summaryrefslogtreecommitdiff
path: root/test/CodeGen/X86/vector-shuffle-128-v16.ll
diff options
context:
space:
mode:
Diffstat (limited to 'test/CodeGen/X86/vector-shuffle-128-v16.ll')
-rw-r--r--test/CodeGen/X86/vector-shuffle-128-v16.ll104
1 files changed, 50 insertions, 54 deletions
diff --git a/test/CodeGen/X86/vector-shuffle-128-v16.ll b/test/CodeGen/X86/vector-shuffle-128-v16.ll
index 2f5a2b116115..2ff7ef4328f5 100644
--- a/test/CodeGen/X86/vector-shuffle-128-v16.ll
+++ b/test/CodeGen/X86/vector-shuffle-128-v16.ll
@@ -12,8 +12,8 @@ define <16 x i8> @shuffle_v16i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00(
; SSE2-LABEL: shuffle_v16i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00:
; SSE2: # %bb.0:
; SSE2-NEXT: punpcklbw {{.*#+}} xmm0 = xmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
-; SSE2-NEXT: pshuflw {{.*#+}} xmm0 = xmm0[0,0,0,0,4,5,6,7]
-; SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,0,1,1]
+; SSE2-NEXT: pshuflw {{.*#+}} xmm0 = xmm0[0,0,2,3,4,5,6,7]
+; SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,0,0,0]
; SSE2-NEXT: retq
;
; SSSE3-LABEL: shuffle_v16i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00:
@@ -206,14 +206,14 @@ define <16 x i8> @shuffle_v16i8_00_00_01_01_02_02_03_03_04_04_05_05_06_06_07_07(
define <16 x i8> @shuffle_v16i8_0101010101010101(<16 x i8> %a, <16 x i8> %b) {
; SSE-LABEL: shuffle_v16i8_0101010101010101:
; SSE: # %bb.0:
-; SSE-NEXT: pshuflw {{.*#+}} xmm0 = xmm0[0,0,0,0,4,5,6,7]
-; SSE-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,0,1,1]
+; SSE-NEXT: pshuflw {{.*#+}} xmm0 = xmm0[0,0,2,3,4,5,6,7]
+; SSE-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,0,0,0]
; SSE-NEXT: retq
;
; AVX1-LABEL: shuffle_v16i8_0101010101010101:
; AVX1: # %bb.0:
-; AVX1-NEXT: vpshuflw {{.*#+}} xmm0 = xmm0[0,0,0,0,4,5,6,7]
-; AVX1-NEXT: vpshufd {{.*#+}} xmm0 = xmm0[0,0,1,1]
+; AVX1-NEXT: vpshuflw {{.*#+}} xmm0 = xmm0[0,0,2,3,4,5,6,7]
+; AVX1-NEXT: vpshufd {{.*#+}} xmm0 = xmm0[0,0,0,0]
; AVX1-NEXT: retq
;
; AVX2OR512VL-LABEL: shuffle_v16i8_0101010101010101:
@@ -257,8 +257,8 @@ define <16 x i8> @shuffle_v16i8_16_00_16_01_16_02_16_03_16_04_16_05_16_06_16_07(
; SSE2: # %bb.0:
; SSE2-NEXT: pxor %xmm2, %xmm2
; SSE2-NEXT: punpcklbw {{.*#+}} xmm2 = xmm2[0],xmm0[0],xmm2[1],xmm0[1],xmm2[2],xmm0[2],xmm2[3],xmm0[3],xmm2[4],xmm0[4],xmm2[5],xmm0[5],xmm2[6],xmm0[6],xmm2[7],xmm0[7]
-; SSE2-NEXT: pshuflw {{.*#+}} xmm0 = xmm1[0,0,0,0,4,5,6,7]
-; SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,0,1,1]
+; SSE2-NEXT: pshuflw {{.*#+}} xmm0 = xmm1[0,0,2,3,4,5,6,7]
+; SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,0,0,0]
; SSE2-NEXT: pand {{.*}}(%rip), %xmm0
; SSE2-NEXT: por %xmm2, %xmm0
; SSE2-NEXT: retq
@@ -372,14 +372,12 @@ define <16 x i8> @shuffle_v16i8_03_02_01_00_31_30_29_28_11_10_09_08_23_22_21_20(
; SSE2-NEXT: punpckhbw {{.*#+}} xmm2 = xmm2[8],xmm1[8],xmm2[9],xmm1[9],xmm2[10],xmm1[10],xmm2[11],xmm1[11],xmm2[12],xmm1[12],xmm2[13],xmm1[13],xmm2[14],xmm1[14],xmm2[15],xmm1[15]
; SSE2-NEXT: pshuflw {{.*#+}} xmm3 = xmm2[3,2,1,0,4,5,6,7]
; SSE2-NEXT: punpcklbw {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1],xmm0[2],xmm1[2],xmm0[3],xmm1[3],xmm0[4],xmm1[4],xmm0[5],xmm1[5],xmm0[6],xmm1[6],xmm0[7],xmm1[7]
-; SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm0[2,3,2,3]
-; SSE2-NEXT: pshuflw {{.*#+}} xmm1 = xmm1[3,2,1,0,4,5,6,7]
-; SSE2-NEXT: punpcklqdq {{.*#+}} xmm3 = xmm3[0],xmm1[0]
-; SSE2-NEXT: pshuflw {{.*#+}} xmm0 = xmm0[3,2,1,0,4,5,6,7]
-; SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm2[2,3,2,3]
-; SSE2-NEXT: pshuflw {{.*#+}} xmm1 = xmm1[3,2,1,0,4,5,6,7]
-; SSE2-NEXT: punpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm1[0]
-; SSE2-NEXT: packuswb %xmm3, %xmm0
+; SSE2-NEXT: pshufhw {{.*#+}} xmm1 = xmm0[0,1,2,3,7,6,5,4]
+; SSE2-NEXT: movsd {{.*#+}} xmm1 = xmm3[0],xmm1[1]
+; SSE2-NEXT: pshuflw {{.*#+}} xmm3 = xmm0[3,2,1,0,4,5,6,7]
+; SSE2-NEXT: pshufhw {{.*#+}} xmm0 = xmm2[0,1,2,3,7,6,5,4]
+; SSE2-NEXT: movsd {{.*#+}} xmm0 = xmm3[0],xmm0[1]
+; SSE2-NEXT: packuswb %xmm1, %xmm0
; SSE2-NEXT: retq
;
; SSSE3-LABEL: shuffle_v16i8_03_02_01_00_31_30_29_28_11_10_09_08_23_22_21_20:
@@ -1181,21 +1179,20 @@ define <16 x i8> @shuffle_v16i8_uu_10_02_07_22_14_07_02_18_03_01_14_18_09_11_00(
; SSE2: # %bb.0: # %entry
; SSE2-NEXT: pxor %xmm2, %xmm2
; SSE2-NEXT: movdqa %xmm0, %xmm3
-; SSE2-NEXT: punpckhbw {{.*#+}} xmm3 = xmm3[8],xmm2[8],xmm3[9],xmm2[9],xmm3[10],xmm2[10],xmm3[11],xmm2[11],xmm3[12],xmm2[12],xmm3[13],xmm2[13],xmm3[14],xmm2[14],xmm3[15],xmm2[15]
-; SSE2-NEXT: pshufd {{.*#+}} xmm4 = xmm3[0,3,0,1]
-; SSE2-NEXT: pshuflw {{.*#+}} xmm4 = xmm4[0,1,2,2,4,5,6,7]
-; SSE2-NEXT: pshufhw {{.*#+}} xmm4 = xmm4[0,1,2,3,4,5,7,7]
+; SSE2-NEXT: punpcklbw {{.*#+}} xmm3 = xmm3[0],xmm2[0],xmm3[1],xmm2[1],xmm3[2],xmm2[2],xmm3[3],xmm2[3],xmm3[4],xmm2[4],xmm3[5],xmm2[5],xmm3[6],xmm2[6],xmm3[7],xmm2[7]
+; SSE2-NEXT: pshuflw {{.*#+}} xmm4 = xmm3[1,3,2,0,4,5,6,7]
+; SSE2-NEXT: pshufd {{.*#+}} xmm4 = xmm4[0,0,2,1]
; SSE2-NEXT: movdqa {{.*#+}} xmm5 = [65535,65535,65535,0,65535,0,0,65535]
