$w[0-9]+]], [[R2]], [[R2]] + store <16 x i8> %3, <16 x i8>* %c + ; CHECK-DAG: st.b [[R3]], 0($4) + + ret void +} + +define void @pckod_v8i16_1(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind { + ; CHECK-LABEL: pckod_v8i16_1: + + %1 = load <8 x i16>, <8 x i16>* %a + %2 = load <8 x i16>, <8 x i16>* %b + ; CHECK-DAG: ld.h [[R2:\$w[0-9]+]], 0($6) + %3 = shufflevector <8 x i16> %1, <8 x i16> %2, <8 x i32> <i32 9, i32 11, i32 13, i32 15, i32 9, i32 11, i32 13, i32 15> + ; CHECK-DAG: pckod.h [[R3:\$w[0-9]+]], [[R2]], [[R2]] + store <8 x i16> %3, <8 x i16>* %c + ; CHECK-DAG: st.h [[R3]], 0($4) + + ret void +} + +define void @pckod_v4i32_1(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind { + ; CHECK-LABEL: pckod_v4i32_1: + + %1 = load <4 x i32>, <4 x i32>* %a + %2 = load <4 x i32>, <4 x i32>* %b + ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6) + %3 = shufflevector <4 x i32> %1, <4 x i32> %2, <4 x i32> <i32 5, i32 7, i32 5, i32 7> + ; CHECK-DAG: pckod.w [[R3:\$w[0-9]+]], [[R2]], [[R2]] + store <4 x i32> %3, <4 x i32>* %c + ; CHECK-DAG: st.w [[R3]], 0($4) + + ret void +} + +define void @pckod_v2i64_1(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind { + ; CHECK-LABEL: pckod_v2i64_1: + + %1 = load <2 x i64>, <2 x i64>* %a + %2 = load <2 x i64>, <2 x i64>* %b + ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6) + %3 = shufflevector <2 x i64> %1, <2 x i64> %2, <2 x i32> <i32 3, i32 3> + ; pckod.d and splati.d are equivalent for v2i64 + ; CHECK-DAG: splati.d [[R3:\$w[0-9]+]], [[R2]][1] + store <2 x i64> %3, <2 x i64>* %c + ; CHECK-DAG: st.d [[R3]], 0($4) + + ret void +} + +define void @pckod_v16i8_2(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind { + ; CHECK-LABEL: pckod_v16i8_2: + + %1 = load <16 x i8>, <16 x i8>* %a + ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5) + %2 = load <16 x i8>, <16 x i8>* %b + %3 = shufflevector <16 x i8> %1, <16 x i8> %2, + <16 x i32> <i32 1, i32 3, i32 5, i32 7, i32 9, i32 11, i32 13, i32 15, i32 1, i32 3, i32 5, i32 7, i32 9, i32 11, i32 13, i32 15> + ; CHECK-DAG: pckod.b [[R3:\$w[0-9]+]], [[R1]], [[R1]] + store <16 x i8> %3, <16 x i8>* %c + ; CHECK-DAG: st.b [[R3]], 0($4) + + ret void +} + +define void @pckod_v8i16_2(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind { + ; CHECK-LABEL: pckod_v8i16_2: + + %1 = load <8 x i16>, <8 x i16>* %a + ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5) + %2 = load <8 x i16>, <8 x i16>* %b + %3 = shufflevector <8 x i16> %1, <8 x i16> %2, <8 x i32> <i32 1, i32 3, i32 5, i32 7, i32 1, i32 3, i32 5, i32 7> + ; CHECK-DAG: pckod.h [[R3:\$w[0-9]+]], [[R1]], [[R1]] + store <8 x i16> %3, <8 x i16>* %c + ; CHECK-DAG: st.h [[R3]], 0($4) + + ret void +} + +define void @pckod_v4i32_2(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind { + ; CHECK-LABEL: pckod_v4i32_2: + + %1 = load <4 x i32>, <4 x i32>* %a + ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5) + %2 = load <4 x i32>, <4 x i32>* %b + %3 = shufflevector <4 x i32> %1, <4 x i32> %2, <4 x i32> <i32 1, i32 3, i32 1, i32 3> + ; CHECK-DAG: pckod.w [[R3:\$w[0-9]+]], [[R1]], [[R1]] + store <4 x i32> %3, <4 x i32>* %c + ; CHECK-DAG: st.w [[R3]], 0($4) + + ret void +} + +define void @pckod_v2i64_2(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind { + ; CHECK-LABEL: pckod_v2i64_2: + + %1 = load <2 x i64>, <2 x i64>* %a + ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5) + %2 = load <2 x i64>, <2 x i64>* %b + %3 = shufflevector <2 x i64> %1, <2 x i64> %2, <2 x i32> <i32 1, i32 1> + ; pckod.d and splati.d are equivalent for v2i64 + ; CHECK-DAG: splati.d [[R3:\$w[0-9]+]], [[R1]][1] + store <2 x i64> %3, <2 x i64>* %c + ; CHECK-DAG: st.d [[R3]], 0($4) + + ret void } define void @splati_v16i8_0(<16 x i8>* %c, <16 x i8>* %a) nounwind { - ; CHECK: splati_v16i8_0: + ; CHECK-LABEL: splati_v16i8_0: - %1 = load <16 x i8>* %a + %1 = load <16 x i8>, <16 x i8>* %a ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5) %2 = shufflevector <16 x i8> %1, <16 x i8> undef, <16 x i32> <i32 4, i32 4, i32 4, i32 4, i32 4, i32 4, i32 4, i32 4, i32 4, i32 4, i32 4, i32 4, i32 4, i32 4, i32 4, i32 4> @@ -774,13 +1425,12 @@ define void @splati_v16i8_0(<16 x i8>* %c, <16 x i8>* %a) nounwind { ; CHECK-DAG: st.b [[R3]], 0($4) ret void - ; CHECK: .size splati_v16i8_0 } define void @splati_v8i16_0(<8 x i16>* %c, <8 x i16>* %a) nounwind { - ; CHECK: splati_v8i16_0: + ; CHECK-LABEL: splati_v8i16_0: - %1 = load <8 x i16>* %a + %1 = load <8 x i16>, <8 x i16>* %a ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5) %2 = shufflevector <8 x i16> %1, <8 x i16> undef, <8 x i32> <i32 4, i32 4, i32 4, i32 4, i32 4, i32 4, i32 4, i32 4> ; CHECK-DAG: splati.h [[R3:\$w[0-9]+]], [[R1]][4] @@ -788,28 +1438,25 @@ define void @splati_v8i16_0(<8 x i16>* %c, <8 x i16>* %a) nounwind { ; CHECK-DAG: st.h [[R3]], 0($4) ret void - ; CHECK: .size splati_v8i16_0 } define void @splati_v4i32_0(<4 x i32>* %c, <4 x i32>* %a) nounwind { - ; CHECK: splati_v4i32_0: + ; CHECK-LABEL: splati_v4i32_0: - %1 = load <4 x i32>* %a + %1 = load <4 x i32>, <4 x i32>* %a ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5) %2 = shufflevector <4 x i32> %1, <4 x i32> undef, <4 x i32> <i32 3, i32 3, i32 3, i32 3> - ; shf.w and splati.w are equivalent - ; CHECK-DAG: shf.w [[R3:\$w[0-9]+]], [[R1]], 255 + ; CHECK-DAG: splati.w [[R3:\$w[0-9]+]], [[R1]][3] store <4 x i32> %2, <4 x i32>* %c ; CHECK-DAG: st.w [[R3]], 0($4) ret void - ; CHECK: .size splati_v4i32_0 } define void @splati_v2i64_0(<2 x i64>* %c, <2 x i64>* %a) nounwind { - ; CHECK: splati_v2i64_0: + ; CHECK-LABEL: splati_v2i64_0: - %1 = load <2 x i64>* %a + %1 = load <2 x i64>, <2 x i64>* %a ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5) %2 = shufflevector <2 x i64> %1, <2 x i64> undef, <2 x i32> <i32 1, i32 1> ; CHECK-DAG: splati.d [[R3:\$w[0-9]+]], [[R1]][1] @@ -817,5 +1464,4 @@ define void @splati_v2i64_0(<2 x i64>* %c, <2 x i64>* %a) nounwind { ; CHECK-DAG: st.d [[R3]], 0($4) ret void - ; CHECK: .size splati_v2i64_0 } diff --git a/test/CodeGen/Mips/msa/spill.ll b/test/CodeGen/Mips/msa/spill.ll index 66f896ac4684c..8c9a799123513 100644 --- a/test/CodeGen/Mips/msa/spill.ll +++ b/test/CodeGen/Mips/msa/spill.ll @@ -6,73 +6,73 @@ define i32 @test_i8(<16 x i8>* %p0, <16 x i8>* %q1) nounwind { entry: - %p1 = getelementptr <16 x i8>* %p0, i32 1 - %p2 = getelementptr <16 x i8>* %p0, i32 2 - %p3 = getelementptr <16 x i8>* %p0, i32 3 - %p4 = getelementptr <16 x