[AVX-512] Add 512-bit logical operations to load folding tables. Add avx512f stack...

llvm-svn: 277961

diff --git a/llvm/lib/Target/X86/X86InstrInfo.cpp b/llvm/lib/Target/X86/X86InstrInfo.cpp
index cabeb5e..99ea37d 100644 (file)
--- a/llvm/lib/Target/X86/X86InstrInfo.cpp
+++ b/llvm/lib/Target/X86/X86InstrInfo.cpp
@@ -1720,6 +1720,22 @@ X86InstrInfo::X86InstrInfo(X86Subtarget &STI)
     { X86::VDIVSSZrr_Int,     X86::VDIVSSZrm_Int,       0 },
     { X86::VDIVSDZrr,         X86::VDIVSDZrm,           0 },
     { X86::VDIVSDZrr_Int,     X86::VDIVSDZrm_Int,       0 },
+    { X86::VANDPDZrr,         X86::VANDPDZrm,           0 },
+    { X86::VANDPSZrr,         X86::VANDPSZrm,           0 },
+    { X86::VANDNPDZrr,        X86::VANDNPDZrm,          0 },
+    { X86::VANDNPSZrr,        X86::VANDNPSZrm,          0 },
+    { X86::VORPDZrr,          X86::VORPDZrm,            0 },
+    { X86::VORPSZrr,          X86::VORPSZrm,            0 },
+    { X86::VXORPDZrr,         X86::VXORPDZrm,           0 },
+    { X86::VXORPSZrr,         X86::VXORPSZrm,           0 },
+    { X86::VPANDDZrr,         X86::VPANDDZrm,           0 },
+    { X86::VPANDQZrr,         X86::VPANDQZrm,           0 },
+    { X86::VPANDNDZrr,        X86::VPANDNDZrm,          0 },
+    { X86::VPANDNQZrr,        X86::VPANDNQZrm,          0 },
+    { X86::VPORDZrr,          X86::VPORDZrm,            0 },
+    { X86::VPORQZrr,          X86::VPORQZrm,            0 },
+    { X86::VPXORDZrr,         X86::VPXORDZrm,           0 },
+    { X86::VPXORQZrr,         X86::VPXORQZrm,           0 },
     { X86::VMAXPDZrr,         X86::VMAXPDZrm,           0 },
     { X86::VMAXCPDZrr,        X86::VMAXCPDZrm,          0 },
     { X86::VMAXPSZrr,         X86::VMAXPSZrm,           0 },

diff --git a/llvm/test/CodeGen/X86/stack-folding-fp-avx512.ll b/llvm/test/CodeGen/X86/stack-folding-fp-avx512.ll
new file mode 100644 (file)
index 0000000..a192b5b
--- /dev/null
+++ b/llvm/test/CodeGen/X86/stack-folding-fp-avx512.ll
@@ -0,0 +1,372 @@
+; RUN: llc -O3 -disable-peephole -mtriple=x86_64-unknown-unknown -mattr=+avx512f < %s | FileCheck %s
+
+target datalayout = "e-m:e-i64:64-f80:128-n8:16:32:64-S128"
+target triple = "x86_64-unknown-unknown"
+
+; Stack reload folding tests.