-; SSE2-NEXT: punpcklbw {{.*#+}} xmm0 = xmm0[0],xmm2[0],xmm0[1],xmm2[1],xmm0[2],xmm2[2],xmm0[3],xmm2[3],xmm0[4],xmm2[4],xmm0[5],xmm2[5],xmm0[6],xmm2[6],xmm0[7],xmm2[7]
-; SSE2-NEXT: pshufd {{.*#+}} xmm2 = xmm0[0,1,0,3]
-; SSE2-NEXT: pshuflw {{.*#+}} xmm2 = xmm2[0,3,1,3,4,5,6,7]
-; SSE2-NEXT: pshufhw {{.*#+}} xmm2 = xmm2[0,1,2,3,4,5,6,4]
-; SSE2-NEXT: pand %xmm5, %xmm2
-; SSE2-NEXT: pandn %xmm4, %xmm5
-; SSE2-NEXT: por %xmm2, %xmm5
-; SSE2-NEXT: psrlq $16, %xmm3
-; SSE2-NEXT: pshufd {{.*#+}} xmm2 = xmm3[0,2,2,3]
-; SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm0[3,1,1,3]
+; SSE2-NEXT: pand %xmm5, %xmm4
+; SSE2-NEXT: punpckhbw {{.*#+}} xmm0 = xmm0[8],xmm2[8],xmm0[9],xmm2[9],xmm0[10],xmm2[10],xmm0[11],xmm2[11],xmm0[12],xmm2[12],xmm0[13],xmm2[13],xmm0[14],xmm2[14],xmm0[15],xmm2[15]
+; SSE2-NEXT: pshufd {{.*#+}} xmm2 = xmm0[0,3,0,1]
+; SSE2-NEXT: pshuflw {{.*#+}} xmm2 = xmm2[0,1,2,2,4,5,6,7]
+; SSE2-NEXT: pshufhw {{.*#+}} xmm2 = xmm2[0,1,2,3,4,5,7,7]
+; SSE2-NEXT: pandn %xmm2, %xmm5
+; SSE2-NEXT: por %xmm4, %xmm5
+; SSE2-NEXT: psrlq $16, %xmm0
+; SSE2-NEXT: pshufd {{.*#+}} xmm2 = xmm0[0,2,2,3]
+; SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm3[3,1,1,3]
; SSE2-NEXT: pshuflw {{.*#+}} xmm0 = xmm0[0,1,2,1,4,5,6,7]
; SSE2-NEXT: pshufhw {{.*#+}} xmm0 = xmm0[0,1,2,3,4,5,7,4]
; SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm0[1,3,2,3]
@@ -1203,7 +1200,7 @@ define <16 x i8> @shuffle_v16i8_uu_10_02_07_22_14_07_02_18_03_01_14_18_09_11_00(
; SSE2-NEXT: packuswb %xmm5, %xmm2
; SSE2-NEXT: movdqa {{.*#+}} xmm0 = [255,255,255,255,0,255,255,255,0,255,255,255,0,255,255,255]
; SSE2-NEXT: pand %xmm0, %xmm2
-; SSE2-NEXT: pshuflw {{.*#+}} xmm1 = xmm1[3,3,1,1,4,5,6,7]
+; SSE2-NEXT: pshuflw {{.*#+}} xmm1 = xmm1[3,1,1,3,4,5,6,7]
; SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm1[0,0,1,1]
; SSE2-NEXT: pandn %xmm1, %xmm0
; SSE2-NEXT: por %xmm2, %xmm0
@@ -1489,8 +1486,8 @@ define <16 x i8> @insert_dup_mem_v16i8_i32(i32* %ptr) {
; SSE2: # %bb.0:
; SSE2-NEXT: movd {{.*#+}} xmm0 = mem[0],zero,zero,zero
; SSE2-NEXT: punpcklbw {{.*#+}} xmm0 = xmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
-; SSE2-NEXT: pshuflw {{.*#+}} xmm0 = xmm0[0,0,0,0,4,5,6,7]
-; SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,0,1,1]
+; SSE2-NEXT: pshuflw {{.*#+}} xmm0 = xmm0[0,0,2,3,4,5,6,7]
+; SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,0,0,0]
; SSE2-NEXT: retq
;
; SSSE3-LABEL: insert_dup_mem_v16i8_i32:
@@ -1531,8 +1528,8 @@ define <16 x i8> @insert_dup_mem_v16i8_sext_i8(i8* %ptr) {
; SSE2-NEXT: movsbl (%rdi), %eax
; SSE2-NEXT: movd %eax, %xmm0
