[llvm] 42fc785 - [X86] Print k-mask in FMA3 comments.
Craig Topper via llvm-commits
llvm-commits at lists.llvm.org
Sun Apr 12 13:17:18 PDT 2020
Author: Craig Topper
Date: 2020-04-12T13:16:53-07:00
New Revision: 42fc7852f53bfd3549442f6ae9087e7e08c2f0cb
URL: https://github.com/llvm/llvm-project/commit/42fc7852f53bfd3549442f6ae9087e7e08c2f0cb
DIFF: https://github.com/llvm/llvm-project/commit/42fc7852f53bfd3549442f6ae9087e7e08c2f0cb.diff
LOG: [X86] Print k-mask in FMA3 comments.
Added:
Modified:
llvm/lib/Target/X86/MCTargetDesc/X86InstComments.cpp
llvm/test/CodeGen/X86/avx512-fma-intrinsics-upgrade.ll
llvm/test/CodeGen/X86/avx512-fma-intrinsics.ll
llvm/test/CodeGen/X86/avx512-fma.ll
llvm/test/CodeGen/X86/avx512-intrinsics-canonical.ll
llvm/test/CodeGen/X86/avx512-intrinsics-fast-isel.ll
llvm/test/CodeGen/X86/avx512-intrinsics-upgrade.ll
llvm/test/CodeGen/X86/avx512-intrinsics.ll
llvm/test/CodeGen/X86/avx512-memfold.ll
llvm/test/CodeGen/X86/avx512-scalar_mask.ll
llvm/test/CodeGen/X86/avx512vl-intrinsics-canonical.ll
llvm/test/CodeGen/X86/avx512vl-intrinsics-fast-isel.ll
llvm/test/CodeGen/X86/avx512vl-intrinsics-upgrade.ll
llvm/test/CodeGen/X86/avx512vl-intrinsics.ll
llvm/test/CodeGen/X86/fma-fneg-combine.ll
llvm/test/CodeGen/X86/fma-scalar-combine.ll
Removed:
################################################################################
diff --git a/llvm/lib/Target/X86/MCTargetDesc/X86InstComments.cpp b/llvm/lib/Target/X86/MCTargetDesc/X86InstComments.cpp
index e14b20a653ed..6d5bb1d76d3c 100644
--- a/llvm/lib/Target/X86/MCTargetDesc/X86InstComments.cpp
+++ b/llvm/lib/Target/X86/MCTargetDesc/X86InstComments.cpp
@@ -281,7 +281,8 @@ static void printMasking(raw_ostream &OS, const MCInst *MI,
OS << " {z}";
}
-static bool printFMAComments(const MCInst *MI, raw_ostream &OS) {
+static bool printFMAComments(const MCInst *MI, raw_ostream &OS,
+ const MCInstrInfo &MCII) {
const char *Mul1Name = nullptr, *Mul2Name = nullptr, *AccName = nullptr;
unsigned NumOperands = MI->getNumOperands();
bool RegForm = false;
@@ -613,8 +614,9 @@ static bool printFMAComments(const MCInst *MI, raw_ostream &OS) {
if (!Mul2Name) Mul2Name = "mem";
if (!AccName) AccName = "mem";
- OS << DestName << " = ";
- // TODO: Print masking information?
+ OS << DestName;
+ printMasking(OS, MI, MCII);
+ OS << " = ";
if (Negate)
OS << '-';
@@ -641,7 +643,7 @@ bool llvm::EmitAnyX86InstComments(const MCInst *MI, raw_ostream &OS,
unsigned NumOperands = MI->getNumOperands();
bool RegForm = false;
- if (printFMAComments(MI, OS))
+ if (printFMAComments(MI, OS, MCII))
return true;
switch (MI->getOpcode()) {
diff --git a/llvm/test/CodeGen/X86/avx512-fma-intrinsics-upgrade.ll b/llvm/test/CodeGen/X86/avx512-fma-intrinsics-upgrade.ll
index 0ab60913c8f3..71ba6d48d0a7 100644
--- a/llvm/test/CodeGen/X86/avx512-fma-intrinsics-upgrade.ll
+++ b/llvm/test/CodeGen/X86/avx512-fma-intrinsics-upgrade.ll
@@ -18,7 +18,7 @@ define <16 x float> @test_mask_vfnmadd_ps(<16 x float> %a0, <16 x float> %a1, <1
; CHECK-LABEL: test_mask_vfnmadd_ps:
; CHECK: ## %bb.0:
; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vfnmadd132ps {{.*#+}} zmm0 = -(zmm0 * zmm1) + zmm2
+; CHECK-NEXT: vfnmadd132ps {{.*#+}} zmm0 {%k1} = -(zmm0 * zmm1) + zmm2
; CHECK-NEXT: retq
%res = call <16 x float> @llvm.x86.avx512.mask.vfnmadd.ps.512(<16 x float> %a0, <16 x float> %a1, <16 x float> %a2, i16 %mask, i32 4) nounwind
ret <16 x float> %res
@@ -38,7 +38,7 @@ define <8 x double> @test_mask_vfnmadd_pd(<8 x double> %a0, <8 x double> %a1, <8
; CHECK-LABEL: test_mask_vfnmadd_pd:
; CHECK: ## %bb.0:
; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vfnmadd132pd {{.*#+}} zmm0 = -(zmm0 * zmm1) + zmm2
+; CHECK-NEXT: vfnmadd132pd {{.*#+}} zmm0 {%k1} = -(zmm0 * zmm1) + zmm2
; CHECK-NEXT: retq
%res = call <8 x double> @llvm.x86.avx512.mask.vfnmadd.pd.512(<8 x double> %a0, <8 x double> %a1, <8 x double> %a2, i8 %mask, i32 4) nounwind
ret <8 x double> %res
@@ -58,7 +58,7 @@ define <16 x float> @test_mask_vfnmsub_ps(<16 x float> %a0, <16 x float> %a1, <1
; CHECK-LABEL: test_mask_vfnmsub_ps:
; CHECK: ## %bb.0:
; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vfnmsub132ps {{.*#+}} zmm0 = -(zmm0 * zmm1) - zmm2
+; CHECK-NEXT: vfnmsub132ps {{.*#+}} zmm0 {%k1} = -(zmm0 * zmm1) - zmm2
; CHECK-NEXT: retq
%res = call <16 x float> @llvm.x86.avx512.mask.vfnmsub.ps.512(<16 x float> %a0, <16 x float> %a1, <16 x float> %a2, i16 %mask, i32 4) nounwind
ret <16 x float> %res
@@ -78,7 +78,7 @@ define <8 x double> @test_mask_vfnmsub_pd(<8 x double> %a0, <8 x double> %a1, <8
; CHECK-LABEL: test_mask_vfnmsub_pd:
; CHECK: ## %bb.0:
; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vfnmsub132pd {{.*#+}} zmm0 = -(zmm0 * zmm1) - zmm2
+; CHECK-NEXT: vfnmsub132pd {{.*#+}} zmm0 {%k1} = -(zmm0 * zmm1) - zmm2
; CHECK-NEXT: retq
%res = call <8 x double> @llvm.x86.avx512.mask.vfnmsub.pd.512(<8 x double> %a0, <8 x double> %a1, <8 x double> %a2, i8 %mask, i32 4) nounwind
ret <8 x double> %res
@@ -97,7 +97,7 @@ define <16 x float> @test_mask_fmaddsub_ps(<16 x float> %a, <16 x float> %b, <16
; CHECK-LABEL: test_mask_fmaddsub_ps:
; CHECK: ## %bb.0:
; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vfmaddsub132ps {{.*#+}} zmm0 = (zmm0 * zmm1) +/- zmm2
+; CHECK-NEXT: vfmaddsub132ps {{.*#+}} zmm0 {%k1} = (zmm0 * zmm1) +/- zmm2
; CHECK-NEXT: retq
%res = call <16 x float> @llvm.x86.avx512.mask.vfmaddsub.ps.512(<16 x float> %a, <16 x float> %b, <16 x float> %c, i16 %mask, i32 4)
ret <16 x float> %res
@@ -119,7 +119,7 @@ define <8 x double> @test_mask_vfmaddsub_pd(<8 x double> %a0, <8 x double> %a1,
; CHECK-LABEL: test_mask_vfmaddsub_pd:
; CHECK: ## %bb.0:
; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vfmaddsub132pd {{.*#+}} zmm0 = (zmm0 * zmm1) +/- zmm2
+; CHECK-NEXT: vfmaddsub132pd {{.*#+}} zmm0 {%k1} = (zmm0 * zmm1) +/- zmm2
; CHECK-NEXT: retq
%res = call <8 x double> @llvm.x86.avx512.mask.vfmaddsub.pd.512(<8 x double> %a0, <8 x double> %a1, <8 x double> %a2, i8 %mask, i32 4) nounwind
ret <8 x double> %res
@@ -129,7 +129,7 @@ define <8 x double>@test_int_x86_avx512_mask_vfmaddsub_pd_512(<8 x double> %x0,
; CHECK-LABEL: test_int_x86_avx512_mask_vfmaddsub_pd_512:
; CHECK: ## %bb.0:
; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vfmaddsub132pd {{.*#+}} zmm0 = (zmm0 * zmm1) +/- zmm2
+; CHECK-NEXT: vfmaddsub132pd {{.*#+}} zmm0 {%k1} = (zmm0 * zmm1) +/- zmm2
; CHECK-NEXT: retq
%res = call <8 x double> @llvm.x86.avx512.mask.vfmaddsub.pd.512(<8 x double> %x0, <8 x double> %x1, <8 x double> %x2, i8 %x3, i32 4)
ret <8 x double> %res
@@ -141,7 +141,7 @@ define <8 x double>@test_int_x86_avx512_mask3_vfmaddsub_pd_512(<8 x double> %x0,
; CHECK-LABEL: test_int_x86_avx512_mask3_vfmaddsub_pd_512:
; CHECK: ## %bb.0:
; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vfmaddsub231pd {{.*#+}} zmm2 = (zmm0 * zmm1) +/- zmm2
+; CHECK-NEXT: vfmaddsub231pd {{.*#+}} zmm2 {%k1} = (zmm0 * zmm1) +/- zmm2
; CHECK-NEXT: vmovapd %zmm2, %zmm0
; CHECK-NEXT: retq
%res = call <8 x double> @llvm.x86.avx512.mask3.vfmaddsub.pd.512(<8 x double> %x0, <8 x double> %x1, <8 x double> %x2, i8 %x3, i32 4)
@@ -154,7 +154,7 @@ define <8 x double>@test_int_x86_avx512_maskz_vfmaddsub_pd_512(<8 x double> %x0,
; CHECK-LABEL: test_int_x86_avx512_maskz_vfmaddsub_pd_512:
; CHECK: ## %bb.0:
; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vfmaddsub213pd {{.*#+}} zmm0 = (zmm1 * zmm0) +/- zmm2
+; CHECK-NEXT: vfmaddsub213pd {{.*#+}} zmm0 {%k1} {z} = (zmm1 * zmm0) +/- zmm2
; CHECK-NEXT: retq
%res = call <8 x double> @llvm.x86.avx512.maskz.vfmaddsub.pd.512(<8 x double> %x0, <8 x double> %x1, <8 x double> %x2, i8 %x3, i32 4)
ret <8 x double> %res
@@ -164,7 +164,7 @@ define <16 x float>@test_int_x86_avx512_mask_vfmaddsub_ps_512(<16 x float> %x0,
; CHECK-LABEL: test_int_x86_avx512_mask_vfmaddsub_ps_512:
; CHECK: ## %bb.0:
; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vfmaddsub132ps {{.*#+}} zmm0 = (zmm0 * zmm1) +/- zmm2
+; CHECK-NEXT: vfmaddsub132ps {{.*#+}} zmm0 {%k1} = (zmm0 * zmm1) +/- zmm2
; CHECK-NEXT: retq
%res = call <16 x float> @llvm.x86.avx512.mask.vfmaddsub.ps.512(<16 x float> %x0, <16 x float> %x1, <16 x float> %x2, i16 %x3, i32 4)
ret <16 x float> %res
@@ -176,7 +176,7 @@ define <16 x float>@test_int_x86_avx512_mask3_vfmaddsub_ps_512(<16 x float> %x0,
; CHECK-LABEL: test_int_x86_avx512_mask3_vfmaddsub_ps_512:
; CHECK: ## %bb.0:
; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vfmaddsub231ps {{.*#+}} zmm2 = (zmm0 * zmm1) +/- zmm2
+; CHECK-NEXT: vfmaddsub231ps {{.*#+}} zmm2 {%k1} = (zmm0 * zmm1) +/- zmm2
; CHECK-NEXT: vmovaps %zmm2, %zmm0
; CHECK-NEXT: retq
%res = call <16 x float> @llvm.x86.avx512.mask3.vfmaddsub.ps.512(<16 x float> %x0, <16 x float> %x1, <16 x float> %x2, i16 %x3, i32 4)
@@ -189,7 +189,7 @@ define <16 x float>@test_int_x86_avx512_maskz_vfmaddsub_ps_512(<16 x float> %x0,
; CHECK-LABEL: test_int_x86_avx512_maskz_vfmaddsub_ps_512:
; CHECK: ## %bb.0:
; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vfmaddsub213ps {{.*#+}} zmm0 = (zmm1 * zmm0) +/- zmm2
+; CHECK-NEXT: vfmaddsub213ps {{.*#+}} zmm0 {%k1} {z} = (zmm1 * zmm0) +/- zmm2
; CHECK-NEXT: retq
%res = call <16 x float> @llvm.x86.avx512.maskz.vfmaddsub.ps.512(<16 x float> %x0, <16 x float> %x1, <16 x float> %x2, i16 %x3, i32 4)
ret <16 x float> %res
@@ -201,7 +201,7 @@ define <8 x double>@test_int_x86_avx512_mask3_vfmsubadd_pd_512(<8 x double> %x0,
; CHECK-LABEL: test_int_x86_avx512_mask3_vfmsubadd_pd_512:
; CHECK: ## %bb.0:
; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vfmsubadd231pd {{.*#+}} zmm2 = (zmm0 * zmm1) -/+ zmm2
+; CHECK-NEXT: vfmsubadd231pd {{.*#+}} zmm2 {%k1} = (zmm0 * zmm1) -/+ zmm2
; CHECK-NEXT: vmovapd %zmm2, %zmm0
; CHECK-NEXT: retq
%res = call <8 x double> @llvm.x86.avx512.mask3.vfmsubadd.pd.512(<8 x double> %x0, <8 x double> %x1, <8 x double> %x2, i8 %x3, i32 4)
@@ -214,7 +214,7 @@ define <16 x float>@test_int_x86_avx512_mask3_vfmsubadd_ps_512(<16 x float> %x0,
; CHECK-LABEL: test_int_x86_avx512_mask3_vfmsubadd_ps_512:
; CHECK: ## %bb.0:
; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vfmsubadd231ps {{.*#+}} zmm2 = (zmm0 * zmm1) -/+ zmm2
+; CHECK-NEXT: vfmsubadd231ps {{.*#+}} zmm2 {%k1} = (zmm0 * zmm1) -/+ zmm2
; CHECK-NEXT: vmovaps %zmm2, %zmm0
; CHECK-NEXT: retq
%res = call <16 x float> @llvm.x86.avx512.mask3.vfmsubadd.ps.512(<16 x float> %x0, <16 x float> %x1, <16 x float> %x2, i16 %x3, i32 4)
@@ -265,7 +265,7 @@ define <16 x float> @test_mask_round_vfmadd512_ps_rrb_current(<16 x float> %a0,
; CHECK-LABEL: test_mask_round_vfmadd512_ps_rrb_current:
; CHECK: ## %bb.0:
; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vfmadd132ps {{.*#+}} zmm0 = (zmm0 * zmm1) + zmm2
+; CHECK-NEXT: vfmadd132ps {{.*#+}} zmm0 {%k1} = (zmm0 * zmm1) + zmm2
; CHECK-NEXT: retq
%res = call <16 x float> @llvm.x86.avx512.mask.vfmadd.ps.512(<16 x float> %a0, <16 x float> %a1, <16 x float> %a2, i16 %mask, i32 4) nounwind
ret <16 x float> %res
@@ -322,7 +322,7 @@ define <8 x double>@test_int_x86_avx512_mask3_vfmsub_pd_512(<8 x double> %x0, <8
; CHECK-LABEL: test_int_x86_avx512_mask3_vfmsub_pd_512:
; CHECK: ## %bb.0:
; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vfmsub231pd {{.*#+}} zmm2 = (zmm0 * zmm1) - zmm2
+; CHECK-NEXT: vfmsub231pd {{.*#+}} zmm2 {%k1} = (zmm0 * zmm1) - zmm2
; CHECK-NEXT: vmovapd %zmm2, %zmm0
; CHECK-NEXT: retq
%res = call <8 x double> @llvm.x86.avx512.mask3.vfmsub.pd.512(<8 x double> %x0, <8 x double> %x1, <8 x double> %x2, i8 %x3, i32 4)
@@ -335,7 +335,7 @@ define <16 x float>@test_int_x86_avx512_mask3_vfmsub_ps_512(<16 x float> %x0, <1
; CHECK-LABEL: test_int_x86_avx512_mask3_vfmsub_ps_512:
; CHECK: ## %bb.0:
; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vfmsub231ps {{.*#+}} zmm2 = (zmm0 * zmm1) - zmm2
+; CHECK-NEXT: vfmsub231ps {{.*#+}} zmm2 {%k1} = (zmm0 * zmm1) - zmm2
; CHECK-NEXT: vmovaps %zmm2, %zmm0
; CHECK-NEXT: retq
%res = call <16 x float> @llvm.x86.avx512.mask3.vfmsub.ps.512(<16 x float> %x0, <16 x float> %x1, <16 x float> %x2, i16 %x3, i32 4)
@@ -386,7 +386,7 @@ define <8 x double> @test_mask_round_vfmadd512_pd_rrb_current(<8 x double> %a0,
; CHECK-LABEL: test_mask_round_vfmadd512_pd_rrb_current:
; CHECK: ## %bb.0:
; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vfmadd132pd {{.*#+}} zmm0 = (zmm0 * zmm1) + zmm2
+; CHECK-NEXT: vfmadd132pd {{.*#+}} zmm0 {%k1} = (zmm0 * zmm1) + zmm2
; CHECK-NEXT: retq
%res = call <8 x double> @llvm.x86.avx512.mask.vfmadd.pd.512(<8 x double> %a0, <8 x double> %a1, <8 x double> %a2, i8 %mask, i32 4) nounwind
ret <8 x double> %res
@@ -441,7 +441,7 @@ define <8 x double>@test_int_x86_avx512_mask_vfmadd_pd_512(<8 x double> %x0, <8
; CHECK-LABEL: test_int_x86_avx512_mask_vfmadd_pd_512:
; CHECK: ## %bb.0:
; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vfmadd132pd {{.*#+}} zmm0 = (zmm0 * zmm1) + zmm2
+; CHECK-NEXT: vfmadd132pd {{.*#+}} zmm0 {%k1} = (zmm0 * zmm1) + zmm2
; CHECK-NEXT: retq
%res = call <8 x double> @llvm.x86.avx512.mask.vfmadd.pd.512(<8 x double> %x0, <8 x double> %x1, <8 x double> %x2, i8 %x3, i32 4)
ret <8 x double> %res
@@ -453,7 +453,7 @@ define <8 x double>@test_int_x86_avx512_mask3_vfmadd_pd_512(<8 x double> %x0, <8
; CHECK-LABEL: test_int_x86_avx512_mask3_vfmadd_pd_512:
; CHECK: ## %bb.0:
; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vfmadd231pd {{.*#+}} zmm2 = (zmm0 * zmm1) + zmm2
+; CHECK-NEXT: vfmadd231pd {{.*#+}} zmm2 {%k1} = (zmm0 * zmm1) + zmm2
; CHECK-NEXT: vmovapd %zmm2, %zmm0
; CHECK-NEXT: retq
%res = call <8 x double> @llvm.x86.avx512.mask3.vfmadd.pd.512(<8 x double> %x0, <8 x double> %x1, <8 x double> %x2, i8 %x3, i32 4)
@@ -466,7 +466,7 @@ define <8 x double>@test_int_x86_avx512_maskz_vfmadd_pd_512(<8 x double> %x0, <8
; CHECK-LABEL: test_int_x86_avx512_maskz_vfmadd_pd_512:
; CHECK: ## %bb.0:
; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vfmadd213pd {{.*#+}} zmm0 = (zmm1 * zmm0) + zmm2
+; CHECK-NEXT: vfmadd213pd {{.*#+}} zmm0 {%k1} {z} = (zmm1 * zmm0) + zmm2
; CHECK-NEXT: retq
%res = call <8 x double> @llvm.x86.avx512.maskz.vfmadd.pd.512(<8 x double> %x0, <8 x double> %x1, <8 x double> %x2, i8 %x3, i32 4)
ret <8 x double> %res
@@ -476,7 +476,7 @@ define <16 x float>@test_int_x86_avx512_mask_vfmadd_ps_512(<16 x float> %x0, <16
; CHECK-LABEL: test_int_x86_avx512_mask_vfmadd_ps_512:
; CHECK: ## %bb.0:
; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vfmadd132ps {{.*#+}} zmm0 = (zmm0 * zmm1) + zmm2
+; CHECK-NEXT: vfmadd132ps {{.*#+}} zmm0 {%k1} = (zmm0 * zmm1) + zmm2
; CHECK-NEXT: retq
%res = call <16 x float> @llvm.x86.avx512.mask.vfmadd.ps.512(<16 x float> %x0, <16 x float> %x1, <16 x float> %x2, i16 %x3, i32 4)
ret <16 x float> %res
@@ -488,7 +488,7 @@ define <16 x float>@test_int_x86_avx512_mask3_vfmadd_ps_512(<16 x float> %x0, <1
; CHECK-LABEL: test_int_x86_avx512_mask3_vfmadd_ps_512:
; CHECK: ## %bb.0:
; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vfmadd231ps {{.*#+}} zmm2 = (zmm0 * zmm1) + zmm2
+; CHECK-NEXT: vfmadd231ps {{.*#+}} zmm2 {%k1} = (zmm0 * zmm1) + zmm2
; CHECK-NEXT: vmovaps %zmm2, %zmm0
; CHECK-NEXT: retq
%res = call <16 x float> @llvm.x86.avx512.mask3.vfmadd.ps.512(<16 x float> %x0, <16 x float> %x1, <16 x float> %x2, i16 %x3, i32 4)
@@ -501,7 +501,7 @@ define <16 x float>@test_int_x86_avx512_maskz_vfmadd_ps_512(<16 x float> %x0, <1
; CHECK-LABEL: test_int_x86_avx512_maskz_vfmadd_ps_512:
; CHECK: ## %bb.0:
; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vfmadd213ps {{.*#+}} zmm0 = (zmm1 * zmm0) + zmm2
+; CHECK-NEXT: vfmadd213ps {{.*#+}} zmm0 {%k1} {z} = (zmm1 * zmm0) + zmm2
; CHECK-NEXT: retq
%res = call <16 x float> @llvm.x86.avx512.maskz.vfmadd.ps.512(<16 x float> %x0, <16 x float> %x1, <16 x float> %x2, i16 %x3, i32 4)
ret <16 x float> %res
@@ -552,7 +552,7 @@ define <8 x double> @test_mask_round_vfnmsub512_pd_rrb_current(<8 x double> %a0,
; CHECK-LABEL: test_mask_round_vfnmsub512_pd_rrb_current:
; CHECK: ## %bb.0:
; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vfnmsub132pd {{.*#+}} zmm0 = -(zmm0 * zmm1) - zmm2
+; CHECK-NEXT: vfnmsub132pd {{.*#+}} zmm0 {%k1} = -(zmm0 * zmm1) - zmm2
; CHECK-NEXT: retq
%res = call <8 x double> @llvm.x86.avx512.mask.vfnmsub.pd.512(<8 x double> %a0, <8 x double> %a1, <8 x double> %a2, i8 %mask, i32 4) nounwind
ret <8 x double> %res
@@ -607,7 +607,7 @@ define <8 x double>@test_int_x86_avx512_mask_vfnmsub_pd_512(<8 x double> %x0, <8
; CHECK-LABEL: test_int_x86_avx512_mask_vfnmsub_pd_512:
; CHECK: ## %bb.0:
; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vfnmsub132pd {{.*#+}} zmm0 = -(zmm0 * zmm1) - zmm2
+; CHECK-NEXT: vfnmsub132pd {{.*#+}} zmm0 {%k1} = -(zmm0 * zmm1) - zmm2
; CHECK-NEXT: retq
%res = call <8 x double> @llvm.x86.avx512.mask.vfnmsub.pd.512(<8 x double> %x0, <8 x double> %x1, <8 x double> %x2, i8 %x3, i32 4)
ret <8 x double> %res
@@ -619,7 +619,7 @@ define <8 x double>@test_int_x86_avx512_mask3_vfnmsub_pd_512(<8 x double> %x0, <
; CHECK-LABEL: test_int_x86_avx512_mask3_vfnmsub_pd_512:
; CHECK: ## %bb.0:
; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vfnmsub231pd {{.*#+}} zmm2 = -(zmm0 * zmm1) - zmm2
+; CHECK-NEXT: vfnmsub231pd {{.*#+}} zmm2 {%k1} = -(zmm0 * zmm1) - zmm2