+;
+; 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 <8 x double> @stack_fold_addpd_zmm(<8 x double> %a0, <8 x double> %a1) {
+  ;CHECK-LABEL: stack_fold_addpd_zmm
+  ;CHECK:       vaddpd {{-?[0-9]*}}(%rsp), {{%zmm[0-9][0-9]*}}, {{%zmm[0-9][0-9]*}} {{.*#+}} 64-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 double> %a0, %a1
+  ret <8 x double> %2
+}
+
+define <16 x float> @stack_fold_addps_zmm(<16 x float> %a0, <16 x float> %a1) {
+  ;CHECK-LABEL: stack_fold_addps_zmm
+  ;CHECK:       vaddps {{-?[0-9]*}}(%rsp), {{%zmm[0-9][0-9]*}}, {{%zmm[0-9][0-9]*}} {{.*#+}} 64-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 <16 x float> %a0, %a1
+  ret <16 x float> %2
+}
+
+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},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %2 = fadd double %a0, %a1
+  ret double %2
+}
+
+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},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %2 = extractelement <2 x double> %a0, i32 0
+  %3 = extractelement <2 x double> %a1, i32 0
+  %4 = fadd double %2, %3
+  %5 = insertelement <2 x double> %a0, double %4, i32 0
+  ret <2 x double> %5
+}
+
+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},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %2 = fadd float %a0, %a1
+  ret float %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},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %2 = extractelement <4 x float> %a0, i32 0
+  %3 = extractelement <4 x float> %a1, i32 0
+  %4 = fadd float %2, %3
+  %5 = insertelement <4 x float> %a0, float %4, i32 0
+  ret <4 x float> %5
+}
+
+define <8 x double> @stack_fold_andnpd_zmm(<8 x double> %a0, <8 x double> %a1) {
+  ;CHECK-LABEL: stack_fold_andnpd_zmm
+  ;CHECK:       vpandnq {{-?[0-9]*}}(%rsp), {{%zmm[0-9][0-9]*}}, {{%zmm[0-9][0-9]*}} {{.*#+}} 64-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 double> %a0 to <8 x i64>
+  %3 = bitcast <8 x double> %a1 to <8 x i64>
+  %4 = xor <8 x i64> %2, <i64 -1, i64 -1, i64 -1, i64 -1, i64 -1, i64 -1, i64 -1, i64 -1>
+  %5 = and <8 x i64> %4, %3
+  %6 = bitcast <8 x i64> %5 to <8 x double>
+  ; fadd forces execution domain
+  %7 = fadd <8 x double> %6, <double 0x0, double 0x0, double 0x0, double 0x0, double 0x0, double 0x0, double 0x0, double 0x0>
+  ret <8 x double> %7
+}
+
+define <16 x float> @stack_fold_andnps_zmm(<16 x float> %a0, <16 x float> %a1) {
+  ;CHECK-LABEL: stack_fold_andnps_zmm
+  ;CHECK:       vpandnd {{-?[0-9]*}}(%rsp), {{%zmm[0-9][0-9]*}}, {{%zmm[0-9][0-9]*}} {{.*#+}} 64-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 <16 x float> %a0 to <16 x i32>
+  %3 = bitcast <16 x float> %a1 to <16 x i32>
+  %4 = xor <16 x i32> %2, <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>
+  %5 = and <16 x i32> %4, %3
+  %6 = bitcast <16 x i32> %5 to <16 x float>
+  ; fadd forces execution domain
+  %7 = fadd <16 x float> %6, <float 0x0, float 0x0, float 0x0, float 0x0, float 0x0, float 0x0, float 0x0, float 0x0, float 0x0, float 0x0, float 0x0, float 0x0, float 0x0, float 0x0, float 0x0, float 0x0>
+  ret <16 x float> %7
+}
+
+define <8 x double> @stack_fold_andpd_zmm(<8 x double> %a0, <8 x double> %a1) {
+  ;CHECK-LABEL: stack_fold_andpd_zmm
+  ;CHECK:       vpandq {{-?[0-9]*}}(%rsp), {{%zmm[0-9][0-9]*}}, {{%zmm[0-9][0-9]*}} {{.*#+}} 64-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 double> %a0 to <8 x i64>
+  %3 = bitcast <8 x double> %a1 to <8 x i64>
+  %4 = and <8 x i64> %2, %3