; SSE2-NEXT: punpcklbw {{.*#+}} xmm0 = xmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
-; SSE2-NEXT: pshuflw {{.*#+}} xmm0 = xmm0[0,0,0,0,4,5,6,7]
-; SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,0,1,1]
+; SSE2-NEXT: pshuflw {{.*#+}} xmm0 = xmm0[0,0,2,3,4,5,6,7]
+; SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,0,0,0]
; SSE2-NEXT: retq
;
; SSSE3-LABEL: insert_dup_mem_v16i8_sext_i8:
@@ -1576,8 +1573,8 @@ define <16 x i8> @insert_dup_elt1_mem_v16i8_i32(i32* %ptr) {
; SSE2: # %bb.0:
; SSE2-NEXT: movd {{.*#+}} xmm0 = mem[0],zero,zero,zero
; SSE2-NEXT: punpcklbw {{.*#+}} xmm0 = xmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
-; SSE2-NEXT: pshuflw {{.*#+}} xmm0 = xmm0[1,1,1,1,4,5,6,7]
-; SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,0,1,1]
+; SSE2-NEXT: pshuflw {{.*#+}} xmm0 = xmm0[1,1,2,3,4,5,6,7]
+; SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,0,0,0]
; SSE2-NEXT: retq
;
; SSSE3-LABEL: insert_dup_elt1_mem_v16i8_i32:
@@ -1614,8 +1611,8 @@ define <16 x i8> @insert_dup_elt2_mem_v16i8_i32(i32* %ptr) {
; SSE2: # %bb.0:
; SSE2-NEXT: movd {{.*#+}} xmm0 = mem[0],zero,zero,zero
; SSE2-NEXT: punpcklbw {{.*#+}} xmm0 = xmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
-; SSE2-NEXT: pshuflw {{.*#+}} xmm0 = xmm0[2,2,2,2,4,5,6,7]
-; SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,0,1,1]
+; SSE2-NEXT: pshuflw {{.*#+}} xmm0 = xmm0[2,2,2,3,4,5,6,7]
+; SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,0,0,0]
; SSE2-NEXT: retq
;
; SSSE3-LABEL: insert_dup_elt2_mem_v16i8_i32:
@@ -1653,8 +1650,8 @@ define <16 x i8> @insert_dup_elt1_mem_v16i8_sext_i8(i8* %ptr) {
; SSE2-NEXT: movsbl (%rdi), %eax
; SSE2-NEXT: movd %eax, %xmm0
; SSE2-NEXT: punpcklbw {{.*#+}} xmm0 = xmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
-; SSE2-NEXT: pshuflw {{.*#+}} xmm0 = xmm0[1,1,1,1,4,5,6,7]
-; SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,0,1,1]
+; SSE2-NEXT: pshuflw {{.*#+}} xmm0 = xmm0[1,1,2,3,4,5,6,7]
+; SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,0,0,0]
; SSE2-NEXT: retq
;
; SSSE3-LABEL: insert_dup_elt1_mem_v16i8_sext_i8:
@@ -1706,8 +1703,8 @@ define <16 x i8> @insert_dup_elt2_mem_v16i8_sext_i8(i8* %ptr) {
; SSE2-NEXT: movsbl (%rdi), %eax
; SSE2-NEXT: movd %eax, %xmm0
; SSE2-NEXT: punpcklbw {{.*#+}} xmm0 = xmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
-; SSE2-NEXT: pshuflw {{.*#+}} xmm0 = xmm0[2,2,2,2,4,5,6,7]
-; SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,0,1,1]
+; SSE2-NEXT: pshuflw {{.*#+}} xmm0 = xmm0[2,2,2,3,4,5,6,7]
+; SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,0,0,0]