i8>* %p0, i32 4 - %p5 = getelementptr <16 x i8>* %p0, i32 5 - %p6 = getelementptr <16 x i8>* %p0, i32 6 - %p7 = getelementptr <16 x i8>* %p0, i32 7 - %p8 = getelementptr <16 x i8>* %p0, i32 8 - %p9 = getelementptr <16 x i8>* %p0, i32 9 - %p10 = getelementptr <16 x i8>* %p0, i32 10 - %p11 = getelementptr <16 x i8>* %p0, i32 11 - %p12 = getelementptr <16 x i8>* %p0, i32 12 - %p13 = getelementptr <16 x i8>* %p0, i32 13 - %p14 = getelementptr <16 x i8>* %p0, i32 14 - %p15 = getelementptr <16 x i8>* %p0, i32 15 - %p16 = getelementptr <16 x i8>* %p0, i32 16 - %p17 = getelementptr <16 x i8>* %p0, i32 17 - %p18 = getelementptr <16 x i8>* %p0, i32 18 - %p19 = getelementptr <16 x i8>* %p0, i32 19 - %p20 = getelementptr <16 x i8>* %p0, i32 20 - %p21 = getelementptr <16 x i8>* %p0, i32 21 - %p22 = getelementptr <16 x i8>* %p0, i32 22 - %p23 = getelementptr <16 x i8>* %p0, i32 23 - %p24 = getelementptr <16 x i8>* %p0, i32 24 - %p25 = getelementptr <16 x i8>* %p0, i32 25 - %p26 = getelementptr <16 x i8>* %p0, i32 26 - %p27 = getelementptr <16 x i8>* %p0, i32 27 - %p28 = getelementptr <16 x i8>* %p0, i32 28 - %p29 = getelementptr <16 x i8>* %p0, i32 29 - %p30 = getelementptr <16 x i8>* %p0, i32 30 - %p31 = getelementptr <16 x i8>* %p0, i32 31 - %p32 = getelementptr <16 x i8>* %p0, i32 32 - %p33 = getelementptr <16 x i8>* %p0, i32 33 - %0 = load <16 x i8>* %p0, align 16 - %1 = load <16 x i8>* %p1, align 16 - %2 = load <16 x i8>* %p2, align 16 - %3 = load <16 x i8>* %p3, align 16 - %4 = load <16 x i8>* %p4, align 16 - %5 = load <16 x i8>* %p5, align 16 - %6 = load <16 x i8>* %p6, align 16 - %7 = load <16 x i8>* %p7, align 16 - %8 = load <16 x i8>* %p8, align 16 - %9 = load <16 x i8>* %p9, align 16 - %10 = load <16 x i8>* %p10, align 16 - %11 = load <16 x i8>* %p11, align 16 - %12 = load <16 x i8>* %p12, align 16 - %13 = load <16 x i8>* %p13, align 16 - %14 = load <16 x i8>* %p14, align 16 - %15 = load <16 x i8>* %p15, align 16 - %16 = load <16 x i8>* %p16, align 16 - %17 = load <16 x i8>* %p17, align 16 - %18 = load <16 x i8>* %p18, align 16 - %19 = load <16 x i8>* %p19, align 16 - %20 = load <16 x i8>* %p20, align 16 - %21 = load <16 x i8>* %p21, align 16 - %22 = load <16 x i8>* %p22, align 16 - %23 = load <16 x i8>* %p23, align 16 - %24 = load <16 x i8>* %p24, align 16 - %25 = load <16 x i8>* %p25, align 16 - %26 = load <16 x i8>* %p26, align 16 - %27 = load <16 x i8>* %p27, align 16 - %28 = load <16 x i8>* %p28, align 16 - %29 = load <16 x i8>* %p29, align 16 - %30 = load <16 x i8>* %p30, align 16 - %31 = load <16 x i8>* %p31, align 16 - %32 = load <16 x i8>* %p32, align 16 - %33 = load <16 x i8>* %p33, align 16 + %p1 = getelementptr <16 x i8>, <16 x i8>* %p0, i32 1 + %p2 = getelementptr <16 x i8>, <16 x i8>* %p0, i32 2 + %p3 = getelementptr <16 x i8>, <16 x i8>* %p0, i32 3 + %p4 = getelementptr <16 x i8>, <16 x i8>* %p0, i32 4 + %p5 = getelementptr <16 x i8>, <16 x i8>* %p0, i32 5 + %p6 = getelementptr <16 x i8>, <16 x i8>* %p0, i32 6 + %p7 = getelementptr <16 x i8>, <16 x i8>* %p0, i32 7 + %p8 = getelementptr <16 x i8>, <16 x i8>* %p0, i32 8 + %p9 = getelementptr <16 x i8>, <16 x i8>* %p0, i32 9 + %p10 = getelementptr <16 x i8>, <16 x i8>* %p0, i32 10 + %p11 = getelementptr <16 x i8>, <16 x i8>* %p0, i32 11 + %p12 = getelementptr <16 x i8>, <16 x i8>* %p0, i32 12 + %p13 = getelementptr <16 x i8>, <16 x i8>* %p0, i32 13 + %p14 = getelementptr <16 x i8>, <16 x i8>* %p0, i32 14 + %p15 = getelementptr <16 x i8>, <16 x i8>* %p0, i32 15 + %p16 = getelementptr <16 x i8>, <16 x i8>* %p0, i32 16 + %p17 = getelementptr <16 x i8>, <16 x i8>* %p0, i32 17 + %p18 = getelementptr <16 x i8>, <16 x i8>* %p0, i32 18 + %p19 = getelementptr <16 x i8>, <16 x i8>* %p0, i32 19 + %p20 = getelementptr <16 x i8>, <16 x i8>* %p0, i32 20 + %p21 = getelementptr <16 x i8>, <16 x i8>* %p0, i32 21 + %p22 = getelementptr <16 x i8>, <16 x i8>* %p0, i32 22 + %p23 = getelementptr <16 x i8>, <16 x i8>* %p0, i32 23 + %p24 = getelementptr <16 x i8>, <16 x i8>* %p0, i32 24 + %p25 = getelementptr <16 x i8>, <16 x i8>* %p0, i32 25 + %p26 = getelementptr <16 x i8>, <16 x i8>* %p0, i32 26 + %p27 = getelementptr <16 x i8>, <16 x i8>* %p0, i32 27 + %p28 = getelementptr <16 x i8>, <16 x i8>* %p0, i32 28 + %p29 = getelementptr <16 x i8>, <16 x i8>* %p0, i32 29 + %p30 = getelementptr <16 x i8>, <16 x i8>* %p0, i32 30 + %p31 = getelementptr <16 x i8>, <16 x i8>* %p0, i32 31 + %p32 = getelementptr <16 x i8>, <16 x i8>* %p0, i32 32 + %p33 = getelementptr <16 x i8>, <16 x i8>* %p0, i32 33 + %0 = load <16 x i8>, <16 x i8>* %p0, align 16 + %1 = load <16 x i8>, <16 x i8>* %p1, align 16 + %2 = load <16 x i8>, <16 x i8>* %p2, align 16 + %3 = load <16 x i8>, <16 x i8>* %p3, align 16 + %4 = load <16 x i8>, <16 x i8>* %p4, align 16 + %5 = load <16 x i8>, <16 x i8>* %p5, align 16 + %6 = load <16 x i8>, <16 x i8>* %p6, align 16 + %7 = load <16 x i8>, <16 x i8>* %p7, align 16 + %8 = load <16 x i8>, <16 x i8>* %p8, align 16 + %9 = load <16 x i8>, <16 x i8>* %p9, align 16 + %10 = load <16 x i8>, <16 x i8>* %p10, align 16 + %11 = load <16 x i8>, <16 x i8>* %p11, align 16 + %12 = load <16 x i8>, <16 x i8>* %p12, align 16 + %13 = load <16 x i8>, <16 x i8>* %p13, align 16 + %14 = load <16 x i8>, <16 x i8>* %p14, align 16 + %15 = load <16 x i8>, <16 x i8>* %p15, align 16 + %16 = load <16 x i8>, <16 x i8>* %p16, align 16 + %17 = load <16 x i8>, <16 x i8>* %p17, align 16 + %18 = load <16 x i8>, <16 x i8>* %p18, align 16 + %19 = load <16 x i8>, <16 x i8>* %p19, align 16 + %20 = load <16 x i8>, <16 x i8>* %p20, align 16 + %21 = load <16 x i8>, <16 x i8>* %p21, align 16 + %22 = load <16 x i8>, <16 x i8>* %p22, align 16 + %23 = load <16 x i8>, <16 x i8>* %p23, align 16 + %24 = load <16 x i8>, <16 x i8>* %p24, align 16 + %25 = load <16 x i8>, <16 x i8>* %p25, align 16 + %26 = load <16 x i8>, <16 x i8>* %p26, align 16 + %27 = load <16 x i8>, <16 x i8>* %p27, align 16 + %28 = load <16 x i8>, <16 x i8>* %p28, align 16 + %29 = load <16 x i8>, <16 x i8>* %p29, align 16 + %30 = load <16 x i8>, <16 x i8>* %p30, align 16 + %31 = load <16 x i8>, <16 x i8>* %p31, align 16 + %32 = load <16 x i8>, <16 x i8>* %p32, align 16 + %33 = load <16 x i8>, <16 x i8>* %p33, align 16 %r1 = call <16 x i8> @llvm.mips.addv.b(<16 x i8> %0, <16 x i8> %1) %r2 = call <16 x i8> @llvm.mips.addv.b(<16 x i8> %r1, <16 x i8> %2) %r3 = call <16 x i8> @llvm.mips.addv.b(<16 x i8> %r2, <16 x i8> %3) @@ -155,73 +155,73 @@ declare i32 @llvm.mips.copy.s.b(<16 x