vendor/llvm/llvm-trunk-r290819

1 files changed, 656 insertions, 91 deletions

diff --git a/test/CodeGen/X86/stack-folding-fp-avx512vl.ll b/test/CodeGen/X86/stack-folding-fp-avx512vl.ll
index 3ab96e3f46295..198a96df6b1fa 100644
--- a/test/CodeGen/X86/stack-folding-fp-avx512vl.ll
+++ b/test/CodeGen/X86/stack-folding-fp-avx512vl.ll
@@ -1,4 +1,4 @@
-; RUN: llc -O3 -disable-peephole -mtriple=x86_64-unknown-unknown -mattr=+avx512vl < %s | FileCheck %s
+; RUN: llc -O3 -disable-peephole -mtriple=x86_64-unknown-unknown -mattr=+avx512vl,+avx512dq < %s | FileCheck %s
 
 target datalayout = "e-m:e-i64:64-f80:128-n8:16:32:64-S128"
 target triple = "x86_64-unknown-unknown"
@@ -8,130 +8,695 @@ target triple = "x86_64-unknown-unknown"
 ; By including a nop call with sideeffects we can force a partial register spill of the
 ; relevant registers and check that the reload is correctly folded into the instruction.
 
-define double @stack_fold_addsd(double %a0, double %a1) {
-  ;CHECK-LABEL: stack_fold_addsd
-  ;CHECK:       vaddsd {{-?[0-9]*}}(%rsp), {{%xmm[0-9][0-9]*}}, {{%xmm[0-9][0-9]*}} {{.*#+}} 8-byte Folded Reload
-  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{flags}"()
-  %2 = fadd double %a0, %a1
-  ret double %2
+define <2 x double> @stack_fold_addpd(<2 x double> %a0, <2 x double> %a1) {
+  ;CHECK-LABEL: stack_fold_addpd
+  ;CHECK:       vaddpd {{-?[0-9]*}}(%rsp), {{%xmm[0-9][0-9]*}}, {{%xmm[0-9][0-9]*}} {{.*#+}} 16-byte Folded Reload
+  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %2 = fadd <2 x double> %a0, %a1
+  ret <2 x double> %2
+}
+
+define <4 x double> @stack_fold_addpd_ymm(<4 x double> %a0, <4 x double> %a1) {
+  ;CHECK-LABEL: stack_fold_addpd_ymm
+  ;CHECK:       vaddpd {{-?[0-9]*}}(%rsp), {{%ymm[0-9][0-9]*}}, {{%ymm[0-9][0-9]*}} {{.*#+}} 32-byte Folded Reload
+  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %2 = fadd <4 x double> %a0, %a1
+  ret <4 x double> %2
+}
+
+define <4 x float> @stack_fold_addps(<4 x float> %a0, <4 x float> %a1) {
+  ;CHECK-LABEL: stack_fold_addps
+  ;CHECK:       vaddps {{-?[0-9]*}}(%rsp), {{%xmm[0-9][0-9]*}}, {{%xmm[0-9][0-9]*}} {{.*#+}} 16-byte Folded Reload
+  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %2 = fadd <4 x float> %a0, %a1
+  ret <4 x float> %2
+}
+
+define <8 x float> @stack_fold_addps_ymm(<8 x float> %a0, <8 x float> %a1) {
+  ;CHECK-LABEL: stack_fold_addps_ymm
+  ;CHECK:       vaddps {{-?[0-9]*}}(%rsp), {{%ymm[0-9][0-9]*}}, {{%ymm[0-9][0-9]*}} {{.*#+}} 32-byte Folded Reload
+  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %2 = fadd <8 x float> %a0, %a1
+  ret <8 x float> %2
+}
+
+define <2 x double> @stack_fold_andnpd(<2 x double> %a0, <2 x double> %a1) {
+  ;CHECK-LABEL: stack_fold_andnpd
+  ;CHECK:       vandnpd {{-?[0-9]*}}(%rsp), {{%xmm[0-9][0-9]*}}, {{%xmm[0-9][0-9]*}} {{.*#+}} 16-byte Folded Reload
+  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %2 = bitcast <2 x double> %a0 to <2 x i64>
+  %3 = bitcast <2 x double> %a1 to <2 x i64>
+  %4 = xor <2 x i64> %2, <i64 -1, i64 -1>
+  %5 = and <2 x i64> %4, %3
+  %6 = bitcast <2 x i64> %5 to <2 x double>
+  ; fadd forces execution domain
+  %7 = fadd <2 x double> %6, <double 0x0, double 0x0>
+  ret <2 x double> %7
+}
+
+define <4 x double> @stack_fold_andnpd_ymm(<4 x double> %a0, <4 x double> %a1) {
+  ;CHECK-LABEL: stack_fold_andnpd_ymm
+  ;CHECK:       vandnpd {{-?[0-9]*}}(%rsp), {{%ymm[0-9][0-9]*}}, {{%ymm[0-9][0-9]*}} {{.*#+}} 32-byte Folded Reload
+  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %2 = bitcast <4 x double> %a0 to <4 x i64>
+  %3 = bitcast <4 x double> %a1 to <4 x i64>
+  %4 = xor <4 x i64> %2, <i64 -1, i64 -1, i64 -1, i64 -1>
+  %5 = and <4 x i64> %4, %3
+  %6 = bitcast <4 x i64> %5 to <4 x double>
+  ; fadd forces execution domain
+  %7 = fadd <4 x double> %6, <double 0x0, double 0x0, double 0x0, double 0x0>
+  ret <4 x double> %7
+}
+
+define <4 x float> @stack_fold_andnps(<4 x float> %a0, <4 x float> %a1) {
+  ;CHECK-LABEL: stack_fold_andnps
+  ;CHECK:       vandnps {{-?[0-9]*}}(%rsp), {{%xmm[0-9][0-9]*}}, {{%xmm[0-9][0-9]*}} {{.*#+}} 16-byte Folded Reload
+  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %2 = bitcast <4 x float> %a0 to <2 x i64>
+  %3 = bitcast <4 x float> %a1 to <2 x i64>
+  %4 = xor <2 x i64> %2, <i64 -1, i64 -1>
+  %5 = and <2 x i64> %4, %3
+  %6 = bitcast <2 x i64> %5 to <4 x float>
+  ; fadd forces execution domain
+  %7 = fadd <4 x float> %6, <float 0x0, float 0x0, float 0x0, float 0x0>
+  ret <4 x float> %7
+}
+
+define <8 x float> @stack_fold_andnps_ymm(<8 x float> %a0, <8 x float> %a1) {
+  ;CHECK-LABEL: stack_fold_andnps_ymm
+  ;CHECK:       vandnps {{-?[0-9]*}}(%rsp), {{%ymm[0-9][0-9]*}}, {{%ymm[0-9][0-9]*}} {{.*#+}} 32-byte Folded Reload
+  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %2 = bitcast <8 x float> %a0 to <4 x i64>
+  %3 = bitcast <8 x float> %a1 to <4 x i64>
+  %4 = xor <4 x i64> %2, <i64 -1, i64 -1, i64 -1, i64 -1>
+  %5 = and <4 x i64> %4, %3
+  %6 = bitcast <4 x i64> %5 to <8 x float>
+  ; fadd forces execution domain
+  %7 = fadd <8 x float> %6, <float 0x0, float 0x0, float 0x0, float 0x0, float 0x0, float 0x0, float 0x0, float 0x0>