; CHECK-NEXT: vmovapd %zmm2, %zmm0
; CHECK-NEXT: retq
%res = call <8 x double> @llvm.x86.avx512.mask3.vfnmsub.pd.512(<8 x double> %x0, <8 x double> %x1, <8 x double> %x2, i8 %x3, i32 4)
@@ -630,7 +630,7 @@ define <16 x float>@test_int_x86_avx512_mask_vfnmsub_ps_512(<16 x float> %x0, <1
; CHECK-LABEL: test_int_x86_avx512_mask_vfnmsub_ps_512:
; CHECK: ## %bb.0:
; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vfnmsub132ps {{.*#+}} zmm0 = -(zmm0 * zmm1) - zmm2
+; CHECK-NEXT: vfnmsub132ps {{.*#+}} zmm0 {%k1} = -(zmm0 * zmm1) - zmm2
; CHECK-NEXT: retq
%res = call <16 x float> @llvm.x86.avx512.mask.vfnmsub.ps.512(<16 x float> %x0, <16 x float> %x1, <16 x float> %x2, i16 %x3, i32 4)
ret <16 x float> %res
@@ -642,7 +642,7 @@ define <16 x float>@test_int_x86_avx512_mask3_vfnmsub_ps_512(<16 x float> %x0, <
; CHECK-LABEL: test_int_x86_avx512_mask3_vfnmsub_ps_512:
; CHECK: ## %bb.0:
; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vfnmsub231ps {{.*#+}} zmm2 = -(zmm0 * zmm1) - zmm2
+; CHECK-NEXT: vfnmsub231ps {{.*#+}} zmm2 {%k1} = -(zmm0 * zmm1) - zmm2
; CHECK-NEXT: vmovaps %zmm2, %zmm0
; CHECK-NEXT: retq
%res = call <16 x float> @llvm.x86.avx512.mask3.vfnmsub.ps.512(<16 x float> %x0, <16 x float> %x1, <16 x float> %x2, i16 %x3, i32 4)
@@ -653,7 +653,7 @@ define <8 x double>@test_int_x86_avx512_mask_vfnmadd_pd_512(<8 x double> %x0, <8
; CHECK-LABEL: test_int_x86_avx512_mask_vfnmadd_pd_512:
; CHECK: ## %bb.0:
; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vfnmadd132pd {{.*#+}} zmm0 = -(zmm0 * zmm1) + zmm2
+; CHECK-NEXT: vfnmadd132pd {{.*#+}} zmm0 {%k1} = -(zmm0 * zmm1) + zmm2
; CHECK-NEXT: retq
%res = call <8 x double> @llvm.x86.avx512.mask.vfnmadd.pd.512(<8 x double> %x0, <8 x double> %x1, <8 x double> %x2, i8 %x3, i32 4)
ret <8 x double> %res
@@ -663,7 +663,7 @@ define <16 x float>@test_int_x86_avx512_mask_vfnmadd_ps_512(<16 x float> %x0, <1
; CHECK-LABEL: test_int_x86_avx512_mask_vfnmadd_ps_512:
; CHECK: ## %bb.0:
; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vfnmadd132ps {{.*#+}} zmm0 = -(zmm0 * zmm1) + zmm2
+; CHECK-NEXT: vfnmadd132ps {{.*#+}} zmm0 {%k1} = -(zmm0 * zmm1) + zmm2
; CHECK-NEXT: retq
%res = call <16 x float> @llvm.x86.avx512.mask.vfnmadd.ps.512(<16 x float> %x0, <16 x float> %x1, <16 x float> %x2, i16 %x3, i32 4)
ret <16 x float> %res
diff --git a/llvm/test/CodeGen/X86/avx512-fma-intrinsics.ll b/llvm/test/CodeGen/X86/avx512-fma-intrinsics.ll
index ff737b91264a..94a1895906f9 100644
--- a/llvm/test/CodeGen/X86/avx512-fma-intrinsics.ll
+++ b/llvm/test/CodeGen/X86/avx512-fma-intrinsics.ll
@@ -21,14 +21,14 @@ define <16 x float> @test_mask_vfnmadd_ps(<16 x float> %a0, <16 x float> %a1, <1
; X86: # %bb.0:
; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1 # encoding: [0xc5,0xf8,0x90,0x4c,0x24,0x04]
; X86-NEXT: vfnmadd132ps %zmm1, %zmm2, %zmm0 {%k1} # encoding: [0x62,0xf2,0x6d,0x49,0x9c,0xc1]
-; X86-NEXT: # zmm0 = -(zmm0 * zmm1) + zmm2
+; X86-NEXT: # zmm0 {%k1} = -(zmm0 * zmm1) + zmm2
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_mask_vfnmadd_ps:
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfnmadd132ps %zmm1, %zmm2, %zmm0 {%k1} # encoding: [0x62,0xf2,0x6d,0x49,0x9c,0xc1]
-; X64-NEXT: # zmm0 = -(zmm0 * zmm1) + zmm2
+; X64-NEXT: # zmm0 {%k1} = -(zmm0 * zmm1) + zmm2
; X64-NEXT: retq # encoding: [0xc3]
%1 = fsub <16 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %a1
%2 = call <16 x float> @llvm.fma.v16f32(<16 x float> %a0, <16 x float> %1, <16 x float> %a2)
@@ -54,14 +54,14 @@ define <8 x double> @test_mask_vfnmadd_pd(<8 x double> %a0, <8 x double> %a1, <8
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfnmadd132pd %zmm1, %zmm2, %zmm0 {%k1} # encoding: [0x62,0xf2,0xed,0x49,0x9c,0xc1]
-; X86-NEXT: # zmm0 = -(zmm0 * zmm1) + zmm2
+; X86-NEXT: # zmm0 {%k1} = -(zmm0 * zmm1) + zmm2
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_mask_vfnmadd_pd:
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfnmadd132pd %zmm1, %zmm2, %zmm0 {%k1} # encoding: [0x62,0xf2,0xed,0x49,0x9c,0xc1]
-; X64-NEXT: # zmm0 = -(zmm0 * zmm1) + zmm2
+; X64-NEXT: # zmm0 {%k1} = -(zmm0 * zmm1) + zmm2
; X64-NEXT: retq # encoding: [0xc3]
%1 = fsub <8 x double> <double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00>, %a1
%2 = call <8 x double> @llvm.fma.v8f64(<8 x double> %a0, <8 x double> %1, <8 x double> %a2)
@@ -87,14 +87,14 @@ define <16 x float> @test_mask_vfnmsub_ps(<16 x float> %a0, <16 x float> %a1, <1
; X86: # %bb.0:
; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1 # encoding: [0xc5,0xf8,0x90,0x4c,0x24,0x04]
; X86-NEXT: vfnmsub132ps %zmm1, %zmm2, %zmm0 {%k1} # encoding: [0x62,0xf2,0x6d,0x49,0x9e,0xc1]
-; X86-NEXT: # zmm0 = -(zmm0 * zmm1) - zmm2
+; X86-NEXT: # zmm0 {%k1} = -(zmm0 * zmm1) - zmm2
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_mask_vfnmsub_ps:
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfnmsub132ps %zmm1, %zmm2, %zmm0 {%k1} # encoding: [0x62,0xf2,0x6d,0x49,0x9e,0xc1]
-; X64-NEXT: # zmm0 = -(zmm0 * zmm1) - zmm2
+; X64-NEXT: # zmm0 {%k1} = -(zmm0 * zmm1) - zmm2
; X64-NEXT: retq # encoding: [0xc3]
%1 = fsub <16 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %a1
%2 = fsub <16 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %a2
@@ -122,14 +122,14 @@ define <8 x double> @test_mask_vfnmsub_pd(<8 x double> %a0, <8 x double> %a1, <8
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfnmsub132pd %zmm1, %zmm2, %zmm0 {%k1} # encoding: [0x62,0xf2,0xed,0x49,0x9e,0xc1]
-; X86-NEXT: # zmm0 = -(zmm0 * zmm1) - zmm2
+; X86-NEXT: # zmm0 {%k1} = -(zmm0 * zmm1) - zmm2
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_mask_vfnmsub_pd:
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfnmsub132pd %zmm1, %zmm2, %zmm0 {%k1} # encoding: [0x62,0xf2,0xed,0x49,0x9e,0xc1]
-; X64-NEXT: # zmm0 = -(zmm0 * zmm1) - zmm2
+; X64-NEXT: # zmm0 {%k1} = -(zmm0 * zmm1) - zmm2
; X64-NEXT: retq # encoding: [0xc3]
%1 = fsub <8 x double> <double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00>, %a1
%2 = fsub <8 x double> <double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00>, %a2
@@ -154,14 +154,14 @@ define <16 x float> @test_mask_fmaddsub_ps(<16 x float> %a, <16 x float> %b, <16
; X86: # %bb.0:
; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1 # encoding: [0xc5,0xf8,0x90,0x4c,0x24,0x04]
; X86-NEXT: vfmaddsub132ps %zmm1, %zmm2, %zmm0 {%k1} # encoding: [0x62,0xf2,0x6d,0x49,0x96,0xc1]
-; X86-NEXT: # zmm0 = (zmm0 * zmm1) +/- zmm2
+; X86-NEXT: # zmm0 {%k1} = (zmm0 * zmm1) +/- zmm2
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_mask_fmaddsub_ps:
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmaddsub132ps %zmm1, %zmm2, %zmm0 {%k1} # encoding: [0x62,0xf2,0x6d,0x49,0x96,0xc1]
-; X64-NEXT: # zmm0 = (zmm0 * zmm1) +/- zmm2
+; X64-NEXT: # zmm0 {%k1} = (zmm0 * zmm1) +/- zmm2
; X64-NEXT: retq # encoding: [0xc3]
%res = call <16 x float> @llvm.x86.avx512.vfmaddsub.ps.512(<16 x float> %a, <16 x float> %b, <16 x float> %c, i32 4)
%bc = bitcast i16 %mask to <16 x i1>
@@ -188,14 +188,14 @@ define <8 x double> @test_mask_vfmaddsub_pd(<8 x double> %a0, <8 x double> %a1,
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmaddsub132pd %zmm1, %zmm2, %zmm0 {%k1} # encoding: [0x62,0xf2,0xed,0x49,0x96,0xc1]
-; X86-NEXT: # zmm0 = (zmm0 * zmm1) +/- zmm2
+; X86-NEXT: # zmm0 {%k1} = (zmm0 * zmm1) +/- zmm2
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_mask_vfmaddsub_pd:
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmaddsub132pd %zmm1, %zmm2, %zmm0 {%k1} # encoding: [0x62,0xf2,0xed,0x49,0x96,0xc1]
-; X64-NEXT: # zmm0 = (zmm0 * zmm1) +/- zmm2
+; X64-NEXT: # zmm0 {%k1} = (zmm0 * zmm1) +/- zmm2
; X64-NEXT: retq # encoding: [0xc3]
%res = call <8 x double> @llvm.x86.avx512.vfmaddsub.pd.512(<8 x double> %a0, <8 x double> %a1, <8 x double> %a2, i32 4) #2
%bc = bitcast i8 %mask to <8 x i1>
@@ -209,14 +209,14 @@ define <8 x double>@test_int_x86_avx512_mask_vfmaddsub_pd_512(<8 x double> %x0,
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmaddsub132pd %zmm1, %zmm2, %zmm0 {%k1} # encoding: [0x62,0xf2,0xed,0x49,0x96,0xc1]
-; X86-NEXT: # zmm0 = (zmm0 * zmm1) +/- zmm2
+; X86-NEXT: # zmm0 {%k1} = (zmm0 * zmm1) +/- zmm2
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_vfmaddsub_pd_512:
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmaddsub132pd %zmm1, %zmm2, %zmm0 {%k1} # encoding: [0x62,0xf2,0xed,0x49,0x96,0xc1]
-; X64-NEXT: # zmm0 = (zmm0 * zmm1) +/- zmm2
+; X64-NEXT: # zmm0 {%k1} = (zmm0 * zmm1) +/- zmm2
; X64-NEXT: retq # encoding: [0xc3]
%res = call <8 x double> @llvm.x86.avx512.vfmaddsub.pd.512(<8 x double> %x0, <8 x double> %x1, <8 x double> %x2, i32 4)
%bc = bitcast i8 %x3 to <8 x i1>
@@ -230,7 +230,7 @@ define <8 x double>@test_int_x86_avx512_mask3_vfmaddsub_pd_512(<8 x double> %x0,
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmaddsub231pd %zmm1, %zmm0, %zmm2 {%k1} # encoding: [0x62,0xf2,0xfd,0x49,0xb6,0xd1]
-; X86-NEXT: # zmm2 = (zmm0 * zmm1) +/- zmm2
+; X86-NEXT: # zmm2 {%k1} = (zmm0 * zmm1) +/- zmm2
; X86-NEXT: vmovapd %zmm2, %zmm0 # encoding: [0x62,0xf1,0xfd,0x48,0x28,0xc2]
; X86-NEXT: retl # encoding: [0xc3]
;
@@ -238,7 +238,7 @@ define <8 x double>@test_int_x86_avx512_mask3_vfmaddsub_pd_512(<8 x double> %x0,
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmaddsub231pd %zmm1, %zmm0, %zmm2 {%k1} # encoding: [0x62,0xf2,0xfd,0x49,0xb6,0xd1]
-; X64-NEXT: # zmm2 = (zmm0 * zmm1) +/- zmm2
+; X64-NEXT: # zmm2 {%k1} = (zmm0 * zmm1) +/- zmm2
; X64-NEXT: vmovapd %zmm2, %zmm0 # encoding: [0x62,0xf1,0xfd,0x48,0x28,0xc2]
; X64-NEXT: retq # encoding: [0xc3]
%1 = call <8 x double> @llvm.fma.v8f64(<8 x double> %x0, <8 x double> %x1, <8 x double> %x2)
@@ -256,14 +256,14 @@ define <8 x double>@test_int_x86_avx512_maskz_vfmaddsub_pd_512(<8 x double> %x0,
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmaddsub213pd %zmm2, %zmm1, %zmm0 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0xc9,0xa6,0xc2]
-; X86-NEXT: # zmm0 = (zmm1 * zmm0) +/- zmm2
+; X86-NEXT: # zmm0 {%k1} {z} = (zmm1 * zmm0) +/- zmm2
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_maskz_vfmaddsub_pd_512:
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmaddsub213pd %zmm2, %zmm1, %zmm0 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0xc9,0xa6,0xc2]
-; X64-NEXT: # zmm0 = (zmm1 * zmm0) +/- zmm2
+; X64-NEXT: # zmm0 {%k1} {z} = (zmm1 * zmm0) +/- zmm2
; X64-NEXT: retq # encoding: [0xc3]
%1 = call <8 x double> @llvm.fma.v8f64(<8 x double> %x0, <8 x double> %x1, <8 x double> %x2)
%2 = fsub <8 x double> <double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00>, %x2
@@ -279,14 +279,14 @@ define <16 x float>@test_int_x86_avx512_mask_vfmaddsub_ps_512(<16 x float> %x0,
; X86: # %bb.0:
; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1 # encoding: [0xc5,0xf8,0x90,0x4c,0x24,0x04]
; X86-NEXT: vfmaddsub132ps %zmm1, %zmm2, %zmm0 {%k1} # encoding: [0x62,0xf2,0x6d,0x49,0x96,0xc1]
-; X86-NEXT: # zmm0 = (zmm0 * zmm1) +/- zmm2
+; X86-NEXT: # zmm0 {%k1} = (zmm0 * zmm1) +/- zmm2
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_vfmaddsub_ps_512:
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmaddsub132ps %zmm1, %zmm2, %zmm0 {%k1} # encoding: [0x62,0xf2,0x6d,0x49,0x96,0xc1]
-; X64-NEXT: # zmm0 = (zmm0 * zmm1) +/- zmm2
+; X64-NEXT: # zmm0 {%k1} = (zmm0 * zmm1) +/- zmm2
; X64-NEXT: retq # encoding: [0xc3]
%res = call <16 x float> @llvm.x86.avx512.vfmaddsub.ps.512(<16 x float> %x0, <16 x float> %x1, <16 x float> %x2, i32 4)
%bc = bitcast i16 %x3 to <16 x i1>
@@ -299,7 +299,7 @@ define <16 x float>@test_int_x86_avx512_mask3_vfmaddsub_ps_512(<16 x float> %x0,
; X86: # %bb.0:
; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1 # encoding: [0xc5,0xf8,0x90,0x4c,0x24,0x04]
; X86-NEXT: vfmaddsub231ps %zmm1, %zmm0, %zmm2 {%k1} # encoding: [0x62,0xf2,0x7d,0x49,0xb6,0xd1]
-; X86-NEXT: # zmm2 = (zmm0 * zmm1) +/- zmm2
+; X86-NEXT: # zmm2 {%k1} = (zmm0 * zmm1) +/- zmm2
; X86-NEXT: vmovaps %zmm2, %zmm0 # encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc2]
; X86-NEXT: retl # encoding: [0xc3]
;
@@ -307,7 +307,7 @@ define <16 x float>@test_int_x86_avx512_mask3_vfmaddsub_ps_512(<16 x float> %x0,
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmaddsub231ps %zmm1, %zmm0, %zmm2 {%k1} # encoding: [0x62,0xf2,0x7d,0x49,0xb6,0xd1]
-; X64-NEXT: # zmm2 = (zmm0 * zmm1) +/- zmm2
+; X64-NEXT: # zmm2 {%k1} = (zmm0 * zmm1) +/- zmm2
; X64-NEXT: vmovaps %zmm2, %zmm0 # encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc2]
; X64-NEXT: retq # encoding: [0xc3]
%1 = call <16 x float> @llvm.fma.v16f32(<16 x float> %x0, <16 x float> %x1, <16 x float> %x2)
@@ -324,14 +324,14 @@ define <16 x float>@test_int_x86_avx512_maskz_vfmaddsub_ps_512(<16 x float> %x0,
; X86: # %bb.0:
; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1 # encoding: [0xc5,0xf8,0x90,0x4c,0x24,0x04]
; X86-NEXT: vfmaddsub213ps %zmm2, %zmm1, %zmm0 {%k1} {z} # encoding: [0x62,0xf2,0x75,0xc9,0xa6,0xc2]
-; X86-NEXT: # zmm0 = (zmm1 * zmm0) +/- zmm2
+; X86-NEXT: # zmm0 {%k1} {z} = (zmm1 * zmm0) +/- zmm2
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_maskz_vfmaddsub_ps_512:
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmaddsub213ps %zmm2, %zmm1, %zmm0 {%k1} {z} # encoding: [0x62,0xf2,0x75,0xc9,0xa6,0xc2]
-; X64-NEXT: # zmm0 = (zmm1 * zmm0) +/- zmm2
+; X64-NEXT: # zmm0 {%k1} {z} = (zmm1 * zmm0) +/- zmm2
; X64-NEXT: retq # encoding: [0xc3]
%1 = call <16 x float> @llvm.fma.v16f32(<16 x float> %x0, <16 x float> %x1, <16 x float> %x2)
%2 = fsub <16 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %x2
@@ -348,7 +348,7 @@ define <8 x double>@test_int_x86_avx512_mask3_vfmsubadd_pd_512(<8 x double> %x0,
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmsubadd231pd %zmm1, %zmm0, %zmm2 {%k1} # encoding: [0x62,0xf2,0xfd,0x49,0xb7,0xd1]
-; X86-NEXT: # zmm2 = (zmm0 * zmm1) -/+ zmm2
+; X86-NEXT: # zmm2 {%k1} = (zmm0 * zmm1) -/+ zmm2
; X86-NEXT: vmovapd %zmm2, %zmm0 # encoding: [0x62,0xf1,0xfd,0x48,0x28,0xc2]
; X86-NEXT: retl # encoding: [0xc3]
;
@@ -356,7 +356,7 @@ define <8 x double>@test_int_x86_avx512_mask3_vfmsubadd_pd_512(<8 x double> %x0,
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmsubadd231pd %zmm1, %zmm0, %zmm2 {%k1} # encoding: [0x62,0xf2,0xfd,0x49,0xb7,0xd1]
-; X64-NEXT: # zmm2 = (zmm0 * zmm1) -/+ zmm2
+; X64-NEXT: # zmm2 {%k1} = (zmm0 * zmm1) -/+ zmm2
; X64-NEXT: vmovapd %zmm2, %zmm0 # encoding: [0x62,0xf1,0xfd,0x48,0x28,0xc2]
; X64-NEXT: retq # encoding: [0xc3]
%1 = call <8 x double> @llvm.fma.v8f64(<8 x double> %x0, <8 x double> %x1, <8 x double> %x2)
@@ -373,7 +373,7 @@ define <16 x float>@test_int_x86_avx512_mask3_vfmsubadd_ps_512(<16 x float> %x0,
; X86: # %bb.0:
; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1 # encoding: [0xc5,0xf8,0x90,0x4c,0x24,0x04]
; X86-NEXT: vfmsubadd231ps %zmm1, %zmm0, %zmm2 {%k1} # encoding: [0x62,0xf2,0x7d,0x49,0xb7,0xd1]
-; X86-NEXT: # zmm2 = (zmm0 * zmm1) -/+ zmm2
+; X86-NEXT: # zmm2 {%k1} = (zmm0 * zmm1) -/+ zmm2
; X86-NEXT: vmovaps %zmm2, %zmm0 # encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc2]
; X86-NEXT: retl # encoding: [0xc3]
;
@@ -381,7 +381,7 @@ define <16 x float>@test_int_x86_avx512_mask3_vfmsubadd_ps_512(<16 x float> %x0,
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmsubadd231ps %zmm1, %zmm0, %zmm2 {%k1} # encoding: [0x62,0xf2,0x7d,0x49,0xb7,0xd1]
-; X64-NEXT: # zmm2 = (zmm0 * zmm1) -/+ zmm2
+; X64-NEXT: # zmm2 {%k1} = (zmm0 * zmm1) -/+ zmm2
; X64-NEXT: vmovaps %zmm2, %zmm0 # encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc2]
; X64-NEXT: retq # encoding: [0xc3]
%1 = call <16 x float> @llvm.fma.v16f32(<16 x float> %x0, <16 x float> %x1, <16 x float> %x2)
@@ -470,14 +470,14 @@ define <16 x float> @test_mask_round_vfmadd512_ps_rrb_current(<16 x float> %a0,
; X86: # %bb.0:
; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1 # encoding: [0xc5,0xf8,0x90,0x4c,0x24,0x04]
; X86-NEXT: vfmadd132ps %zmm1, %zmm2, %zmm0 {%k1} # encoding: [0x62,0xf2,0x6d,0x49,0x98,0xc1]
-; X86-NEXT: # zmm0 = (zmm0 * zmm1) + zmm2
+; X86-NEXT: # zmm0 {%k1} = (zmm0 * zmm1) + zmm2
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_mask_round_vfmadd512_ps_rrb_current:
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmadd132ps %zmm1, %zmm2, %zmm0 {%k1} # encoding: [0x62,0xf2,0x6d,0x49,0x98,0xc1]
-; X64-NEXT: # zmm0 = (zmm0 * zmm1) + zmm2
+; X64-NEXT: # zmm0 {%k1} = (zmm0 * zmm1) + zmm2
; X64-NEXT: retq # encoding: [0xc3]
%res = call <16 x float> @llvm.x86.avx512.vfmadd.ps.512(<16 x float> %a0, <16 x float> %a1, <16 x float> %a2, i32 4) nounwind
%bc = bitcast i16 %mask to <16 x i1>
@@ -537,7 +537,7 @@ define <8 x double>@test_int_x86_avx512_mask3_vfmsub_pd_512(<8 x double> %x0, <8
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmsub231pd %zmm1, %zmm0, %zmm2 {%k1} # encoding: [0x62,0xf2,0xfd,0x49,0xba,0xd1]
-; X86-NEXT: # zmm2 = (zmm0 * zmm1) - zmm2
+; X86-NEXT: # zmm2 {%k1} = (zmm0 * zmm1) - zmm2
; X86-NEXT: vmovapd %zmm2, %zmm0 # encoding: [0x62,0xf1,0xfd,0x48,0x28,0xc2]
; X86-NEXT: retl # encoding: [0xc3]
;
@@ -545,7 +545,7 @@ define <8 x double>@test_int_x86_avx512_mask3_vfmsub_pd_512(<8 x double> %x0, <8