i8>, i32) nounwind define i32 @test_i16(<8 x i16>* %p0, <8 x i16>* %q1) nounwind { entry: - %p1 = getelementptr <8 x i16>* %p0, i32 1 - %p2 = getelementptr <8 x i16>* %p0, i32 2 - %p3 = getelementptr <8 x i16>* %p0, i32 3 - %p4 = getelementptr <8 x i16>* %p0, i32 4 - %p5 = getelementptr <8 x i16>* %p0, i32 5 - %p6 = getelementptr <8 x i16>* %p0, i32 6 - %p7 = getelementptr <8 x i16>* %p0, i32 7 - %p8 = getelementptr <8 x i16>* %p0, i32 8 - %p9 = getelementptr <8 x i16>* %p0, i32 9 - %p10 = getelementptr <8 x i16>* %p0, i32 10 - %p11 = getelementptr <8 x i16>* %p0, i32 11 - %p12 = getelementptr <8 x i16>* %p0, i32 12 - %p13 = getelementptr <8 x i16>* %p0, i32 13 - %p14 = getelementptr <8 x i16>* %p0, i32 14 - %p15 = getelementptr <8 x i16>* %p0, i32 15 - %p16 = getelementptr <8 x i16>* %p0, i32 16 - %p17 = getelementptr <8 x i16>* %p0, i32 17 - %p18 = getelementptr <8 x i16>* %p0, i32 18 - %p19 = getelementptr <8 x i16>* %p0, i32 19 - %p20 = getelementptr <8 x i16>* %p0, i32 20 - %p21 = getelementptr <8 x i16>* %p0, i32 21 - %p22 = getelementptr <8 x i16>* %p0, i32 22 - %p23 = getelementptr <8 x i16>* %p0, i32 23 - %p24 = getelementptr <8 x i16>* %p0, i32 24 - %p25 = getelementptr <8 x i16>* %p0, i32 25 - %p26 = getelementptr <8 x i16>* %p0, i32 26 - %p27 = getelementptr <8 x i16>* %p0, i32 27 - %p28 = getelementptr <8 x i16>* %p0, i32 28 - %p29 = getelementptr <8 x i16>* %p0, i32 29 - %p30 = getelementptr <8 x i16>* %p0, i32 30 - %p31 = getelementptr <8 x i16>* %p0, i32 31 - %p32 = getelementptr <8 x i16>* %p0, i32 32 - %p33 = getelementptr <8 x i16>* %p0, i32 33 - %0 = load <8 x i16>* %p0, align 16 - %1 = load <8 x i16>* %p1, align 16 - %2 = load <8 x i16>* %p2, align 16 - %3 = load <8 x i16>* %p3, align 16 - %4 = load <8 x i16>* %p4, align 16 - %5 = load <8 x i16>* %p5, align 16 - %6 = load <8 x i16>* %p6, align 16 - %7 = load <8 x i16>* %p7, align 16 - %8 = load <8 x i16>* %p8, align 16 - %9 = load <8 x i16>* %p9, align 16 - %10 = load <8 x i16>* %p10, align 16 - %11 = load <8 x i16>* %p11, align 16 - %12 = load <8 x i16>* %p12, align 16 - %13 = load <8 x i16>* %p13, align 16 - %14 = load <8 x i16>* %p14, align 16 - %15 = load <8 x i16>* %p15, align 16 - %16 = load <8 x i16>* %p16, align 16 - %17 = load <8 x i16>* %p17, align 16 - %18 = load <8 x i16>* %p18, align 16 - %19 = load <8 x i16>* %p19, align 16 - %20 = load <8 x i16>* %p20, align 16 - %21 = load <8 x i16>* %p21, align 16 - %22 = load <8 x i16>* %p22, align 16 - %23 = load <8 x i16>* %p23, align 16 - %24 = load <8 x i16>* %p24, align 16 - %25 = load <8 x i16>* %p25, align 16 - %26 = load <8 x i16>* %p26, align 16 - %27 = load <8 x i16>* %p27, align 16 - %28 = load <8 x i16>* %p28, align 16 - %29 = load <8 x i16>* %p29, align 16 - %30 = load <8 x i16>* %p30, align 16 - %31 = load <8 x i16>* %p31, align 16 - %32 = load <8 x i16>* %p32, align 16 - %33 = load <8 x i16>* %p33, align 16 + %p1 = getelementptr <8 x i16>, <8 x i16>* %p0, i32 1 + %p2 = getelementptr <8 x i16>, <8 x i16>* %p0, i32 2 + %p3 = getelementptr <8 x i16>, <8 x i16>* %p0, i32 3 + %p4 = getelementptr <8 x i16>, <8 x i16>* %p0, i32 4 + %p5 = getelementptr <8 x i16>, <8 x i16>* %p0, i32 5 + %p6 = getelementptr <8 x i16>, <8 x i16>* %p0, i32 6 + %p7 = getelementptr <8 x i16>, <8 x i16>* %p0, i32 7 + %p8 = getelementptr <8 x i16>, <8 x i16>* %p0, i32 8 + %p9 = getelementptr <8 x i16>, <8 x i16>* %p0, i32 9 + %p10 = getelementptr <8 x i16>, <8 x i16>* %p0, i32 10 + %p11 = getelementptr <8 x i16>, <8 x i16>* %p0, i32 11 + %p12 = getelementptr <8 x i16>, <8 x i16>* %p0, i32 12 + %p13 = getelementptr <8 x i16>, <8 x i16>* %p0, i32 13 + %p14 = getelementptr <8 x i16>, <8 x i16>* %p0, i32 14 + %p15 = getelementptr <8 x i16>, <8 x i16>* %p0, i32 15 + %p16 = getelementptr <8 x i16>, <8 x i16>* %p0, i32 16 + %p17 = getelementptr <8 x i16>, <8 x i16>* %p0, i32 17 + %p18 = getelementptr <8 x i16>, <8 x i16>* %p0, i32 18 + %p19 = getelementptr <8 x i16>, <8 x i16>* %p0, i32 19 + %p20 = getelementptr <8 x i16>, <8 x i16>* %p0, i32 20 + %p21 = getelementptr <8 x i16>, <8 x i16>* %p0, i32 21 + %p22 = getelementptr <8 x i16>, <8 x i16>* %p0, i32 22 + %p23 = getelementptr <8 x i16>, <8 x i16>* %p0, i32 23 + %p24 = getelementptr <8 x i16>, <8 x i16>* %p0, i32 24 + %p25 = getelementptr <8 x i16>, <8 x i16>* %p0, i32 25 + %p26 = getelementptr <8 x i16>, <8 x i16>* %p0, i32 26 + %p27 = getelementptr <8 x i16>, <8 x i16>* %p0, i32 27 + %p28 = getelementptr <8 x i16>, <8 x i16>* %p0, i32 28 + %p29 = getelementptr <8 x i16>, <8 x i16>* %p0, i32 29 + %p30 = getelementptr <8 x i16>, <8 x i16>* %p0, i32 30 + %p31 = getelementptr <8 x i16>, <8 x i16>* %p0, i32 31 + %p32 = getelementptr <8 x i16>, <8 x i16>* %p0, i32 32 + %p33 = getelementptr <8 x i16>, <8 x i16>* %p0, i32 33 + %0 = load <8 x i16>, <8 x i16>* %p0, align 16 + %1 = load <8 x i16>, <8 x i16>* %p1, align 16 + %2 = load <8 x i16>, <8 x i16>* %p2, align 16 + %3 = load <8 x i16>, <8 x i16>* %p3, align 16 + %4 = load <8 x i16>, <8 x i16>* %p4, align 16 + %5 = load <8 x i16>, <8 x i16>* %p5, align 16 + %6 = load <8 x i16>, <8 x i16>* %p6, align 16 + %7 = load <8 x i16>, <8 x i16>* %p7, align 16 + %8 = load <8 x i16>, <8 x i16>* %p8, align 16 + %9 = load <8 x i16>, <8 x i16>* %p9, align 16 + %10 = load <8 x i16>, <8 x i16>* %p10, align 16 + %11 = load <8 x i16>, <8 x i16>* %p11, align 16 + %12 = load <8 x i16>, <8 x i16>* %p12, align 16 + %13 = load <8 x i16>, <8 x i16>* %p13, align 16 + %14 = load <8 x i16>, <8 x i16>* %p14, align 16 + %15 = load <8 x i16>, <8 x i16>* %p15, align 16 + %16 = load <8 x i16>, <8 x i16>* %p16, align 16 + %17 = load <8 x i16>, <8 x i16>* %p17, align 16 + %18 = load <8 x i16>, <8 x i16>* %p18, align 16 + %19 = load <8 x i16>, <8 x i16>* %p19, align 16 + %20 = load <8 x i16>, <8 x i16>* %p20, align 16 + %21 = load <8 x i16>, <8 x i16>* %p21, align 16 + %22 = load <8 x i16>, <8 x i16>* %p22, align 16 + %23 = load <8 x i16>, <8 x i16>* %p23, align 16 + %24 = load <8 x i16>, <8 x i16>* %p24, align 16 + %25 = load <8 x i16>, <8 x i16>* %p25, align 16 + %26 = load <8 x i16>, <8 x i16>* %p26, align 16 + %27 = load <8 x i16>, <8 x i16>* %p27, align 16 + %28 = load <8 x i16>, <8 x i16>* %p28, align 16 + %29 = load <8 x i16>, <8 x i16>* %p29, align 16 + %30 = load <8 x i16>, <8 x i16>* %p30, align 16 + %31 = load <8 x i16>, <8 x i16>* %p31, align 16 + %32 = load <8 x i16>, <8 x i16>* %p32, align 16 + %33 = load <8 x i16>, <8 x i16>* %p33, align 16 %r1 = call <8 x