+  ret <8 x float> %7
+}
+
+define <2 x double> @stack_fold_andpd(<2 x double> %a0, <2 x double> %a1) {
+  ;CHECK-LABEL: stack_fold_andpd
+  ;CHECK:       vandpd {{-?[0-9]*}}(%rsp), {{%xmm[0-9][0-9]*}}, {{%xmm[0-9][0-9]*}} {{.*#+}} 16-byte Folded Reload
+  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %2 = bitcast <2 x double> %a0 to <2 x i64>
+  %3 = bitcast <2 x double> %a1 to <2 x i64>
+  %4 = and <2 x i64> %2, %3
+  %5 = bitcast <2 x i64> %4 to <2 x double>
+  ; fadd forces execution domain
+  %6 = fadd <2 x double> %5, <double 0x0, double 0x0>
+  ret <2 x double> %6
+}
+
+define <4 x double> @stack_fold_andpd_ymm(<4 x double> %a0, <4 x double> %a1) {
+  ;CHECK-LABEL: stack_fold_andpd_ymm
+  ;CHECK:       vandpd {{-?[0-9]*}}(%rsp), {{%ymm[0-9][0-9]*}}, {{%ymm[0-9][0-9]*}} {{.*#+}} 32-byte Folded Reload
+  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %2 = bitcast <4 x double> %a0 to <4 x i64>
+  %3 = bitcast <4 x double> %a1 to <4 x i64>
+  %4 = and <4 x i64> %2, %3
+  %5 = bitcast <4 x i64> %4 to <4 x double>
+  ; fadd forces execution domain
+  %6 = fadd <4 x double> %5, <double 0x0, double 0x0, double 0x0, double 0x0>
+  ret <4 x double> %6
+}
+
+define <4 x float> @stack_fold_andps(<4 x float> %a0, <4 x float> %a1) {
+  ;CHECK-LABEL: stack_fold_andps
+  ;CHECK:       vandps {{-?[0-9]*}}(%rsp), {{%xmm[0-9][0-9]*}}, {{%xmm[0-9][0-9]*}} {{.*#+}} 16-byte Folded Reload
+  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %2 = bitcast <4 x float> %a0 to <4 x i32>
+  %3 = bitcast <4 x float> %a1 to <4 x i32>
+  %4 = and <4 x i32> %2, %3
+  %5 = bitcast <4 x i32> %4 to <4 x float>
+  ; fadd forces execution domain
+  %6 = fadd <4 x float> %5, <float 0x0, float 0x0, float 0x0, float 0x0>
+  ret <4 x float> %6
+}
+
+define <8 x float> @stack_fold_andps_ymm(<8 x float> %a0, <8 x float> %a1) {
+  ;CHECK-LABEL: stack_fold_andps_ymm
+  ;CHECK:       vandps {{-?[0-9]*}}(%rsp), {{%ymm[0-9][0-9]*}}, {{%ymm[0-9][0-9]*}} {{.*#+}} 32-byte Folded Reload
+  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %2 = bitcast <8 x float> %a0 to <8 x i32>
+  %3 = bitcast <8 x float> %a1 to <8 x i32>
+  %4 = and <8 x i32> %2, %3
+  %5 = bitcast <8 x i32> %4 to <8 x float>
+  ; fadd forces execution domain
+  %6 = fadd <8 x float> %5, <float 0x0, float 0x0, float 0x0, float 0x0, float 0x0, float 0x0, float 0x0, float 0x0>
+  ret <8 x float> %6
+}
+
+define i8 @stack_fold_cmppd(<2 x double> %a0, <2 x double> %a1) {
+  ;CHECK-LABEL: stack_fold_cmppd
+  ;CHECK:       vcmpeqpd {{-?[0-9]*}}(%rsp), {{%xmm[0-9][0-9]*}}, {{%k[0-9]}} {{.*#+}} 16-byte Folded Reload
+  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %res = call i8 @llvm.x86.avx512.mask.cmp.pd.128(<2 x double> %a0, <2 x double> %a1, i32 0, i8 -1)
+  ret i8 %res
+}
+declare i8 @llvm.x86.avx512.mask.cmp.pd.128(<2 x double> , <2 x double> , i32, i8)
+
+define i8 @stack_fold_cmppd_ymm(<4 x double> %a0, <4 x double> %a1) {
+  ;CHECK-LABEL: stack_fold_cmppd_ymm
+  ;CHECK:       vcmpeqpd {{-?[0-9]*}}(%rsp), {{%ymm[0-9][0-9]*}}, {{%k[0-9]}} {{.*#+}} 32-byte Folded Reload
+  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %res = call i8 @llvm.x86.avx512.mask.cmp.pd.256(<4 x double> %a0, <4 x double> %a1, i32 0, i8 -1)
+  ret i8 %res
+}
+declare i8 @llvm.x86.avx512.mask.cmp.pd.256(<4 x double> , <4 x double> , i32, i8)
+
+define i8 @stack_fold_cmpps(<4 x float> %a0, <4 x float> %a1) {
+  ;CHECK-LABEL: stack_fold_cmpps
+  ;CHECK:       vcmpeqps {{-?[0-9]*}}(%rsp), {{%xmm[0-9][0-9]*}}, {{%k[0-9]*}} {{.*#+}} 16-byte Folded Reload
+  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %res = call i8 @llvm.x86.avx512.mask.cmp.ps.128(<4 x float> %a0, <4 x float> %a1, i32 0, i8 -1)
+  ret i8 %res
+}
+declare i8 @llvm.x86.avx512.mask.cmp.ps.128(<4 x float> , <4 x float> , i32, i8)
+
+define i8 @stack_fold_cmpps_ymm(<8 x float> %a0, <8 x float> %a1) {
+  ;CHECK-LABEL: stack_fold_cmpps_ymm
+  ;CHECK:       vcmpeqps {{-?[0-9]*}}(%rsp), {{%ymm[0-9][0-9]*}}, {{%k[0-9]*}} {{.*#+}} 32-byte Folded Reload
+  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %res = call i8 @llvm.x86.avx512.mask.cmp.ps.256(<8 x float> %a0, <8 x float> %a1, i32 0, i8 -1)
+  ret i8 %res
 }
+declare i8 @llvm.x86.avx512.mask.cmp.ps.256(<8 x float> , <8 x float> , i32, i8)
 
-define <2 x double> @stack_fold_addsd_int(<2 x double> %a0, <2 x double> %a1) {
-  ;CHECK-LABEL: stack_fold_addsd_int
-  ;CHECK:       vaddsd {{-?[0-9]*}}(%rsp), {{%xmm[0-9][0-9]*}}, {{%xmm[0-9][0-9]*}} {{.*#+}} 16-byte Folded Reload
-  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{flags}"()
-  %2 = call <2 x double> @llvm.x86.sse2.add.sd(<2 x double> %a0, <2 x double> %a1)
+define <2 x double> @stack_fold_divpd(<2 x double> %a0, <2 x double> %a1) {
+  ;CHECK-LABEL: stack_fold_divpd
+  ;CHECK:       vdivpd {{-?[0-9]*}}(%rsp), {{%xmm[0-9][0-9]*}}, {{%xmm[0-9][0-9]*}} {{.*#+}} 16-byte Folded Reload
+  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %2 = fdiv <2 x double> %a0, %a1
   ret <2 x double> %2
 }
-declare <2 x double> @llvm.x86.sse2.add.sd(<2 x double>, <2 x double>) nounwind readnone
 