+  %5 = bitcast <8 x i64> %4 to <8 x double>
+  ; fadd forces execution domain
+  %6 = fadd <8 x double> %5, <double 0x0, double 0x0, double 0x0, double 0x0, double 0x0, double 0x0, double 0x0, double 0x0>
+  ret <8 x double> %6
+}
+
+define <16 x float> @stack_fold_andps_zmm(<16 x float> %a0, <16 x float> %a1) {
+  ;CHECK-LABEL: stack_fold_andps_zmm
+  ;CHECK:       vpandd {{-?[0-9]*}}(%rsp), {{%zmm[0-9][0-9]*}}, {{%zmm[0-9][0-9]*}} {{.*#+}} 64-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 <16 x float> %a0 to <16 x i32>
+  %3 = bitcast <16 x float> %a1 to <16 x i32>
+  %4 = and <16 x i32> %2, %3
+  %5 = bitcast <16 x i32> %4 to <16 x float>
+  ; fadd forces execution domain
+  %6 = fadd <16 x float> %5, <float 0x0, float 0x0, float 0x0, float 0x0, float 0x0, float 0x0, float 0x0, float 0x0, float 0x0, float 0x0, float 0x0, float 0x0, float 0x0, float 0x0, float 0x0, float 0x0>
+  ret <16 x float> %6
+}
+
+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},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %2 = extractelement <2 x double> %a0, i32 0
+  %3 = extractelement <2 x double> %a1, i32 0
+  %4 = fdiv double %2, %3
+  %5 = insertelement <2 x double> %a0, double %4, i32 0
+  ret <2 x double> %5
+}
+
+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},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %2 = fdiv float %a0, %a1
+  ret float %2
+}
+
+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},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %2 = extractelement <4 x float> %a0, i32 0
+  %3 = extractelement <4 x float> %a1, i32 0
+  %4 = fdiv float %2, %3
+  %5 = insertelement <4 x float> %a0, float %4, i32 0
+  ret <4 x float> %5
+}
+
+define <4 x float> @stack_fold_insertps(<4 x float> %a0, <4 x float> %a1) {
+  ;CHECK-LABEL: stack_fold_insertps
+  ;CHECK:       vinsertps $17, {{-?[0-9]*}}(%rsp), {{%xmm[0-9][0-9]*}}, {{%xmm[0-9][0-9]*}} {{.*#+}} 16-byte Folded Reload
+  ;CHECK-NEXT:                                                                              {{.*#+}} xmm0 = zero,mem[0],xmm0[2,3]
+  %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.sse41.insertps(<4 x float> %a0, <4 x float> %a1, i8 209)
+  ret <4 x float> %2
+}
+declare <4 x float> @llvm.x86.sse41.insertps(<4 x float>, <4 x float>, i8) nounwind readnone
+
+define <8 x double> @stack_fold_maxpd_zmm(<8 x double> %a0, <8 x double> %a1) #0 {
+  ;CHECK-LABEL: stack_fold_maxpd_zmm
+  ;CHECK:       vmaxpd {{-?[0-9]*}}(%rsp), {{%zmm[0-9][0-9]*}}, {{%zmm[0-9][0-9]*}} {{.*#+}} 64-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 double> @llvm.x86.avx512.mask.max.pd.512(<8 x double> %a0, <8 x double> %a1, <8 x double> zeroinitializer, i8 -1, i32 4)
+  ret <8 x double> %2
+}
+declare <8 x double> @llvm.x86.avx512.mask.max.pd.512(<8 x double>, <8 x double>, <8 x double>, i8, i32) nounwind readnone
+
+define <8 x double> @stack_fold_maxpd_zmm_commutable(<8 x double> %a0, <8 x double> %a1) #1 {
+  ;CHECK-LABEL: stack_fold_maxpd_zmm_commutable
+  ;CHECK:       vmaxpd {{-?[0-9]*}}(%rsp), {{%zmm[0-9][0-9]*}}, {{%zmm[0-9][0-9]*}} {{.*#+}} 64-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 double> @llvm.x86.avx512.mask.max.pd.512(<8 x double> %a0, <8 x double> %a1, <8 x double> zeroinitializer, i8 -1, i32 4)
+  ret <8 x double> %2
+}
+
+define <16 x float> @stack_fold_maxps_zmm(<16 x float> %a0, <16 x float> %a1) #0 {
+  ;CHECK-LABEL: stack_fold_maxps_zmm