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmsub231pd %zmm1, %zmm0, %zmm2 {%k1} # encoding: [0x62,0xf2,0xfd,0x49,0xba,0xd1]
-; X64-NEXT: # zmm2 = (zmm0 * zmm1) - zmm2
+; X64-NEXT: # zmm2 {%k1} = (zmm0 * zmm1) - zmm2
; X64-NEXT: vmovapd %zmm2, %zmm0 # encoding: [0x62,0xf1,0xfd,0x48,0x28,0xc2]
; X64-NEXT: retq # encoding: [0xc3]
%1 = fsub <8 x double> <double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00>, %x2
@@ -560,7 +560,7 @@ define <16 x float>@test_int_x86_avx512_mask3_vfmsub_ps_512(<16 x float> %x0, <1
; X86: # %bb.0:
; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1 # encoding: [0xc5,0xf8,0x90,0x4c,0x24,0x04]
; X86-NEXT: vfmsub231ps %zmm1, %zmm0, %zmm2 {%k1} # encoding: [0x62,0xf2,0x7d,0x49,0xba,0xd1]
-; X86-NEXT: # zmm2 = (zmm0 * zmm1) - zmm2
+; X86-NEXT: # zmm2 {%k1} = (zmm0 * zmm1) - zmm2
; X86-NEXT: vmovaps %zmm2, %zmm0 # encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc2]
; X86-NEXT: retl # encoding: [0xc3]
;
@@ -568,7 +568,7 @@ define <16 x float>@test_int_x86_avx512_mask3_vfmsub_ps_512(<16 x float> %x0, <1
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmsub231ps %zmm1, %zmm0, %zmm2 {%k1} # encoding: [0x62,0xf2,0x7d,0x49,0xba,0xd1]
-; X64-NEXT: # zmm2 = (zmm0 * zmm1) - zmm2
+; X64-NEXT: # zmm2 {%k1} = (zmm0 * zmm1) - zmm2
; X64-NEXT: vmovaps %zmm2, %zmm0 # encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc2]
; X64-NEXT: retq # encoding: [0xc3]
%1 = fsub <16 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %x2
@@ -660,14 +660,14 @@ define <8 x double> @test_mask_round_vfmadd512_pd_rrb_current(<8 x double> %a0,
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmadd132pd %zmm1, %zmm2, %zmm0 {%k1} # encoding: [0x62,0xf2,0xed,0x49,0x98,0xc1]
-; X86-NEXT: # zmm0 = (zmm0 * zmm1) + zmm2
+; X86-NEXT: # zmm0 {%k1} = (zmm0 * zmm1) + zmm2
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_mask_round_vfmadd512_pd_rrb_current:
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmadd132pd %zmm1, %zmm2, %zmm0 {%k1} # encoding: [0x62,0xf2,0xed,0x49,0x98,0xc1]
-; X64-NEXT: # zmm0 = (zmm0 * zmm1) + zmm2
+; X64-NEXT: # zmm0 {%k1} = (zmm0 * zmm1) + zmm2
; X64-NEXT: retq # encoding: [0xc3]
%res = call <8 x double> @llvm.x86.avx512.vfmadd.pd.512(<8 x double> %a0, <8 x double> %a1, <8 x double> %a2, i32 4) nounwind
%bc = bitcast i8 %mask to <8 x i1>
@@ -727,14 +727,14 @@ define <8 x double>@test_int_x86_avx512_mask_vfmadd_pd_512(<8 x double> %x0, <8
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmadd132pd %zmm1, %zmm2, %zmm0 {%k1} # encoding: [0x62,0xf2,0xed,0x49,0x98,0xc1]
-; X86-NEXT: # zmm0 = (zmm0 * zmm1) + zmm2
+; X86-NEXT: # zmm0 {%k1} = (zmm0 * zmm1) + zmm2
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_vfmadd_pd_512:
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmadd132pd %zmm1, %zmm2, %zmm0 {%k1} # encoding: [0x62,0xf2,0xed,0x49,0x98,0xc1]
-; X64-NEXT: # zmm0 = (zmm0 * zmm1) + zmm2
+; X64-NEXT: # zmm0 {%k1} = (zmm0 * zmm1) + zmm2
; X64-NEXT: retq # encoding: [0xc3]
%res = call <8 x double> @llvm.x86.avx512.vfmadd.pd.512(<8 x double> %x0, <8 x double> %x1, <8 x double> %x2, i32 4)
%bc = bitcast i8 %x3 to <8 x i1>
@@ -748,7 +748,7 @@ define <8 x double>@test_int_x86_avx512_mask3_vfmadd_pd_512(<8 x double> %x0, <8
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmadd231pd %zmm1, %zmm0, %zmm2 {%k1} # encoding: [0x62,0xf2,0xfd,0x49,0xb8,0xd1]
-; X86-NEXT: # zmm2 = (zmm0 * zmm1) + zmm2
+; X86-NEXT: # zmm2 {%k1} = (zmm0 * zmm1) + zmm2
; X86-NEXT: vmovapd %zmm2, %zmm0 # encoding: [0x62,0xf1,0xfd,0x48,0x28,0xc2]
; X86-NEXT: retl # encoding: [0xc3]
;
@@ -756,7 +756,7 @@ define <8 x double>@test_int_x86_avx512_mask3_vfmadd_pd_512(<8 x double> %x0, <8
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmadd231pd %zmm1, %zmm0, %zmm2 {%k1} # encoding: [0x62,0xf2,0xfd,0x49,0xb8,0xd1]
-; X64-NEXT: # zmm2 = (zmm0 * zmm1) + zmm2
+; X64-NEXT: # zmm2 {%k1} = (zmm0 * zmm1) + zmm2
; X64-NEXT: vmovapd %zmm2, %zmm0 # encoding: [0x62,0xf1,0xfd,0x48,0x28,0xc2]
; X64-NEXT: retq # encoding: [0xc3]
%1 = call <8 x double> @llvm.fma.v8f64(<8 x double> %x0, <8 x double> %x1, <8 x double> %x2)
@@ -771,14 +771,14 @@ define <8 x double>@test_int_x86_avx512_maskz_vfmadd_pd_512(<8 x double> %x0, <8
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmadd213pd %zmm2, %zmm1, %zmm0 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0xc9,0xa8,0xc2]
-; X86-NEXT: # zmm0 = (zmm1 * zmm0) + zmm2
+; X86-NEXT: # zmm0 {%k1} {z} = (zmm1 * zmm0) + zmm2
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_maskz_vfmadd_pd_512:
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmadd213pd %zmm2, %zmm1, %zmm0 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0xc9,0xa8,0xc2]
-; X64-NEXT: # zmm0 = (zmm1 * zmm0) + zmm2
+; X64-NEXT: # zmm0 {%k1} {z} = (zmm1 * zmm0) + zmm2
; X64-NEXT: retq # encoding: [0xc3]
%1 = call <8 x double> @llvm.fma.v8f64(<8 x double> %x0, <8 x double> %x1, <8 x double> %x2)
%2 = bitcast i8 %x3 to <8 x i1>
@@ -791,14 +791,14 @@ define <16 x float>@test_int_x86_avx512_mask_vfmadd_ps_512(<16 x float> %x0, <16
; X86: # %bb.0:
; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1 # encoding: [0xc5,0xf8,0x90,0x4c,0x24,0x04]
; X86-NEXT: vfmadd132ps %zmm1, %zmm2, %zmm0 {%k1} # encoding: [0x62,0xf2,0x6d,0x49,0x98,0xc1]
-; X86-NEXT: # zmm0 = (zmm0 * zmm1) + zmm2
+; X86-NEXT: # zmm0 {%k1} = (zmm0 * zmm1) + zmm2
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_vfmadd_ps_512:
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmadd132ps %zmm1, %zmm2, %zmm0 {%k1} # encoding: [0x62,0xf2,0x6d,0x49,0x98,0xc1]
-; X64-NEXT: # zmm0 = (zmm0 * zmm1) + zmm2
+; X64-NEXT: # zmm0 {%k1} = (zmm0 * zmm1) + zmm2
; X64-NEXT: retq # encoding: [0xc3]
%res = call <16 x float> @llvm.x86.avx512.vfmadd.ps.512(<16 x float> %x0, <16 x float> %x1, <16 x float> %x2, i32 4)
%bc = bitcast i16 %x3 to <16 x i1>
@@ -811,7 +811,7 @@ define <16 x float>@test_int_x86_avx512_mask3_vfmadd_ps_512(<16 x float> %x0, <1
; X86: # %bb.0:
; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1 # encoding: [0xc5,0xf8,0x90,0x4c,0x24,0x04]
; X86-NEXT: vfmadd231ps %zmm1, %zmm0, %zmm2 {%k1} # encoding: [0x62,0xf2,0x7d,0x49,0xb8,0xd1]
-; X86-NEXT: # zmm2 = (zmm0 * zmm1) + zmm2
+; X86-NEXT: # zmm2 {%k1} = (zmm0 * zmm1) + zmm2
; X86-NEXT: vmovaps %zmm2, %zmm0 # encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc2]
; X86-NEXT: retl # encoding: [0xc3]
;
@@ -819,7 +819,7 @@ define <16 x float>@test_int_x86_avx512_mask3_vfmadd_ps_512(<16 x float> %x0, <1
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmadd231ps %zmm1, %zmm0, %zmm2 {%k1} # encoding: [0x62,0xf2,0x7d,0x49,0xb8,0xd1]
-; X64-NEXT: # zmm2 = (zmm0 * zmm1) + zmm2
+; X64-NEXT: # zmm2 {%k1} = (zmm0 * zmm1) + zmm2
; X64-NEXT: vmovaps %zmm2, %zmm0 # encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc2]
; X64-NEXT: retq # encoding: [0xc3]
%1 = call <16 x float> @llvm.fma.v16f32(<16 x float> %x0, <16 x float> %x1, <16 x float> %x2)
@@ -833,14 +833,14 @@ define <16 x float> @test_int_x86_avx512_maskz_vfmadd_ps_512(<16 x float> %x0, <
; X86: # %bb.0:
; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1 # encoding: [0xc5,0xf8,0x90,0x4c,0x24,0x04]
; X86-NEXT: vfmadd213ps %zmm2, %zmm1, %zmm0 {%k1} {z} # encoding: [0x62,0xf2,0x75,0xc9,0xa8,0xc2]
-; X86-NEXT: # zmm0 = (zmm1 * zmm0) + zmm2
+; X86-NEXT: # zmm0 {%k1} {z} = (zmm1 * zmm0) + zmm2
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_maskz_vfmadd_ps_512:
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmadd213ps %zmm2, %zmm1, %zmm0 {%k1} {z} # encoding: [0x62,0xf2,0x75,0xc9,0xa8,0xc2]
-; X64-NEXT: # zmm0 = (zmm1 * zmm0) + zmm2
+; X64-NEXT: # zmm0 {%k1} {z} = (zmm1 * zmm0) + zmm2
; X64-NEXT: retq # encoding: [0xc3]
%1 = call <16 x float> @llvm.fma.v16f32(<16 x float> %x0, <16 x float> %x1, <16 x float> %x2)
%2 = bitcast i16 %x3 to <16 x i1>
@@ -938,14 +938,14 @@ define <8 x double> @test_mask_round_vfnmsub512_pd_rrb_current(<8 x double> %a0,
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfnmsub132pd %zmm1, %zmm2, %zmm0 {%k1} # encoding: [0x62,0xf2,0xed,0x49,0x9e,0xc1]
-; X86-NEXT: # zmm0 = -(zmm0 * zmm1) - zmm2
+; X86-NEXT: # zmm0 {%k1} = -(zmm0 * zmm1) - zmm2
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_mask_round_vfnmsub512_pd_rrb_current:
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfnmsub132pd %zmm1, %zmm2, %zmm0 {%k1} # encoding: [0x62,0xf2,0xed,0x49,0x9e,0xc1]
-; X64-NEXT: # zmm0 = -(zmm0 * zmm1) - zmm2
+; X64-NEXT: # zmm0 {%k1} = -(zmm0 * zmm1) - zmm2
; X64-NEXT: retq # encoding: [0xc3]
%1 = fsub <8 x double> <double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00>, %a1
%2 = fsub <8 x double> <double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00>, %a2
@@ -1017,14 +1017,14 @@ define <8 x double>@test_int_x86_avx512_mask_vfnmsub_pd_512(<8 x double> %x0, <8
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfnmsub132pd %zmm1, %zmm2, %zmm0 {%k1} # encoding: [0x62,0xf2,0xed,0x49,0x9e,0xc1]
-; X86-NEXT: # zmm0 = -(zmm0 * zmm1) - zmm2
+; X86-NEXT: # zmm0 {%k1} = -(zmm0 * zmm1) - zmm2
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_vfnmsub_pd_512:
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfnmsub132pd %zmm1, %zmm2, %zmm0 {%k1} # encoding: [0x62,0xf2,0xed,0x49,0x9e,0xc1]
-; X64-NEXT: # zmm0 = -(zmm0 * zmm1) - zmm2
+; X64-NEXT: # zmm0 {%k1} = -(zmm0 * zmm1) - zmm2
; X64-NEXT: retq # encoding: [0xc3]
%1 = fsub <8 x double> <double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00>, %x1
%2 = fsub <8 x double> <double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00>, %x2
@@ -1040,7 +1040,7 @@ define <8 x double>@test_int_x86_avx512_mask3_vfnmsub_pd_512(<8 x double> %x0, <
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfnmsub231pd %zmm1, %zmm0, %zmm2 {%k1} # encoding: [0x62,0xf2,0xfd,0x49,0xbe,0xd1]
-; X86-NEXT: # zmm2 = -(zmm0 * zmm1) - zmm2
+; X86-NEXT: # zmm2 {%k1} = -(zmm0 * zmm1) - zmm2
; X86-NEXT: vmovapd %zmm2, %zmm0 # encoding: [0x62,0xf1,0xfd,0x48,0x28,0xc2]
; X86-NEXT: retl # encoding: [0xc3]
;
@@ -1048,7 +1048,7 @@ define <8 x double>@test_int_x86_avx512_mask3_vfnmsub_pd_512(<8 x double> %x0, <
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfnmsub231pd %zmm1, %zmm0, %zmm2 {%k1} # encoding: [0x62,0xf2,0xfd,0x49,0xbe,0xd1]
-; X64-NEXT: # zmm2 = -(zmm0 * zmm1) - zmm2
+; X64-NEXT: # zmm2 {%k1} = -(zmm0 * zmm1) - zmm2
; X64-NEXT: vmovapd %zmm2, %zmm0 # encoding: [0x62,0xf1,0xfd,0x48,0x28,0xc2]
; X64-NEXT: retq # encoding: [0xc3]
%1 = fsub <8 x double> <double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00>, %x0
@@ -1064,14 +1064,14 @@ define <16 x float>@test_int_x86_avx512_mask_vfnmsub_ps_512(<16 x float> %x0, <1
; X86: # %bb.0:
; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1 # encoding: [0xc5,0xf8,0x90,0x4c,0x24,0x04]
; X86-NEXT: vfnmsub132ps %zmm1, %zmm2, %zmm0 {%k1} # encoding: [0x62,0xf2,0x6d,0x49,0x9e,0xc1]
-; X86-NEXT: # zmm0 = -(zmm0 * zmm1) - zmm2
+; X86-NEXT: # zmm0 {%k1} = -(zmm0 * zmm1) - zmm2
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_vfnmsub_ps_512:
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfnmsub132ps %zmm1, %zmm2, %zmm0 {%k1} # encoding: [0x62,0xf2,0x6d,0x49,0x9e,0xc1]
-; X64-NEXT: # zmm0 = -(zmm0 * zmm1) - zmm2
+; X64-NEXT: # zmm0 {%k1} = -(zmm0 * zmm1) - zmm2
; X64-NEXT: retq # encoding: [0xc3]
%1 = fsub <16 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %x1
%2 = fsub <16 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %x2
@@ -1086,7 +1086,7 @@ define <16 x float>@test_int_x86_avx512_mask3_vfnmsub_ps_512(<16 x float> %x0, <
; X86: # %bb.0:
; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1 # encoding: [0xc5,0xf8,0x90,0x4c,0x24,0x04]
; X86-NEXT: vfnmsub231ps %zmm1, %zmm0, %zmm2 {%k1} # encoding: [0x62,0xf2,0x7d,0x49,0xbe,0xd1]
-; X86-NEXT: # zmm2 = -(zmm0 * zmm1) - zmm2
+; X86-NEXT: # zmm2 {%k1} = -(zmm0 * zmm1) - zmm2
; X86-NEXT: vmovaps %zmm2, %zmm0 # encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc2]
; X86-NEXT: retl # encoding: [0xc3]
;
@@ -1094,7 +1094,7 @@ define <16 x float>@test_int_x86_avx512_mask3_vfnmsub_ps_512(<16 x float> %x0, <
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfnmsub231ps %zmm1, %zmm0, %zmm2 {%k1} # encoding: [0x62,0xf2,0x7d,0x49,0xbe,0xd1]
-; X64-NEXT: # zmm2 = -(zmm0 * zmm1) - zmm2
+; X64-NEXT: # zmm2 {%k1} = -(zmm0 * zmm1) - zmm2
; X64-NEXT: vmovaps %zmm2, %zmm0 # encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc2]
; X64-NEXT: retq # encoding: [0xc3]
%1 = fsub <16 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %x0
@@ -1111,14 +1111,14 @@ define <8 x double>@test_int_x86_avx512_mask_vfnmadd_pd_512(<8 x double> %x0, <8
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfnmadd132pd %zmm1, %zmm2, %zmm0 {%k1} # encoding: [0x62,0xf2,0xed,0x49,0x9c,0xc1]
-; X86-NEXT: # zmm0 = -(zmm0 * zmm1) + zmm2
+; X86-NEXT: # zmm0 {%k1} = -(zmm0 * zmm1) + zmm2
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_vfnmadd_pd_512:
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfnmadd132pd %zmm1, %zmm2, %zmm0 {%k1} # encoding: [0x62,0xf2,0xed,0x49,0x9c,0xc1]
-; X64-NEXT: # zmm0 = -(zmm0 * zmm1) + zmm2
+; X64-NEXT: # zmm0 {%k1} = -(zmm0 * zmm1) + zmm2
; X64-NEXT: retq # encoding: [0xc3]
%1 = fsub <8 x double> <double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00>, %x1
%2 = call <8 x double> @llvm.fma.v8f64(<8 x double> %x0, <8 x double> %1, <8 x double> %x2)
@@ -1132,14 +1132,14 @@ define <16 x float>@test_int_x86_avx512_mask_vfnmadd_ps_512(<16 x float> %x0, <1
; X86: # %bb.0:
; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1 # encoding: [0xc5,0xf8,0x90,0x4c,0x24,0x04]
; X86-NEXT: vfnmadd132ps %zmm1, %zmm2, %zmm0 {%k1} # encoding: [0x62,0xf2,0x6d,0x49,0x9c,0xc1]
-; X86-NEXT: # zmm0 = -(zmm0 * zmm1) + zmm2
+; X86-NEXT: # zmm0 {%k1} = -(zmm0 * zmm1) + zmm2
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_vfnmadd_ps_512:
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfnmadd132ps %zmm1, %zmm2, %zmm0 {%k1} # encoding: [0x62,0xf2,0x6d,0x49,0x9c,0xc1]
-; X64-NEXT: # zmm0 = -(zmm0 * zmm1) + zmm2
+; X64-NEXT: # zmm0 {%k1} = -(zmm0 * zmm1) + zmm2
; X64-NEXT: retq # encoding: [0xc3]
%1 = fsub <16 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %x1
%2 = call <16 x float> @llvm.fma.v16f32(<16 x float> %x0, <16 x float> %1, <16 x float> %x2)
diff --git a/llvm/test/CodeGen/X86/avx512-fma.ll b/llvm/test/CodeGen/X86/avx512-fma.ll
index 07eeeb4c662a..5fdb56e3ff7c 100644
--- a/llvm/test/CodeGen/X86/avx512-fma.ll
+++ b/llvm/test/CodeGen/X86/avx512-fma.ll
@@ -125,14 +125,14 @@ define <16 x float> @test_x86_fmadd132_ps(<16 x float> %a0, <16 x float> %a1, <1
; KNL-NEXT: vpmovsxbd %xmm2, %zmm2
; KNL-NEXT: vpslld $31, %zmm2, %zmm2
; KNL-NEXT: vptestmd %zmm2, %zmm2, %k1
-; KNL-NEXT: vfmadd132ps {{.*#+}} zmm0 = (zmm0 * mem) + zmm1
+; KNL-NEXT: vfmadd132ps {{.*#+}} zmm0 {%k1} = (zmm0 * mem) + zmm1
; KNL-NEXT: retq
;
; SKX-LABEL: test_x86_fmadd132_ps:
; SKX: ## %bb.0:
; SKX-NEXT: vpsllw $7, %xmm2, %xmm2
; SKX-NEXT: vpmovb2m %xmm2, %k1
-; SKX-NEXT: vfmadd132ps {{.*#+}} zmm0 = (zmm0 * mem) + zmm1
+; SKX-NEXT: vfmadd132ps {{.*#+}} zmm0 {%k1} = (zmm0 * mem) + zmm1
; SKX-NEXT: retq
%a2 = load <16 x float>,<16 x float> *%a2_ptrt,align 1
%x = fmul <16 x float> %a0, %a2
@@ -148,7 +148,7 @@ define <16 x float> @test_x86_fmadd231_ps(<16 x float> %a0, <16 x float> %a1, <1
; KNL-NEXT: vpmovsxbd %xmm2, %zmm2
; KNL-NEXT: vpslld $31, %zmm2, %zmm2
; KNL-NEXT: vptestmd %zmm2, %zmm2, %k1
-; KNL-NEXT: vfmadd231ps {{.*#+}} zmm1 = (zmm0 * mem) + zmm1
+; KNL-NEXT: vfmadd231ps {{.*#+}} zmm1 {%k1} = (zmm0 * mem) + zmm1
; KNL-NEXT: vmovaps %zmm1, %zmm0
; KNL-NEXT: retq
;
@@ -156,7 +156,7 @@ define <16 x float> @test_x86_fmadd231_ps(<16 x float> %a0, <16 x float> %a1, <1
; SKX: ## %bb.0:
; SKX-NEXT: vpsllw $7, %xmm2, %xmm2
; SKX-NEXT: vpmovb2m %xmm2, %k1
-; SKX-NEXT: vfmadd231ps {{.*#+}} zmm1 = (zmm0 * mem) + zmm1
+; SKX-NEXT: vfmadd231ps {{.*#+}} zmm1 {%k1} = (zmm0 * mem) + zmm1
; SKX-NEXT: vmovaps %zmm1, %zmm0
; SKX-NEXT: retq
%a2 = load <16 x float>,<16 x float> *%a2_ptrt,align 1
@@ -173,7 +173,7 @@ define <16 x float> @test_x86_fmadd213_ps(<16 x float> %a0, <16 x float> %a1, <1
; KNL-NEXT: vpmovsxbd %xmm2, %zmm2
; KNL-NEXT: vpslld $31, %zmm2, %zmm2
; KNL-NEXT: vptestmd %zmm2, %zmm2, %k1
-; KNL-NEXT: vfmadd213ps {{.*#+}} zmm1 = (zmm0 * zmm1) + mem
+; KNL-NEXT: vfmadd213ps {{.*#+}} zmm1 {%k1} = (zmm0 * zmm1) + mem
; KNL-NEXT: vmovaps %zmm1, %zmm0
; KNL-NEXT: retq
;
@@ -181,7 +181,7 @@ define <16 x float> @test_x86_fmadd213_ps(<16 x float> %a0, <16 x float> %a1, <1
; SKX: ## %bb.0:
; SKX-NEXT: vpsllw $7, %xmm2, %xmm2
; SKX-NEXT: vpmovb2m %xmm2, %k1
-; SKX-NEXT: vfmadd213ps {{.*#+}} zmm1 = (zmm0 * zmm1) + mem
+; SKX-NEXT: vfmadd213ps {{.*#+}} zmm1 {%k1} = (zmm0 * zmm1) + mem
; SKX-NEXT: vmovaps %zmm1, %zmm0
; SKX-NEXT: retq
%a2 = load <16 x float>,<16 x float> *%a2_ptrt,align 1
diff --git a/llvm/test/CodeGen/X86/avx512-intrinsics-canonical.ll b/llvm/test/CodeGen/X86/avx512-intrinsics-canonical.ll
index b020fdd01eab..3e9225315228 100644
--- a/llvm/test/CodeGen/X86/avx512-intrinsics-canonical.ll
+++ b/llvm/test/CodeGen/X86/avx512-intrinsics-canonical.ll