-define float @stack_fold_addss(float %a0, float %a1) {
-  ;CHECK-LABEL: stack_fold_addss
-  ;CHECK:       vaddss {{-?[0-9]*}}(%rsp), {{%xmm[0-9][0-9]*}}, {{%xmm[0-9][0-9]*}} {{.*#+}} 4-byte Folded Reload
-  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{flags}"()
-  %2 = fadd float %a0, %a1
-  ret float %2
+define <4 x double> @stack_fold_divpd_ymm(<4 x double> %a0, <4 x double> %a1) {
+  ;CHECK-LABEL: stack_fold_divpd_ymm
+  ;CHECK:       vdivpd {{-?[0-9]*}}(%rsp), {{%ymm[0-9][0-9]*}}, {{%ymm[0-9][0-9]*}} {{.*#+}} 32-byte Folded Reload
+  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %2 = fdiv <4 x double> %a0, %a1
+  ret <4 x double> %2
 }
 
-define <4 x float> @stack_fold_addss_int(<4 x float> %a0, <4 x float> %a1) {
-  ;CHECK-LABEL: stack_fold_addss_int
-  ;CHECK:       vaddss {{-?[0-9]*}}(%rsp), {{%xmm[0-9][0-9]*}}, {{%xmm[0-9][0-9]*}} {{.*#+}} 16-byte Folded Reload
-  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{flags}"()
-  %2 = call <4 x float> @llvm.x86.sse.add.ss(<4 x float> %a0, <4 x float> %a1)
+define <4 x float> @stack_fold_divps(<4 x float> %a0, <4 x float> %a1) {
+  ;CHECK-LABEL: stack_fold_divps
+  ;CHECK:       vdivps {{-?[0-9]*}}(%rsp), {{%xmm[0-9][0-9]*}}, {{%xmm[0-9][0-9]*}} {{.*#+}} 16-byte Folded Reload
+  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %2 = fdiv <4 x float> %a0, %a1
   ret <4 x float> %2
 }
-declare <4 x float> @llvm.x86.sse.add.ss(<4 x float>, <4 x float>) nounwind readnone
 
-define <2 x double> @stack_fold_divsd_int(<2 x double> %a0, <2 x double> %a1) {
-  ;CHECK-LABEL: stack_fold_divsd_int
-  ;CHECK:       vdivsd {{-?[0-9]*}}(%rsp), {{%xmm[0-9][0-9]*}}, {{%xmm[0-9][0-9]*}} {{.*#+}} 16-byte Folded Reload
-  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{flags}"()
-  %2 = call <2 x double> @llvm.x86.sse2.div.sd(<2 x double> %a0, <2 x double> %a1)
+define <8 x float> @stack_fold_divps_ymm(<8 x float> %a0, <8 x float> %a1) {
+  ;CHECK-LABEL: stack_fold_divps_ymm
+  ;CHECK:       vdivps {{-?[0-9]*}}(%rsp), {{%ymm[0-9][0-9]*}}, {{%ymm[0-9][0-9]*}} {{.*#+}} 32-byte Folded Reload
+  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %2 = fdiv <8 x float> %a0, %a1
+  ret <8 x float> %2
+}
+
+define <2 x double> @stack_fold_maxpd(<2 x double> %a0, <2 x double> %a1) #0 {
+  ;CHECK-LABEL: stack_fold_maxpd
+  ;CHECK:       vmaxpd {{-?[0-9]*}}(%rsp), {{%xmm[0-9][0-9]*}}, {{%xmm[0-9][0-9]*}} {{.*#+}} 16-byte Folded Reload
+  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %2 = call <2 x double> @llvm.x86.sse2.max.pd(<2 x double> %a0, <2 x double> %a1)
+  ret <2 x double> %2
+}
+declare <2 x double> @llvm.x86.sse2.max.pd(<2 x double>, <2 x double>) nounwind readnone
+
+define <2 x double> @stack_fold_maxpd_commutable(<2 x double> %a0, <2 x double> %a1) #1 {
+  ;CHECK-LABEL: stack_fold_maxpd_commutable
+  ;CHECK:       vmaxpd {{-?[0-9]*}}(%rsp), {{%xmm[0-9][0-9]*}}, {{%xmm[0-9][0-9]*}} {{.*#+}} 16-byte Folded Reload
+  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %2 = call <2 x double> @llvm.x86.sse2.max.pd(<2 x double> %a0, <2 x double> %a1)
   ret <2 x double> %2
 }
-declare <2 x double> @llvm.x86.sse2.div.sd(<2 x double>, <2 x double>) nounwind readnone
 
-define float @stack_fold_divss(float %a0, float %a1) {
-  ;CHECK-LABEL: stack_fold_divss
-  ;CHECK:       vdivss {{-?[0-9]*}}(%rsp), {{%xmm[0-9][0-9]*}}, {{%xmm[0-9][0-9]*}} {{.*#+}} 4-byte Folded Reload
-  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{flags}"()
-  %2 = fdiv float %a0, %a1
-  ret float %2
+define <4 x double> @stack_fold_maxpd_ymm(<4 x double> %a0, <4 x double> %a1) #0 {
+  ;CHECK-LABEL: stack_fold_maxpd_ymm
+  ;CHECK:       vmaxpd {{-?[0-9]*}}(%rsp), {{%ymm[0-9][0-9]*}}, {{%ymm[0-9][0-9]*}} {{.*#+}} 32-byte Folded Reload
+  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %2 = call <4 x double> @llvm.x86.avx.max.pd.256(<4 x double> %a0, <4 x double> %a1)
+  ret <4 x double> %2
+}
+declare <4 x double> @llvm.x86.avx.max.pd.256(<4 x double>, <4 x double>) nounwind readnone
+
+define <4 x double> @stack_fold_maxpd_ymm_commutable(<4 x double> %a0, <4 x double> %a1) #1 {
+  ;CHECK-LABEL: stack_fold_maxpd_ymm_commutable
+  ;CHECK:       vmaxpd {{-?[0-9]*}}(%rsp), {{%ymm[0-9][0-9]*}}, {{%ymm[0-9][0-9]*}} {{.*#+}} 32-byte Folded Reload
+  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %2 = call <4 x double> @llvm.x86.avx.max.pd.256(<4 x double> %a0, <4 x double> %a1)
+  ret <4 x double> %2
+}
+
+define <4 x float> @stack_fold_maxps(<4 x float> %a0, <4 x float> %a1) #0 {
+  ;CHECK-LABEL: stack_fold_maxps
+  ;CHECK:       vmaxps {{-?[0-9]*}}(%rsp), {{%xmm[0-9][0-9]*}}, {{%xmm[0-9][0-9]*}} {{.*#+}} 16-byte Folded Reload
+  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %2 = call <4 x float> @llvm.x86.sse.max.ps(<4 x float> %a0, <4 x float> %a1)
+  ret <4 x float> %2
 }
+declare <4 x float> @llvm.x86.sse.max.ps(<4 x float>, <4 x float>) nounwind readnone
 