+  ;CHECK:       vmaxps {{-?[0-9]*}}(%rsp), {{%zmm[0-9][0-9]*}}, {{%zmm[0-9][0-9]*}} {{.*#+}} 64-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 <16 x float> @llvm.x86.avx512.mask.max.ps.512(<16 x float> %a0, <16 x float> %a1, <16 x float> zeroinitializer, i16 -1, i32 4)
+  ret <16 x float> %2
+}
+declare <16 x float> @llvm.x86.avx512.mask.max.ps.512(<16 x float>, <16 x float>, <16 x float>, i16, i32) nounwind readnone
+
+define <16 x float> @stack_fold_maxps_zmm_commutable(<16 x float> %a0, <16 x float> %a1) #1 {
+  ;CHECK-LABEL: stack_fold_maxps_zmm_commutable
+  ;CHECK:       vmaxps {{-?[0-9]*}}(%rsp), {{%zmm[0-9][0-9]*}}, {{%zmm[0-9][0-9]*}} {{.*#+}} 64-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 <16 x float> @llvm.x86.avx512.mask.max.ps.512(<16 x float> %a0, <16 x float> %a1, <16 x float> zeroinitializer, i16 -1, i32 4)
+  ret <16 x float> %2
+}
+
+define <8 x double> @stack_fold_minpd_zmm(<8 x double> %a0, <8 x double> %a1) #0 {
+  ;CHECK-LABEL: stack_fold_minpd_zmm
+  ;CHECK:       vminpd {{-?[0-9]*}}(%rsp), {{%zmm[0-9][0-9]*}}, {{%zmm[0-9][0-9]*}} {{.*#+}} 64-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 double> @llvm.x86.avx512.mask.min.pd.512(<8 x double> %a0, <8 x double> %a1, <8 x double> zeroinitializer, i8 -1, i32 4)
+  ret <8 x double> %2
+}
+declare <8 x double> @llvm.x86.avx512.mask.min.pd.512(<8 x double>, <8 x double>, <8 x double>, i8, i32) nounwind readnone
+
+define <8 x double> @stack_fold_minpd_zmm_commutable(<8 x double> %a0, <8 x double> %a1) #1 {
+  ;CHECK-LABEL: stack_fold_minpd_zmm_commutable
+  ;CHECK:       vminpd {{-?[0-9]*}}(%rsp), {{%zmm[0-9][0-9]*}}, {{%zmm[0-9][0-9]*}} {{.*#+}} 64-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 double> @llvm.x86.avx512.mask.min.pd.512(<8 x double> %a0, <8 x double> %a1, <8 x double> zeroinitializer, i8 -1, i32 4)
+  ret <8 x double> %2
+}
+
+define <16 x float> @stack_fold_minps_zmm(<16 x float> %a0, <16 x float> %a1) #0 {
+  ;CHECK-LABEL: stack_fold_minps_zmm
+  ;CHECK:       vminps {{-?[0-9]*}}(%rsp), {{%zmm[0-9][0-9]*}}, {{%zmm[0-9][0-9]*}} {{.*#+}} 64-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 <16 x float> @llvm.x86.avx512.mask.min.ps.512(<16 x float> %a0, <16 x float> %a1, <16 x float> zeroinitializer, i16 -1, i32 4)
+  ret <16 x float> %2
+}
+declare <16 x float> @llvm.x86.avx512.mask.min.ps.512(<16 x float>, <16 x float>, <16 x float>, i16, i32) nounwind readnone
+
+define <16 x float> @stack_fold_minps_zmm_commutable(<16 x float> %a0, <16 x float> %a1) #1 {
+  ;CHECK-LABEL: stack_fold_minps_zmm_commutable
+  ;CHECK:       vminps {{-?[0-9]*}}(%rsp), {{%zmm[0-9][0-9]*}}, {{%zmm[0-9][0-9]*}} {{.*#+}} 64-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 <16 x float> @llvm.x86.avx512.mask.min.ps.512(<16 x float> %a0, <16 x float> %a1, <16 x float> zeroinitializer, i16 -1, i32 4)
+  ret <16 x float> %2
+}
+
+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},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %2 = fmul double %a0, %a1
+  ret double %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},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %2 = extractelement <2 x double> %a0, i32 0
+  %3 = extractelement <2 x double> %a1, i32 0
+  %4 = fmul double %2, %3
+  %5 = insertelement <2 x double> %a0, double %4, i32 0
+  ret <2 x double> %5
+}
+
+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},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %2 = fmul float %a0, %a1
+  ret float %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},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %2 = extractelement <4 x float> %a0, i32 0