@@ -237,14 +237,14 @@ define <8 x double> @test_mm512_mask_fmadd_pd(<8 x double> %__A, i8 zeroext %__U
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax ## encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmadd132pd %zmm1, %zmm2, %zmm0 {%k1} ## encoding: [0x62,0xf2,0xed,0x49,0x98,0xc1]
-; X86-NEXT: ## zmm0 = (zmm0 * zmm1) + zmm2
+; X86-NEXT: ## zmm0 {%k1} = (zmm0 * zmm1) + zmm2
; X86-NEXT: retl ## encoding: [0xc3]
;
; X64-LABEL: test_mm512_mask_fmadd_pd:
; X64: ## %bb.0: ## %entry
; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmadd132pd %zmm1, %zmm2, %zmm0 {%k1} ## encoding: [0x62,0xf2,0xed,0x49,0x98,0xc1]
-; X64-NEXT: ## zmm0 = (zmm0 * zmm1) + zmm2
+; X64-NEXT: ## zmm0 {%k1} = (zmm0 * zmm1) + zmm2
; X64-NEXT: retq ## encoding: [0xc3]
entry:
%0 = tail call <8 x double> @llvm.fma.v8f64(<8 x double> %__A, <8 x double> %__B, <8 x double> %__C) #10
@@ -259,7 +259,7 @@ define <8 x double> @test_mm512_mask3_fmadd_pd(<8 x double> %__A, <8 x double> %
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax ## encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmadd231pd %zmm1, %zmm0, %zmm2 {%k1} ## encoding: [0x62,0xf2,0xfd,0x49,0xb8,0xd1]
-; X86-NEXT: ## zmm2 = (zmm0 * zmm1) + zmm2
+; X86-NEXT: ## zmm2 {%k1} = (zmm0 * zmm1) + zmm2
; X86-NEXT: vmovapd %zmm2, %zmm0 ## encoding: [0x62,0xf1,0xfd,0x48,0x28,0xc2]
; X86-NEXT: retl ## encoding: [0xc3]
;
@@ -267,7 +267,7 @@ define <8 x double> @test_mm512_mask3_fmadd_pd(<8 x double> %__A, <8 x double> %
; X64: ## %bb.0: ## %entry
; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmadd231pd %zmm1, %zmm0, %zmm2 {%k1} ## encoding: [0x62,0xf2,0xfd,0x49,0xb8,0xd1]
-; X64-NEXT: ## zmm2 = (zmm0 * zmm1) + zmm2
+; X64-NEXT: ## zmm2 {%k1} = (zmm0 * zmm1) + zmm2
; X64-NEXT: vmovapd %zmm2, %zmm0 ## encoding: [0x62,0xf1,0xfd,0x48,0x28,0xc2]
; X64-NEXT: retq ## encoding: [0xc3]
entry:
@@ -283,14 +283,14 @@ define <8 x double> @test_mm512_maskz_fmadd_pd(i8 zeroext %__U, <8 x double> %__
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax ## encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmadd213pd %zmm2, %zmm1, %zmm0 {%k1} {z} ## encoding: [0x62,0xf2,0xf5,0xc9,0xa8,0xc2]
-; X86-NEXT: ## zmm0 = (zmm1 * zmm0) + zmm2
+; X86-NEXT: ## zmm0 {%k1} {z} = (zmm1 * zmm0) + zmm2
; X86-NEXT: retl ## encoding: [0xc3]
;
; X64-LABEL: test_mm512_maskz_fmadd_pd:
; X64: ## %bb.0: ## %entry
; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmadd213pd %zmm2, %zmm1, %zmm0 {%k1} {z} ## encoding: [0x62,0xf2,0xf5,0xc9,0xa8,0xc2]
-; X64-NEXT: ## zmm0 = (zmm1 * zmm0) + zmm2
+; X64-NEXT: ## zmm0 {%k1} {z} = (zmm1 * zmm0) + zmm2
; X64-NEXT: retq ## encoding: [0xc3]
entry:
%0 = tail call <8 x double> @llvm.fma.v8f64(<8 x double> %__A, <8 x double> %__B, <8 x double> %__C) #10
@@ -317,14 +317,14 @@ define <8 x double> @test_mm512_mask_fmsub_pd(<8 x double> %__A, i8 zeroext %__U
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax ## encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmsub132pd %zmm1, %zmm2, %zmm0 {%k1} ## encoding: [0x62,0xf2,0xed,0x49,0x9a,0xc1]
-; X86-NEXT: ## zmm0 = (zmm0 * zmm1) - zmm2
+; X86-NEXT: ## zmm0 {%k1} = (zmm0 * zmm1) - zmm2
; X86-NEXT: retl ## encoding: [0xc3]
;
; X64-LABEL: test_mm512_mask_fmsub_pd:
; X64: ## %bb.0: ## %entry
; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmsub132pd %zmm1, %zmm2, %zmm0 {%k1} ## encoding: [0x62,0xf2,0xed,0x49,0x9a,0xc1]
-; X64-NEXT: ## zmm0 = (zmm0 * zmm1) - zmm2
+; X64-NEXT: ## zmm0 {%k1} = (zmm0 * zmm1) - zmm2
; X64-NEXT: retq ## encoding: [0xc3]
entry:
%sub.i = fsub <8 x double> <double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00>, %__C
@@ -340,14 +340,14 @@ define <8 x double> @test_mm512_maskz_fmsub_pd(i8 zeroext %__U, <8 x double> %__
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax ## encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmsub213pd %zmm2, %zmm1, %zmm0 {%k1} {z} ## encoding: [0x62,0xf2,0xf5,0xc9,0xaa,0xc2]
-; X86-NEXT: ## zmm0 = (zmm1 * zmm0) - zmm2
+; X86-NEXT: ## zmm0 {%k1} {z} = (zmm1 * zmm0) - zmm2
; X86-NEXT: retl ## encoding: [0xc3]
;
; X64-LABEL: test_mm512_maskz_fmsub_pd:
; X64: ## %bb.0: ## %entry
; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmsub213pd %zmm2, %zmm1, %zmm0 {%k1} {z} ## encoding: [0x62,0xf2,0xf5,0xc9,0xaa,0xc2]
-; X64-NEXT: ## zmm0 = (zmm1 * zmm0) - zmm2
+; X64-NEXT: ## zmm0 {%k1} {z} = (zmm1 * zmm0) - zmm2
; X64-NEXT: retq ## encoding: [0xc3]
entry:
%sub.i = fsub <8 x double> <double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00>, %__C
@@ -375,7 +375,7 @@ define <8 x double> @test_mm512_mask3_fnmadd_pd(<8 x double> %__A, <8 x double>
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax ## encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfnmadd231pd %zmm1, %zmm0, %zmm2 {%k1} ## encoding: [0x62,0xf2,0xfd,0x49,0xbc,0xd1]
-; X86-NEXT: ## zmm2 = -(zmm0 * zmm1) + zmm2
+; X86-NEXT: ## zmm2 {%k1} = -(zmm0 * zmm1) + zmm2
; X86-NEXT: vmovapd %zmm2, %zmm0 ## encoding: [0x62,0xf1,0xfd,0x48,0x28,0xc2]
; X86-NEXT: retl ## encoding: [0xc3]
;
@@ -383,7 +383,7 @@ define <8 x double> @test_mm512_mask3_fnmadd_pd(<8 x double> %__A, <8 x double>
; X64: ## %bb.0: ## %entry
; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfnmadd231pd %zmm1, %zmm0, %zmm2 {%k1} ## encoding: [0x62,0xf2,0xfd,0x49,0xbc,0xd1]
-; X64-NEXT: ## zmm2 = -(zmm0 * zmm1) + zmm2
+; X64-NEXT: ## zmm2 {%k1} = -(zmm0 * zmm1) + zmm2
; X64-NEXT: vmovapd %zmm2, %zmm0 ## encoding: [0x62,0xf1,0xfd,0x48,0x28,0xc2]
; X64-NEXT: retq ## encoding: [0xc3]
entry:
@@ -400,14 +400,14 @@ define <8 x double> @test_mm512_maskz_fnmadd_pd(i8 zeroext %__U, <8 x double> %_
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax ## encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfnmadd213pd %zmm2, %zmm1, %zmm0 {%k1} {z} ## encoding: [0x62,0xf2,0xf5,0xc9,0xac,0xc2]
-; X86-NEXT: ## zmm0 = -(zmm1 * zmm0) + zmm2
+; X86-NEXT: ## zmm0 {%k1} {z} = -(zmm1 * zmm0) + zmm2
; X86-NEXT: retl ## encoding: [0xc3]
;
; X64-LABEL: test_mm512_maskz_fnmadd_pd:
; X64: ## %bb.0: ## %entry
; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfnmadd213pd %zmm2, %zmm1, %zmm0 {%k1} {z} ## encoding: [0x62,0xf2,0xf5,0xc9,0xac,0xc2]
-; X64-NEXT: ## zmm0 = -(zmm1 * zmm0) + zmm2
+; X64-NEXT: ## zmm0 {%k1} {z} = -(zmm1 * zmm0) + zmm2
; X64-NEXT: retq ## encoding: [0xc3]
entry:
%sub.i = fsub <8 x double> <double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00>, %__A
@@ -436,14 +436,14 @@ define <8 x double> @test_mm512_maskz_fnmsub_pd(i8 zeroext %__U, <8 x double> %_
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax ## encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfnmsub213pd %zmm2, %zmm1, %zmm0 {%k1} {z} ## encoding: [0x62,0xf2,0xf5,0xc9,0xae,0xc2]
-; X86-NEXT: ## zmm0 = -(zmm1 * zmm0) - zmm2
+; X86-NEXT: ## zmm0 {%k1} {z} = -(zmm1 * zmm0) - zmm2
; X86-NEXT: retl ## encoding: [0xc3]
;
; X64-LABEL: test_mm512_maskz_fnmsub_pd:
; X64: ## %bb.0: ## %entry
; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfnmsub213pd %zmm2, %zmm1, %zmm0 {%k1} {z} ## encoding: [0x62,0xf2,0xf5,0xc9,0xae,0xc2]
-; X64-NEXT: ## zmm0 = -(zmm1 * zmm0) - zmm2
+; X64-NEXT: ## zmm0 {%k1} {z} = -(zmm1 * zmm0) - zmm2
; X64-NEXT: retq ## encoding: [0xc3]
entry:
%sub.i = fsub <8 x double> <double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00>, %__A
@@ -678,14 +678,14 @@ define <16 x float> @test_mm512_mask_fmadd_ps(<16 x float> %__A, i16 zeroext %__
; X86: ## %bb.0: ## %entry
; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1 ## encoding: [0xc5,0xf8,0x90,0x4c,0x24,0x04]
; X86-NEXT: vfmadd132ps %zmm1, %zmm2, %zmm0 {%k1} ## encoding: [0x62,0xf2,0x6d,0x49,0x98,0xc1]
-; X86-NEXT: ## zmm0 = (zmm0 * zmm1) + zmm2
+; X86-NEXT: ## zmm0 {%k1} = (zmm0 * zmm1) + zmm2
; X86-NEXT: retl ## encoding: [0xc3]
;
; X64-LABEL: test_mm512_mask_fmadd_ps:
; X64: ## %bb.0: ## %entry
; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmadd132ps %zmm1, %zmm2, %zmm0 {%k1} ## encoding: [0x62,0xf2,0x6d,0x49,0x98,0xc1]
-; X64-NEXT: ## zmm0 = (zmm0 * zmm1) + zmm2
+; X64-NEXT: ## zmm0 {%k1} = (zmm0 * zmm1) + zmm2
; X64-NEXT: retq ## encoding: [0xc3]
entry:
%0 = tail call <16 x float> @llvm.fma.v16f32(<16 x float> %__A, <16 x float> %__B, <16 x float> %__C) #10
@@ -699,7 +699,7 @@ define <16 x float> @test_mm512_mask3_fmadd_ps(<16 x float> %__A, <16 x float> %
; X86: ## %bb.0: ## %entry
; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1 ## encoding: [0xc5,0xf8,0x90,0x4c,0x24,0x04]
; X86-NEXT: vfmadd231ps %zmm1, %zmm0, %zmm2 {%k1} ## encoding: [0x62,0xf2,0x7d,0x49,0xb8,0xd1]
-; X86-NEXT: ## zmm2 = (zmm0 * zmm1) + zmm2
+; X86-NEXT: ## zmm2 {%k1} = (zmm0 * zmm1) + zmm2
; X86-NEXT: vmovaps %zmm2, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc2]
; X86-NEXT: retl ## encoding: [0xc3]
;
@@ -707,7 +707,7 @@ define <16 x float> @test_mm512_mask3_fmadd_ps(<16 x float> %__A, <16 x float> %
; X64: ## %bb.0: ## %entry
; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmadd231ps %zmm1, %zmm0, %zmm2 {%k1} ## encoding: [0x62,0xf2,0x7d,0x49,0xb8,0xd1]
-; X64-NEXT: ## zmm2 = (zmm0 * zmm1) + zmm2
+; X64-NEXT: ## zmm2 {%k1} = (zmm0 * zmm1) + zmm2
; X64-NEXT: vmovaps %zmm2, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc2]
; X64-NEXT: retq ## encoding: [0xc3]
entry:
@@ -722,14 +722,14 @@ define <16 x float> @test_mm512_maskz_fmadd_ps(i16 zeroext %__U, <16 x float> %_
; X86: ## %bb.0: ## %entry
; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1 ## encoding: [0xc5,0xf8,0x90,0x4c,0x24,0x04]
; X86-NEXT: vfmadd213ps %zmm2, %zmm1, %zmm0 {%k1} {z} ## encoding: [0x62,0xf2,0x75,0xc9,0xa8,0xc2]
-; X86-NEXT: ## zmm0 = (zmm1 * zmm0) + zmm2
+; X86-NEXT: ## zmm0 {%k1} {z} = (zmm1 * zmm0) + zmm2
; X86-NEXT: retl ## encoding: [0xc3]
;
; X64-LABEL: test_mm512_maskz_fmadd_ps:
; X64: ## %bb.0: ## %entry
; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmadd213ps %zmm2, %zmm1, %zmm0 {%k1} {z} ## encoding: [0x62,0xf2,0x75,0xc9,0xa8,0xc2]
-; X64-NEXT: ## zmm0 = (zmm1 * zmm0) + zmm2
+; X64-NEXT: ## zmm0 {%k1} {z} = (zmm1 * zmm0) + zmm2
; X64-NEXT: retq ## encoding: [0xc3]
entry:
%0 = tail call <16 x float> @llvm.fma.v16f32(<16 x float> %__A, <16 x float> %__B, <16 x float> %__C) #10
@@ -755,14 +755,14 @@ define <16 x float> @test_mm512_mask_fmsub_ps(<16 x float> %__A, i16 zeroext %__
; X86: ## %bb.0: ## %entry
; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1 ## encoding: [0xc5,0xf8,0x90,0x4c,0x24,0x04]
; X86-NEXT: vfmsub132ps %zmm1, %zmm2, %zmm0 {%k1} ## encoding: [0x62,0xf2,0x6d,0x49,0x9a,0xc1]
-; X86-NEXT: ## zmm0 = (zmm0 * zmm1) - zmm2
+; X86-NEXT: ## zmm0 {%k1} = (zmm0 * zmm1) - zmm2
; X86-NEXT: retl ## encoding: [0xc3]
;
; X64-LABEL: test_mm512_mask_fmsub_ps:
; X64: ## %bb.0: ## %entry
; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmsub132ps %zmm1, %zmm2, %zmm0 {%k1} ## encoding: [0x62,0xf2,0x6d,0x49,0x9a,0xc1]
-; X64-NEXT: ## zmm0 = (zmm0 * zmm1) - zmm2
+; X64-NEXT: ## zmm0 {%k1} = (zmm0 * zmm1) - zmm2
; X64-NEXT: retq ## encoding: [0xc3]
entry:
%sub.i = fsub <16 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %__C
@@ -777,14 +777,14 @@ define <16 x float> @test_mm512_maskz_fmsub_ps(i16 zeroext %__U, <16 x float> %_
; X86: ## %bb.0: ## %entry
; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1 ## encoding: [0xc5,0xf8,0x90,0x4c,0x24,0x04]
; X86-NEXT: vfmsub213ps %zmm2, %zmm1, %zmm0 {%k1} {z} ## encoding: [0x62,0xf2,0x75,0xc9,0xaa,0xc2]
-; X86-NEXT: ## zmm0 = (zmm1 * zmm0) - zmm2
+; X86-NEXT: ## zmm0 {%k1} {z} = (zmm1 * zmm0) - zmm2
; X86-NEXT: retl ## encoding: [0xc3]
;
; X64-LABEL: test_mm512_maskz_fmsub_ps:
; X64: ## %bb.0: ## %entry
; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmsub213ps %zmm2, %zmm1, %zmm0 {%k1} {z} ## encoding: [0x62,0xf2,0x75,0xc9,0xaa,0xc2]
-; X64-NEXT: ## zmm0 = (zmm1 * zmm0) - zmm2
+; X64-NEXT: ## zmm0 {%k1} {z} = (zmm1 * zmm0) - zmm2
; X64-NEXT: retq ## encoding: [0xc3]
entry:
%sub.i = fsub <16 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %__C
@@ -811,7 +811,7 @@ define <16 x float> @test_mm512_mask3_fnmadd_ps(<16 x float> %__A, <16 x float>
; X86: ## %bb.0: ## %entry
; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1 ## encoding: [0xc5,0xf8,0x90,0x4c,0x24,0x04]
; X86-NEXT: vfnmadd231ps %zmm1, %zmm0, %zmm2 {%k1} ## encoding: [0x62,0xf2,0x7d,0x49,0xbc,0xd1]
-; X86-NEXT: ## zmm2 = -(zmm0 * zmm1) + zmm2
+; X86-NEXT: ## zmm2 {%k1} = -(zmm0 * zmm1) + zmm2
; X86-NEXT: vmovaps %zmm2, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc2]
; X86-NEXT: retl ## encoding: [0xc3]
;
@@ -819,7 +819,7 @@ define <16 x float> @test_mm512_mask3_fnmadd_ps(<16 x float> %__A, <16 x float>
; X64: ## %bb.0: ## %entry
; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfnmadd231ps %zmm1, %zmm0, %zmm2 {%k1} ## encoding: [0x62,0xf2,0x7d,0x49,0xbc,0xd1]
-; X64-NEXT: ## zmm2 = -(zmm0 * zmm1) + zmm2
+; X64-NEXT: ## zmm2 {%k1} = -(zmm0 * zmm1) + zmm2
; X64-NEXT: vmovaps %zmm2, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc2]
; X64-NEXT: retq ## encoding: [0xc3]
entry:
@@ -835,14 +835,14 @@ define <16 x float> @test_mm512_maskz_fnmadd_ps(i16 zeroext %__U, <16 x float> %
; X86: ## %bb.0: ## %entry
; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1 ## encoding: [0xc5,0xf8,0x90,0x4c,0x24,0x04]
; X86-NEXT: vfnmadd213ps %zmm2, %zmm1, %zmm0 {%k1} {z} ## encoding: [0x62,0xf2,0x75,0xc9,0xac,0xc2]
-; X86-NEXT: ## zmm0 = -(zmm1 * zmm0) + zmm2
+; X86-NEXT: ## zmm0 {%k1} {z} = -(zmm1 * zmm0) + zmm2
; X86-NEXT: retl ## encoding: [0xc3]
;
; X64-LABEL: test_mm512_maskz_fnmadd_ps:
; X64: ## %bb.0: ## %entry
; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfnmadd213ps %zmm2, %zmm1, %zmm0 {%k1} {z} ## encoding: [0x62,0xf2,0x75,0xc9,0xac,0xc2]
-; X64-NEXT: ## zmm0 = -(zmm1 * zmm0) + zmm2
+; X64-NEXT: ## zmm0 {%k1} {z} = -(zmm1 * zmm0) + zmm2
; X64-NEXT: retq ## encoding: [0xc3]
entry:
%sub.i = fsub <16 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %__A
@@ -870,14 +870,14 @@ define <16 x float> @test_mm512_maskz_fnmsub_ps(i16 zeroext %__U, <16 x float> %
; X86: ## %bb.0: ## %entry
; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1 ## encoding: [0xc5,0xf8,0x90,0x4c,0x24,0x04]
; X86-NEXT: vfnmsub213ps %zmm2, %zmm1, %zmm0 {%k1} {z} ## encoding: [0x62,0xf2,0x75,0xc9,0xae,0xc2]
-; X86-NEXT: ## zmm0 = -(zmm1 * zmm0) - zmm2
+; X86-NEXT: ## zmm0 {%k1} {z} = -(zmm1 * zmm0) - zmm2
; X86-NEXT: retl ## encoding: [0xc3]
;
; X64-LABEL: test_mm512_maskz_fnmsub_ps:
; X64: ## %bb.0: ## %entry
; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfnmsub213ps %zmm2, %zmm1, %zmm0 {%k1} {z} ## encoding: [0x62,0xf2,0x75,0xc9,0xae,0xc2]
-; X64-NEXT: ## zmm0 = -(zmm1 * zmm0) - zmm2
+; X64-NEXT: ## zmm0 {%k1} {z} = -(zmm1 * zmm0) - zmm2
; X64-NEXT: retq ## encoding: [0xc3]
entry:
%sub.i = fsub <16 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %__A
@@ -1035,14 +1035,14 @@ define <8 x double> @test_mm512_mask_fmaddsub_pd(<8 x double> %__A, i8 zeroext %
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax ## encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmaddsub132pd %zmm1, %zmm2, %zmm0 {%k1} ## encoding: [0x62,0xf2,0xed,0x49,0x96,0xc1]
-; X86-NEXT: ## zmm0 = (zmm0 * zmm1) +/- zmm2
+; X86-NEXT: ## zmm0 {%k1} = (zmm0 * zmm1) +/- zmm2
; X86-NEXT: retl ## encoding: [0xc3]
;
; X64-LABEL: test_mm512_mask_fmaddsub_pd:
; X64: ## %bb.0: ## %entry
; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmaddsub132pd %zmm1, %zmm2, %zmm0 {%k1} ## encoding: [0x62,0xf2,0xed,0x49,0x96,0xc1]
-; X64-NEXT: ## zmm0 = (zmm0 * zmm1) +/- zmm2
+; X64-NEXT: ## zmm0 {%k1} = (zmm0 * zmm1) +/- zmm2
; X64-NEXT: retq ## encoding: [0xc3]
entry:
%0 = tail call <8 x double> @llvm.fma.v8f64(<8 x double> %__A, <8 x double> %__B, <8 x double> %__C) #10
@@ -1060,7 +1060,7 @@ define <8 x double> @test_mm512_mask3_fmaddsub_pd(<8 x double> %__A, <8 x double
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax ## encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmaddsub231pd %zmm1, %zmm0, %zmm2 {%k1} ## encoding: [0x62,0xf2,0xfd,0x49,0xb6,0xd1]
-; X86-NEXT: ## zmm2 = (zmm0 * zmm1) +/- zmm2
+; X86-NEXT: ## zmm2 {%k1} = (zmm0 * zmm1) +/- zmm2
; X86-NEXT: vmovapd %zmm2, %zmm0 ## encoding: [0x62,0xf1,0xfd,0x48,0x28,0xc2]
; X86-NEXT: retl ## encoding: [0xc3]
;
@@ -1068,7 +1068,7 @@ define <8 x double> @test_mm512_mask3_fmaddsub_pd(<8 x double> %__A, <8 x double
; X64: ## %bb.0: ## %entry
; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmaddsub231pd %zmm1, %zmm0, %zmm2 {%k1} ## encoding: [0x62,0xf2,0xfd,0x49,0xb6,0xd1]
-; X64-NEXT: ## zmm2 = (zmm0 * zmm1) +/- zmm2
+; X64-NEXT: ## zmm2 {%k1} = (zmm0 * zmm1) +/- zmm2
; X64-NEXT: vmovapd %zmm2, %zmm0 ## encoding: [0x62,0xf1,0xfd,0x48,0x28,0xc2]
; X64-NEXT: retq ## encoding: [0xc3]
entry:
@@ -1087,14 +1087,14 @@ define <8 x double> @test_mm512_maskz_fmaddsub_pd(i8 zeroext %__U, <8 x double>
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax ## encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmaddsub213pd %zmm2, %zmm1, %zmm0 {%k1} {z} ## encoding: [0x62,0xf2,0xf5,0xc9,0xa6,0xc2]
-; X86-NEXT: ## zmm0 = (zmm1 * zmm0) +/- zmm2
+; X86-NEXT: ## zmm0 {%k1} {z} = (zmm1 * zmm0) +/- zmm2
; X86-NEXT: retl ## encoding: [0xc3]
;
; X64-LABEL: test_mm512_maskz_fmaddsub_pd:
; X64: ## %bb.0: ## %entry
; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmaddsub213pd %zmm2, %zmm1, %zmm0 {%k1} {z} ## encoding: [0x62,0xf2,0xf5,0xc9,0xa6,0xc2]
-; X64-NEXT: ## zmm0 = (zmm1 * zmm0) +/- zmm2
+; X64-NEXT: ## zmm0 {%k1} {z} = (zmm1 * zmm0) +/- zmm2
; X64-NEXT: retq ## encoding: [0xc3]
entry:
%0 = tail call <8 x double> @llvm.fma.v8f64(<8 x double> %__A, <8 x double> %__B, <8 x double> %__C) #10
@@ -1126,14 +1126,14 @@ define <8 x double> @test_mm512_mask_fmsubadd_pd(<8 x double> %__A, i8 zeroext %
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax ## encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmsubadd132pd %zmm1, %zmm2, %zmm0 {%k1} ## encoding: [0x62,0xf2,0xed,0x49,0x97,0xc1]
-; X86-NEXT: ## zmm0 = (zmm0 * zmm1) -/+ zmm2
+; X86-NEXT: ## zmm0 {%k1} = (zmm0 * zmm1) -/+ zmm2
; X86-NEXT: retl ## encoding: [0xc3]
;
; X64-LABEL: test_mm512_mask_fmsubadd_pd:
; X64: ## %bb.0: ## %entry
; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmsubadd132pd %zmm1, %zmm2, %zmm0 {%k1} ## encoding: [0x62,0xf2,0xed,0x49,0x97,0xc1]
-; X64-NEXT: ## zmm0 = (zmm0 * zmm1) -/+ zmm2
+; X64-NEXT: ## zmm0 {%k1} = (zmm0 * zmm1) -/+ zmm2
; X64-NEXT: retq ## encoding: [0xc3]
entry:
%sub.i = fsub <8 x double> <double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00>, %__C
@@ -1151,14 +1151,14 @@ define <8 x double> @test_mm512_maskz_fmsubadd_pd(i8 zeroext %__U, <8 x double>
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax ## encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmsubadd213pd %zmm2, %zmm1, %zmm0 {%k1} {z} ## encoding: [0x62,0xf2,0xf5,0xc9,0xa7,0xc2]
-; X86-NEXT: ## zmm0 = (zmm1 * zmm0) -/+ zmm2
+; X86-NEXT: ## zmm0 {%k1} {z} = (zmm1 * zmm0) -/+ zmm2
; X86-NEXT: retl ## encoding: [0xc3]
;
; X64-LABEL: test_mm512_maskz_fmsubadd_pd:
; X64: ## %bb.0: ## %entry
; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmsubadd213pd %zmm2, %zmm1, %zmm0 {%k1} {z} ## encoding: [0x62,0xf2,0xf5,0xc9,0xa7,0xc2]
-; X64-NEXT: ## zmm0 = (zmm1 * zmm0) -/+ zmm2
+; X64-NEXT: ## zmm0 {%k1} {z} = (zmm1 * zmm0) -/+ zmm2
; X64-NEXT: retq ## encoding: [0xc3]
entry:
%sub.i = fsub <8 x double> <double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00>, %__C
@@ -1311,14 +1311,14 @@ define <16 x float> @test_mm512_mask_fmaddsub_ps(<16 x float> %__A, i16 zeroext
; X86: ## %bb.0: ## %entry
; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1 ## encoding: [0xc5,0xf8,0x90,0x4c,0x24,0x04]
; X86-NEXT: vfmaddsub132ps %zmm1, %zmm2, %zmm0 {%k1} ## encoding: [0x62,0xf2,0x6d,0x49,0x96,0xc1]
-; X86-NEXT: ## zmm0 = (zmm0 * zmm1) +/- zmm2
+; X86-NEXT: ## zmm0 {%k1} = (zmm0 * zmm1) +/- zmm2
; X86-NEXT: retl ## encoding: [0xc3]
;
; X64-LABEL: test_mm512_mask_fmaddsub_ps:
; X64: ## %bb.0: ## %entry
; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmaddsub132ps %zmm1, %zmm2, %zmm0 {%k1} ## encoding: [0x62,0xf2,0x6d,0x49,0x96,0xc1]
-; X64-NEXT: ## zmm0 = (zmm0 * zmm1) +/- zmm2
+; X64-NEXT: ## zmm0 {%k1} = (zmm0 * zmm1) +/- zmm2
; X64-NEXT: retq ## encoding: [0xc3]
entry:
%0 = tail call <16 x float> @llvm.fma.v16f32(<16 x float> %__A, <16 x float> %__B, <16 x float> %__C) #10
@@ -1335,7 +1335,7 @@ define <16 x float> @test_mm512_mask3_fmaddsub_ps(<16 x float> %__A, <16 x float
; X86: ## %bb.0: ## %entry
; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1 ## encoding: [0xc5,0xf8,0x90,0x4c,0x24,0x04]
; X86-NEXT: vfmaddsub231ps %zmm1, %zmm0, %zmm2 {%k1} ## encoding: [0x62,0xf2,0x7d,0x49,0xb6,0xd1]
-; X86-NEXT: ## zmm2 = (zmm0 * zmm1) +/- zmm2
+; X86-NEXT: ## zmm2 {%k1} = (zmm0 * zmm1) +/- zmm2
; X86-NEXT: vmovaps %zmm2, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc2]
; X86-NEXT: retl ## encoding: [0xc3]
;
@@ -1343,7 +1343,7 @@ define <16 x float> @test_mm512_mask3_fmaddsub_ps(<16 x float> %__A, <16 x float
; X64: ## %bb.0: ## %entry
; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmaddsub231ps %zmm1, %zmm0, %zmm2 {%k1} ## encoding: [0x62,0xf2,0x7d,0x49,0xb6,0xd1]
-; X64-NEXT: ## zmm2 = (zmm0 * zmm1) +/- zmm2
+; X64-NEXT: ## zmm2 {%k1} = (zmm0 * zmm1) +/- zmm2
; X64-NEXT: vmovaps %zmm2, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc2]
; X64-NEXT: retq ## encoding: [0xc3]
entry:
@@ -1361,14 +1361,14 @@ define <16 x float> @test_mm512_maskz_fmaddsub_ps(i16 zeroext %__U, <16 x float>
; X86: ## %bb.0: ## %entry
; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1 ## encoding: [0xc5,0xf8,0x90,0x4c,0x24,0x04]
; X86-NEXT: vfmaddsub213ps %zmm2, %zmm1, %zmm0 {%k1} {z} ## encoding: [0x62,0xf2,0x75,0xc9,0xa6,0xc2]
-; X86-NEXT: ## zmm0 = (zmm1 * zmm0) +/- zmm2
+; X86-NEXT: ## zmm0 {%k1} {z} = (zmm1 * zmm0) +/- zmm2
; X86-NEXT: retl ## encoding: [0xc3]
;
; X64-LABEL: test_mm512_maskz_fmaddsub_ps:
; X64: ## %bb.0: ## %entry
; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmaddsub213ps %zmm2, %zmm1, %zmm0 {%k1} {z} ## encoding: [0x62,0xf2,0x75,0xc9,0xa6,0xc2]
-; X64-NEXT: ## zmm0 = (zmm1 * zmm0) +/- zmm2
+; X64-NEXT: ## zmm0 {%k1} {z} = (zmm1 * zmm0) +/- zmm2
; X64-NEXT: retq ## encoding: [0xc3]
entry:
%0 = tail call <16 x float> @llvm.fma.v16f32(<16 x float> %__A, <16 x float> %__B, <16 x float> %__C) #10
@@ -1399,14 +1399,14 @@ define <16 x float> @test_mm512_mask_fmsubadd_ps(<16 x float> %__A, i16 zeroext
; X86: ## %bb.0: ## %entry
; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1 ## encoding: [0xc5,0xf8,0x90,0x4c,0x24,0x04]
; X86-NEXT: vfmsubadd132ps %zmm1, %zmm2, %zmm0 {%k1} ## encoding: [0x62,0xf2,0x6d,0x49,0x97,0xc1]
-; X86-NEXT: ## zmm0 = (zmm0 * zmm1) -/+ zmm2
+; X86-NEXT: ## zmm0 {%k1} = (zmm0 * zmm1) -/+ zmm2
; X86-NEXT: retl ## encoding: [0xc3]
;
; X64-LABEL: test_mm512_mask_fmsubadd_ps:
; X64: ## %bb.0: ## %entry
; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmsubadd132ps %zmm1, %zmm2, %zmm0 {%k1} ## encoding: [0x62,0xf2,0x6d,0x49,0x97,0xc1]
-; X64-NEXT: ## zmm0 = (zmm0 * zmm1) -/+ zmm2
+; X64-NEXT: ## zmm0 {%k1} = (zmm0 * zmm1) -/+ zmm2
; X64-NEXT: retq ## encoding: [0xc3]
entry:
%sub.i = fsub <16 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %__C
@@ -1423,14 +1423,14 @@ define <16 x float> @test_mm512_maskz_fmsubadd_ps(i16 zeroext %__U, <16 x float>
; X86: ## %bb.0: ## %entry
; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1 ## encoding: [0xc5,0xf8,0x90,0x4c,0x24,0x04]
; X86-NEXT: vfmsubadd213ps %zmm2, %zmm1, %zmm0 {%k1} {z} ## encoding: [0x62,0xf2,0x75,0xc9,0xa7,0xc2]
-; X86-NEXT: ## zmm0 = (zmm1 * zmm0) -/+ zmm2
+; X86-NEXT: ## zmm0 {%k1} {z} = (zmm1 * zmm0) -/+ zmm2
; X86-NEXT: retl ## encoding: [0xc3]
;
; X64-LABEL: test_mm512_maskz_fmsubadd_ps:
; X64: ## %bb.0: ## %entry
; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmsubadd213ps %zmm2, %zmm1, %zmm0 {%k1} {z} ## encoding: [0x62,0xf2,0x75,0xc9,0xa7,0xc2]
-; X64-NEXT: ## zmm0 = (zmm1 * zmm0) -/+ zmm2
+; X64-NEXT: ## zmm0 {%k1} {z} = (zmm1 * zmm0) -/+ zmm2
; X64-NEXT: retq ## encoding: [0xc3]
entry:
%sub.i = fsub <16 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %__C
@@ -1471,7 +1471,7 @@ define <8 x double> @test_mm512_mask3_fmsub_pd(<8 x double> %__A, <8 x double> %
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax ## encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmsub231pd %zmm1, %zmm0, %zmm2 {%k1} ## encoding: [0x62,0xf2,0xfd,0x49,0xba,0xd1]
-; X86-NEXT: ## zmm2 = (zmm0 * zmm1) - zmm2
+; X86-NEXT: ## zmm2 {%k1} = (zmm0 * zmm1) - zmm2
; X86-NEXT: vmovapd %zmm2, %zmm0 ## encoding: [0x62,0xf1,0xfd,0x48,0x28,0xc2]
; X86-NEXT: retl ## encoding: [0xc3]
;
@@ -1479,7 +1479,7 @@ define <8 x double> @test_mm512_mask3_fmsub_pd(<8 x double> %__A, <8 x double> %
; X64: ## %bb.0: ## %entry
; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmsub231pd %zmm1, %zmm0, %zmm2 {%k1} ## encoding: [0x62,0xf2,0xfd,0x49,0xba,0xd1]
-; X64-NEXT: ## zmm2 = (zmm0 * zmm1) - zmm2
+; X64-NEXT: ## zmm2 {%k1} = (zmm0 * zmm1) - zmm2
; X64-NEXT: vmovapd %zmm2, %zmm0 ## encoding: [0x62,0xf1,0xfd,0x48,0x28,0xc2]
; X64-NEXT: retq ## encoding: [0xc3]
entry:
@@ -1517,7 +1517,7 @@ define <16 x float> @test_mm512_mask3_fmsub_ps(<16 x float> %__A, <16 x float> %
; X86: ## %bb.0: ## %entry
; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1 ## encoding: [0xc5,0xf8,0x90,0x4c,0x24,0x04]
; X86-NEXT: vfmsub231ps %zmm1, %zmm0, %zmm2 {%k1} ## encoding: [0x62,0xf2,0x7d,0x49,0xba,0xd1]
-; X86-NEXT: ## zmm2 = (zmm0 * zmm1) - zmm2
+; X86-NEXT: ## zmm2 {%k1} = (zmm0 * zmm1) - zmm2
; X86-NEXT: vmovaps %zmm2, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc2]
; X86-NEXT: retl ## encoding: [0xc3]
;
@@ -1525,7 +1525,7 @@ define <16 x float> @test_mm512_mask3_fmsub_ps(<16 x float> %__A, <16 x float> %
; X64: ## %bb.0: ## %entry
; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmsub231ps %zmm1, %zmm0, %zmm2 {%k1} ## encoding: [0x62,0xf2,0x7d,0x49,0xba,0xd1]
-; X64-NEXT: ## zmm2 = (zmm0 * zmm1) - zmm2
+; X64-NEXT: ## zmm2 {%k1} = (zmm0 * zmm1) - zmm2
; X64-NEXT: vmovaps %zmm2, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc2]
; X64-NEXT: retq ## encoding: [0xc3]
entry:
@@ -1565,7 +1565,7 @@ define <8 x double> @test_mm512_mask3_fmsubadd_pd(<8 x double> %__A, <8 x double
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax ## encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmsubadd231pd %zmm1, %zmm0, %zmm2 {%k1} ## encoding: [0x62,0xf2,0xfd,0x49,0xb7,0xd1]
-; X86-NEXT: ## zmm2 = (zmm0 * zmm1) -/+ zmm2
+; X86-NEXT: ## zmm2 {%k1} = (zmm0 * zmm1) -/+ zmm2
; X86-NEXT: vmovapd %zmm2, %zmm0 ## encoding: [0x62,0xf1,0xfd,0x48,0x28,0xc2]
; X86-NEXT: retl ## encoding: [0xc3]
;
@@ -1573,7 +1573,7 @@ define <8 x double> @test_mm512_mask3_fmsubadd_pd(<8 x double> %__A, <8 x double
; X64: ## %bb.0: ## %entry
; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmsubadd231pd %zmm1, %zmm0, %zmm2 {%k1} ## encoding: [0x62,0xf2,0xfd,0x49,0xb7,0xd1]
-; X64-NEXT: ## zmm2 = (zmm0 * zmm1) -/+ zmm2
+; X64-NEXT: ## zmm2 {%k1} = (zmm0 * zmm1) -/+ zmm2
; X64-NEXT: vmovapd %zmm2, %zmm0 ## encoding: [0x62,0xf1,0xfd,0x48,0x28,0xc2]
; X64-NEXT: retq ## encoding: [0xc3]
entry:
@@ -1613,7 +1613,7 @@ define <16 x float> @test_mm512_mask3_fmsubadd_ps(<16 x float> %__A, <16 x float
; X86: ## %bb.0: ## %entry
; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1 ## encoding: [0xc5,0xf8,0x90,0x4c,0x24,0x04]
; X86-NEXT: vfmsubadd231ps %zmm1, %zmm0, %zmm2 {%k1} ## encoding: [0x62,0xf2,0x7d,0x49,0xb7,0xd1]
-; X86-NEXT: ## zmm2 = (zmm0 * zmm1) -/+ zmm2
+; X86-NEXT: ## zmm2 {%k1} = (zmm0 * zmm1) -/+ zmm2
; X86-NEXT: vmovaps %zmm2, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc2]
; X86-NEXT: retl ## encoding: [0xc3]
;
@@ -1621,7 +1621,7 @@ define <16 x float> @test_mm512_mask3_fmsubadd_ps(<16 x float> %__A, <16 x float
; X64: ## %bb.0: ## %entry
; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmsubadd231ps %zmm1, %zmm0, %zmm2 {%k1} ## encoding: [0x62,0xf2,0x7d,0x49,0xb7,0xd1]
-; X64-NEXT: ## zmm2 = (zmm0 * zmm1) -/+ zmm2
+; X64-NEXT: ## zmm2 {%k1} = (zmm0 * zmm1) -/+ zmm2
; X64-NEXT: vmovaps %zmm2, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc2]
; X64-NEXT: retq ## encoding: [0xc3]
entry:
@@ -1661,14 +1661,14 @@ define <8 x double> @test_mm512_mask_fnmadd_pd(<8 x double> %__A, i8 zeroext %__
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax ## encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfnmadd132pd %zmm1, %zmm2, %zmm0 {%k1} ## encoding: [0x62,0xf2,0xed,0x49,0x9c,0xc1]
-; X86-NEXT: ## zmm0 = -(zmm0 * zmm1) + zmm2
+; X86-NEXT: ## zmm0 {%k1} = -(zmm0 * zmm1) + zmm2
; X86-NEXT: retl ## encoding: [0xc3]
;
; X64-LABEL: test_mm512_mask_fnmadd_pd:
; X64: ## %bb.0: ## %entry
; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfnmadd132pd %zmm1, %zmm2, %zmm0 {%k1} ## encoding: [0x62,0xf2,0xed,0x49,0x9c,0xc1]
-; X64-NEXT: ## zmm0 = -(zmm0 * zmm1) + zmm2
+; X64-NEXT: ## zmm0 {%k1} = -(zmm0 * zmm1) + zmm2
; X64-NEXT: retq ## encoding: [0xc3]
entry:
%sub.i = fsub <8 x double> <double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00>, %__A
@@ -1703,14 +1703,14 @@ define <16 x float> @test_mm512_mask_fnmadd_ps(<16 x float> %__A, i16 zeroext %_
; X86: ## %bb.0: ## %entry
; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1 ## encoding: [0xc5,0xf8,0x90,0x4c,0x24,0x04]
; X86-NEXT: vfnmadd132ps %zmm1, %zmm2, %zmm0 {%k1} ## encoding: [0x62,0xf2,0x6d,0x49,0x9c,0xc1]
-; X86-NEXT: ## zmm0 = -(zmm0 * zmm1) + zmm2
+; X86-NEXT: ## zmm0 {%k1} = -(zmm0 * zmm1) + zmm2
; X86-NEXT: retl ## encoding: [0xc3]
;
; X64-LABEL: test_mm512_mask_fnmadd_ps:
; X64: ## %bb.0: ## %entry
; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfnmadd132ps %zmm1, %zmm2, %zmm0 {%k1} ## encoding: [0x62,0xf2,0x6d,0x49,0x9c,0xc1]
-; X64-NEXT: ## zmm0 = -(zmm0 * zmm1) + zmm2
+; X64-NEXT: ## zmm0 {%k1} = -(zmm0 * zmm1) + zmm2
; X64-NEXT: retq ## encoding: [0xc3]
entry:
%sub.i = fsub <16 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %__A
@@ -1772,14 +1772,14 @@ define <8 x double> @test_mm512_mask_fnmsub_pd(<8 x double> %__A, i8 zeroext %__
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax ## encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfnmsub132pd %zmm1, %zmm2, %zmm0 {%k1} ## encoding: [0x62,0xf2,0xed,0x49,0x9e,0xc1]
-; X86-NEXT: ## zmm0 = -(zmm0 * zmm1) - zmm2
+; X86-NEXT: ## zmm0 {%k1} = -(zmm0 * zmm1) - zmm2
; X86-NEXT: retl ## encoding: [0xc3]
;
; X64-LABEL: test_mm512_mask_fnmsub_pd:
; X64: ## %bb.0: ## %entry
; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfnmsub132pd %zmm1, %zmm2, %zmm0 {%k1} ## encoding: [0x62,0xf2,0xed,0x49,0x9e,0xc1]
-; X64-NEXT: ## zmm0 = -(zmm0 * zmm1) - zmm2
+; X64-NEXT: ## zmm0 {%k1} = -(zmm0 * zmm1) - zmm2
; X64-NEXT: retq ## encoding: [0xc3]
entry:
%sub.i = fsub <8 x double> <double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00>, %__B
@@ -1796,7 +1796,7 @@ define <8 x double> @test_mm512_mask3_fnmsub_pd(<8 x double> %__A, <8 x double>
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax ## encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfnmsub231pd %zmm1, %zmm0, %zmm2 {%k1} ## encoding: [0x62,0xf2,0xfd,0x49,0xbe,0xd1]
-; X86-NEXT: ## zmm2 = -(zmm0 * zmm1) - zmm2
+; X86-NEXT: ## zmm2 {%k1} = -(zmm0 * zmm1) - zmm2
; X86-NEXT: vmovapd %zmm2, %zmm0 ## encoding: [0x62,0xf1,0xfd,0x48,0x28,0xc2]
; X86-NEXT: retl ## encoding: [0xc3]
;
@@ -1804,7 +1804,7 @@ define <8 x double> @test_mm512_mask3_fnmsub_pd(<8 x double> %__A, <8 x double>
; X64: ## %bb.0: ## %entry
; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfnmsub231pd %zmm1, %zmm0, %zmm2 {%k1} ## encoding: [0x62,0xf2,0xfd,0x49,0xbe,0xd1]
-; X64-NEXT: ## zmm2 = -(zmm0 * zmm1) - zmm2
+; X64-NEXT: ## zmm2 {%k1} = -(zmm0 * zmm1) - zmm2
; X64-NEXT: vmovapd %zmm2, %zmm0 ## encoding: [0x62,0xf1,0xfd,0x48,0x28,0xc2]
; X64-NEXT: retq ## encoding: [0xc3]
entry:
@@ -1865,14 +1865,14 @@ define <16 x float> @test_mm512_mask_fnmsub_ps(<16 x float> %__A, i16 zeroext %_
; X86: ## %bb.0: ## %entry
; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1 ## encoding: [0xc5,0xf8,0x90,0x4c,0x24,0x04]
; X86-NEXT: vfnmsub132ps %zmm1, %zmm2, %zmm0 {%k1} ## encoding: [0x62,0xf2,0x6d,0x49,0x9e,0xc1]
-; X86-NEXT: ## zmm0 = -(zmm0 * zmm1) - zmm2
+; X86-NEXT: ## zmm0 {%k1} = -(zmm0 * zmm1) - zmm2
; X86-NEXT: retl ## encoding: [0xc3]
;
; X64-LABEL: test_mm512_mask_fnmsub_ps:
; X64: ## %bb.0: ## %entry
; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfnmsub132ps %zmm1, %zmm2, %zmm0 {%k1} ## encoding: [0x62,0xf2,0x6d,0x49,0x9e,0xc1]
-; X64-NEXT: ## zmm0 = -(zmm0 * zmm1) - zmm2
+; X64-NEXT: ## zmm0 {%k1} = -(zmm0 * zmm1) - zmm2
; X64-NEXT: retq ## encoding: [0xc3]
entry:
%sub.i = fsub <16 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %__B
@@ -1888,7 +1888,7 @@ define <16 x float> @test_mm512_mask3_fnmsub_ps(<16 x float> %__A, <16 x float>
; X86: ## %bb.0: ## %entry
; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1 ## encoding: [0xc5,0xf8,0x90,0x4c,0x24,0x04]
; X86-NEXT: vfnmsub231ps %zmm1, %zmm0, %zmm2 {%k1} ## encoding: [0x62,0xf2,0x7d,0x49,0xbe,0xd1]
-; X86-NEXT: ## zmm2 = -(zmm0 * zmm1) - zmm2
+; X86-NEXT: ## zmm2 {%k1} = -(zmm0 * zmm1) - zmm2
; X86-NEXT: vmovaps %zmm2, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc2]
; X86-NEXT: retl ## encoding: [0xc3]
;
@@ -1896,7 +1896,7 @@ define <16 x float> @test_mm512_mask3_fnmsub_ps(<16 x float> %__A, <16 x float>
; X64: ## %bb.0: ## %entry
; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfnmsub231ps %zmm1, %zmm0, %zmm2 {%k1} ## encoding: [0x62,0xf2,0x7d,0x49,0xbe,0xd1]
-; X64-NEXT: ## zmm2 = -(zmm0 * zmm1) - zmm2
+; X64-NEXT: ## zmm2 {%k1} = -(zmm0 * zmm1) - zmm2
; X64-NEXT: vmovaps %zmm2, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc2]
; X64-NEXT: retq ## encoding: [0xc3]
entry:
@@ -1914,14 +1914,14 @@ define <4 x float> @test_mm_mask_fmadd_ss(<4 x float> %__W, i8 zeroext %__U, <4
; X86-NEXT: movb {{[0-9]+}}(%esp), %al ## encoding: [0x8a,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmadd213ss %xmm2, %xmm1, %xmm0 {%k1} ## encoding: [0x62,0xf2,0x75,0x09,0xa9,0xc2]
-; X86-NEXT: ## xmm0 = (xmm1 * xmm0) + xmm2
+; X86-NEXT: ## xmm0 {%k1} = (xmm1 * xmm0) + xmm2
; X86-NEXT: retl ## encoding: [0xc3]
;
; X64-LABEL: test_mm_mask_fmadd_ss:
; X64: ## %bb.0: ## %entry
; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmadd213ss %xmm2, %xmm1, %xmm0 {%k1} ## encoding: [0x62,0xf2,0x75,0x09,0xa9,0xc2]
-; X64-NEXT: ## xmm0 = (xmm1 * xmm0) + xmm2
+; X64-NEXT: ## xmm0 {%k1} = (xmm1 * xmm0) + xmm2
; X64-NEXT: retq ## encoding: [0xc3]
entry:
%0 = extractelement <4 x float> %__W, i64 0
@@ -1942,14 +1942,14 @@ define <4 x float> @test_mm_mask_fmadd_round_ss(<4 x float> %__W, i8 zeroext %__
; X86-NEXT: movb {{[0-9]+}}(%esp), %al ## encoding: [0x8a,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmadd213ss %xmm2, %xmm1, %xmm0 {%k1} ## encoding: [0x62,0xf2,0x75,0x09,0xa9,0xc2]
-; X86-NEXT: ## xmm0 = (xmm1 * xmm0) + xmm2
+; X86-NEXT: ## xmm0 {%k1} = (xmm1 * xmm0) + xmm2
; X86-NEXT: retl ## encoding: [0xc3]
;
; X64-LABEL: test_mm_mask_fmadd_round_ss:
; X64: ## %bb.0: ## %entry
; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmadd213ss %xmm2, %xmm1, %xmm0 {%k1} ## encoding: [0x62,0xf2,0x75,0x09,0xa9,0xc2]
-; X64-NEXT: ## xmm0 = (xmm1 * xmm0) + xmm2
+; X64-NEXT: ## xmm0 {%k1} = (xmm1 * xmm0) + xmm2
; X64-NEXT: retq ## encoding: [0xc3]
entry:
%0 = tail call <4 x float> @llvm.x86.avx512.mask.vfmadd.ss(<4 x float> %__W, <4 x float> %__A, <4 x float> %__B, i8 %__U, i32 4)
@@ -1964,14 +1964,14 @@ define <4 x float> @test_mm_maskz_fmadd_ss(i8 zeroext %__U, <4 x float> %__A, <4
; X86-NEXT: movb {{[0-9]+}}(%esp), %al ## encoding: [0x8a,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmadd213ss %xmm2, %xmm1, %xmm0 {%k1} {z} ## encoding: [0x62,0xf2,0x75,0x89,0xa9,0xc2]
-; X86-NEXT: ## xmm0 = (xmm1 * xmm0) + xmm2
+; X86-NEXT: ## xmm0 {%k1} {z} = (xmm1 * xmm0) + xmm2
; X86-NEXT: retl ## encoding: [0xc3]
;
; X64-LABEL: test_mm_maskz_fmadd_ss:
; X64: ## %bb.0: ## %entry
; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmadd213ss %xmm2, %xmm1, %xmm0 {%k1} {z} ## encoding: [0x62,0xf2,0x75,0x89,0xa9,0xc2]
-; X64-NEXT: ## xmm0 = (xmm1 * xmm0) + xmm2
+; X64-NEXT: ## xmm0 {%k1} {z} = (xmm1 * xmm0) + xmm2
; X64-NEXT: retq ## encoding: [0xc3]
entry:
%0 = extractelement <4 x float> %__A, i64 0
@@ -1991,14 +1991,14 @@ define <4 x float> @test_mm_maskz_fmadd_round_ss(i8 zeroext %__U, <4 x float> %_
; X86-NEXT: movb {{[0-9]+}}(%esp), %al ## encoding: [0x8a,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmadd213ss %xmm2, %xmm1, %xmm0 {%k1} {z} ## encoding: [0x62,0xf2,0x75,0x89,0xa9,0xc2]
-; X86-NEXT: ## xmm0 = (xmm1 * xmm0) + xmm2
+; X86-NEXT: ## xmm0 {%k1} {z} = (xmm1 * xmm0) + xmm2
; X86-NEXT: retl ## encoding: [0xc3]
;
; X64-LABEL: test_mm_maskz_fmadd_round_ss:
; X64: ## %bb.0: ## %entry
; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmadd213ss %xmm2, %xmm1, %xmm0 {%k1} {z} ## encoding: [0x62,0xf2,0x75,0x89,0xa9,0xc2]
-; X64-NEXT: ## xmm0 = (xmm1 * xmm0) + xmm2
+; X64-NEXT: ## xmm0 {%k1} {z} = (xmm1 * xmm0) + xmm2
; X64-NEXT: retq ## encoding: [0xc3]
entry:
%0 = tail call <4 x float> @llvm.x86.avx512.maskz.vfmadd.ss(<4 x float> %__A, <4 x float> %__B, <4 x float> %__C, i8 %__U, i32 4)
@@ -2013,7 +2013,7 @@ define <4 x float> @test_mm_mask3_fmadd_ss(<4 x float> %__W, <4 x float> %__X, <
; X86-NEXT: movb {{[0-9]+}}(%esp), %al ## encoding: [0x8a,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmadd231ss %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf2,0x7d,0x09,0xb9,0xd1]
-; X86-NEXT: ## xmm2 = (xmm0 * xmm1) + xmm2
+; X86-NEXT: ## xmm2 {%k1} = (xmm0 * xmm1) + xmm2
; X86-NEXT: vmovaps %xmm2, %xmm0 ## encoding: [0xc5,0xf8,0x28,0xc2]
; X86-NEXT: retl ## encoding: [0xc3]
;
@@ -2021,7 +2021,7 @@ define <4 x float> @test_mm_mask3_fmadd_ss(<4 x float> %__W, <4 x float> %__X, <
; X64: ## %bb.0: ## %entry
; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmadd231ss %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf2,0x7d,0x09,0xb9,0xd1]
-; X64-NEXT: ## xmm2 = (xmm0 * xmm1) + xmm2
+; X64-NEXT: ## xmm2 {%k1} = (xmm0 * xmm1) + xmm2
; X64-NEXT: vmovaps %xmm2, %xmm0 ## encoding: [0xc5,0xf8,0x28,0xc2]
; X64-NEXT: retq ## encoding: [0xc3]
entry:
@@ -2043,7 +2043,7 @@ define <4 x float> @test_mm_mask3_fmadd_round_ss(<4 x float> %__W, <4 x float> %
; X86-NEXT: movb {{[0-9]+}}(%esp), %al ## encoding: [0x8a,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmadd231ss %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf2,0x7d,0x09,0xb9,0xd1]
-; X86-NEXT: ## xmm2 = (xmm0 * xmm1) + xmm2
+; X86-NEXT: ## xmm2 {%k1} = (xmm0 * xmm1) + xmm2
; X86-NEXT: vmovaps %xmm2, %xmm0 ## encoding: [0xc5,0xf8,0x28,0xc2]
; X86-NEXT: retl ## encoding: [0xc3]
;
@@ -2051,7 +2051,7 @@ define <4 x float> @test_mm_mask3_fmadd_round_ss(<4 x float> %__W, <4 x float> %
; X64: ## %bb.0: ## %entry
; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmadd231ss %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf2,0x7d,0x09,0xb9,0xd1]
-; X64-NEXT: ## xmm2 = (xmm0 * xmm1) + xmm2
+; X64-NEXT: ## xmm2 {%k1} = (xmm0 * xmm1) + xmm2
; X64-NEXT: vmovaps %xmm2, %xmm0 ## encoding: [0xc5,0xf8,0x28,0xc2]
; X64-NEXT: retq ## encoding: [0xc3]
entry:
@@ -2067,14 +2067,14 @@ define <4 x float> @test_mm_mask_fmsub_ss(<4 x float> %__W, i8 zeroext %__U, <4
; X86-NEXT: movb {{[0-9]+}}(%esp), %al ## encoding: [0x8a,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmsub213ss %xmm2, %xmm1, %xmm0 {%k1} ## encoding: [0x62,0xf2,0x75,0x09,0xab,0xc2]
-; X86-NEXT: ## xmm0 = (xmm1 * xmm0) - xmm2
+; X86-NEXT: ## xmm0 {%k1} = (xmm1 * xmm0) - xmm2
; X86-NEXT: retl ## encoding: [0xc3]
;
; X64-LABEL: test_mm_mask_fmsub_ss:
; X64: ## %bb.0: ## %entry
; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmsub213ss %xmm2, %xmm1, %xmm0 {%k1} ## encoding: [0x62,0xf2,0x75,0x09,0xab,0xc2]
-; X64-NEXT: ## xmm0 = (xmm1 * xmm0) - xmm2
+; X64-NEXT: ## xmm0 {%k1} = (xmm1 * xmm0) - xmm2
; X64-NEXT: retq ## encoding: [0xc3]
entry:
%0 = extractelement <4 x float> %__W, i64 0
@@ -2096,14 +2096,14 @@ define <4 x float> @test_mm_mask_fmsub_round_ss(<4 x float> %__W, i8 zeroext %__
; X86-NEXT: movb {{[0-9]+}}(%esp), %al ## encoding: [0x8a,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmsub213ss %xmm2, %xmm1, %xmm0 {%k1} ## encoding: [0x62,0xf2,0x75,0x09,0xab,0xc2]
-; X86-NEXT: ## xmm0 = (xmm1 * xmm0) - xmm2
+; X86-NEXT: ## xmm0 {%k1} = (xmm1 * xmm0) - xmm2
; X86-NEXT: retl ## encoding: [0xc3]
;
; X64-LABEL: test_mm_mask_fmsub_round_ss:
; X64: ## %bb.0: ## %entry
; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmsub213ss %xmm2, %xmm1, %xmm0 {%k1} ## encoding: [0x62,0xf2,0x75,0x09,0xab,0xc2]
-; X64-NEXT: ## xmm0 = (xmm1 * xmm0) - xmm2
+; X64-NEXT: ## xmm0 {%k1} = (xmm1 * xmm0) - xmm2
; X64-NEXT: retq ## encoding: [0xc3]
entry:
%sub = fsub <4 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %__B
@@ -2117,14 +2117,14 @@ define <4 x float> @test_mm_maskz_fmsub_ss(i8 zeroext %__U, <4 x float> %__A, <4
; X86-NEXT: movb {{[0-9]+}}(%esp), %al ## encoding: [0x8a,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmsub213ss %xmm2, %xmm1, %xmm0 {%k1} {z} ## encoding: [0x62,0xf2,0x75,0x89,0xab,0xc2]
-; X86-NEXT: ## xmm0 = (xmm1 * xmm0) - xmm2
+; X86-NEXT: ## xmm0 {%k1} {z} = (xmm1 * xmm0) - xmm2
; X86-NEXT: retl ## encoding: [0xc3]
;
; X64-LABEL: test_mm_maskz_fmsub_ss:
; X64: ## %bb.0: ## %entry
; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmsub213ss %xmm2, %xmm1, %xmm0 {%k1} {z} ## encoding: [0x62,0xf2,0x75,0x89,0xab,0xc2]
-; X64-NEXT: ## xmm0 = (xmm1 * xmm0) - xmm2
+; X64-NEXT: ## xmm0 {%k1} {z} = (xmm1 * xmm0) - xmm2
; X64-NEXT: retq ## encoding: [0xc3]
entry:
%0 = extractelement <4 x float> %__A, i64 0
@@ -2145,14 +2145,14 @@ define <4 x float> @test_mm_maskz_fmsub_round_ss(i8 zeroext %__U, <4 x float> %_
; X86-NEXT: movb {{[0-9]+}}(%esp), %al ## encoding: [0x8a,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmsub213ss %xmm2, %xmm1, %xmm0 {%k1} {z} ## encoding: [0x62,0xf2,0x75,0x89,0xab,0xc2]
-; X86-NEXT: ## xmm0 = (xmm1 * xmm0) - xmm2
+; X86-NEXT: ## xmm0 {%k1} {z} = (xmm1 * xmm0) - xmm2
; X86-NEXT: retl ## encoding: [0xc3]
;
; X64-LABEL: test_mm_maskz_fmsub_round_ss:
; X64: ## %bb.0: ## %entry
; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmsub213ss %xmm2, %xmm1, %xmm0 {%k1} {z} ## encoding: [0x62,0xf2,0x75,0x89,0xab,0xc2]
-; X64-NEXT: ## xmm0 = (xmm1 * xmm0) - xmm2
+; X64-NEXT: ## xmm0 {%k1} {z} = (xmm1 * xmm0) - xmm2
; X64-NEXT: retq ## encoding: [0xc3]
entry:
%sub = fsub <4 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %__C
@@ -2166,7 +2166,7 @@ define <4 x float> @test_mm_mask3_fmsub_ss(<4 x float> %__W, <4 x float> %__X, <
; X86-NEXT: movb {{[0-9]+}}(%esp), %al ## encoding: [0x8a,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmsub231ss %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf2,0x7d,0x09,0xbb,0xd1]
-; X86-NEXT: ## xmm2 = (xmm0 * xmm1) - xmm2
+; X86-NEXT: ## xmm2 {%k1} = (xmm0 * xmm1) - xmm2
; X86-NEXT: vmovaps %xmm2, %xmm0 ## encoding: [0xc5,0xf8,0x28,0xc2]
; X86-NEXT: retl ## encoding: [0xc3]
;
@@ -2174,7 +2174,7 @@ define <4 x float> @test_mm_mask3_fmsub_ss(<4 x float> %__W, <4 x float> %__X, <
; X64: ## %bb.0: ## %entry
; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmsub231ss %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf2,0x7d,0x09,0xbb,0xd1]
-; X64-NEXT: ## xmm2 = (xmm0 * xmm1) - xmm2
+; X64-NEXT: ## xmm2 {%k1} = (xmm0 * xmm1) - xmm2
; X64-NEXT: vmovaps %xmm2, %xmm0 ## encoding: [0xc5,0xf8,0x28,0xc2]
; X64-NEXT: retq ## encoding: [0xc3]
entry:
@@ -2197,7 +2197,7 @@ define <4 x float> @test_mm_mask3_fmsub_round_ss(<4 x float> %__W, <4 x float> %
; X86-NEXT: movb {{[0-9]+}}(%esp), %al ## encoding: [0x8a,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmsub231ss %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf2,0x7d,0x09,0xbb,0xd1]
-; X86-NEXT: ## xmm2 = (xmm0 * xmm1) - xmm2
+; X86-NEXT: ## xmm2 {%k1} = (xmm0 * xmm1) - xmm2
; X86-NEXT: vmovaps %xmm2, %xmm0 ## encoding: [0xc5,0xf8,0x28,0xc2]
; X86-NEXT: retl ## encoding: [0xc3]
;
@@ -2205,7 +2205,7 @@ define <4 x float> @test_mm_mask3_fmsub_round_ss(<4 x float> %__W, <4 x float> %
; X64: ## %bb.0: ## %entry
; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmsub231ss %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf2,0x7d,0x09,0xbb,0xd1]
-; X64-NEXT: ## xmm2 = (xmm0 * xmm1) - xmm2
+; X64-NEXT: ## xmm2 {%k1} = (xmm0 * xmm1) - xmm2
; X64-NEXT: vmovaps %xmm2, %xmm0 ## encoding: [0xc5,0xf8,0x28,0xc2]
; X64-NEXT: retq ## encoding: [0xc3]
entry:
@@ -2221,14 +2221,14 @@ define <4 x float> @test_mm_mask_fnmadd_ss(<4 x float> %__W, i8 zeroext %__U, <4
; X86-NEXT: movb {{[0-9]+}}(%esp), %al ## encoding: [0x8a,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfnmadd213ss %xmm2, %xmm1, %xmm0 {%k1} ## encoding: [0x62,0xf2,0x75,0x09,0xad,0xc2]
-; X86-NEXT: ## xmm0 = -(xmm1 * xmm0) + xmm2
+; X86-NEXT: ## xmm0 {%k1} = -(xmm1 * xmm0) + xmm2
; X86-NEXT: retl ## encoding: [0xc3]
;
; X64-LABEL: test_mm_mask_fnmadd_ss:
; X64: ## %bb.0: ## %entry
; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfnmadd213ss %xmm2, %xmm1, %xmm0 {%k1} ## encoding: [0x62,0xf2,0x75,0x09,0xad,0xc2]
-; X64-NEXT: ## xmm0 = -(xmm1 * xmm0) + xmm2
+; X64-NEXT: ## xmm0 {%k1} = -(xmm1 * xmm0) + xmm2
; X64-NEXT: retq ## encoding: [0xc3]
entry:
%0 = extractelement <4 x float> %__W, i64 0
@@ -2250,14 +2250,14 @@ define <4 x float> @test_mm_mask_fnmadd_round_ss(<4 x float> %__W, i8 zeroext %_
; X86-NEXT: movb {{[0-9]+}}(%esp), %al ## encoding: [0x8a,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfnmadd213ss %xmm2, %xmm1, %xmm0 {%k1} ## encoding: [0x62,0xf2,0x75,0x09,0xad,0xc2]
-; X86-NEXT: ## xmm0 = -(xmm1 * xmm0) + xmm2
+; X86-NEXT: ## xmm0 {%k1} = -(xmm1 * xmm0) + xmm2
; X86-NEXT: retl ## encoding: [0xc3]
;
; X64-LABEL: test_mm_mask_fnmadd_round_ss:
; X64: ## %bb.0: ## %entry
; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfnmadd213ss %xmm2, %xmm1, %xmm0 {%k1} ## encoding: [0x62,0xf2,0x75,0x09,0xad,0xc2]
-; X64-NEXT: ## xmm0 = -(xmm1 * xmm0) + xmm2
+; X64-NEXT: ## xmm0 {%k1} = -(xmm1 * xmm0) + xmm2
; X64-NEXT: retq ## encoding: [0xc3]
entry:
%sub = fsub <4 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %__A
@@ -2271,14 +2271,14 @@ define <4 x float> @test_mm_maskz_fnmadd_ss(i8 zeroext %__U, <4 x float> %__A, <
; X86-NEXT: movb {{[0-9]+}}(%esp), %al ## encoding: [0x8a,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfnmadd213ss %xmm2, %xmm1, %xmm0 {%k1} {z} ## encoding: [0x62,0xf2,0x75,0x89,0xad,0xc2]
-; X86-NEXT: ## xmm0 = -(xmm1 * xmm0) + xmm2
+; X86-NEXT: ## xmm0 {%k1} {z} = -(xmm1 * xmm0) + xmm2
; X86-NEXT: retl ## encoding: [0xc3]
;
; X64-LABEL: test_mm_maskz_fnmadd_ss:
; X64: ## %bb.0: ## %entry
; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfnmadd213ss %xmm2, %xmm1, %xmm0 {%k1} {z} ## encoding: [0x62,0xf2,0x75,0x89,0xad,0xc2]
-; X64-NEXT: ## xmm0 = -(xmm1 * xmm0) + xmm2
+; X64-NEXT: ## xmm0 {%k1} {z} = -(xmm1 * xmm0) + xmm2
; X64-NEXT: retq ## encoding: [0xc3]
entry:
%0 = extractelement <4 x float> %__A, i64 0
@@ -2299,14 +2299,14 @@ define <4 x float> @test_mm_maskz_fnmadd_round_ss(i8 zeroext %__U, <4 x float> %
; X86-NEXT: movb {{[0-9]+}}(%esp), %al ## encoding: [0x8a,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfnmadd213ss %xmm2, %xmm1, %xmm0 {%k1} {z} ## encoding: [0x62,0xf2,0x75,0x89,0xad,0xc2]
-; X86-NEXT: ## xmm0 = -(xmm1 * xmm0) + xmm2
+; X86-NEXT: ## xmm0 {%k1} {z} = -(xmm1 * xmm0) + xmm2
; X86-NEXT: retl ## encoding: [0xc3]
;
; X64-LABEL: test_mm_maskz_fnmadd_round_ss:
; X64: ## %bb.0: ## %entry
; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfnmadd213ss %xmm2, %xmm1, %xmm0 {%k1} {z} ## encoding: [0x62,0xf2,0x75,0x89,0xad,0xc2]
-; X64-NEXT: ## xmm0 = -(xmm1 * xmm0) + xmm2
+; X64-NEXT: ## xmm0 {%k1} {z} = -(xmm1 * xmm0) + xmm2
; X64-NEXT: retq ## encoding: [0xc3]
entry:
%sub = fsub <4 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %__B
@@ -2320,7 +2320,7 @@ define <4 x float> @test_mm_mask3_fnmadd_ss(<4 x float> %__W, <4 x float> %__X,
; X86-NEXT: movb {{[0-9]+}}(%esp), %al ## encoding: [0x8a,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfnmadd231ss %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf2,0x7d,0x09,0xbd,0xd1]
-; X86-NEXT: ## xmm2 = -(xmm0 * xmm1) + xmm2
+; X86-NEXT: ## xmm2 {%k1} = -(xmm0 * xmm1) + xmm2
; X86-NEXT: vmovaps %xmm2, %xmm0 ## encoding: [0xc5,0xf8,0x28,0xc2]
; X86-NEXT: retl ## encoding: [0xc3]
;
@@ -2328,7 +2328,7 @@ define <4 x float> @test_mm_mask3_fnmadd_ss(<4 x float> %__W, <4 x float> %__X,
; X64: ## %bb.0: ## %entry
; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfnmadd231ss %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf2,0x7d,0x09,0xbd,0xd1]
-; X64-NEXT: ## xmm2 = -(xmm0 * xmm1) + xmm2
+; X64-NEXT: ## xmm2 {%k1} = -(xmm0 * xmm1) + xmm2
; X64-NEXT: vmovaps %xmm2, %xmm0 ## encoding: [0xc5,0xf8,0x28,0xc2]
; X64-NEXT: retq ## encoding: [0xc3]
entry:
@@ -2351,7 +2351,7 @@ define <4 x float> @test_mm_mask3_fnmadd_round_ss(<4 x float> %__W, <4 x float>
; X86-NEXT: movb {{[0-9]+}}(%esp), %al ## encoding: [0x8a,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfnmadd231ss %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf2,0x7d,0x09,0xbd,0xd1]
-; X86-NEXT: ## xmm2 = -(xmm0 * xmm1) + xmm2
+; X86-NEXT: ## xmm2 {%k1} = -(xmm0 * xmm1) + xmm2
; X86-NEXT: vmovaps %xmm2, %xmm0 ## encoding: [0xc5,0xf8,0x28,0xc2]
; X86-NEXT: retl ## encoding: [0xc3]
;
@@ -2359,7 +2359,7 @@ define <4 x float> @test_mm_mask3_fnmadd_round_ss(<4 x float> %__W, <4 x float>
; X64: ## %bb.0: ## %entry
; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfnmadd231ss %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf2,0x7d,0x09,0xbd,0xd1]
-; X64-NEXT: ## xmm2 = -(xmm0 * xmm1) + xmm2
+; X64-NEXT: ## xmm2 {%k1} = -(xmm0 * xmm1) + xmm2
; X64-NEXT: vmovaps %xmm2, %xmm0 ## encoding: [0xc5,0xf8,0x28,0xc2]
; X64-NEXT: retq ## encoding: [0xc3]
entry:
@@ -2374,14 +2374,14 @@ define <4 x float> @test_mm_mask_fnmsub_ss(<4 x float> %__W, i8 zeroext %__U, <4
; X86-NEXT: movb {{[0-9]+}}(%esp), %al ## encoding: [0x8a,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfnmsub213ss %xmm2, %xmm1, %xmm0 {%k1} ## encoding: [0x62,0xf2,0x75,0x09,0xaf,0xc2]
-; X86-NEXT: ## xmm0 = -(xmm1 * xmm0) - xmm2
+; X86-NEXT: ## xmm0 {%k1} = -(xmm1 * xmm0) - xmm2
; X86-NEXT: retl ## encoding: [0xc3]
;
; X64-LABEL: test_mm_mask_fnmsub_ss:
; X64: ## %bb.0: ## %entry
; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfnmsub213ss %xmm2, %xmm1, %xmm0 {%k1} ## encoding: [0x62,0xf2,0x75,0x09,0xaf,0xc2]
-; X64-NEXT: ## xmm0 = -(xmm1 * xmm0) - xmm2
+; X64-NEXT: ## xmm0 {%k1} = -(xmm1 * xmm0) - xmm2
; X64-NEXT: retq ## encoding: [0xc3]
entry:
%0 = extractelement <4 x float> %__W, i64 0
@@ -2404,14 +2404,14 @@ define <4 x float> @test_mm_mask_fnmsub_round_ss(<4 x float> %__W, i8 zeroext %_
; X86-NEXT: movb {{[0-9]+}}(%esp), %al ## encoding: [0x8a,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfnmsub213ss %xmm2, %xmm1, %xmm0 {%k1} ## encoding: [0x62,0xf2,0x75,0x09,0xaf,0xc2]
-; X86-NEXT: ## xmm0 = -(xmm1 * xmm0) - xmm2
+; X86-NEXT: ## xmm0 {%k1} = -(xmm1 * xmm0) - xmm2
; X86-NEXT: retl ## encoding: [0xc3]
;
; X64-LABEL: test_mm_mask_fnmsub_round_ss:
; X64: ## %bb.0: ## %entry
; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfnmsub213ss %xmm2, %xmm1, %xmm0 {%k1} ## encoding: [0x62,0xf2,0x75,0x09,0xaf,0xc2]