-define <4 x float> @stack_fold_divss_int(<4 x float> %a0, <4 x float> %a1) {
-  ;CHECK-LABEL: stack_fold_divss_int
-  ;CHECK:       vdivss {{-?[0-9]*}}(%rsp), {{%xmm[0-9][0-9]*}}, {{%xmm[0-9][0-9]*}} {{.*#+}} 16-byte Folded Reload
-  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{flags}"()
-  %2 = call <4 x float> @llvm.x86.sse.div.ss(<4 x float> %a0, <4 x float> %a1)
+define <4 x float> @stack_fold_maxps_commutable(<4 x float> %a0, <4 x float> %a1) #1 {
+  ;CHECK-LABEL: stack_fold_maxps_commutable
+  ;CHECK:       vmaxps {{-?[0-9]*}}(%rsp), {{%xmm[0-9][0-9]*}}, {{%xmm[0-9][0-9]*}} {{.*#+}} 16-byte Folded Reload
+  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %2 = call <4 x float> @llvm.x86.sse.max.ps(<4 x float> %a0, <4 x float> %a1)
   ret <4 x float> %2
 }
-declare <4 x float> @llvm.x86.sse.div.ss(<4 x float>, <4 x float>) nounwind readnone
 
-define double @stack_fold_mulsd(double %a0, double %a1) {
-  ;CHECK-LABEL: stack_fold_mulsd
-  ;CHECK:       vmulsd {{-?[0-9]*}}(%rsp), {{%xmm[0-9][0-9]*}}, {{%xmm[0-9][0-9]*}} {{.*#+}} 8-byte Folded Reload
-  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{flags}"()
-  %2 = fmul double %a0, %a1
-  ret double %2
+define <8 x float> @stack_fold_maxps_ymm(<8 x float> %a0, <8 x float> %a1) #0 {
+  ;CHECK-LABEL: stack_fold_maxps_ymm
+  ;CHECK:       vmaxps {{-?[0-9]*}}(%rsp), {{%ymm[0-9][0-9]*}}, {{%ymm[0-9][0-9]*}} {{.*#+}} 32-byte Folded Reload
+  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %2 = call <8 x float> @llvm.x86.avx.max.ps.256(<8 x float> %a0, <8 x float> %a1)
+  ret <8 x float> %2
+}
+declare <8 x float> @llvm.x86.avx.max.ps.256(<8 x float>, <8 x float>) nounwind readnone
+
+define <8 x float> @stack_fold_maxps_ymm_commutable(<8 x float> %a0, <8 x float> %a1) #1 {
+  ;CHECK-LABEL: stack_fold_maxps_ymm_commutable
+  ;CHECK:       vmaxps {{-?[0-9]*}}(%rsp), {{%ymm[0-9][0-9]*}}, {{%ymm[0-9][0-9]*}} {{.*#+}} 32-byte Folded Reload
+  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %2 = call <8 x float> @llvm.x86.avx.max.ps.256(<8 x float> %a0, <8 x float> %a1)
+  ret <8 x float> %2
 }
 
-define <2 x double> @stack_fold_mulsd_int(<2 x double> %a0, <2 x double> %a1) {
-  ;CHECK-LABEL: stack_fold_mulsd_int
-  ;CHECK:       vmulsd {{-?[0-9]*}}(%rsp), {{%xmm[0-9][0-9]*}}, {{%xmm[0-9][0-9]*}} {{.*#+}} 16-byte Folded Reload
-  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{flags}"()
-  %2 = call <2 x double> @llvm.x86.sse2.mul.sd(<2 x double> %a0, <2 x double> %a1)
+define <4 x float> @stack_fold_minps(<4 x float> %a0, <4 x float> %a1) #0 {
+  ;CHECK-LABEL: stack_fold_minps
+  ;CHECK:       vminps {{-?[0-9]*}}(%rsp), {{%xmm[0-9][0-9]*}}, {{%xmm[0-9][0-9]*}} {{.*#+}} 16-byte Folded Reload
+  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %2 = call <4 x float> @llvm.x86.sse.min.ps(<4 x float> %a0, <4 x float> %a1)
+  ret <4 x float> %2
+}
+declare <4 x float> @llvm.x86.sse.min.ps(<4 x float>, <4 x float>) nounwind readnone
+
+define <4 x float> @stack_fold_minps_commutable(<4 x float> %a0, <4 x float> %a1) #1 {
+  ;CHECK-LABEL: stack_fold_minps_commutable
+  ;CHECK:       vminps {{-?[0-9]*}}(%rsp), {{%xmm[0-9][0-9]*}}, {{%xmm[0-9][0-9]*}} {{.*#+}} 16-byte Folded Reload
+  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %2 = call <4 x float> @llvm.x86.sse.min.ps(<4 x float> %a0, <4 x float> %a1)
+  ret <4 x float> %2
+}
+
+define <8 x float> @stack_fold_minps_ymm(<8 x float> %a0, <8 x float> %a1) #0 {
+  ;CHECK-LABEL: stack_fold_minps_ymm
+  ;CHECK:       vminps {{-?[0-9]*}}(%rsp), {{%ymm[0-9][0-9]*}}, {{%ymm[0-9][0-9]*}} {{.*#+}} 32-byte Folded Reload
+  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %2 = call <8 x float> @llvm.x86.avx.min.ps.256(<8 x float> %a0, <8 x float> %a1)
+  ret <8 x float> %2
+}
+declare <8 x float> @llvm.x86.avx.min.ps.256(<8 x float>, <8 x float>) nounwind readnone
+
+define <8 x float> @stack_fold_minps_ymm_commutable(<8 x float> %a0, <8 x float> %a1) #1 {
+  ;CHECK-LABEL: stack_fold_minps_ymm_commutable
+  ;CHECK:       vminps {{-?[0-9]*}}(%rsp), {{%ymm[0-9][0-9]*}}, {{%ymm[0-9][0-9]*}} {{.*#+}} 32-byte Folded Reload
+  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %2 = call <8 x float> @llvm.x86.avx.min.ps.256(<8 x float> %a0, <8 x float> %a1)
+  ret <8 x float> %2
+}
+
+define <2 x double> @stack_fold_mulpd(<2 x double> %a0, <2 x double> %a1) {
+  ;CHECK-LABEL: stack_fold_mulpd
+  ;CHECK:       vmulpd {{-?[0-9]*}}(%rsp), {{%xmm[0-9][0-9]*}}, {{%xmm[0-9][0-9]*}} {{.*#+}} 16-byte Folded Reload
+  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %2 = fmul <2 x double> %a0, %a1
   ret <2 x double> %2
 }
-declare <2 x double> @llvm.x86.sse2.mul.sd(<2 x double>, <2 x double>) nounwind readnone
 
-define float @stack_fold_mulss(float %a0, float %a1) {
-  ;CHECK-LABEL: stack_fold_mulss
-  ;CHECK:       vmulss {{-?[0-9]*}}(%rsp), {{%xmm[0-9][0-9]*}}, {{%xmm[0-9][0-9]*}} {{.*#+}} 4-byte Folded Reload
-  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{flags}"()
-  %2 = fmul float %a0, %a1
-  ret float %2
+define <4 x double> @stack_fold_mulpd_ymm(<4 x double> %a0, <4 x double> %a1) {
+  ;CHECK-LABEL: stack_fold_mulpd_ymm
+  ;CHECK:       vmulpd {{-?[0-9]*}}(%rsp), {{%ymm[0-9][0-9]*}}, {{%ymm[0-9][0-9]*}} {{.*#+}} 32-byte Folded Reload
+  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %2 = fmul <4 x double> %a0, %a1
+  ret <4 x double> %2
 }
 