+  %3 = extractelement <4 x float> %a1, i32 0
+  %4 = fmul float %2, %3
+  %5 = insertelement <4 x float> %a0, float %4, i32 0
+  ret <4 x float> %5
+}
+
+define <8 x double> @stack_fold_orpd_zmm(<8 x double> %a0, <8 x double> %a1) {
+  ;CHECK-LABEL: stack_fold_orpd_zmm
+  ;CHECK:       vporq {{-?[0-9]*}}(%rsp), {{%zmm[0-9][0-9]*}}, {{%zmm[0-9][0-9]*}} {{.*#+}} 64-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 double> %a0 to <8 x i64>
+  %3 = bitcast <8 x double> %a1 to <8 x i64>
+  %4 = or <8 x i64> %2, %3
+  %5 = bitcast <8 x i64> %4 to <8 x double>
+  ; fadd forces execution domain
+  %6 = fadd <8 x double> %5, <double 0x0, double 0x0, double 0x0, double 0x0, double 0x0, double 0x0, double 0x0, double 0x0>
+  ret <8 x double> %6
+}
+
+define <16 x float> @stack_fold_orps_zmm(<16 x float> %a0, <16 x float> %a1) {
+  ;CHECK-LABEL: stack_fold_orps_zmm
+  ;CHECK:       vpord {{-?[0-9]*}}(%rsp), {{%zmm[0-9][0-9]*}}, {{%zmm[0-9][0-9]*}} {{.*#+}} 64-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 <16 x float> %a0 to <16 x i32>
+  %3 = bitcast <16 x float> %a1 to <16 x i32>
+  %4 = or <16 x i32> %2, %3
+  %5 = bitcast <16 x i32> %4 to <16 x float>
+  ; fadd forces execution domain
+  %6 = fadd <16 x float> %5, <float 0x0, float 0x0, float 0x0, float 0x0, float 0x0, float 0x0, float 0x0, float 0x0, float 0x0, float 0x0, float 0x0, float 0x0, float 0x0, float 0x0, float 0x0, float 0x0>
+  ret <16 x float> %6
+}
+
+define <8 x double> @stack_fold_subpd_zmm(<8 x double> %a0, <8 x double> %a1) {
+  ;CHECK-LABEL: stack_fold_subpd_zmm
+  ;CHECK:       vsubpd {{-?[0-9]*}}(%rsp), {{%zmm[0-9][0-9]*}}, {{%zmm[0-9][0-9]*}} {{.*#+}} 64-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 double> %a0, %a1
+  ret <8 x double> %2
+}
+
+define <16 x float> @stack_fold_subps_zmm(<16 x float> %a0, <16 x float> %a1) {
+  ;CHECK-LABEL: stack_fold_subps_zmm
+  ;CHECK:       vsubps {{-?[0-9]*}}(%rsp), {{%zmm[0-9][0-9]*}}, {{%zmm[0-9][0-9]*}} {{.*#+}} 64-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 <16 x float> %a0, %a1
+  ret <16 x float> %2
+}
+
+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},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %2 = fsub double %a0, %a1
+  ret double %2
+}
+
+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},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %2 = extractelement <2 x double> %a0, i32 0
+  %3 = extractelement <2 x double> %a1, i32 0
+  %4 = fsub double %2, %3
+  %5 = insertelement <2 x double> %a0, double %4, i32 0
+  ret <2 x double> %5
+}
+
+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},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %2 = fsub float %a0, %a1
+  ret float %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},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %2 = extractelement <4 x float> %a0, i32 0
+  %3 = extractelement <4 x float> %a1, i32 0
+  %4 = fsub float %2, %3
+  %5 = insertelement <4 x float> %a0, float %4, i32 0
+  ret <4 x float> %5
+}
+
+define <8 x double> @stack_fold_xorpd_zmm(<8 x double> %a0, <8 x double> %a1) {
+  ;CHECK-LABEL: stack_fold_xorpd_zmm
+  ;CHECK:       vpxorq {{-?[0-9]*}}(%rsp), {{%zmm[0-9][0-9]*}}, {{%zmm[0-9][0-9]*}} {{.*#+}} 64-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 double> %a0 to <8 x i64>
+  %3 = bitcast <8 x double> %a1 to <8 x i64>
+  %4 = xor <8 x i64> %2, %3
+  %5 = bitcast <8 x i64> %4 to <8 x double>