-; X64-NEXT: ## xmm0 = -(xmm1 * xmm0) - xmm2
+; X64-NEXT: ## xmm0 {%k1} = -(xmm1 * xmm0) - xmm2
; X64-NEXT: retq ## encoding: [0xc3]
entry:
%sub = fsub <4 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %__A
@@ -2426,14 +2426,14 @@ define <4 x float> @test_mm_maskz_fnmsub_ss(i8 zeroext %__U, <4 x float> %__A, <
; X86-NEXT: movb {{[0-9]+}}(%esp), %al ## encoding: [0x8a,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfnmsub213ss %xmm2, %xmm1, %xmm0 {%k1} {z} ## encoding: [0x62,0xf2,0x75,0x89,0xaf,0xc2]
-; X86-NEXT: ## xmm0 = -(xmm1 * xmm0) - xmm2
+; X86-NEXT: ## xmm0 {%k1} {z} = -(xmm1 * xmm0) - xmm2
; X86-NEXT: retl ## encoding: [0xc3]
;
; X64-LABEL: test_mm_maskz_fnmsub_ss:
; X64: ## %bb.0: ## %entry
; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfnmsub213ss %xmm2, %xmm1, %xmm0 {%k1} {z} ## encoding: [0x62,0xf2,0x75,0x89,0xaf,0xc2]
-; X64-NEXT: ## xmm0 = -(xmm1 * xmm0) - xmm2
+; X64-NEXT: ## xmm0 {%k1} {z} = -(xmm1 * xmm0) - xmm2
; X64-NEXT: retq ## encoding: [0xc3]
entry:
%0 = extractelement <4 x float> %__A, i64 0
@@ -2455,14 +2455,14 @@ define <4 x float> @test_mm_maskz_fnmsub_round_ss(i8 zeroext %__U, <4 x float> %
; X86-NEXT: movb {{[0-9]+}}(%esp), %al ## encoding: [0x8a,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfnmsub213ss %xmm2, %xmm1, %xmm0 {%k1} {z} ## encoding: [0x62,0xf2,0x75,0x89,0xaf,0xc2]
-; X86-NEXT: ## xmm0 = -(xmm1 * xmm0) - xmm2
+; X86-NEXT: ## xmm0 {%k1} {z} = -(xmm1 * xmm0) - xmm2
; X86-NEXT: retl ## encoding: [0xc3]
;
; X64-LABEL: test_mm_maskz_fnmsub_round_ss:
; X64: ## %bb.0: ## %entry
; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfnmsub213ss %xmm2, %xmm1, %xmm0 {%k1} {z} ## encoding: [0x62,0xf2,0x75,0x89,0xaf,0xc2]
-; X64-NEXT: ## xmm0 = -(xmm1 * xmm0) - xmm2
+; X64-NEXT: ## xmm0 {%k1} {z} = -(xmm1 * xmm0) - xmm2
; X64-NEXT: retq ## encoding: [0xc3]
entry:
%sub = fsub <4 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %__B
@@ -2477,7 +2477,7 @@ define <4 x float> @test_mm_mask3_fnmsub_ss(<4 x float> %__W, <4 x float> %__X,
; X86-NEXT: movb {{[0-9]+}}(%esp), %al ## encoding: [0x8a,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfnmsub231ss %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf2,0x7d,0x09,0xbf,0xd1]
-; X86-NEXT: ## xmm2 = -(xmm0 * xmm1) - xmm2
+; X86-NEXT: ## xmm2 {%k1} = -(xmm0 * xmm1) - xmm2
; X86-NEXT: vmovaps %xmm2, %xmm0 ## encoding: [0xc5,0xf8,0x28,0xc2]
; X86-NEXT: retl ## encoding: [0xc3]
;
@@ -2485,7 +2485,7 @@ define <4 x float> @test_mm_mask3_fnmsub_ss(<4 x float> %__W, <4 x float> %__X,
; X64: ## %bb.0: ## %entry
; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfnmsub231ss %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf2,0x7d,0x09,0xbf,0xd1]
-; X64-NEXT: ## xmm2 = -(xmm0 * xmm1) - xmm2
+; X64-NEXT: ## xmm2 {%k1} = -(xmm0 * xmm1) - xmm2
; X64-NEXT: vmovaps %xmm2, %xmm0 ## encoding: [0xc5,0xf8,0x28,0xc2]
; X64-NEXT: retq ## encoding: [0xc3]
entry:
@@ -2509,7 +2509,7 @@ define <4 x float> @test_mm_mask3_fnmsub_round_ss(<4 x float> %__W, <4 x float>
; X86-NEXT: movb {{[0-9]+}}(%esp), %al ## encoding: [0x8a,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfnmsub231ss %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf2,0x7d,0x09,0xbf,0xd1]
-; X86-NEXT: ## xmm2 = -(xmm0 * xmm1) - xmm2
+; X86-NEXT: ## xmm2 {%k1} = -(xmm0 * xmm1) - xmm2
; X86-NEXT: vmovaps %xmm2, %xmm0 ## encoding: [0xc5,0xf8,0x28,0xc2]
; X86-NEXT: retl ## encoding: [0xc3]
;
@@ -2517,7 +2517,7 @@ define <4 x float> @test_mm_mask3_fnmsub_round_ss(<4 x float> %__W, <4 x float>
; X64: ## %bb.0: ## %entry
; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfnmsub231ss %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf2,0x7d,0x09,0xbf,0xd1]
-; X64-NEXT: ## xmm2 = -(xmm0 * xmm1) - xmm2
+; X64-NEXT: ## xmm2 {%k1} = -(xmm0 * xmm1) - xmm2
; X64-NEXT: vmovaps %xmm2, %xmm0 ## encoding: [0xc5,0xf8,0x28,0xc2]
; X64-NEXT: retq ## encoding: [0xc3]
entry:
@@ -2532,14 +2532,14 @@ define <2 x double> @test_mm_mask_fmadd_sd(<2 x double> %__W, i8 zeroext %__U, <
; X86-NEXT: movb {{[0-9]+}}(%esp), %al ## encoding: [0x8a,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmadd213sd %xmm2, %xmm1, %xmm0 {%k1} ## encoding: [0x62,0xf2,0xf5,0x09,0xa9,0xc2]
-; X86-NEXT: ## xmm0 = (xmm1 * xmm0) + xmm2
+; X86-NEXT: ## xmm0 {%k1} = (xmm1 * xmm0) + xmm2
; X86-NEXT: retl ## encoding: [0xc3]
;
; X64-LABEL: test_mm_mask_fmadd_sd:
; X64: ## %bb.0: ## %entry
; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmadd213sd %xmm2, %xmm1, %xmm0 {%k1} ## encoding: [0x62,0xf2,0xf5,0x09,0xa9,0xc2]
-; X64-NEXT: ## xmm0 = (xmm1 * xmm0) + xmm2
+; X64-NEXT: ## xmm0 {%k1} = (xmm1 * xmm0) + xmm2
; X64-NEXT: retq ## encoding: [0xc3]
entry:
%0 = extractelement <2 x double> %__W, i64 0
@@ -2560,14 +2560,14 @@ define <2 x double> @test_mm_mask_fmadd_round_sd(<2 x double> %__W, i8 zeroext %
; X86-NEXT: movb {{[0-9]+}}(%esp), %al ## encoding: [0x8a,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmadd213sd %xmm2, %xmm1, %xmm0 {%k1} ## encoding: [0x62,0xf2,0xf5,0x09,0xa9,0xc2]
-; X86-NEXT: ## xmm0 = (xmm1 * xmm0) + xmm2
+; X86-NEXT: ## xmm0 {%k1} = (xmm1 * xmm0) + xmm2
; X86-NEXT: retl ## encoding: [0xc3]
;
; X64-LABEL: test_mm_mask_fmadd_round_sd:
; X64: ## %bb.0: ## %entry
; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmadd213sd %xmm2, %xmm1, %xmm0 {%k1} ## encoding: [0x62,0xf2,0xf5,0x09,0xa9,0xc2]
-; X64-NEXT: ## xmm0 = (xmm1 * xmm0) + xmm2
+; X64-NEXT: ## xmm0 {%k1} = (xmm1 * xmm0) + xmm2
; X64-NEXT: retq ## encoding: [0xc3]
entry:
%0 = tail call <2 x double> @llvm.x86.avx512.mask.vfmadd.sd(<2 x double> %__W, <2 x double> %__A, <2 x double> %__B, i8 %__U, i32 4)
@@ -2582,14 +2582,14 @@ define <2 x double> @test_mm_maskz_fmadd_sd(i8 zeroext %__U, <2 x double> %__A,
; X86-NEXT: movb {{[0-9]+}}(%esp), %al ## encoding: [0x8a,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmadd213sd %xmm2, %xmm1, %xmm0 {%k1} {z} ## encoding: [0x62,0xf2,0xf5,0x89,0xa9,0xc2]
-; X86-NEXT: ## xmm0 = (xmm1 * xmm0) + xmm2
+; X86-NEXT: ## xmm0 {%k1} {z} = (xmm1 * xmm0) + xmm2
; X86-NEXT: retl ## encoding: [0xc3]
;
; X64-LABEL: test_mm_maskz_fmadd_sd:
; X64: ## %bb.0: ## %entry
; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmadd213sd %xmm2, %xmm1, %xmm0 {%k1} {z} ## encoding: [0x62,0xf2,0xf5,0x89,0xa9,0xc2]
-; X64-NEXT: ## xmm0 = (xmm1 * xmm0) + xmm2
+; X64-NEXT: ## xmm0 {%k1} {z} = (xmm1 * xmm0) + xmm2
; X64-NEXT: retq ## encoding: [0xc3]
entry:
%0 = extractelement <2 x double> %__A, i64 0
@@ -2609,14 +2609,14 @@ define <2 x double> @test_mm_maskz_fmadd_round_sd(i8 zeroext %__U, <2 x double>
; X86-NEXT: movb {{[0-9]+}}(%esp), %al ## encoding: [0x8a,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmadd213sd %xmm2, %xmm1, %xmm0 {%k1} {z} ## encoding: [0x62,0xf2,0xf5,0x89,0xa9,0xc2]
-; X86-NEXT: ## xmm0 = (xmm1 * xmm0) + xmm2
+; X86-NEXT: ## xmm0 {%k1} {z} = (xmm1 * xmm0) + xmm2
; X86-NEXT: retl ## encoding: [0xc3]
;
; X64-LABEL: test_mm_maskz_fmadd_round_sd:
; X64: ## %bb.0: ## %entry
; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmadd213sd %xmm2, %xmm1, %xmm0 {%k1} {z} ## encoding: [0x62,0xf2,0xf5,0x89,0xa9,0xc2]
-; X64-NEXT: ## xmm0 = (xmm1 * xmm0) + xmm2
+; X64-NEXT: ## xmm0 {%k1} {z} = (xmm1 * xmm0) + xmm2
; X64-NEXT: retq ## encoding: [0xc3]
entry:
%0 = tail call <2 x double> @llvm.x86.avx512.maskz.vfmadd.sd(<2 x double> %__A, <2 x double> %__B, <2 x double> %__C, i8 %__U, i32 4)
@@ -2631,7 +2631,7 @@ define <2 x double> @test_mm_mask3_fmadd_sd(<2 x double> %__W, <2 x double> %__X
; X86-NEXT: movb {{[0-9]+}}(%esp), %al ## encoding: [0x8a,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmadd231sd %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf2,0xfd,0x09,0xb9,0xd1]
-; X86-NEXT: ## xmm2 = (xmm0 * xmm1) + xmm2
+; X86-NEXT: ## xmm2 {%k1} = (xmm0 * xmm1) + xmm2
; X86-NEXT: vmovapd %xmm2, %xmm0 ## encoding: [0xc5,0xf9,0x28,0xc2]
; X86-NEXT: retl ## encoding: [0xc3]
;
@@ -2639,7 +2639,7 @@ define <2 x double> @test_mm_mask3_fmadd_sd(<2 x double> %__W, <2 x double> %__X
; X64: ## %bb.0: ## %entry
; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmadd231sd %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf2,0xfd,0x09,0xb9,0xd1]
-; X64-NEXT: ## xmm2 = (xmm0 * xmm1) + xmm2
+; X64-NEXT: ## xmm2 {%k1} = (xmm0 * xmm1) + xmm2
; X64-NEXT: vmovapd %xmm2, %xmm0 ## encoding: [0xc5,0xf9,0x28,0xc2]
; X64-NEXT: retq ## encoding: [0xc3]
entry:
@@ -2661,7 +2661,7 @@ define <2 x double> @test_mm_mask3_fmadd_round_sd(<2 x double> %__W, <2 x double
; X86-NEXT: movb {{[0-9]+}}(%esp), %al ## encoding: [0x8a,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmadd231sd %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf2,0xfd,0x09,0xb9,0xd1]
-; X86-NEXT: ## xmm2 = (xmm0 * xmm1) + xmm2
+; X86-NEXT: ## xmm2 {%k1} = (xmm0 * xmm1) + xmm2
; X86-NEXT: vmovapd %xmm2, %xmm0 ## encoding: [0xc5,0xf9,0x28,0xc2]
; X86-NEXT: retl ## encoding: [0xc3]
;
@@ -2669,7 +2669,7 @@ define <2 x double> @test_mm_mask3_fmadd_round_sd(<2 x double> %__W, <2 x double
; X64: ## %bb.0: ## %entry
; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmadd231sd %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf2,0xfd,0x09,0xb9,0xd1]
-; X64-NEXT: ## xmm2 = (xmm0 * xmm1) + xmm2
+; X64-NEXT: ## xmm2 {%k1} = (xmm0 * xmm1) + xmm2
; X64-NEXT: vmovapd %xmm2, %xmm0 ## encoding: [0xc5,0xf9,0x28,0xc2]
; X64-NEXT: retq ## encoding: [0xc3]
entry:
@@ -2685,14 +2685,14 @@ define <2 x double> @test_mm_mask_fmsub_sd(<2 x double> %__W, i8 zeroext %__U, <
; X86-NEXT: movb {{[0-9]+}}(%esp), %al ## encoding: [0x8a,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmsub213sd %xmm2, %xmm1, %xmm0 {%k1} ## encoding: [0x62,0xf2,0xf5,0x09,0xab,0xc2]
-; X86-NEXT: ## xmm0 = (xmm1 * xmm0) - xmm2
+; X86-NEXT: ## xmm0 {%k1} = (xmm1 * xmm0) - xmm2
; X86-NEXT: retl ## encoding: [0xc3]
;
; X64-LABEL: test_mm_mask_fmsub_sd:
; X64: ## %bb.0: ## %entry
; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmsub213sd %xmm2, %xmm1, %xmm0 {%k1} ## encoding: [0x62,0xf2,0xf5,0x09,0xab,0xc2]
-; X64-NEXT: ## xmm0 = (xmm1 * xmm0) - xmm2
+; X64-NEXT: ## xmm0 {%k1} = (xmm1 * xmm0) - xmm2
; X64-NEXT: retq ## encoding: [0xc3]
entry:
%0 = extractelement <2 x double> %__W, i64 0
@@ -2714,14 +2714,14 @@ define <2 x double> @test_mm_mask_fmsub_round_sd(<2 x double> %__W, i8 zeroext %
; X86-NEXT: movb {{[0-9]+}}(%esp), %al ## encoding: [0x8a,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmsub213sd %xmm2, %xmm1, %xmm0 {%k1} ## encoding: [0x62,0xf2,0xf5,0x09,0xab,0xc2]
-; X86-NEXT: ## xmm0 = (xmm1 * xmm0) - xmm2
+; X86-NEXT: ## xmm0 {%k1} = (xmm1 * xmm0) - xmm2
; X86-NEXT: retl ## encoding: [0xc3]
;
; X64-LABEL: test_mm_mask_fmsub_round_sd:
; X64: ## %bb.0: ## %entry
; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmsub213sd %xmm2, %xmm1, %xmm0 {%k1} ## encoding: [0x62,0xf2,0xf5,0x09,0xab,0xc2]
-; X64-NEXT: ## xmm0 = (xmm1 * xmm0) - xmm2
+; X64-NEXT: ## xmm0 {%k1} = (xmm1 * xmm0) - xmm2
; X64-NEXT: retq ## encoding: [0xc3]
entry:
%sub = fsub <2 x double> <double -0.000000e+00, double -0.000000e+00>, %__B
@@ -2735,14 +2735,14 @@ define <2 x double> @test_mm_maskz_fmsub_sd(i8 zeroext %__U, <2 x double> %__A,
; X86-NEXT: movb {{[0-9]+}}(%esp), %al ## encoding: [0x8a,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmsub213sd %xmm2, %xmm1, %xmm0 {%k1} {z} ## encoding: [0x62,0xf2,0xf5,0x89,0xab,0xc2]
-; X86-NEXT: ## xmm0 = (xmm1 * xmm0) - xmm2
+; X86-NEXT: ## xmm0 {%k1} {z} = (xmm1 * xmm0) - xmm2
; X86-NEXT: retl ## encoding: [0xc3]
;
; X64-LABEL: test_mm_maskz_fmsub_sd:
; X64: ## %bb.0: ## %entry
; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmsub213sd %xmm2, %xmm1, %xmm0 {%k1} {z} ## encoding: [0x62,0xf2,0xf5,0x89,0xab,0xc2]
-; X64-NEXT: ## xmm0 = (xmm1 * xmm0) - xmm2
+; X64-NEXT: ## xmm0 {%k1} {z} = (xmm1 * xmm0) - xmm2
; X64-NEXT: retq ## encoding: [0xc3]
entry:
%0 = extractelement <2 x double> %__A, i64 0
@@ -2763,14 +2763,14 @@ define <2 x double> @test_mm_maskz_fmsub_round_sd(i8 zeroext %__U, <2 x double>
; X86-NEXT: movb {{[0-9]+}}(%esp), %al ## encoding: [0x8a,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmsub213sd %xmm2, %xmm1, %xmm0 {%k1} {z} ## encoding: [0x62,0xf2,0xf5,0x89,0xab,0xc2]
-; X86-NEXT: ## xmm0 = (xmm1 * xmm0) - xmm2
+; X86-NEXT: ## xmm0 {%k1} {z} = (xmm1 * xmm0) - xmm2
; X86-NEXT: retl ## encoding: [0xc3]
;
; X64-LABEL: test_mm_maskz_fmsub_round_sd:
; X64: ## %bb.0: ## %entry
; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmsub213sd %xmm2, %xmm1, %xmm0 {%k1} {z} ## encoding: [0x62,0xf2,0xf5,0x89,0xab,0xc2]
-; X64-NEXT: ## xmm0 = (xmm1 * xmm0) - xmm2
+; X64-NEXT: ## xmm0 {%k1} {z} = (xmm1 * xmm0) - xmm2
; X64-NEXT: retq ## encoding: [0xc3]
entry:
%sub = fsub <2 x double> <double -0.000000e+00, double -0.000000e+00>, %__C
@@ -2784,7 +2784,7 @@ define <2 x double> @test_mm_mask3_fmsub_sd(<2 x double> %__W, <2 x double> %__X
; X86-NEXT: movb {{[0-9]+}}(%esp), %al ## encoding: [0x8a,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmsub231sd %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf2,0xfd,0x09,0xbb,0xd1]
-; X86-NEXT: ## xmm2 = (xmm0 * xmm1) - xmm2
+; X86-NEXT: ## xmm2 {%k1} = (xmm0 * xmm1) - xmm2
; X86-NEXT: vmovapd %xmm2, %xmm0 ## encoding: [0xc5,0xf9,0x28,0xc2]
; X86-NEXT: retl ## encoding: [0xc3]
;
@@ -2792,7 +2792,7 @@ define <2 x double> @test_mm_mask3_fmsub_sd(<2 x double> %__W, <2 x double> %__X
; X64: ## %bb.0: ## %entry
; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmsub231sd %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf2,0xfd,0x09,0xbb,0xd1]
-; X64-NEXT: ## xmm2 = (xmm0 * xmm1) - xmm2
+; X64-NEXT: ## xmm2 {%k1} = (xmm0 * xmm1) - xmm2
; X64-NEXT: vmovapd %xmm2, %xmm0 ## encoding: [0xc5,0xf9,0x28,0xc2]
; X64-NEXT: retq ## encoding: [0xc3]
entry:
@@ -2815,7 +2815,7 @@ define <2 x double> @test_mm_mask3_fmsub_round_sd(<2 x double> %__W, <2 x double
; X86-NEXT: movb {{[0-9]+}}(%esp), %al ## encoding: [0x8a,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmsub231sd %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf2,0xfd,0x09,0xbb,0xd1]
-; X86-NEXT: ## xmm2 = (xmm0 * xmm1) - xmm2
+; X86-NEXT: ## xmm2 {%k1} = (xmm0 * xmm1) - xmm2
; X86-NEXT: vmovapd %xmm2, %xmm0 ## encoding: [0xc5,0xf9,0x28,0xc2]
; X86-NEXT: retl ## encoding: [0xc3]
;
@@ -2823,7 +2823,7 @@ define <2 x double> @test_mm_mask3_fmsub_round_sd(<2 x double> %__W, <2 x double
; X64: ## %bb.0: ## %entry
; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmsub231sd %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf2,0xfd,0x09,0xbb,0xd1]
-; X64-NEXT: ## xmm2 = (xmm0 * xmm1) - xmm2
+; X64-NEXT: ## xmm2 {%k1} = (xmm0 * xmm1) - xmm2
; X64-NEXT: vmovapd %xmm2, %xmm0 ## encoding: [0xc5,0xf9,0x28,0xc2]
; X64-NEXT: retq ## encoding: [0xc3]
entry:
@@ -2839,14 +2839,14 @@ define <2 x double> @test_mm_mask_fnmadd_sd(<2 x double> %__W, i8 zeroext %__U,
; X86-NEXT: movb {{[0-9]+}}(%esp), %al ## encoding: [0x8a,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfnmadd213sd %xmm2, %xmm1, %xmm0 {%k1} ## encoding: [0x62,0xf2,0xf5,0x09,0xad,0xc2]
-; X86-NEXT: ## xmm0 = -(xmm1 * xmm0) + xmm2
+; X86-NEXT: ## xmm0 {%k1} = -(xmm1 * xmm0) + xmm2
; X86-NEXT: retl ## encoding: [0xc3]
;
; X64-LABEL: test_mm_mask_fnmadd_sd:
; X64: ## %bb.0: ## %entry
; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfnmadd213sd %xmm2, %xmm1, %xmm0 {%k1} ## encoding: [0x62,0xf2,0xf5,0x09,0xad,0xc2]
-; X64-NEXT: ## xmm0 = -(xmm1 * xmm0) + xmm2
+; X64-NEXT: ## xmm0 {%k1} = -(xmm1 * xmm0) + xmm2
; X64-NEXT: retq ## encoding: [0xc3]
entry:
%0 = extractelement <2 x double> %__W, i64 0
@@ -2868,14 +2868,14 @@ define <2 x double> @test_mm_mask_fnmadd_round_sd(<2 x double> %__W, i8 zeroext
; X86-NEXT: movb {{[0-9]+}}(%esp), %al ## encoding: [0x8a,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfnmadd213sd %xmm2, %xmm1, %xmm0 {%k1} ## encoding: [0x62,0xf2,0xf5,0x09,0xad,0xc2]
-; X86-NEXT: ## xmm0 = -(xmm1 * xmm0) + xmm2
+; X86-NEXT: ## xmm0 {%k1} = -(xmm1 * xmm0) + xmm2
; X86-NEXT: retl ## encoding: [0xc3]
;
; X64-LABEL: test_mm_mask_fnmadd_round_sd:
; X64: ## %bb.0: ## %entry
; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfnmadd213sd %xmm2, %xmm1, %xmm0 {%k1} ## encoding: [0x62,0xf2,0xf5,0x09,0xad,0xc2]
-; X64-NEXT: ## xmm0 = -(xmm1 * xmm0) + xmm2
+; X64-NEXT: ## xmm0 {%k1} = -(xmm1 * xmm0) + xmm2
; X64-NEXT: retq ## encoding: [0xc3]
entry:
%sub = fsub <2 x double> <double -0.000000e+00, double -0.000000e+00>, %__A
@@ -2889,14 +2889,14 @@ define <2 x double> @test_mm_maskz_fnmadd_sd(i8 zeroext %__U, <2 x double> %__A,
; X86-NEXT: movb {{[0-9]+}}(%esp), %al ## encoding: [0x8a,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfnmadd213sd %xmm2, %xmm1, %xmm0 {%k1} {z} ## encoding: [0x62,0xf2,0xf5,0x89,0xad,0xc2]
-; X86-NEXT: ## xmm0 = -(xmm1 * xmm0) + xmm2
+; X86-NEXT: ## xmm0 {%k1} {z} = -(xmm1 * xmm0) + xmm2
; X86-NEXT: retl ## encoding: [0xc3]
;
; X64-LABEL: test_mm_maskz_fnmadd_sd:
; X64: ## %bb.0: ## %entry
; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfnmadd213sd %xmm2, %xmm1, %xmm0 {%k1} {z} ## encoding: [0x62,0xf2,0xf5,0x89,0xad,0xc2]
-; X64-NEXT: ## xmm0 = -(xmm1 * xmm0) + xmm2
+; X64-NEXT: ## xmm0 {%k1} {z} = -(xmm1 * xmm0) + xmm2