-define <4 x float> @stack_fold_mulss_int(<4 x float> %a0, <4 x float> %a1) {
-  ;CHECK-LABEL: stack_fold_mulss_int
-  ;CHECK:       vmulss {{-?[0-9]*}}(%rsp), {{%xmm[0-9][0-9]*}}, {{%xmm[0-9][0-9]*}} {{.*#+}} 16-byte Folded Reload
-  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{flags}"()
-  %2 = call <4 x float> @llvm.x86.sse.mul.ss(<4 x float> %a0, <4 x float> %a1)
+define <4 x float> @stack_fold_mulps(<4 x float> %a0, <4 x float> %a1) {
+  ;CHECK-LABEL: stack_fold_mulps
+  ;CHECK:       vmulps {{-?[0-9]*}}(%rsp), {{%xmm[0-9][0-9]*}}, {{%xmm[0-9][0-9]*}} {{.*#+}} 16-byte Folded Reload
+  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %2 = fmul <4 x float> %a0, %a1
   ret <4 x float> %2
 }
-declare <4 x float> @llvm.x86.sse.mul.ss(<4 x float>, <4 x float>) nounwind readnone
 
-define double @stack_fold_subsd(double %a0, double %a1) {
-  ;CHECK-LABEL: stack_fold_subsd
-  ;CHECK:       vsubsd {{-?[0-9]*}}(%rsp), {{%xmm[0-9][0-9]*}}, {{%xmm[0-9][0-9]*}} {{.*#+}} 8-byte Folded Reload
-  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{flags}"()
-  %2 = fsub double %a0, %a1
-  ret double %2
+define <8 x float> @stack_fold_mulps_ymm(<8 x float> %a0, <8 x float> %a1) {
+  ;CHECK-LABEL: stack_fold_mulps_ymm
+  ;CHECK:       vmulps {{-?[0-9]*}}(%rsp), {{%ymm[0-9][0-9]*}}, {{%ymm[0-9][0-9]*}} {{.*#+}} 32-byte Folded Reload
+  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %2 = fmul <8 x float> %a0, %a1
+  ret <8 x float> %2
+}
+
+define <2 x double> @stack_fold_orpd(<2 x double> %a0, <2 x double> %a1) #0 {
+  ;CHECK-LABEL: stack_fold_orpd
+  ;CHECK:       vorpd {{-?[0-9]*}}(%rsp), {{%xmm[0-9][0-9]*}}, {{%xmm[0-9][0-9]*}} {{.*#+}} 16-byte Folded Reload
+  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %2 = bitcast <2 x double> %a0 to <2 x i64>
+  %3 = bitcast <2 x double> %a1 to <2 x i64>
+  %4 = or <2 x i64> %2, %3
+  %5 = bitcast <2 x i64> %4 to <2 x double>
+  ; fadd forces execution domain
+  %6 = fadd <2 x double> %5, <double 0x0, double 0x0>
+  ret <2 x double> %6
 }
 
-define <2 x double> @stack_fold_subsd_int(<2 x double> %a0, <2 x double> %a1) {
-  ;CHECK-LABEL: stack_fold_subsd_int
-  ;CHECK:       vsubsd {{-?[0-9]*}}(%rsp), {{%xmm[0-9][0-9]*}}, {{%xmm[0-9][0-9]*}} {{.*#+}} 16-byte Folded Reload
-  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{flags}"()
-  %2 = call <2 x double> @llvm.x86.sse2.sub.sd(<2 x double> %a0, <2 x double> %a1)
+define <4 x double> @stack_fold_orpd_ymm(<4 x double> %a0, <4 x double> %a1) #0 {
+  ;CHECK-LABEL: stack_fold_orpd_ymm
+  ;CHECK:       vorpd {{-?[0-9]*}}(%rsp), {{%ymm[0-9][0-9]*}}, {{%ymm[0-9][0-9]*}} {{.*#+}} 32-byte Folded Reload
+  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %2 = bitcast <4 x double> %a0 to <4 x i64>
+  %3 = bitcast <4 x double> %a1 to <4 x i64>
+  %4 = or <4 x i64> %2, %3
+  %5 = bitcast <4 x i64> %4 to <4 x double>
+  ; fadd forces execution domain
+  %6 = fadd <4 x double> %5, <double 0x0, double 0x0, double 0x0, double 0x0>
+  ret <4 x double> %6
+}
+
+define <4 x float> @stack_fold_orps(<4 x float> %a0, <4 x float> %a1) {
+  ;CHECK-LABEL: stack_fold_orps
+  ;CHECK:       vorps {{-?[0-9]*}}(%rsp), {{%xmm[0-9][0-9]*}}, {{%xmm[0-9][0-9]*}} {{.*#+}} 16-byte Folded Reload
+  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %2 = bitcast <4 x float> %a0 to <4 x i32>
+  %3 = bitcast <4 x float> %a1 to <4 x i32>
+  %4 = or <4 x i32> %2, %3
+  %5 = bitcast <4 x i32> %4 to <4 x float>
+  ; fadd forces execution domain
+  %6 = fadd <4 x float> %5, <float 0x0, float 0x0, float 0x0, float 0x0>
+  ret <4 x float> %6
+}
+
+define <8 x float> @stack_fold_orps_ymm(<8 x float> %a0, <8 x float> %a1) {
+  ;CHECK-LABEL: stack_fold_orps_ymm
+  ;CHECK:       vorps {{-?[0-9]*}}(%rsp), {{%ymm[0-9][0-9]*}}, {{%ymm[0-9][0-9]*}} {{.*#+}} 32-byte Folded Reload
+  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %2 = bitcast <8 x float> %a0 to <8 x i32>
+  %3 = bitcast <8 x float> %a1 to <8 x i32>
+  %4 = or <8 x i32> %2, %3
+  %5 = bitcast <8 x i32> %4 to <8 x float>
+  ; fadd forces execution domain
+  %6 = fadd <8 x float> %5, <float 0x0, float 0x0, float 0x0, float 0x0, float 0x0, float 0x0, float 0x0, float 0x0>
+  ret <8 x float> %6
+}
+
+define <2 x double> @stack_fold_subpd(<2 x double> %a0, <2 x double> %a1) {
+  ;CHECK-LABEL: stack_fold_subpd
+  ;CHECK:       vsubpd {{-?[0-9]*}}(%rsp), {{%xmm[0-9][0-9]*}}, {{%xmm[0-9][0-9]*}} {{.*#+}} 16-byte Folded Reload
+  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %2 = fsub <2 x double> %a0, %a1
   ret <2 x double> %2
 }
-declare <2 x double> @llvm.x86.sse2.sub.sd(<2 x double>, <2 x double>) nounwind readnone
 
-define float @stack_fold_subss(float %a0, float %a1) {
-  ;CHECK-LABEL: stack_fold_subss
-  ;CHECK:       vsubss {{-?[0-9]*}}(%rsp), {{%xmm[0-9][0-9]*}}, {{%xmm[0-9][0-9]*}} {{.*#+}} 4-byte Folded Reload
-  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{flags}"()
-  %2 = fsub float %a0, %a1
-  ret float %2
+define <4 x double> @stack_fold_subpd_ymm(<4 x double> %a0, <4 x double> %a1) {
+  ;CHECK-LABEL: stack_fold_subpd_ymm
+  ;CHECK:       vsubpd {{-?[0-9]*}}(%rsp), {{%ymm[0-9][0-9]*}}, {{%ymm[0-9][0-9]*}} {{.*#+}} 32-byte Folded Reload
+  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %2 = fsub <4 x double> %a0, %a1
+  ret <4 x double> %2
 }
 