+  ; fadd forces execution domain
+  %6 = fadd <8 x double> %5, <double 0x0, double 0x0, double 0x0, double 0x0, double 0x0, double 0x0, double 0x0, double 0x0>
+  ret <8 x double> %6
+}
+
+define <16 x float> @stack_fold_xorps_zmm(<16 x float> %a0, <16 x float> %a1) {
+  ;CHECK-LABEL: stack_fold_xorps_zmm
+  ;CHECK:       vpxord {{-?[0-9]*}}(%rsp), {{%zmm[0-9][0-9]*}}, {{%zmm[0-9][0-9]*}} {{.*#+}} 64-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 <16 x float> %a0 to <16 x i32>
+  %3 = bitcast <16 x float> %a1 to <16 x i32>
+  %4 = xor <16 x i32> %2, %3
+  %5 = bitcast <16 x i32> %4 to <16 x float>
+  ; fadd forces execution domain
+  %6 = fadd <16 x float> %5, <float 0x0, float 0x0, float 0x0, float 0x0, float 0x0, float 0x0, float 0x0, float 0x0, float 0x0, float 0x0, float 0x0, float 0x0, float 0x0, float 0x0, float 0x0, float 0x0>
+  ret <16 x float> %6
+}
+
+attributes #0 = { "unsafe-fp-math"="false" }
+attributes #1 = { "unsafe-fp-math"="true" }

diff --git a/llvm/test/CodeGen/X86/stack-folding-fp-avx512vl.ll b/llvm/test/CodeGen/X86/stack-folding-fp-avx512vl.ll
index 357c98a..8273879 100644 (file)
--- a/llvm/test/CodeGen/X86/stack-folding-fp-avx512vl.ll
+++ b/llvm/test/CodeGen/X86/stack-folding-fp-avx512vl.ll
@@ -40,44 +40,6 @@ define <8 x float> @stack_fold_addps_ymm(<8 x float> %a0, <8 x float> %a1) {
   ret <8 x float> %2
 }
 
-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},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
-  %2 = fadd double %a0, %a1
-  ret double %2
-}
-
-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},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
-  %2 = extractelement <2 x double> %a0, i32 0
-  %3 = extractelement <2 x double> %a1, i32 0
-  %4 = fadd double %2, %3
-  %5 = insertelement <2 x double> %a0, double %4, i32 0
-  ret <2 x double> %5
-}
-
-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},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
-  %2 = fadd float %a0, %a1
-  ret float %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},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
-  %2 = extractelement <4 x float> %a0, i32 0
-  %3 = extractelement <4 x float> %a1, i32 0
-  %4 = fadd float %2, %3
-  %5 = insertelement <4 x float> %a0, float %4, i32 0
-  ret <4 x float> %5
-}
-
 define <2 x double> @stack_fold_andnpd(<2 x double> %a0, <2 x double> %a1) {
   ;CHECK-LABEL: stack_fold_andnpd
   ;CHECK:       vpandnq {{-?[0-9]*}}(%rsp), {{%xmm[0-9][0-9]*}}, {{%xmm[0-9][0-9]*}} {{.*#+}} 16-byte Folded Reload
@@ -186,46 +148,6 @@ define <8 x float> @stack_fold_andps_ymm(<8 x float> %a0, <8 x float> %a1) {
   ret <8 x float> %6
 }
 
-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},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
-  %2 = extractelement <2 x double> %a0, i32 0
-  %3 = extractelement <2 x double> %a1, i32 0
-  %4 = fdiv double %2, %3
-  %5 = insertelement <2 x double> %a0, double %4, i32 0
-  ret <2 x double> %5
-}
-
-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},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
-  %2 = fdiv float %a0, %a1
-  ret float %2
-}
-
-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},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
-  %2 = extractelement <4 x float> %a0, i32 0
-  %3 = extractelement <4 x float> %a1, i32 0
-  %4 = fdiv float %2, %3
-  %5 = insertelement <4 x float> %a0, float %4, i32 0
-  ret <4 x float> %5
-}
-