i16> @llvm.mips.addv.h(<8 x i16> %0, <8 x i16> %1) %r2 = call <8 x i16> @llvm.mips.addv.h(<8 x i16> %r1, <8 x i16> %2) %r3 = call <8 x i16> @llvm.mips.addv.h(<8 x i16> %r2, <8 x i16> %3) @@ -304,73 +304,73 @@ declare i32 @llvm.mips.copy.s.h(<8 x i16>, i32) nounwind define i32 @test_i32(<4 x i32>* %p0, <4 x i32>* %q1) nounwind { entry: - %p1 = getelementptr <4 x i32>* %p0, i32 1 - %p2 = getelementptr <4 x i32>* %p0, i32 2 - %p3 = getelementptr <4 x i32>* %p0, i32 3 - %p4 = getelementptr <4 x i32>* %p0, i32 4 - %p5 = getelementptr <4 x i32>* %p0, i32 5 - %p6 = getelementptr <4 x i32>* %p0, i32 6 - %p7 = getelementptr <4 x i32>* %p0, i32 7 - %p8 = getelementptr <4 x i32>* %p0, i32 8 - %p9 = getelementptr <4 x i32>* %p0, i32 9 - %p10 = getelementptr <4 x i32>* %p0, i32 10 - %p11 = getelementptr <4 x i32>* %p0, i32 11 - %p12 = getelementptr <4 x i32>* %p0, i32 12 - %p13 = getelementptr <4 x i32>* %p0, i32 13 - %p14 = getelementptr <4 x i32>* %p0, i32 14 - %p15 = getelementptr <4 x i32>* %p0, i32 15 - %p16 = getelementptr <4 x i32>* %p0, i32 16 - %p17 = getelementptr <4 x i32>* %p0, i32 17 - %p18 = getelementptr <4 x i32>* %p0, i32 18 - %p19 = getelementptr <4 x i32>* %p0, i32 19 - %p20 = getelementptr <4 x i32>* %p0, i32 20 - %p21 = getelementptr <4 x i32>* %p0, i32 21 - %p22 = getelementptr <4 x i32>* %p0, i32 22 - %p23 = getelementptr <4 x i32>* %p0, i32 23 - %p24 = getelementptr <4 x i32>* %p0, i32 24 - %p25 = getelementptr <4 x i32>* %p0, i32 25 - %p26 = getelementptr <4 x i32>* %p0, i32 26 - %p27 = getelementptr <4 x i32>* %p0, i32 27 - %p28 = getelementptr <4 x i32>* %p0, i32 28 - %p29 = getelementptr <4 x i32>* %p0, i32 29 - %p30 = getelementptr <4 x i32>* %p0, i32 30 - %p31 = getelementptr <4 x i32>* %p0, i32 31 - %p32 = getelementptr <4 x i32>* %p0, i32 32 - %p33 = getelementptr <4 x i32>* %p0, i32 33 - %0 = load <4 x i32>* %p0, align 16 - %1 = load <4 x i32>* %p1, align 16 - %2 = load <4 x i32>* %p2, align 16 - %3 = load <4 x i32>* %p3, align 16 - %4 = load <4 x i32>* %p4, align 16 - %5 = load <4 x i32>* %p5, align 16 - %6 = load <4 x i32>* %p6, align 16 - %7 = load <4 x i32>* %p7, align 16 - %8 = load <4 x i32>* %p8, align 16 - %9 = load <4 x i32>* %p9, align 16 - %10 = load <4 x i32>* %p10, align 16 - %11 = load <4 x i32>* %p11, align 16 - %12 = load <4 x i32>* %p12, align 16 - %13 = load <4 x i32>* %p13, align 16 - %14 = load <4 x i32>* %p14, align 16 - %15 = load <4 x i32>* %p15, align 16 - %16 = load <4 x i32>* %p16, align 16 - %17 = load <4 x i32>* %p17, align 16 - %18 = load <4 x i32>* %p18, align 16 - %19 = load <4 x i32>* %p19, align 16 - %20 = load <4 x i32>* %p20, align 16 - %21 = load <4 x i32>* %p21, align 16 - %22 = load <4 x i32>* %p22, align 16 - %23 = load <4 x i32>* %p23, align 16 - %24 = load <4 x i32>* %p24, align 16 - %25 = load <4 x i32>* %p25, align 16 - %26 = load <4 x i32>* %p26, align 16 - %27 = load <4 x i32>* %p27, align 16 - %28 = load <4 x i32>* %p28, align 16 - %29 = load <4 x i32>* %p29, align 16 - %30 = load <4 x i32>* %p30, align 16 - %31 = load <4 x i32>* %p31, align 16 - %32 = load <4 x i32>* %p32, align 16 - %33 = load <4 x i32>* %p33, align 16 + %p1 = getelementptr <4 x i32>, <4 x i32>* %p0, i32 1 + %p2 = getelementptr <4 x i32>, <4 x i32>* %p0, i32 2 + %p3 = getelementptr <4 x i32>, <4 x i32>* %p0, i32 3 + %p4 = getelementptr <4 x i32>, <4 x i32>* %p0, i32 4 + %p5 = getelementptr <4 x i32>, <4 x i32>* %p0, i32 5 + %p6 = getelementptr <4 x i32>, <4 x i32>* %p0, i32 6 + %p7 = getelementptr <4 x i32>, <4 x i32>* %p0, i32 7 + %p8 = getelementptr <4 x i32>, <4 x i32>* %p0, i32 8 + %p9 = getelementptr <4 x i32>, <4 x i32>* %p0, i32 9 + %p10 = getelementptr <4 x i32>, <4 x i32>* %p0, i32 10 + %p11 = getelementptr <4 x i32>, <4 x i32>* %p0, i32 11 + %p12 = getelementptr <4 x i32>, <4 x i32>* %p0, i32 12 + %p13 = getelementptr <4 x i32>, <4 x i32>* %p0, i32 13 + %p14 = getelementptr <4 x i32>, <4 x i32>* %p0, i32 14 + %p15 = getelementptr <4 x i32>, <4 x i32>* %p0, i32 15 + %p16 = getelementptr <4 x i32>, <4 x i32>* %p0, i32 16 + %p17 = getelementptr <4 x i32>, <4 x i32>* %p0, i32 17 + %p18 = getelementptr <4 x i32>, <4 x i32>* %p0, i32 18 + %p19 = getelementptr <4 x i32>, <4 x i32>* %p0, i32 19 + %p20 = getelementptr <4 x i32>, <4 x i32>* %p0, i32 20 + %p21 = getelementptr <4 x i32>, <4 x i32>* %p0, i32 21 + %p22 = getelementptr <4 x i32>, <4 x i32>* %p0, i32 22 + %p23 = getelementptr <4 x i32>, <4 x i32>* %p0, i32 23 + %p24 = getelementptr <4 x i32>, <4 x i32>* %p0, i32 24 + %p25 = getelementptr <4 x i32>, <4 x i32>* %p0, i32 25 + %p26 = getelementptr <4 x i32>, <4 x i32>* %p0, i32 26 + %p27 = getelementptr <4 x i32>, <4 x i32>* %p0, i32 27 + %p28 = getelementptr <4 x i32>, <4 x i32>* %p0, i32 28 + %p29 = getelementptr <4 x i32>, <4 x i32>* %p0, i32 29 + %p30 = getelementptr <4 x i32>, <4 x i32>* %p0, i32 30 + %p31 = getelementptr <4 x i32>, <4 x i32>* %p0, i32 31 + %p32 = getelementptr <4 x i32>, <4 x i32>* %p0, i32 32 + %p33 = getelementptr <4 x i32>, <4 x i32>* %p0, i32 33 + %0 = load <4 x i32>, <4 x i32>* %p0, align 16 + %1 = load <4 x i32>, <4 x i32>* %p1, align 16 + %2 = load <4 x i32>, <4 x i32>* %p2, align 16 + %3 = load <4 x i32>, <4 x i32>* %p3, align 16 + %4 = load <4 x i32>, <4 x i32>* %p4, align 16 + %5 = load <4 x i32>, <4 x i32>* %p5, align 16 + %6 = load <4 x i32>, <4 x i32>* %p6, align 16 + %7 = load <4 x i32>, <4 x i32>* %p7, align 16 + %8 = load <4 x i32>, <4 x i32>* %p8, align 16 + %9 = load <4 x i32>, <4 x i32>* %p9, align 16 + %10 = load <4 x i32>, <4 x i32>* %p10, align 16 + %11 = load <4 x i32>, <4 x i32>* %p11, align 16 + %12 = load <4 x i32>, <4 x i32>* %p12, align 16 + %13 = load <4 x i32>, <4 x i32>* %p13, align 16 + %14 = load <4 x i32>, <4 x i32>* %p14, align 16 + %15 = load <4 x i32>, <4 x i32>* %p15, align 16 + %16 = load <4 x i32>, <4 x i32>* %p16, align 16 + %17 = load <4 x i32>, <4 x i32>* %p17, align 16 + %18 = load <4 x i32>, <4 x i32>* %p18, align 16 + %19 = load <4 x i32>, <4 x i32>* %p19, align 16 + %20 = load <4 x i32>, <4 x i32>* %p20, align 16 + %21 = load <4 x i32>, <4 x i32>* %p21, align 16 + %22 = load <4 x i32>, <4 x i32>* %p22, align 16 + %23 = load <4 x i32>, <4 x i32>* %p23, align 16 + %24 = load <4 x i32>, <4 x i32>* %p24, align 16 + %25 = load <4 x i32>, <4 x i32>* %p25, align 16 + %26 = load <4 x i32>, <4 x i32>* %p26, align 16 + %27 = load <4 x i32>, <4 x i32>* %p27, align 16 + %28 = load <4 x i32>, <4 x i32>* %p28, align 16 + %29 = load <4 x i32>, <4 