-define <4 x float> @stack_fold_subss_int(<4 x float> %a0, <4 x float> %a1) {
-  ;CHECK-LABEL: stack_fold_subss_int
-  ;CHECK:       vsubss {{-?[0-9]*}}(%rsp), {{%xmm[0-9][0-9]*}}, {{%xmm[0-9][0-9]*}} {{.*#+}} 16-byte Folded Reload
-  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{flags}"()
-  %2 = call <4 x float> @llvm.x86.sse.sub.ss(<4 x float> %a0, <4 x float> %a1)
+define <4 x float> @stack_fold_subps(<4 x float> %a0, <4 x float> %a1) {
+  ;CHECK-LABEL: stack_fold_subps
+  ;CHECK:       vsubps {{-?[0-9]*}}(%rsp), {{%xmm[0-9][0-9]*}}, {{%xmm[0-9][0-9]*}} {{.*#+}} 16-byte Folded Reload
+  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %2 = fsub <4 x float> %a0, %a1
   ret <4 x float> %2
 }
-declare <4 x float> @llvm.x86.sse.sub.ss(<4 x float>, <4 x float>) nounwind readnone
+
+define <8 x float> @stack_fold_subps_ymm(<8 x float> %a0, <8 x float> %a1) {
+  ;CHECK-LABEL: stack_fold_subps_ymm
+  ;CHECK:       vsubps {{-?[0-9]*}}(%rsp), {{%ymm[0-9][0-9]*}}, {{%ymm[0-9][0-9]*}} {{.*#+}} 32-byte Folded Reload
+  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %2 = fsub <8 x float> %a0, %a1
+  ret <8 x float> %2
+}
+
+define <2 x double> @stack_fold_xorpd(<2 x double> %a0, <2 x double> %a1) #0 {
+  ;CHECK-LABEL: stack_fold_xorpd
+  ;CHECK:       vxorpd {{-?[0-9]*}}(%rsp), {{%xmm[0-9][0-9]*}}, {{%xmm[0-9][0-9]*}} {{.*#+}} 16-byte Folded Reload
+  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %2 = bitcast <2 x double> %a0 to <2 x i64>
+  %3 = bitcast <2 x double> %a1 to <2 x i64>
+  %4 = xor <2 x i64> %2, %3
+  %5 = bitcast <2 x i64> %4 to <2 x double>
+  ; fadd forces execution domain
+  %6 = fadd <2 x double> %5, <double 0x0, double 0x0>
+  ret <2 x double> %6
+}
+
+define <4 x double> @stack_fold_xorpd_ymm(<4 x double> %a0, <4 x double> %a1) #0 {
+  ;CHECK-LABEL: stack_fold_xorpd_ymm
+  ;CHECK:       vxorpd {{-?[0-9]*}}(%rsp), {{%ymm[0-9][0-9]*}}, {{%ymm[0-9][0-9]*}} {{.*#+}} 32-byte Folded Reload
+  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %2 = bitcast <4 x double> %a0 to <4 x i64>
+  %3 = bitcast <4 x double> %a1 to <4 x i64>
+  %4 = xor <4 x i64> %2, %3
+  %5 = bitcast <4 x i64> %4 to <4 x double>
+  ; fadd forces execution domain
+  %6 = fadd <4 x double> %5, <double 0x0, double 0x0, double 0x0, double 0x0>
+  ret <4 x double> %6
+}
+
+define <4 x float> @stack_fold_xorps(<4 x float> %a0, <4 x float> %a1) {
+  ;CHECK-LABEL: stack_fold_xorps
+  ;CHECK:       vxorps {{-?[0-9]*}}(%rsp), {{%xmm[0-9][0-9]*}}, {{%xmm[0-9][0-9]*}} {{.*#+}} 16-byte Folded Reload
+  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %2 = bitcast <4 x float> %a0 to <4 x i32>
+  %3 = bitcast <4 x float> %a1 to <4 x i32>
+  %4 = xor <4 x i32> %2, %3
+  %5 = bitcast <4 x i32> %4 to <4 x float>
+  ; fadd forces execution domain
+  %6 = fadd <4 x float> %5, <float 0x0, float 0x0, float 0x0, float 0x0>
+  ret <4 x float> %6
+}
+
+define <8 x float> @stack_fold_xorps_ymm(<8 x float> %a0, <8 x float> %a1) {
+  ;CHECK-LABEL: stack_fold_xorps_ymm
+  ;CHECK:       vxorps {{-?[0-9]*}}(%rsp), {{%ymm[0-9][0-9]*}}, {{%ymm[0-9][0-9]*}} {{.*#+}} 32-byte Folded Reload
+  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %2 = bitcast <8 x float> %a0 to <8 x i32>
+  %3 = bitcast <8 x float> %a1 to <8 x i32>
+  %4 = xor <8 x i32> %2, %3
+  %5 = bitcast <8 x i32> %4 to <8 x float>
+  ; fadd forces execution domain
+  %6 = fadd <8 x float> %5, <float 0x0, float 0x0, float 0x0, float 0x0, float 0x0, float 0x0, float 0x0, float 0x0>
+  ret <8 x float> %6
+}
+
+define <4 x float> @stack_fold_extractf32x4(<8 x float> %a0, <8 x float> %a1) {
+  ;CHECK-LABEL: stack_fold_extractf32x4
+  ;CHECK:       vextractf32x4 $1, {{%ymm[0-9][0-9]*}}, {{-?[0-9]*}}(%rsp) {{.*#+}} 16-byte Folded Spill
+  %1 = shufflevector <8 x float> %a0, <8 x float> %a1, <4 x i32> <i32 4, i32 5, i32 6, i32 7>
+  %2 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm1},~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  ret <4 x float> %1
+}
+
+define <2 x double> @stack_fold_extractf64x2(<4 x double> %a0, <4 x double> %a1) {
+  ;CHECK-LABEL: stack_fold_extractf64x2
+  ;CHECK:       vextractf64x2 $1, {{%ymm[0-9][0-9]*}}, {{-?[0-9]*}}(%rsp) {{.*#+}} 16-byte Folded Spill
+  %1 = shufflevector <4 x double> %a0, <4 x double> %a1, <2 x i32> <i32 2, i32 3>
+  %2 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm1},~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  ret <2 x double> %1
+}
+
+define <8 x float> @stack_fold_insertf32x4(<4 x float> %a0, <4 x float> %a1) {
+  ;CHECK-LABEL: stack_fold_insertf32x4
+  ;CHECK:       vinsertf32x4 $1, {{-?[0-9]*}}(%rsp), {{%ymm[0-9][0-9]*}}, {{%ymm[0-9][0-9]*}} {{.*#+}} 16-byte Folded Reload
+  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %2 = shufflevector <4 x float> %a0, <4 x float> %a1, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
+  ret <8 x float> %2
+}
+
+define <4 x double> @stack_fold_insertf64x2(<2 x double> %a0, <2 x double> %a1) {
+  ;CHECK-LABEL: stack_fold_insertf64x2
+  ;CHECK:       vinsertf64x2 $1, {{-?[0-9]*}}(%rsp), {{%ymm[0-9][0-9]*}}, {{%ymm[0-9][0-9]*}} {{.*#+}} 16-byte Folded Reload
+  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %2 = shufflevector <2 x double> %a0, <2 x double> %a1, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
+  ret <4 x double> %2
+}
+