-define <4 x float> @stack_fold_insertps(<4 x float> %a0, <4 x float> %a1) {
-  ;CHECK-LABEL: stack_fold_insertps
-  ;CHECK:       vinsertps $17, {{-?[0-9]*}}(%rsp), {{%xmm[0-9][0-9]*}}, {{%xmm[0-9][0-9]*}} {{.*#+}} 16-byte Folded Reload
-  ;CHECK-NEXT:                                                                              {{.*#+}} xmm0 = zero,mem[0],xmm0[2,3]
-  %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.sse41.insertps(<4 x float> %a0, <4 x float> %a1, i8 209)
-  ret <4 x float> %2
-}
-declare <4 x float> @llvm.x86.sse41.insertps(<4 x float>, <4 x float>, i8) nounwind readnone
-
 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
@@ -294,42 +216,38 @@ define <8 x float> @stack_fold_maxps_ymm_commutable(<8 x float> %a0, <8 x float>
   ret <8 x float> %2
 }
 
-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},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
-  %2 = fmul double %a0, %a1
-  ret double %2
+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},~{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 <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},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
-  %2 = extractelement <2 x double> %a0, i32 0
-  %3 = extractelement <2 x double> %a1, i32 0
-  %4 = fmul double %2, %3
-  %5 = insertelement <2 x double> %a0, double %4, i32 0
-  ret <2 x double> %5
+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},~{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 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},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
-  %2 = fmul float %a0, %a1
-  ret 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},~{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 <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},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
-  %2 = extractelement <4 x float> %a0, i32 0
-  %3 = extractelement <4 x float> %a1, i32 0
-  %4 = fmul float %2, %3
-  %5 = insertelement <4 x float> %a0, float %4, i32 0
-  ret <4 x float> %5
+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},~{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_orpd(<2 x double> %a0, <2 x double> %a1) {
@@ -416,44 +334,6 @@ define <8 x float> @stack_fold_subps_ymm(<8 x float> %a0, <8 x float> %a1) {
   ret <8 x float> %2
 }
 
-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},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
-  %2 = fsub double %a0, %a1
-  ret double %2
-}
-
-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},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
-  %2 = extractelement <2 x double> %a0, i32 0
-  %3 = extractelement <2 x double> %a1, i32 0
-  %4 = fsub double %2, %3
-  %5 = insertelement <2 x double> %a0, double %4, i32 0
-  ret <2 x double> %5
-}
-
-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},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
-  %2 = fsub float %a0, %a1
-  ret float %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},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
-  %2 = extractelement <4 x float> %a0, i32 0
-  %3 = extractelement <4 x float> %a1, i32 0
-  %4 = fsub float %2, %3
-  %5 = insertelement <4 x float> %a0, float %4, i32 0
-  ret <4 x float> %5
-}
-
 define <2 x double> @stack_fold_xorpd(<2 x double> %a0, <2 x double> %a1) {
   ;CHECK-LABEL: stack_fold_xorpd
   ;CHECK:       vpxorq {{-?[0-9]*}}(%rsp), {{%xmm[0-9][0-9]*}}, {{%xmm[0-9][0-9]*}} {{.*#+}} 16-byte Folded Reload