x i32>* %p29, align 16 + %30 = load <4 x i32>, <4 x i32>* %p30, align 16 + %31 = load <4 x i32>, <4 x i32>* %p31, align 16 + %32 = load <4 x i32>, <4 x i32>* %p32, align 16 + %33 = load <4 x i32>, <4 x i32>* %p33, align 16 %r1 = call <4 x i32> @llvm.mips.addv.w(<4 x i32> %0, <4 x i32> %1) %r2 = call <4 x i32> @llvm.mips.addv.w(<4 x i32> %r1, <4 x i32> %2) %r3 = call <4 x i32> @llvm.mips.addv.w(<4 x i32> %r2, <4 x i32> %3) @@ -453,73 +453,73 @@ declare i32 @llvm.mips.copy.s.w(<4 x i32>, i32) nounwind define i32 @test_i64(<2 x i64>* %p0, <2 x i64>* %q1) nounwind { entry: - %p1 = getelementptr <2 x i64>* %p0, i32 1 - %p2 = getelementptr <2 x i64>* %p0, i32 2 - %p3 = getelementptr <2 x i64>* %p0, i32 3 - %p4 = getelementptr <2 x i64>* %p0, i32 4 - %p5 = getelementptr <2 x i64>* %p0, i32 5 - %p6 = getelementptr <2 x i64>* %p0, i32 6 - %p7 = getelementptr <2 x i64>* %p0, i32 7 - %p8 = getelementptr <2 x i64>* %p0, i32 8 - %p9 = getelementptr <2 x i64>* %p0, i32 9 - %p10 = getelementptr <2 x i64>* %p0, i32 10 - %p11 = getelementptr <2 x i64>* %p0, i32 11 - %p12 = getelementptr <2 x i64>* %p0, i32 12 - %p13 = getelementptr <2 x i64>* %p0, i32 13 - %p14 = getelementptr <2 x i64>* %p0, i32 14 - %p15 = getelementptr <2 x i64>* %p0, i32 15 - %p16 = getelementptr <2 x i64>* %p0, i32 16 - %p17 = getelementptr <2 x i64>* %p0, i32 17 - %p18 = getelementptr <2 x i64>* %p0, i32 18 - %p19 = getelementptr <2 x i64>* %p0, i32 19 - %p20 = getelementptr <2 x i64>* %p0, i32 20 - %p21 = getelementptr <2 x i64>* %p0, i32 21 - %p22 = getelementptr <2 x i64>* %p0, i32 22 - %p23 = getelementptr <2 x i64>* %p0, i32 23 - %p24 = getelementptr <2 x i64>* %p0, i32 24 - %p25 = getelementptr <2 x i64>* %p0, i32 25 - %p26 = getelementptr <2 x i64>* %p0, i32 26 - %p27 = getelementptr <2 x i64>* %p0, i32 27 - %p28 = getelementptr <2 x i64>* %p0, i32 28 - %p29 = getelementptr <2 x i64>* %p0, i32 29 - %p30 = getelementptr <2 x i64>* %p0, i32 30 - %p31 = getelementptr <2 x i64>* %p0, i32 31 - %p32 = getelementptr <2 x i64>* %p0, i32 32 - %p33 = getelementptr <2 x i64>* %p0, i32 33 - %0 = load <2 x i64>* %p0, align 16 - %1 = load <2 x i64>* %p1, align 16 - %2 = load <2 x i64>* %p2, align 16 - %3 = load <2 x i64>* %p3, align 16 - %4 = load <2 x i64>* %p4, align 16 - %5 = load <2 x i64>* %p5, align 16 - %6 = load <2 x i64>* %p6, align 16 - %7 = load <2 x i64>* %p7, align 16 - %8 = load <2 x i64>* %p8, align 16 - %9 = load <2 x i64>* %p9, align 16 - %10 = load <2 x i64>* %p10, align 16 - %11 = load <2 x i64>* %p11, align 16 - %12 = load <2 x i64>* %p12, align 16 - %13 = load <2 x i64>* %p13, align 16 - %14 = load <2 x i64>* %p14, align 16 - %15 = load <2 x i64>* %p15, align 16 - %16 = load <2 x i64>* %p16, align 16 - %17 = load <2 x i64>* %p17, align 16 - %18 = load <2 x i64>* %p18, align 16 - %19 = load <2 x i64>* %p19, align 16 - %20 = load <2 x i64>* %p20, align 16 - %21 = load <2 x i64>* %p21, align 16 - %22 = load <2 x i64>* %p22, align 16 - %23 = load <2 x i64>* %p23, align 16 - %24 = load <2 x i64>* %p24, align 16 - %25 = load <2 x i64>* %p25, align 16 - %26 = load <2 x i64>* %p26, align 16 - %27 = load <2 x i64>* %p27, align 16 - %28 = load <2 x i64>* %p28, align 16 - %29 = load <2 x i64>* %p29, align 16 - %30 = load <2 x i64>* %p30, align 16 - %31 = load <2 x i64>* %p31, align 16 - %32 = load <2 x i64>* %p32, align 16 - %33 = load <2 x i64>* %p33, align 16 + %p1 = getelementptr <2 x i64>, <2 x i64>* %p0, i32 1 + %p2 = getelementptr <2 x i64>, <2 x i64>* %p0, i32 2 + %p3 = getelementptr <2 x i64>, <2 x i64>* %p0, i32 3 + %p4 = getelementptr <2 x i64>, <2 x i64>* %p0, i32 4 + %p5 = getelementptr <2 x i64>, <2 x i64>* %p0, i32 5 + %p6 = getelementptr <2 x i64>, <2 x i64>* %p0, i32 6 + %p7 = getelementptr <2 x i64>, <2 x i64>* %p0, i32 7 + %p8 = getelementptr <2 x i64>, <2 x i64>* %p0, i32 8 + %p9 = getelementptr <2 x i64>, <2 x i64>* %p0, i32 9 + %p10 = getelementptr <2 x i64>, <2 x i64>* %p0, i32 10 + %p11 = getelementptr <2 x i64>, <2 x i64>* %p0, i32 11 + %p12 = getelementptr <2 x i64>, <2 x i64>* %p0, i32 12 + %p13 = getelementptr <2 x i64>, <2 x i64>* %p0, i32 13 + %p14 = getelementptr <2 x i64>, <2 x i64>* %p0, i32 14 + %p15 = getelementptr <2 x i64>, <2 x i64>* %p0, i32 15 + %p16 = getelementptr <2 x i64>, <2 x i64>* %p0, i32 16 + %p17 = getelementptr <2 x i64>, <2 x i64>* %p0, i32 17 + %p18 = getelementptr <2 x i64>, <2 x i64>* %p0, i32 18 + %p19 = getelementptr <2 x i64>, <2 x i64>* %p0, i32 19 + %p20 = getelementptr <2 x i64>, <2 x i64>* %p0, i32 20 + %p21 = getelementptr <2 x i64>, <2 x i64>* %p0, i32 21 + %p22 = getelementptr <2 x i64>, <2 x i64>* %p0, i32 22 + %p23 = getelementptr <2 x i64>, <2 x i64>* %p0, i32 23 + %p24 = getelementptr <2 x i64>, <2 x i64>* %p0, i32 24 + %p25 = getelementptr <2 x i64>, <2 x i64>* %p0, i32 25 + %p26 = getelementptr <2 x i64>, <2 x i64>* %p0, i32 26 + %p27 = getelementptr <2 x i64>, <2 x i64>* %p0, i32 27 + %p28 = getelementptr <2 x i64>, <2 x i64>* %p0, i32 28 + %p29 = getelementptr <2 x i64>, <2 x i64>* %p0, i32 29 + %p30 = getelementptr <2 x i64>, <2 x i64>* %p0, i32 30 + %p31 = getelementptr <2 x i64>, <2 x i64>* %p0, i32 31 + %p32 = getelementptr <2 x i64>, <2 x i64>* %p0, i32 32 + %p33 = getelementptr <2 x i64>, <2 x i64>* %p0, i32 33 + %0 = load <2 x i64>, <2 x i64>* %p0, align 16 + %1 = load <2 x i64>, <2 x i64>* %p1, align 16 + %2 = load <2 x i64>, <2 x i64>* %p2, align 16 + %3 = load <2 x i64>, <2 x i64>* %p3, align 16 + %4 = load <2 x i64>, <2 x i64>* %p4, align 16 + %5 = load <2 x i64>, <2 x i64>* %p5, align 16 + %6 = load <2 x i64>, <2 x i64>* %p6, align 16 + %7 = load <2 x i64>, <2 x i64>* %p7, align 16 + %8 = load <2 x i64>, <2 x i64>* %p8, align 16 + %9 = load <2 x i64>, <2 x i64>* %p9, align 16 + %10 = load <2 x i64>, <2 x i64>* %p10, align 16 + %11 = load <2 x i64>, <2 x i64>* %p11, align 16 + %12 = load <2 x i64>, <2 x i64>* %p12, align 16 + %13 = load <2 x i64>, <2 x i64>* %p13, align 16 + %14 = load <2 x i64>, <2 x i64>* %p14, align 16 + %15 = load <2 x i64>, <2 x i64>* %p15, align 16 + %16 = load <2 x i64>, <2 x i64>* %p16, align 16 + %17 = load <2 x i64>, <2 x i64>* %p17, align 16 + %18 = load <2 x i64>, <2 x i64>* %p18, align 16 + %19 = load <2 x i64>, <2 x i64>* %p19, align 16 + %20 = load <2 x i64>, <2 x i64>* %p20, align 16 + %21 = load <2 x i64>, <2 x i64>* %p21, align 16 + %22 = load <2 x i64>, <2 x i64>* %p22, align 16 + %23 = load <2 x i64>, <2 x i64>* %p23, align 16 + %24 = load <2 x i64>, <2 