+define <4 x float> @stack_fold_vpermt2ps(<4 x float> %x0, <4 x i32> %x1, <4 x float> %x2) {
+  ;CHECK-LABEL: stack_fold_vpermt2ps
+  ;CHECK:       vpermt2ps {{-?[0-9]*}}(%rsp), %xmm1, %xmm0 # 16-byte Folded Reload
+  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %res = call <4 x float> @llvm.x86.avx512.mask.vpermi2var.ps.128(<4 x float> %x0, <4 x i32> %x1, <4 x float> %x2, i8 -1)
+  ret <4 x float> %res
+}
+declare <4 x float> @llvm.x86.avx512.mask.vpermi2var.ps.128(<4 x float>, <4 x i32>, <4 x float>, i8)
+
+define <4 x float> @stack_fold_vpermi2ps(<4 x i32> %x0, <4 x float> %x1, <4 x float> %x2) {
+  ;CHECK-LABEL: stack_fold_vpermi2ps
+  ;CHECK:       vpermi2ps {{-?[0-9]*}}(%rsp), %xmm1, %xmm0 # 16-byte Folded Reload
+  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %res = call <4 x float> @llvm.x86.avx512.mask.vpermt2var.ps.128(<4 x i32> %x0, <4 x float> %x1, <4 x float> %x2, i8 -1)
+  ret <4 x float> %res
+}
+declare <4 x float> @llvm.x86.avx512.mask.vpermt2var.ps.128(<4 x i32>, <4 x float>, <4 x float>, i8)
+
+define <2 x double> @stack_fold_vpermt2pd(<2 x double> %x0, <2 x i64> %x1, <2 x double> %x2) {
+  ;CHECK-LABEL: stack_fold_vpermt2pd
+  ;CHECK:       vpermt2pd {{-?[0-9]*}}(%rsp), %xmm1, %xmm0 # 16-byte Folded Reload
+  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %res = call <2 x double> @llvm.x86.avx512.mask.vpermi2var.pd.128(<2 x double> %x0, <2 x i64> %x1, <2 x double> %x2, i8 -1)
+  ret <2 x double> %res
+}
+declare <2 x double> @llvm.x86.avx512.mask.vpermi2var.pd.128(<2 x double>, <2 x i64>, <2 x double>, i8)
+
+define <2 x double> @stack_fold_vpermi2pd(<2 x i64> %x0, <2 x double> %x1, <2 x double> %x2) {
+  ;CHECK-LABEL: stack_fold_vpermi2pd
+  ;CHECK:       vpermi2pd {{-?[0-9]*}}(%rsp), %xmm1, %xmm0 # 16-byte Folded Reload
+  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %res = call <2 x double> @llvm.x86.avx512.mask.vpermt2var.pd.128(<2 x i64> %x0, <2 x double> %x1, <2 x double> %x2, i8 -1)
+  ret <2 x double> %res
+}
+declare <2 x double> @llvm.x86.avx512.mask.vpermt2var.pd.128(<2 x i64>, <2 x double>, <2 x double>, i8)
+
+define <8 x float> @stack_fold_vpermt2ps_ymm(<8 x float> %x0, <8 x i32> %x1, <8 x float> %x2) {
+  ;CHECK-LABEL: stack_fold_vpermt2ps_ymm
+  ;CHECK:       vpermt2ps {{-?[0-9]*}}(%rsp), %ymm1, %ymm0 # 32-byte Folded Reload
+  %1 = tail call <4 x i64> asm sideeffect "nop", "=x,~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %res = call <8 x float> @llvm.x86.avx512.mask.vpermi2var.ps.256(<8 x float> %x0, <8 x i32> %x1, <8 x float> %x2, i8 -1)
+  ret <8 x float> %res
+}
+declare <8 x float> @llvm.x86.avx512.mask.vpermi2var.ps.256(<8 x float>, <8 x i32>, <8 x float>, i8)
+
+define <8 x float> @stack_fold_vpermi2ps_ymm(<8 x i32> %x0, <8 x float> %x1, <8 x float> %x2) {
+  ;CHECK-LABEL: stack_fold_vpermi2ps_ymm
+  ;CHECK:       vpermi2ps {{-?[0-9]*}}(%rsp), %ymm1, %ymm0 # 32-byte Folded Reload
+  %1 = tail call <4 x i64> asm sideeffect "nop", "=x,~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %res = call <8 x float> @llvm.x86.avx512.mask.vpermt2var.ps.256(<8 x i32> %x0, <8 x float> %x1, <8 x float> %x2, i8 -1)
+  ret <8 x float> %res
+}
+declare <8 x float> @llvm.x86.avx512.mask.vpermt2var.ps.256(<8 x i32>, <8 x float>, <8 x float>, i8)
+
+define <4 x double> @stack_fold_vpermt2pd_ymm(<4 x double> %x0, <4 x i64> %x1, <4 x double> %x2) {
+  ;CHECK-LABEL: stack_fold_vpermt2pd_ymm
+  ;CHECK:       vpermt2pd {{-?[0-9]*}}(%rsp), %ymm1, %ymm0 # 32-byte Folded Reload
+  %1 = tail call <4 x i64> asm sideeffect "nop", "=x,~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %res = call <4 x double> @llvm.x86.avx512.mask.vpermi2var.pd.256(<4 x double> %x0, <4 x i64> %x1, <4 x double> %x2, i8 -1)
+  ret <4 x double> %res
+}
+declare <4 x double> @llvm.x86.avx512.mask.vpermi2var.pd.256(<4 x double>, <4 x i64>, <4 x double>, i8)
+
+define <4 x double> @stack_fold_vpermi2pd_ymm(<4 x i64> %x0, <4 x double> %x1, <4 x double> %x2) {
+  ;CHECK-LABEL: stack_fold_vpermi2pd_ymm
+  ;CHECK:       vpermi2pd {{-?[0-9]*}}(%rsp), %ymm1, %ymm0 # 32-byte Folded Reload
+  %1 = tail call <4 x i64> asm sideeffect "nop", "=x,~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %res = call <4 x double> @llvm.x86.avx512.mask.vpermt2var.pd.256(<4 x i64> %x0, <4 x double> %x1, <4 x double> %x2, i8 -1)
+  ret <4 x double> %res
+}
+declare <4 x double> @llvm.x86.avx512.mask.vpermt2var.pd.256(<4 x i64>, <4 x double>, <4 x double>, i8)
+
+define <4 x double> @stack_fold_permpd(<4 x double> %a0) {
+  ;CHECK-LABEL: stack_fold_permpd
+  ;CHECK:   vpermpd $235, {{-?[0-9]*}}(%rsp), {{%ymm[0-9][0-9]*}} {{.*#+}} 32-byte Folded Reload
+  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm1},~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %2 = shufflevector <4 x double> %a0, <4 x double> undef, <4 x i32> <i32 3, i32 2, i32 2, i32 3>
+  ; fadd forces execution domain
+  %3 = fadd <4 x double> %2, <double 0x0, double 0x0, double 0x0, double 0x0>
+  ret <4 x double> %3
+}
+
+define <4 x double> @stack_fold_permpdvar(<4 x i64> %a0, <4 x double> %a1) {
+  ;CHECK-LABEL: stack_fold_permpdvar
+  ;CHECK:   vpermpd {{-?[0-9]*}}(%rsp), {{%ymm[0-9][0-9]*}}, {{%ymm[0-9][0-9]*}} {{.*#+}} 32-byte Folded Reload