; SSE2-NEXT: retq
;
; SSSE3-LABEL: insert_dup_elt2_mem_v16i8_sext_i8:
@@ -1761,15 +1758,14 @@ define <16 x i8> @PR31364(i8* nocapture readonly %a, i8* nocapture readonly %b)
; SSE2-NEXT: shll $8, %ecx
; SSE2-NEXT: orl %eax, %ecx
; SSE2-NEXT: movzwl %cx, %eax
-; SSE2-NEXT: movd %eax, %xmm0
-; SSE2-NEXT: pxor %xmm1, %xmm1
-; SSE2-NEXT: punpcklbw {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1],xmm0[2],xmm1[2],xmm0[3],xmm1[3],xmm0[4],xmm1[4],xmm0[5],xmm1[5],xmm0[6],xmm1[6],xmm0[7],xmm1[7]
-; SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm0[0,1,0,3]
+; SSE2-NEXT: movd %eax, %xmm1
+; SSE2-NEXT: pxor %xmm0, %xmm0
+; SSE2-NEXT: punpcklbw {{.*#+}} xmm1 = xmm1[0],xmm0[0],xmm1[1],xmm0[1],xmm1[2],xmm0[2],xmm1[3],xmm0[3],xmm1[4],xmm0[4],xmm1[5],xmm0[5],xmm1[6],xmm0[6],xmm1[7],xmm0[7]
+; SSE2-NEXT: pshuflw {{.*#+}} xmm0 = xmm1[1,1,1,3,4,5,6,7]
+; SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,0,0,1]
+; SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm1[0,1,0,3]
; SSE2-NEXT: pshuflw {{.*#+}} xmm1 = xmm1[1,1,1,1,4,5,6,7]
; SSE2-NEXT: pshufhw {{.*#+}} xmm1 = xmm1[0,1,2,3,5,4,4,4]
-; SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,1,0,1]
-; SSE2-NEXT: pshuflw {{.*#+}} xmm0 = xmm0[1,1,1,1,4,5,6,7]
-; SSE2-NEXT: pshufhw {{.*#+}} xmm0 = xmm0[0,1,2,3,5,5,5,7]
; SSE2-NEXT: packuswb %xmm1, %xmm0
; SSE2-NEXT: retq
;
@@ -1813,13 +1809,13 @@ define <16 x i8> @PR31301(i8* nocapture readonly %x, i8* nocapture readonly %y)
; SSE2-NEXT: movzbl (%rdi), %eax
; SSE2-NEXT: movd %eax, %xmm0
; SSE2-NEXT: punpcklbw {{.*#+}} xmm0 = xmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
-; SSE2-NEXT: pshuflw {{.*#+}} xmm0 = xmm0[0,0,0,0,4,5,6,7]
-; SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,0,1,1]
+; SSE2-NEXT: pshuflw {{.*#+}} xmm0 = xmm0[0,0,2,3,4,5,6,7]
+; SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,0,0,0]
; SSE2-NEXT: movzbl (%rsi), %eax
; SSE2-NEXT: movd %eax, %xmm1
; SSE2-NEXT: punpcklbw {{.*#+}} xmm1 = xmm1[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
-; SSE2-NEXT: pshuflw {{.*#+}} xmm1 = xmm1[0,0,0,0,4,5,6,7]
-; SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm1[0,0,1,1]
+; SSE2-NEXT: pshuflw {{.*#+}} xmm1 = xmm1[0,0,2,3,4,5,6,7]
+; SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm1[0,0,0,0]
; SSE2-NEXT: punpcklbw {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1],xmm0[2],xmm1[2],xmm0[3],xmm1[3],xmm0[4],xmm1[4],xmm0[5],xmm1[5],xmm0[6],xmm1[6],xmm0[7],xmm1[7]
; SSE2-NEXT: retq
;
generated by cgit v1.2.3 (git 2.25.1) at 2024-11-13 16:26:37 +0000