x i64>* %p24, align 16 + %25 = load <2 x i64>, <2 x i64>* %p25, align 16 + %26 = load <2 x i64>, <2 x i64>* %p26, align 16 + %27 = load <2 x i64>, <2 x i64>* %p27, align 16 + %28 = load <2 x i64>, <2 x i64>* %p28, align 16 + %29 = load <2 x i64>, <2 x i64>* %p29, align 16 + %30 = load <2 x i64>, <2 x i64>* %p30, align 16 + %31 = load <2 x i64>, <2 x i64>* %p31, align 16 + %32 = load <2 x i64>, <2 x i64>* %p32, align 16 + %33 = load <2 x i64>, <2 x i64>* %p33, align 16 %r1 = call <2 x i64> @llvm.mips.addv.d(<2 x i64> %0, <2 x i64> %1) %r2 = call <2 x i64> @llvm.mips.addv.d(<2 x i64> %r1, <2 x i64> %2) %r3 = call <2 x i64> @llvm.mips.addv.d(<2 x i64> %r2, <2 x i64> %3) diff --git a/test/CodeGen/Mips/msa/vec.ll b/test/CodeGen/Mips/msa/vec.ll index d5b97f52fb834..8790923ce7273 100644 --- a/test/CodeGen/Mips/msa/vec.ll +++ b/test/CodeGen/Mips/msa/vec.ll @@ -9,8 +9,8 @@ define void @llvm_mips_and_v_b_test() nounwind { entry: - %0 = load <16 x i8>* @llvm_mips_and_v_b_ARG1 - %1 = load <16 x i8>* @llvm_mips_and_v_b_ARG2 + %0 = load <16 x i8>, <16 x i8>* @llvm_mips_and_v_b_ARG1 + %1 = load <16 x i8>, <16 x i8>* @llvm_mips_and_v_b_ARG2 %2 = bitcast <16 x i8> %0 to <16 x i8> %3 = bitcast <16 x i8> %1 to <16 x i8> %4 = tail call <16 x i8> @llvm.mips.and.v(<16 x i8> %2, <16 x i8> %3) @@ -32,8 +32,8 @@ entry: define void @llvm_mips_and_v_h_test() nounwind { entry: - %0 = load <8 x i16>* @llvm_mips_and_v_h_ARG1 - %1 = load <8 x i16>* @llvm_mips_and_v_h_ARG2 + %0 = load <8 x i16>, <8 x i16>* @llvm_mips_and_v_h_ARG1 + %1 = load <8 x i16>, <8 x i16>* @llvm_mips_and_v_h_ARG2 %2 = bitcast <8 x i16> %0 to <16 x i8> %3 = bitcast <8 x i16> %1 to <16 x i8> %4 = tail call <16 x i8> @llvm.mips.and.v(<16 x i8> %2, <16 x i8> %3) @@ -55,8 +55,8 @@ entry: define void @llvm_mips_and_v_w_test() nounwind { entry: - %0 = load <4 x i32>* @llvm_mips_and_v_w_ARG1 - %1 = load <4 x i32>* @llvm_mips_and_v_w_ARG2 + %0 = load <4 x i32>, <4 x i32>* @llvm_mips_and_v_w_ARG1 + %1 = load <4 x i32>, <4 x i32>* @llvm_mips_and_v_w_ARG2 %2 = bitcast <4 x i32> %0 to <16 x i8> %3 = bitcast <4 x i32> %1 to <16 x i8> %4 = tail call <16 x i8> @llvm.mips.and.v(<16 x i8> %2, <16 x i8> %3) @@ -78,8 +78,8 @@ entry: define void @llvm_mips_and_v_d_test() nounwind { entry: - %0 = load <2 x i64>* @llvm_mips_and_v_d_ARG1 - %1 = load <2 x i64>* @llvm_mips_and_v_d_ARG2 + %0 = load <2 x i64>, <2 x i64>* @llvm_mips_and_v_d_ARG1 + %1 = load <2 x i64>, <2 x i64>* @llvm_mips_and_v_d_ARG2 %2 = bitcast <2 x i64> %0 to <16 x i8> %3 = bitcast <2 x i64> %1 to <16 x i8> %4 = tail call <16 x i8> @llvm.mips.and.v(<16 x i8> %2, <16 x i8> %3) @@ -97,8 +97,8 @@ entry: ; define void @and_v_b_test() nounwind { entry: - %0 = load <16 x i8>* @llvm_mips_and_v_b_ARG1 - %1 = load <16 x i8>* @llvm_mips_and_v_b_ARG2 + %0 = load <16 x i8>, <16 x i8>* @llvm_mips_and_v_b_ARG1 + %1 = load <16 x i8>, <16 x i8>* @llvm_mips_and_v_b_ARG2 %2 = and <16 x i8> %0, %1 store <16 x i8> %2, <16 x i8>* @llvm_mips_and_v_b_RES ret void @@ -113,8 +113,8 @@ entry: ; define void @and_v_h_test() nounwind { entry: - %0 = load <8 x i16>* @llvm_mips_and_v_h_ARG1 - %1 = load <8 x i16>* @llvm_mips_and_v_h_ARG2 + %0 = load <8 x i16>, <8 x i16>* @llvm_mips_and_v_h_ARG1 + %1 = load <8 x i16>, <8 x i16>* @llvm_mips_and_v_h_ARG2 %2 = and <8 x i16> %0, %1 store <8 x i16> %2, <8 x i16>* @llvm_mips_and_v_h_RES ret void @@ -130,8 +130,8 @@ entry: define void @and_v_w_test() nounwind { entry: - %0 = load <4 x i32>* @llvm_mips_and_v_w_ARG1 - %1 = load <4 x i32>* @llvm_mips_and_v_w_ARG2 + %0 = load <4 x i32>, <4 x i32>* @llvm_mips_and_v_w_ARG1 + %1 = load <4 x i32>, <4 x i32>* @llvm_mips_and_v_w_ARG2 %2 = and <4 x i32> %0, %1 store <4 x i32> %2, <4 x i32>* @llvm_mips_and_v_w_RES ret void @@ -147,8 +147,8 @@ entry: define void @and_v_d_test() nounwind { entry: - %0 = load <2 x i64>* @llvm_mips_and_v_d_ARG1 - %1 = load <2 x i64>* @llvm_mips_and_v_d_ARG2 + %0 = load <2 x i64>, <2 x i64>* @llvm_mips_and_v_d_ARG1 + %1 = load <2 x i64>, <2 x i64>* @llvm_mips_and_v_d_ARG2 %2 = and <2 x i64> %0, %1 store <2 x i64> %2, <2 x i64>* @llvm_mips_and_v_d_RES ret void @@ -168,9 +168,9 @@ entry: define void @llvm_mips_bmnz_v_b_test() nounwind { entry: - %0 = load <16 x i8>* @llvm_mips_bmnz_v_b_ARG1 - %1 = load <16 x i8>* @llvm_mips_bmnz_v_b_ARG2 - %2 = load <16 x i8>* @llvm_mips_bmnz_v_b_ARG3 + %0 = load <16 x i8>, <16 x i8>* @llvm_mips_bmnz_v_b_ARG1 + %1 = load <16 x i8>, <16 x i8>* @llvm_mips_bmnz_v_b_ARG2 + %2 = load <16 x i8>, <16 x i8>* @llvm_mips_bmnz_v_b_ARG3 %3 = bitcast <16 x i8> %0 to <16 x i8> %4 = bitcast <16 x i8> %1 to <16 x i8> %5 = bitcast <16 x i8> %2 to <16 x i8> @@ -198,9 +198,9 @@ entry: define void @llvm_mips_bmnz_v_h_test() nounwind { entry: - %0 = load <8 x i16>* @llvm_mips_bmnz_v_h_ARG1 - %1 = load <8 x i16>* @llvm_mips_bmnz_v_h_ARG2 - %2 = load <8 x i16>* @llvm_mips_bmnz_v_h_ARG3 + %0 = load <8 x i16>, <8 x i16>* @llvm_mips_bmnz_v_h_ARG1 + %1 = load <8 x i16>, <8 x i16>* @llvm_mips_bmnz_v_h_ARG2 + %2 = load <8 x i16>, <8 x i16>* @llvm_mips_bmnz_v_h_ARG3 %3 = bitcast <8 x i16> %0 to <16 x i8> %4 = bitcast <8 x i16> %1 to <16 x i8> %5 = bitcast <8 x i16> %2 to <16 x i8> @@ -228,9 +228,9 @@ entry: define void @llvm_mips_bmnz_v_w_test() nounwind { entry: - %0 = load <4 x i32>* @llvm_mips_bmnz_v_w_ARG1 - %1 = load <4 x i32>* @llvm_mips_bmnz_v_w_ARG2 - %2 = load <4 x i32>* @llvm_mips_bmnz_v_w_ARG3 + %0 = load <4 x i32>, <4 x i32>* @llvm_mips_bmnz_v_w_ARG1 + %1 = load <4 x i32>, <4 x i32>* @llvm_mips_bmnz_v_w_ARG2 + %2 = load <4 x i32>, <4 x i32>* @llvm_mips_bmnz_v_w_ARG3 %3 = bitcast <4 x i32> %0 to <16 x i8> %4 = bitcast <4 x i32> %1 to <16 x i8> %5 = bitcast <4 x i32> %2 to <16 x i8> @@ -258,9 +258,9 @@ entry: define void @llvm_mips_bmnz_v_d_test() nounwind { entry: - %0 = load <2 x i64>* @llvm_mips_bmnz_v_d_ARG1 - %1 = load <2 x i64>* @llvm_mips_bmnz_v_d_ARG2 - %2 = load <2 x i64>* @llvm_mips_bmnz_v_d_ARG3 + %0 = load <2 x i64>, <2 x i64>* @llvm_mips_bmnz_v_d_ARG1 + %1 = load <2 x i64>, <2 x i64>* @llvm_mips_bmnz_v_d_ARG2 + %2 = load <2 x i64>, <2 x i64>* @llvm_mips_bmnz_v_d_ARG3 %3 = bitcast <2 x i64> %0 to <16 x i8> %4 = bitcast <2 x i64> %1 to <16 x i8> %5 = bitcast <2 x i64> %2 to <16 x i8> @@ -288,9 +288,9 @@ entry: define void @llvm_mips_bmz_v_b_test() nounwind { entry: - %0 = load <16 x i8>* @llvm_mips_bmz_v_b_ARG1 - %1 = load <16 x i8>* @llvm_mips_bmz_v_b_ARG2 - %2 = load <16 x i8>* @llvm_mips_bmz_v_b_ARG3 + %0 = load <16 x i8>, <16 x