+  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm1},~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %2 = call <4 x double> @llvm.x86.avx512.mask.permvar.df.256(<4 x double> %a1, <4 x i64> %a0, <4 x double> undef, i8 -1)
+  ; fadd forces execution domain
+  %3 = fadd <4 x double> %2, <double 0x0, double 0x0, double 0x0, double 0x0>
+  ret <4 x double> %3
+}
+declare <4 x double> @llvm.x86.avx512.mask.permvar.df.256(<4 x double>, <4 x i64>, <4 x double>, i8) nounwind readonly
+
+define <8 x float> @stack_fold_permps(<8 x i32> %a0, <8 x float> %a1) {
+  ;CHECK-LABEL: stack_fold_permps
+  ;CHECK:       vpermps {{-?[0-9]*}}(%rsp), {{%ymm[0-9][0-9]*}}, {{%ymm[0-9][0-9]*}} {{.*#+}} 32-byte Folded Reload
+  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %2 = call <8 x float> @llvm.x86.avx2.permps(<8 x float> %a1, <8 x i32> %a0)
+  ret <8 x float> %2
+}
+declare <8 x float> @llvm.x86.avx2.permps(<8 x float>, <8 x i32>) nounwind readonly
+
+define <2 x double> @stack_fold_permilpd(<2 x double> %a0) {
+  ;CHECK-LABEL: stack_fold_permilpd
+  ;CHECK:   vpermilpd $1, {{-?[0-9]*}}(%rsp), {{%xmm[0-9][0-9]*}} {{.*#+}} 16-byte Folded Reload
+  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm1},~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %2 = shufflevector <2 x double> %a0, <2 x double> undef, <2 x i32> <i32 1, i32 0>
+  ret <2 x double> %2
+}
+
+define <4 x double> @stack_fold_permilpd_ymm(<4 x double> %a0) {
+  ;CHECK-LABEL: stack_fold_permilpd_ymm
+  ;CHECK:   vpermilpd $5, {{-?[0-9]*}}(%rsp), {{%ymm[0-9][0-9]*}} {{.*#+}} 32-byte Folded Reload
+  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm1},~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %2 = shufflevector <4 x double> %a0, <4 x double> undef, <4 x i32> <i32 1, i32 0, i32 3, i32 2>
+  ret <4 x double> %2
+}
+
+define <2 x double> @stack_fold_permilpdvar(<2 x double> %a0, <2 x i64> %a1) {
+  ;CHECK-LABEL: stack_fold_permilpdvar
+  ;CHECK:       vpermilpd {{-?[0-9]*}}(%rsp), {{%xmm[0-9][0-9]*}}, {{%xmm[0-9][0-9]*}} {{.*#+}} 16-byte Folded Reload
+  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %2 = call <2 x double> @llvm.x86.avx.vpermilvar.pd(<2 x double> %a0, <2 x i64> %a1)
+  ret <2 x double> %2
+}
+declare <2 x double> @llvm.x86.avx.vpermilvar.pd(<2 x double>, <2 x i64>) nounwind readnone
+
+define <4 x double> @stack_fold_permilpdvar_ymm(<4 x double> %a0, <4 x i64> %a1) {
+  ;CHECK-LABEL: stack_fold_permilpdvar_ymm
+  ;CHECK:       vpermilpd {{-?[0-9]*}}(%rsp), {{%ymm[0-9][0-9]*}}, {{%ymm[0-9][0-9]*}} {{.*#+}} 32-byte Folded Reload
+  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %2 = call <4 x double> @llvm.x86.avx.vpermilvar.pd.256(<4 x double> %a0, <4 x i64> %a1)
+  ret <4 x double> %2
+}
+declare <4 x double> @llvm.x86.avx.vpermilvar.pd.256(<4 x double>, <4 x i64>) nounwind readnone
+
+define <4 x float> @stack_fold_permilps(<4 x float> %a0) {
+  ;CHECK-LABEL: stack_fold_permilps
+  ;CHECK:   vpermilps $27, {{-?[0-9]*}}(%rsp), {{%xmm[0-9][0-9]*}} {{.*#+}} 16-byte Folded Reload
+  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm1},~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %2 = shufflevector <4 x float> %a0, <4 x float> undef, <4 x i32> <i32 3, i32 2, i32 1, i32 0>
+  ret <4 x float> %2
+}
+
+define <8 x float> @stack_fold_permilps_ymm(<8 x float> %a0) {
+  ;CHECK-LABEL: stack_fold_permilps_ymm
+  ;CHECK:   vpermilps $27, {{-?[0-9]*}}(%rsp), {{%ymm[0-9][0-9]*}} {{.*#+}} 32-byte Folded Reload
+  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm1},~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %2 = shufflevector <8 x float> %a0, <8 x float> undef, <8 x i32> <i32 3, i32 2, i32 1, i32 0, i32 7, i32 6, i32 5, i32 4>
+  ret <8 x float> %2
+}
+
+define <4 x float> @stack_fold_permilpsvar(<4 x float> %a0, <4 x i32> %a1) {
+  ;CHECK-LABEL: stack_fold_permilpsvar
+  ;CHECK:       vpermilps {{-?[0-9]*}}(%rsp), {{%xmm[0-9][0-9]*}}, {{%xmm[0-9][0-9]*}} {{.*#+}} 16-byte Folded Reload
+  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %2 = call <4 x float> @llvm.x86.avx.vpermilvar.ps(<4 x float> %a0, <4 x i32> %a1)
+  ret <4 x float> %2
+}
+declare <4 x float> @llvm.x86.avx.vpermilvar.ps(<4 x float>, <4 x i32>) nounwind readnone
+
+define <8 x float> @stack_fold_permilpsvar_ymm(<8 x float> %a0, <8 x i32> %a1) {
+  ;CHECK-LABEL: stack_fold_permilpsvar_ymm
+  ;CHECK:       vpermilps {{-?[0-9]*}}(%rsp), {{%ymm[0-9][0-9]*}}, {{%ymm[0-9][0-9]*}} {{.*#+}} 32-byte Folded Reload
+  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %2 = call <8 x float> @llvm.x86.avx.vpermilvar.ps.256(<8 x float> %a0, <8 x i32> %a1)
+  ret <8 x float> %2
+}
+declare <8 x float> @llvm.x86.avx.vpermilvar.ps.256(<8 x float>, <8 x i32>) nounwind readnone
+
+define <8 x float> @stack_fold_permilpsvar_ymm_maskz(<8 x float> %a0, <8 x i32> %a1, i8 %mask) {
+  ;CHECK-LABEL: stack_fold_permilpsvar_ymm_maskz
+  ;CHECK:       vpermilps {{-?[0-9]*}}(%rsp), {{%ymm[0-9][0-9]*}}, {{%ymm[0-9][0-9]*}} {{{%k[0-7]}}} {z} {{.*#+}} 32-byte Folded Reload
+  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %2 = call <8 x float> @llvm.x86.avx.vpermilvar.ps.256(<8 x float> %a0, <8 x i32> %a1)
+  %3 = bitcast i8 %mask to <8 x i1>
+  %4 = select <8 x i1> %3, <8 x float> %2, <8 x float> zeroinitializer
+  ret <8 x float> %4
+}
+
+attributes #0 = { "unsafe-fp-math"="false" }
+attributes #1 = { "unsafe-fp-math"="true" }