i8>* @llvm_mips_bmz_v_b_ARG1 + %1 = load <16 x i8>, <16 x i8>* @llvm_mips_bmz_v_b_ARG2 + %2 = load <16 x i8>, <16 x i8>* @llvm_mips_bmz_v_b_ARG3 %3 = bitcast <16 x i8> %0 to <16 x i8> %4 = bitcast <16 x i8> %1 to <16 x i8> %5 = bitcast <16 x i8> %2 to <16 x i8> @@ -319,9 +319,9 @@ entry: define void @llvm_mips_bmz_v_h_test() nounwind { entry: - %0 = load <8 x i16>* @llvm_mips_bmz_v_h_ARG1 - %1 = load <8 x i16>* @llvm_mips_bmz_v_h_ARG2 - %2 = load <8 x i16>* @llvm_mips_bmz_v_h_ARG3 + %0 = load <8 x i16>, <8 x i16>* @llvm_mips_bmz_v_h_ARG1 + %1 = load <8 x i16>, <8 x i16>* @llvm_mips_bmz_v_h_ARG2 + %2 = load <8 x i16>, <8 x i16>* @llvm_mips_bmz_v_h_ARG3 %3 = bitcast <8 x i16> %0 to <16 x i8> %4 = bitcast <8 x i16> %1 to <16 x i8> %5 = bitcast <8 x i16> %2 to <16 x i8> @@ -350,9 +350,9 @@ entry: define void @llvm_mips_bmz_v_w_test() nounwind { entry: - %0 = load <4 x i32>* @llvm_mips_bmz_v_w_ARG1 - %1 = load <4 x i32>* @llvm_mips_bmz_v_w_ARG2 - %2 = load <4 x i32>* @llvm_mips_bmz_v_w_ARG3 + %0 = load <4 x i32>, <4 x i32>* @llvm_mips_bmz_v_w_ARG1 + %1 = load <4 x i32>, <4 x i32>* @llvm_mips_bmz_v_w_ARG2 + %2 = load <4 x i32>, <4 x i32>* @llvm_mips_bmz_v_w_ARG3 %3 = bitcast <4 x i32> %0 to <16 x i8> %4 = bitcast <4 x i32> %1 to <16 x i8> %5 = bitcast <4 x i32> %2 to <16 x i8> @@ -381,9 +381,9 @@ entry: define void @llvm_mips_bmz_v_d_test() nounwind { entry: - %0 = load <2 x i64>* @llvm_mips_bmz_v_d_ARG1 - %1 = load <2 x i64>* @llvm_mips_bmz_v_d_ARG2 - %2 = load <2 x i64>* @llvm_mips_bmz_v_d_ARG3 + %0 = load <2 x i64>, <2 x i64>* @llvm_mips_bmz_v_d_ARG1 + %1 = load <2 x i64>, <2 x i64>* @llvm_mips_bmz_v_d_ARG2 + %2 = load <2 x i64>, <2 x i64>* @llvm_mips_bmz_v_d_ARG3 %3 = bitcast <2 x i64> %0 to <16 x i8> %4 = bitcast <2 x i64> %1 to <16 x i8> %5 = bitcast <2 x i64> %2 to <16 x i8> @@ -412,9 +412,9 @@ entry: define void @llvm_mips_bsel_v_b_test() nounwind { entry: - %0 = load <16 x i8>* @llvm_mips_bsel_v_b_ARG1 - %1 = load <16 x i8>* @llvm_mips_bsel_v_b_ARG2 - %2 = load <16 x i8>* @llvm_mips_bsel_v_b_ARG3 + %0 = load <16 x i8>, <16 x i8>* @llvm_mips_bsel_v_b_ARG1 + %1 = load <16 x i8>, <16 x i8>* @llvm_mips_bsel_v_b_ARG2 + %2 = load <16 x i8>, <16 x i8>* @llvm_mips_bsel_v_b_ARG3 %3 = bitcast <16 x i8> %0 to <16 x i8> %4 = bitcast <16 x i8> %1 to <16 x i8> %5 = bitcast <16 x i8> %2 to <16 x i8> @@ -443,9 +443,9 @@ entry: define void @llvm_mips_bsel_v_h_test() nounwind { entry: - %0 = load <8 x i16>* @llvm_mips_bsel_v_h_ARG1 - %1 = load <8 x i16>* @llvm_mips_bsel_v_h_ARG2 - %2 = load <8 x i16>* @llvm_mips_bsel_v_h_ARG3 + %0 = load <8 x i16>, <8 x i16>* @llvm_mips_bsel_v_h_ARG1 + %1 = load <8 x i16>, <8 x i16>* @llvm_mips_bsel_v_h_ARG2 + %2 = load <8 x i16>, <8 x i16>* @llvm_mips_bsel_v_h_ARG3 %3 = bitcast <8 x i16> %0 to <16 x i8> %4 = bitcast <8 x i16> %1 to <16 x i8> %5 = bitcast <8 x i16> %2 to <16 x i8> @@ -474,9 +474,9 @@ entry: define void @llvm_mips_bsel_v_w_test() nounwind { entry: - %0 = load <4 x i32>* @llvm_mips_bsel_v_w_ARG1 - %1 = load <4 x i32>* @llvm_mips_bsel_v_w_ARG2 - %2 = load <4 x i32>* @llvm_mips_bsel_v_w_ARG3 + %0 = load <4 x i32>, <4 x i32>* @llvm_mips_bsel_v_w_ARG1 + %1 = load <4 x i32>, <4 x i32>* @llvm_mips_bsel_v_w_ARG2 + %2 = load <4 x i32>, <4 x i32>* @llvm_mips_bsel_v_w_ARG3 %3 = bitcast <4 x i32> %0 to <16 x i8> %4 = bitcast <4 x i32> %1 to <16 x i8> %5 = bitcast <4 x i32> %2 to <16 x i8> @@ -505,9 +505,9 @@ entry: define void @llvm_mips_bsel_v_d_test() nounwind { entry: - %0 = load <2 x i64>* @llvm_mips_bsel_v_d_ARG1 - %1 = load <2 x i64>* @llvm_mips_bsel_v_d_ARG2 - %2 = load <2 x i64>* @llvm_mips_bsel_v_d_ARG3 + %0 = load <2 x i64>, <2 x i64>* @llvm_mips_bsel_v_d_ARG1 + %1 = load <2 x i64>, <2 x i64>* @llvm_mips_bsel_v_d_ARG2 + %2 = load <2 x i64>, <2 x i64>* @llvm_mips_bsel_v_d_ARG3 %3 = bitcast <2 x i64> %0 to <16 x i8> %4 = bitcast <2 x i64> %1 to <16 x i8> %5 = bitcast <2 x i64> %2 to <16 x i8> @@ -535,8 +535,8 @@ entry: define void @llvm_mips_nor_v_b_test() nounwind { entry: - %0 = load <16 x i8>* @llvm_mips_nor_v_b_ARG1 - %1 = load <16 x i8>* @llvm_mips_nor_v_b_ARG2 + %0 = load <16 x i8>, <16 x i8>* @llvm_mips_nor_v_b_ARG1 + %1 = load <16 x i8>, <16 x i8>* @llvm_mips_nor_v_b_ARG2 %2 = bitcast <16 x i8> %0 to <16 x i8> %3 = bitcast <16 x i8> %1 to <16 x i8> %4 = tail call <16 x i8> @llvm.mips.nor.v(<16 x i8> %2, <16 x i8> %3) @@ -558,8 +558,8 @@ entry: define void @llvm_mips_nor_v_h_test() nounwind { entry: - %0 = load <8 x i16>* @llvm_mips_nor_v_h_ARG1 - %1 = load <8 x i16>* @llvm_mips_nor_v_h_ARG2 + %0 = load <8 x i16>, <8 x i16>* @llvm_mips_nor_v_h_ARG1 + %1 = load <8 x i16>, <8 x i16>* @llvm_mips_nor_v_h_ARG2 %2 = bitcast <8 x i16> %0 to <16 x i8> %3 = bitcast <8 x i16> %1 to <16 x i8> %4 = tail call <16 x i8> @llvm.mips.nor.v(<16 x i8> %2, <16 x i8> %3) @@ -581,8 +581,8 @@ entry: define void @llvm_mips_nor_v_w_test() nounwind { entry: - %0 = load <4 x i32>* @llvm_mips_nor_v_w_ARG1 - %1 = load <4 x i32>* @llvm_mips_nor_v_w_ARG2 + %0 = load <4 x i32>, <4 x i32>* @llvm_mips_nor_v_w_ARG1 + %1 = load <4 x i32>, <4 x i32>* @llvm_mips_nor_v_w_ARG2 %2 = bitcast <4 x i32> %0 to <16 x i8> %3 = bitcast <4 x i32> %1 to <16 x i8> %4 = tail call <16 x i8> @llvm.mips.nor.v(<16 x i8> %2, <16 x i8> %3) @@ -604,8 +604,8 @@ entry: define void @llvm_mips_nor_v_d_test() nounwind { entry: - %0 = load <2 x i64>* @llvm_mips_nor_v_d_ARG1 - %1 = load <2 x i64>* @llvm_mips_nor_v_d_ARG2 + %0 = load <2 x i64>, <2 x i64>* @llvm_mips_nor_v_d_ARG1 + %1 = load <2 x i64>, <2 x i64>* @llvm_mips_nor_v_d_ARG2 %2 = bitcast <2 x i64> %0 to <16 x i8> %3 = bitcast <2 x i64> %1 to <16 x i8> %4 = tail call <16 x i8> @llvm.mips.nor.v(<16 x i8> %2, <16 x i8> %3) @@ -627,8 +627,8 @@ entry: define void @llvm_mips_or_v_b_test() nounwind { entry: - %0 = load <16 x i8>* @llvm_mips_or_v_b_ARG1 - %1 = load <16 x i8>* @llvm_mips_or_v_b_ARG2 + %0 = load <16 x i8>, <16 x i8>* @llvm_mips_or_v_b_ARG1 + %1 = load <16 x i8>, <16 x i8>* @llvm_mips_or_v_b_ARG2 %2 = bitcast <16 x i8> %0 to <16 x i8> %3 = bitcast <16 x i8> %1 to <16 x i8> %4 = tail call <16 x i8> @llvm.mips.or.v(<16 x i8> %2, <16 x i8> %3) @@ -650,8 +650,8 @@ entry: define void @llvm_mips_or_v_h_test() nounwind { entry: - %0 = load <8 x i16>* @llvm_mips_or_v_h_ARG1 - %1 = load <8 x i16>* @llvm_mips_or_v_h_ARG2 + %0 = load <8 x i16>, <8 x i16>* @llvm_mips_or_v_h_ARG1 + %1 = load <8 x i16>, <8 x i16>* @llvm_mips_or_v_h_ARG2 %2 = bitcast <8 x i16> %0 to <16 x i8> %3 = bitcast <8 x i16> %1 to <16 x i8> %4 = tail call <16 x i8> @llvm.mips.or.v(<16 x i8> %2, <16 x i8> %3) @@ -673,8 +673,8 @@ entry: define void @llvm_mips_or_v_w_test() nounwind { entry: - %0 = load <4 x i32>* @llvm_mips_or_v_w_ARG1 - %1 = load <4 x i32>* @llvm_mips_or_v_w_ARG2 + %0 = load <4 x i32>, <4 x i32>* @llvm_mips_or_v_w_ARG1 + %1 = load <4 x i32>, <4 x i32>* @llvm_mips_or_v_w_ARG2 %2 = bitcast <4 x i32> %0 to <16 x i8> %3 = bitcast <4 x i32> %1 to <16 x i8> %4 = tail call <16 x i8> @llvm.mips.or.v(<16 x i8> %2, <16 x i8> %3) @@ -696,8 +696,8 @@ entry: define void @llvm_mips_or_v_d_test() nounwind { entry: - %0 = load <2 x i64>* @llvm_mips_or_v_d_ARG1 - %1 = load <2 x i64>* @llvm_mips_or_v_d_ARG2 + %0 = load <2 x i64>, <2 x i64>* @llvm_mips_or_v_d_ARG1 + %1 = load <2 x i64>, <2 x i64>* @llvm_mips_or_v_d_ARG2 %2 = bitcast <2 x i64> %0 to <16 x i8> %3 = bitcast <2 x i64> %1 to <16 x i8> %4 = tail call <16 x i8> @llvm.mips.or.v(<16 x i8> %2, <16 x i8> %3) @@ -715,8 +715,8 @@ entry: ; define void @or_v_b_test() nounwind { entry: - %0 = load <16 x i8>* @llvm_mips_or_v_b_ARG1 - %1 = load <16 x i8>* @llvm_mips_or_v_b_ARG2 + %0 = load <16 x i8>, <16 x i8>* @llvm_mips_or_v_b_ARG1 + %1 = load <16 x i8>, <16 x i8>* @llvm_mips_or_v_b_ARG2 %2 = or <16 x i8> %0, %1 store <16 x i8> %2, <16 x i8>* @llvm_mips_or_v_b_RES ret void @@ -731,8 +731,8 @@ entry: ; define void @or_v_h_test() nounwind { entry: - %0 = load <8 x i16>* @llvm_mips_or_v_h_ARG1 - %1 = load <8 x i16>* @llvm_mips_or_v_h_ARG2 + %0 = load <8 x i16>, <8 x i16>* @llvm_mips_or_v_h_ARG1 + %1 = load <8 x i16>, <8 x i16>* @llvm_mips_or_v_h_ARG2 %2 = or <8 x i16> %0, %1 store <8 x i16> %2, <8 x i16>* @llvm_mips_or_v_h_RES ret void @@ -748,8 +748,8 @@ entry: define void @or_v_w_test() nounwind { entry: - %0 = load <4 x i32>* @llvm_mips_or_v_w_ARG1 - %1 = load <4 x i32>* @llvm_mips_or_v_w_ARG2 + %0 = load <4 x i32>, <4 x i32>* @llvm_mips_or_v_w_ARG1 + %1 = load <4 x i32>, <4 x i32>* @llvm_mips_or_v_w_ARG2 %2 = or <4 x i32> %0, %1 store <4 x i32> %2, <4 x i32>* @llvm_mips_or_v_w_RES ret void @@ -765,8 +765,8 @@ entry: define void @or_v_d_test() nounwind { entry: - %0 = load <2 x i64>* @llvm_mips_or_v_d_ARG1 - %1 = load <2 x i64>* @llvm_mips_or_v_d_ARG2 + %0 = load <2 x i64>, <2 x i64>* @llvm_mips_or_v_d_ARG1 + %1 = load <2 x i64>, <2 x i64>* @llvm_mips_or_v_d_ARG2 %2 = or <2 x i64> %0, %1 store <2 x i64> %2, <2 x i64>* @llvm_mips_or_v_d_RES ret void @@ -785,8 +785,8 @@ entry: define void @llvm_mips_xor_v_b_test() nounwind { entry: - %0 = load <16 x i8>* @llvm_mips_xor_v_b_ARG1 - %1 = load <16 x i8>* @llvm_mips_xor_v_b_ARG2 + %0 = load <16 x i8>, <16 x i8>* @llvm_mips_xor_v_b_ARG1 + %1 = load <16 x i8>, <16 x i8>* @llvm_mips_xor_v_b_ARG2 %2 = bitcast <16 x i8> %0 to <16 x i8> %3 = bitcast <16 x i8> %1 to <16 x i8> %4 = tail call <16 x i8> @llvm.mips.xor.v(<16 x i8> %2, <16 x i8> %3) @@ -808,8 +808,8 @@ entry: define void @llvm_mips_xor_v_h_test() nounwind { entry: - %0 = load <8 x i16>* @llvm_mips_xor_v_h_ARG1 - %1 = load <8 x i16>* @llvm_mips_xor_v_h_ARG2 + %0 = load <8 x i16>, <8 x i16>* @llvm_mips_xor_v_h_ARG1 + %1 = load <8 x i16>, <8 x i16>* @llvm_mips_xor_v_h_ARG2 %2 = bitcast <8 x i16> %0 to <16 x i8> %3 = bitcast <8 x i16> %1 to <16 x i8> %4 = tail call <16 x i8> @llvm.mips.xor.v(<16 x i8> %2, <16 x i8> %3) @@ -831,8 +831,8 @@ entry: define void @llvm_mips_xor_v_w_test() nounwind { entry: - %0 = load <4 x i32>* @llvm_mips_xor_v_w_ARG1 - %1 = load <4 x i32>* @llvm_mips_xor_v_w_ARG2 + %0 = load <4 x i32>, <4 x i32>* @llvm_mips_xor_v_w_ARG1 + %1 = load <4 x i32>, <4 x i32>* @llvm_mips_xor_v_w_ARG2 %2 = bitcast <4 x i32> %0 to <16 x i8> %3 = bitcast <4 x i32> %1 to <16 x i8> %4 = tail call <16 x i8> @llvm.mips.xor.v(<16 x i8> %2, <16 x i8> %3) @@ -854,8 +854,8 @@ entry: define void @llvm_mips_xor_v_d_test() nounwind { entry: - %0 = load <2 x i64>* @llvm_mips_xor_v_d_ARG1 - %1 = load <2 x i64>* @llvm_mips_xor_v_d_ARG2 + %0 = load <2 x i64>, <2 x i64>* @llvm_mips_xor_v_d_ARG1 + %1 = load <2 x i64>, <2 x i64>* @llvm_mips_xor_v_d_ARG2 %2 = bitcast <2 x i64> %0 to <16 x i8> %3 = bitcast <2 x i64> %1 to <16 x i8> %4 = tail call <16 x i8> @llvm.mips.xor.v(<16 x i8> %2, <16 x i8> %3) @@ -873,8 +873,8 @@ entry: ; define void @xor_v_b_test() nounwind { entry: - %0 = load <16 x i8>* @llvm_mips_xor_v_b_ARG1 - %1 = load <16 x i8>* @llvm_mips_xor_v_b_ARG2 + %0 = load <16 x i8>, <16 x i8>* @llvm_mips_xor_v_b_ARG1 + %1 = load <16 x i8>, <16 x i8>* @llvm_mips_xor_v_b_ARG2 %2 = xor <16 x i8> %0, %1 store <16 x i8> %2, <16 x i8>* @llvm_mips_xor_v_b_RES ret void @@ -889,8 +889,8 @@ entry: ; define void @xor_v_h_test() nounwind { entry: - %0 = load <8 x i16>* @llvm_mips_xor_v_h_ARG1 - %1 = load <8 x i16>* @llvm_mips_xor_v_h_ARG2 + %0 = load <8 x i16>, <8 x i16>* @llvm_mips_xor_v_h_ARG1 + %1 = load <8 x i16>, <8 x i16>* @llvm_mips_xor_v_h_ARG2 %2 = xor <8 x i16> %0, %1 store <8 x i16> %2, <8 x i16>* @llvm_mips_xor_v_h_RES ret void @@ -906,8 +906,8 @@ entry: define void @xor_v_w_test() nounwind { entry: - %0 = load <4 x i32>* @llvm_mips_xor_v_w_ARG1 - %1 = load <4 x i32>* @llvm_mips_xor_v_w_ARG2 + %0 = load <4 x i32>, <4 x i32>* @llvm_mips_xor_v_w_ARG1 + %1 = load <4 x i32>, <4 x i32>* @llvm_mips_xor_v_w_ARG2 %2 = xor <4 x i32> %0, %1 store <4 x i32> %2, <4 x i32>* @llvm_mips_xor_v_w_RES ret void @@ -923,8 +923,8 @@ entry: define void @xor_v_d_test() nounwind { entry: - %0 = load <2 x i64>* @llvm_mips_xor_v_d_ARG1 - %1 = load <2 x i64>* @llvm_mips_xor_v_d_ARG2 + %0 = load <2 x i64>, <2 x i64>* @llvm_mips_xor_v_d_ARG1 + %1 = load <2 x i64>, <2 x i64>* @llvm_mips_xor_v_d_ARG2 %2 = xor <2 x i64> %0, %1 store <2 x i64> %2, <2 x i64>* @llvm_mips_xor_v_d_RES ret void diff --git a/test/CodeGen/Mips/msa/vecs10.ll b/test/CodeGen/Mips/msa/vecs10.ll index e22e0755ef00f..f442f77274471 100644 --- a/test/CodeGen/Mips/msa/vecs10.ll +++ b/test/CodeGen/Mips/msa/vecs10.ll @@ -7,7 +7,7 @@ define i32 @llvm_mips_bnz_v_test() nounwind { entry: - %0 = load <16 x i8>* @llvm_mips_bnz_v_ARG1 + %0 = load <16 x i8>, <16 x i8>* @llvm_mips_bnz_v_ARG1 %1 = tail call i32 @llvm.mips.bnz.v(<16 x i8> %0) %2 = icmp eq i32 %1, 0 br i1 %2, label %true, label %false @@ -28,7 +28,7 @@ declare i32 @llvm.mips.bnz.v(<16 x i8>) nounwind define i32 @llvm_mips_bz_v_test() nounwind { entry: - %0 = load <16 x i8>* @llvm_mips_bz_v_ARG1 + %0 = load <16 x i8>, <16 x i8>* @llvm_mips_bz_v_ARG1 %1 = tail call i32 @llvm.mips.bz.v(<16 x i8> %0) %2 = icmp eq i32 %1, 0 br i1 %2, label %true, label %false |