; X64-NEXT: retq ## encoding: [0xc3]
entry:
%0 = extractelement <2 x double> %__A, i64 0
@@ -2917,14 +2917,14 @@ define <2 x double> @test_mm_maskz_fnmadd_round_sd(i8 zeroext %__U, <2 x double>
; X86-NEXT: movb {{[0-9]+}}(%esp), %al ## encoding: [0x8a,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfnmadd213sd %xmm2, %xmm1, %xmm0 {%k1} {z} ## encoding: [0x62,0xf2,0xf5,0x89,0xad,0xc2]
-; X86-NEXT: ## xmm0 = -(xmm1 * xmm0) + xmm2
+; X86-NEXT: ## xmm0 {%k1} {z} = -(xmm1 * xmm0) + xmm2
; X86-NEXT: retl ## encoding: [0xc3]
;
; X64-LABEL: test_mm_maskz_fnmadd_round_sd:
; X64: ## %bb.0: ## %entry
; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfnmadd213sd %xmm2, %xmm1, %xmm0 {%k1} {z} ## encoding: [0x62,0xf2,0xf5,0x89,0xad,0xc2]
-; X64-NEXT: ## xmm0 = -(xmm1 * xmm0) + xmm2
+; X64-NEXT: ## xmm0 {%k1} {z} = -(xmm1 * xmm0) + xmm2
; X64-NEXT: retq ## encoding: [0xc3]
entry:
%sub = fsub <2 x double> <double -0.000000e+00, double -0.000000e+00>, %__B
@@ -2938,7 +2938,7 @@ define <2 x double> @test_mm_mask3_fnmadd_sd(<2 x double> %__W, <2 x double> %__
; X86-NEXT: movb {{[0-9]+}}(%esp), %al ## encoding: [0x8a,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfnmadd231sd %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf2,0xfd,0x09,0xbd,0xd1]
-; X86-NEXT: ## xmm2 = -(xmm0 * xmm1) + xmm2
+; X86-NEXT: ## xmm2 {%k1} = -(xmm0 * xmm1) + xmm2
; X86-NEXT: vmovapd %xmm2, %xmm0 ## encoding: [0xc5,0xf9,0x28,0xc2]
; X86-NEXT: retl ## encoding: [0xc3]
;
@@ -2946,7 +2946,7 @@ define <2 x double> @test_mm_mask3_fnmadd_sd(<2 x double> %__W, <2 x double> %__
; X64: ## %bb.0: ## %entry
; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfnmadd231sd %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf2,0xfd,0x09,0xbd,0xd1]
-; X64-NEXT: ## xmm2 = -(xmm0 * xmm1) + xmm2
+; X64-NEXT: ## xmm2 {%k1} = -(xmm0 * xmm1) + xmm2
; X64-NEXT: vmovapd %xmm2, %xmm0 ## encoding: [0xc5,0xf9,0x28,0xc2]
; X64-NEXT: retq ## encoding: [0xc3]
entry:
@@ -2969,7 +2969,7 @@ define <2 x double> @test_mm_mask3_fnmadd_round_sd(<2 x double> %__W, <2 x doubl
; X86-NEXT: movb {{[0-9]+}}(%esp), %al ## encoding: [0x8a,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfnmadd231sd %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf2,0xfd,0x09,0xbd,0xd1]
-; X86-NEXT: ## xmm2 = -(xmm0 * xmm1) + xmm2
+; X86-NEXT: ## xmm2 {%k1} = -(xmm0 * xmm1) + xmm2
; X86-NEXT: vmovapd %xmm2, %xmm0 ## encoding: [0xc5,0xf9,0x28,0xc2]
; X86-NEXT: retl ## encoding: [0xc3]
;
@@ -2977,7 +2977,7 @@ define <2 x double> @test_mm_mask3_fnmadd_round_sd(<2 x double> %__W, <2 x doubl
; X64: ## %bb.0: ## %entry
; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfnmadd231sd %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf2,0xfd,0x09,0xbd,0xd1]
-; X64-NEXT: ## xmm2 = -(xmm0 * xmm1) + xmm2
+; X64-NEXT: ## xmm2 {%k1} = -(xmm0 * xmm1) + xmm2
; X64-NEXT: vmovapd %xmm2, %xmm0 ## encoding: [0xc5,0xf9,0x28,0xc2]
; X64-NEXT: retq ## encoding: [0xc3]
entry:
@@ -2992,14 +2992,14 @@ define <2 x double> @test_mm_mask_fnmsub_sd(<2 x double> %__W, i8 zeroext %__U,
; X86-NEXT: movb {{[0-9]+}}(%esp), %al ## encoding: [0x8a,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfnmsub213sd %xmm2, %xmm1, %xmm0 {%k1} ## encoding: [0x62,0xf2,0xf5,0x09,0xaf,0xc2]
-; X86-NEXT: ## xmm0 = -(xmm1 * xmm0) - xmm2
+; X86-NEXT: ## xmm0 {%k1} = -(xmm1 * xmm0) - xmm2
; X86-NEXT: retl ## encoding: [0xc3]
;
; X64-LABEL: test_mm_mask_fnmsub_sd:
; X64: ## %bb.0: ## %entry
; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfnmsub213sd %xmm2, %xmm1, %xmm0 {%k1} ## encoding: [0x62,0xf2,0xf5,0x09,0xaf,0xc2]
-; X64-NEXT: ## xmm0 = -(xmm1 * xmm0) - xmm2
+; X64-NEXT: ## xmm0 {%k1} = -(xmm1 * xmm0) - xmm2
; X64-NEXT: retq ## encoding: [0xc3]
entry:
%0 = extractelement <2 x double> %__W, i64 0
@@ -3022,14 +3022,14 @@ define <2 x double> @test_mm_mask_fnmsub_round_sd(<2 x double> %__W, i8 zeroext
; X86-NEXT: movb {{[0-9]+}}(%esp), %al ## encoding: [0x8a,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfnmsub213sd %xmm2, %xmm1, %xmm0 {%k1} ## encoding: [0x62,0xf2,0xf5,0x09,0xaf,0xc2]
-; X86-NEXT: ## xmm0 = -(xmm1 * xmm0) - xmm2
+; X86-NEXT: ## xmm0 {%k1} = -(xmm1 * xmm0) - xmm2
; X86-NEXT: retl ## encoding: [0xc3]
;
; X64-LABEL: test_mm_mask_fnmsub_round_sd:
; X64: ## %bb.0: ## %entry
; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfnmsub213sd %xmm2, %xmm1, %xmm0 {%k1} ## encoding: [0x62,0xf2,0xf5,0x09,0xaf,0xc2]
-; X64-NEXT: ## xmm0 = -(xmm1 * xmm0) - xmm2
+; X64-NEXT: ## xmm0 {%k1} = -(xmm1 * xmm0) - xmm2
; X64-NEXT: retq ## encoding: [0xc3]
entry:
%sub = fsub <2 x double> <double -0.000000e+00, double -0.000000e+00>, %__A
@@ -3044,14 +3044,14 @@ define <2 x double> @test_mm_maskz_fnmsub_sd(i8 zeroext %__U, <2 x double> %__A,
; X86-NEXT: movb {{[0-9]+}}(%esp), %al ## encoding: [0x8a,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfnmsub213sd %xmm2, %xmm1, %xmm0 {%k1} {z} ## encoding: [0x62,0xf2,0xf5,0x89,0xaf,0xc2]
-; X86-NEXT: ## xmm0 = -(xmm1 * xmm0) - xmm2
+; X86-NEXT: ## xmm0 {%k1} {z} = -(xmm1 * xmm0) - xmm2
; X86-NEXT: retl ## encoding: [0xc3]
;
; X64-LABEL: test_mm_maskz_fnmsub_sd:
; X64: ## %bb.0: ## %entry
; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfnmsub213sd %xmm2, %xmm1, %xmm0 {%k1} {z} ## encoding: [0x62,0xf2,0xf5,0x89,0xaf,0xc2]
-; X64-NEXT: ## xmm0 = -(xmm1 * xmm0) - xmm2
+; X64-NEXT: ## xmm0 {%k1} {z} = -(xmm1 * xmm0) - xmm2
; X64-NEXT: retq ## encoding: [0xc3]
entry:
%0 = extractelement <2 x double> %__A, i64 0
@@ -3073,14 +3073,14 @@ define <2 x double> @test_mm_maskz_fnmsub_round_sd(i8 zeroext %__U, <2 x double>
; X86-NEXT: movb {{[0-9]+}}(%esp), %al ## encoding: [0x8a,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfnmsub213sd %xmm2, %xmm1, %xmm0 {%k1} {z} ## encoding: [0x62,0xf2,0xf5,0x89,0xaf,0xc2]
-; X86-NEXT: ## xmm0 = -(xmm1 * xmm0) - xmm2
+; X86-NEXT: ## xmm0 {%k1} {z} = -(xmm1 * xmm0) - xmm2
; X86-NEXT: retl ## encoding: [0xc3]
;
; X64-LABEL: test_mm_maskz_fnmsub_round_sd:
; X64: ## %bb.0: ## %entry
; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfnmsub213sd %xmm2, %xmm1, %xmm0 {%k1} {z} ## encoding: [0x62,0xf2,0xf5,0x89,0xaf,0xc2]
-; X64-NEXT: ## xmm0 = -(xmm1 * xmm0) - xmm2
+; X64-NEXT: ## xmm0 {%k1} {z} = -(xmm1 * xmm0) - xmm2
; X64-NEXT: retq ## encoding: [0xc3]
entry:
%sub = fsub <2 x double> <double -0.000000e+00, double -0.000000e+00>, %__B
@@ -3095,7 +3095,7 @@ define <2 x double> @test_mm_mask3_fnmsub_sd(<2 x double> %__W, <2 x double> %__
; X86-NEXT: movb {{[0-9]+}}(%esp), %al ## encoding: [0x8a,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfnmsub231sd %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf2,0xfd,0x09,0xbf,0xd1]
-; X86-NEXT: ## xmm2 = -(xmm0 * xmm1) - xmm2
+; X86-NEXT: ## xmm2 {%k1} = -(xmm0 * xmm1) - xmm2
; X86-NEXT: vmovapd %xmm2, %xmm0 ## encoding: [0xc5,0xf9,0x28,0xc2]
; X86-NEXT: retl ## encoding: [0xc3]
;
@@ -3103,7 +3103,7 @@ define <2 x double> @test_mm_mask3_fnmsub_sd(<2 x double> %__W, <2 x double> %__
; X64: ## %bb.0: ## %entry
; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfnmsub231sd %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf2,0xfd,0x09,0xbf,0xd1]
-; X64-NEXT: ## xmm2 = -(xmm0 * xmm1) - xmm2
+; X64-NEXT: ## xmm2 {%k1} = -(xmm0 * xmm1) - xmm2
; X64-NEXT: vmovapd %xmm2, %xmm0 ## encoding: [0xc5,0xf9,0x28,0xc2]
; X64-NEXT: retq ## encoding: [0xc3]
entry:
@@ -3127,7 +3127,7 @@ define <2 x double> @test_mm_mask3_fnmsub_round_sd(<2 x double> %__W, <2 x doubl
; X86-NEXT: movb {{[0-9]+}}(%esp), %al ## encoding: [0x8a,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfnmsub231sd %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf2,0xfd,0x09,0xbf,0xd1]
-; X86-NEXT: ## xmm2 = -(xmm0 * xmm1) - xmm2
+; X86-NEXT: ## xmm2 {%k1} = -(xmm0 * xmm1) - xmm2
; X86-NEXT: vmovapd %xmm2, %xmm0 ## encoding: [0xc5,0xf9,0x28,0xc2]
; X86-NEXT: retl ## encoding: [0xc3]
;
@@ -3135,7 +3135,7 @@ define <2 x double> @test_mm_mask3_fnmsub_round_sd(<2 x double> %__W, <2 x doubl
; X64: ## %bb.0: ## %entry
; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfnmsub231sd %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf2,0xfd,0x09,0xbf,0xd1]
-; X64-NEXT: ## xmm2 = -(xmm0 * xmm1) - xmm2
+; X64-NEXT: ## xmm2 {%k1} = -(xmm0 * xmm1) - xmm2
; X64-NEXT: vmovapd %xmm2, %xmm0 ## encoding: [0xc5,0xf9,0x28,0xc2]
; X64-NEXT: retq ## encoding: [0xc3]
entry:
diff --git a/llvm/test/CodeGen/X86/avx512-intrinsics-fast-isel.ll b/llvm/test/CodeGen/X86/avx512-intrinsics-fast-isel.ll
index 1b7220562ff2..fee195ae121f 100644
--- a/llvm/test/CodeGen/X86/avx512-intrinsics-fast-isel.ll
+++ b/llvm/test/CodeGen/X86/avx512-intrinsics-fast-isel.ll
@@ -3289,13 +3289,13 @@ define <8 x double> @test_mm512_mask_fmadd_pd(<8 x double> %__A, i8 zeroext %__U
; X86: # %bb.0: # %entry
; X86-NEXT: movb {{[0-9]+}}(%esp), %al
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfmadd132pd {{.*#+}} zmm0 = (zmm0 * zmm1) + zmm2
+; X86-NEXT: vfmadd132pd {{.*#+}} zmm0 {%k1} = (zmm0 * zmm1) + zmm2
; X86-NEXT: retl
;
; X64-LABEL: test_mm512_mask_fmadd_pd:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfmadd132pd {{.*#+}} zmm0 = (zmm0 * zmm1) + zmm2
+; X64-NEXT: vfmadd132pd {{.*#+}} zmm0 {%k1} = (zmm0 * zmm1) + zmm2
; X64-NEXT: retq
entry:
%0 = tail call <8 x double> @llvm.fma.v8f64(<8 x double> %__A, <8 x double> %__B, <8 x double> %__C) #10
@@ -3309,14 +3309,14 @@ define <8 x double> @test_mm512_mask3_fmadd_pd(<8 x double> %__A, <8 x double> %
; X86: # %bb.0: # %entry
; X86-NEXT: movb {{[0-9]+}}(%esp), %al
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfmadd231pd {{.*#+}} zmm2 = (zmm0 * zmm1) + zmm2
+; X86-NEXT: vfmadd231pd {{.*#+}} zmm2 {%k1} = (zmm0 * zmm1) + zmm2
; X86-NEXT: vmovapd %zmm2, %zmm0
; X86-NEXT: retl
;
; X64-LABEL: test_mm512_mask3_fmadd_pd:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfmadd231pd {{.*#+}} zmm2 = (zmm0 * zmm1) + zmm2
+; X64-NEXT: vfmadd231pd {{.*#+}} zmm2 {%k1} = (zmm0 * zmm1) + zmm2
; X64-NEXT: vmovapd %zmm2, %zmm0
; X64-NEXT: retq
entry:
@@ -3331,13 +3331,13 @@ define <8 x double> @test_mm512_maskz_fmadd_pd(i8 zeroext %__U, <8 x double> %__
; X86: # %bb.0: # %entry
; X86-NEXT: movb {{[0-9]+}}(%esp), %al
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfmadd213pd {{.*#+}} zmm0 = (zmm1 * zmm0) + zmm2
+; X86-NEXT: vfmadd213pd {{.*#+}} zmm0 {%k1} {z} = (zmm1 * zmm0) + zmm2
; X86-NEXT: retl
;
; X64-LABEL: test_mm512_maskz_fmadd_pd:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfmadd213pd {{.*#+}} zmm0 = (zmm1 * zmm0) + zmm2
+; X64-NEXT: vfmadd213pd {{.*#+}} zmm0 {%k1} {z} = (zmm1 * zmm0) + zmm2
; X64-NEXT: retq
entry:
%0 = tail call <8 x double> @llvm.fma.v8f64(<8 x double> %__A, <8 x double> %__B, <8 x double> %__C) #10
@@ -3369,13 +3369,13 @@ define <8 x double> @test_mm512_mask_fmsub_pd(<8 x double> %__A, i8 zeroext %__U
; X86: # %bb.0: # %entry
; X86-NEXT: movb {{[0-9]+}}(%esp), %al
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfmsub132pd {{.*#+}} zmm0 = (zmm0 * zmm1) - zmm2
+; X86-NEXT: vfmsub132pd {{.*#+}} zmm0 {%k1} = (zmm0 * zmm1) - zmm2
; X86-NEXT: retl
;
; X64-LABEL: test_mm512_mask_fmsub_pd:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfmsub132pd {{.*#+}} zmm0 = (zmm0 * zmm1) - zmm2
+; X64-NEXT: vfmsub132pd {{.*#+}} zmm0 {%k1} = (zmm0 * zmm1) - zmm2
; X64-NEXT: retq
entry:
%sub.i = fsub <8 x double> <double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00>, %__C
@@ -3390,13 +3390,13 @@ define <8 x double> @test_mm512_maskz_fmsub_pd(i8 zeroext %__U, <8 x double> %__
; X86: # %bb.0: # %entry
; X86-NEXT: movb {{[0-9]+}}(%esp), %al
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfmsub213pd {{.*#+}} zmm0 = (zmm1 * zmm0) - zmm2
+; X86-NEXT: vfmsub213pd {{.*#+}} zmm0 {%k1} {z} = (zmm1 * zmm0) - zmm2
; X86-NEXT: retl
;
; X64-LABEL: test_mm512_maskz_fmsub_pd:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfmsub213pd {{.*#+}} zmm0 = (zmm1 * zmm0) - zmm2
+; X64-NEXT: vfmsub213pd {{.*#+}} zmm0 {%k1} {z} = (zmm1 * zmm0) - zmm2
; X64-NEXT: retq
entry:
%sub.i = fsub <8 x double> <double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00>, %__C
@@ -3429,14 +3429,14 @@ define <8 x double> @test_mm512_mask3_fnmadd_pd(<8 x double> %__A, <8 x double>
; X86: # %bb.0: # %entry
; X86-NEXT: movb {{[0-9]+}}(%esp), %al
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfnmadd231pd {{.*#+}} zmm2 = -(zmm0 * zmm1) + zmm2
+; X86-NEXT: vfnmadd231pd {{.*#+}} zmm2 {%k1} = -(zmm0 * zmm1) + zmm2
; X86-NEXT: vmovapd %zmm2, %zmm0
; X86-NEXT: retl
;
; X64-LABEL: test_mm512_mask3_fnmadd_pd:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfnmadd231pd {{.*#+}} zmm2 = -(zmm0 * zmm1) + zmm2
+; X64-NEXT: vfnmadd231pd {{.*#+}} zmm2 {%k1} = -(zmm0 * zmm1) + zmm2
; X64-NEXT: vmovapd %zmm2, %zmm0
; X64-NEXT: retq
entry:
@@ -3452,13 +3452,13 @@ define <8 x double> @test_mm512_maskz_fnmadd_pd(i8 zeroext %__U, <8 x double> %_
; X86: # %bb.0: # %entry
; X86-NEXT: movb {{[0-9]+}}(%esp), %al
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfnmadd213pd {{.*#+}} zmm0 = -(zmm1 * zmm0) + zmm2
+; X86-NEXT: vfnmadd213pd {{.*#+}} zmm0 {%k1} {z} = -(zmm1 * zmm0) + zmm2
; X86-NEXT: retl
;
; X64-LABEL: test_mm512_maskz_fnmadd_pd:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfnmadd213pd {{.*#+}} zmm0 = -(zmm1 * zmm0) + zmm2
+; X64-NEXT: vfnmadd213pd {{.*#+}} zmm0 {%k1} {z} = -(zmm1 * zmm0) + zmm2
; X64-NEXT: retq
entry:
%sub.i = fsub <8 x double> <double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00>, %__A
@@ -3488,13 +3488,13 @@ define <8 x double> @test_mm512_maskz_fnmsub_pd(i8 zeroext %__U, <8 x double> %_
; X86: # %bb.0: # %entry
; X86-NEXT: movb {{[0-9]+}}(%esp), %al
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfnmsub213pd {{.*#+}} zmm0 = -(zmm1 * zmm0) - zmm2
+; X86-NEXT: vfnmsub213pd {{.*#+}} zmm0 {%k1} {z} = -(zmm1 * zmm0) - zmm2
; X86-NEXT: retl
;
; X64-LABEL: test_mm512_maskz_fnmsub_pd:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfnmsub213pd {{.*#+}} zmm0 = -(zmm1 * zmm0) - zmm2
+; X64-NEXT: vfnmsub213pd {{.*#+}} zmm0 {%k1} {z} = -(zmm1 * zmm0) - zmm2
; X64-NEXT: retq
entry:
%sub.i = fsub <8 x double> <double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00>, %__A
@@ -3753,13 +3753,13 @@ define <16 x float> @test_mm512_mask_fmadd_ps(<16 x float> %__A, i16 zeroext %__
; X86: # %bb.0: # %entry
; X86-NEXT: movzwl {{[0-9]+}}(%esp), %eax
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfmadd132ps {{.*#+}} zmm0 = (zmm0 * zmm1) + zmm2
+; X86-NEXT: vfmadd132ps {{.*#+}} zmm0 {%k1} = (zmm0 * zmm1) + zmm2
; X86-NEXT: retl
;
; X64-LABEL: test_mm512_mask_fmadd_ps:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfmadd132ps {{.*#+}} zmm0 = (zmm0 * zmm1) + zmm2
+; X64-NEXT: vfmadd132ps {{.*#+}} zmm0 {%k1} = (zmm0 * zmm1) + zmm2
; X64-NEXT: retq
entry:
%0 = tail call <16 x float> @llvm.fma.v16f32(<16 x float> %__A, <16 x float> %__B, <16 x float> %__C) #10
@@ -3773,14 +3773,14 @@ define <16 x float> @test_mm512_mask3_fmadd_ps(<16 x float> %__A, <16 x float> %
; X86: # %bb.0: # %entry
; X86-NEXT: movzwl {{[0-9]+}}(%esp), %eax
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfmadd231ps {{.*#+}} zmm2 = (zmm0 * zmm1) + zmm2
+; X86-NEXT: vfmadd231ps {{.*#+}} zmm2 {%k1} = (zmm0 * zmm1) + zmm2
; X86-NEXT: vmovaps %zmm2, %zmm0
; X86-NEXT: retl
;
; X64-LABEL: test_mm512_mask3_fmadd_ps:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfmadd231ps {{.*#+}} zmm2 = (zmm0 * zmm1) + zmm2
+; X64-NEXT: vfmadd231ps {{.*#+}} zmm2 {%k1} = (zmm0 * zmm1) + zmm2
; X64-NEXT: vmovaps %zmm2, %zmm0
; X64-NEXT: retq
entry:
@@ -3795,13 +3795,13 @@ define <16 x float> @test_mm512_maskz_fmadd_ps(i16 zeroext %__U, <16 x float> %_
; X86: # %bb.0: # %entry
; X86-NEXT: movzwl {{[0-9]+}}(%esp), %eax
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfmadd213ps {{.*#+}} zmm0 = (zmm1 * zmm0) + zmm2
+; X86-NEXT: vfmadd213ps {{.*#+}} zmm0 {%k1} {z} = (zmm1 * zmm0) + zmm2
; X86-NEXT: retl
;
; X64-LABEL: test_mm512_maskz_fmadd_ps:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfmadd213ps {{.*#+}} zmm0 = (zmm1 * zmm0) + zmm2
+; X64-NEXT: vfmadd213ps {{.*#+}} zmm0 {%k1} {z} = (zmm1 * zmm0) + zmm2
; X64-NEXT: retq
entry:
%0 = tail call <16 x float> @llvm.fma.v16f32(<16 x float> %__A, <16 x float> %__B, <16 x float> %__C) #10
@@ -3833,13 +3833,13 @@ define <16 x float> @test_mm512_mask_fmsub_ps(<16 x float> %__A, i16 zeroext %__
; X86: # %bb.0: # %entry
; X86-NEXT: movzwl {{[0-9]+}}(%esp), %eax
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfmsub132ps {{.*#+}} zmm0 = (zmm0 * zmm1) - zmm2
+; X86-NEXT: vfmsub132ps {{.*#+}} zmm0 {%k1} = (zmm0 * zmm1) - zmm2
; X86-NEXT: retl
;
; X64-LABEL: test_mm512_mask_fmsub_ps:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfmsub132ps {{.*#+}} zmm0 = (zmm0 * zmm1) - zmm2
+; X64-NEXT: vfmsub132ps {{.*#+}} zmm0 {%k1} = (zmm0 * zmm1) - zmm2
; X64-NEXT: retq
entry:
%sub.i = fsub <16 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %__C
@@ -3854,13 +3854,13 @@ define <16 x float> @test_mm512_maskz_fmsub_ps(i16 zeroext %__U, <16 x float> %_
; X86: # %bb.0: # %entry
; X86-NEXT: movzwl {{[0-9]+}}(%esp), %eax
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfmsub213ps {{.*#+}} zmm0 = (zmm1 * zmm0) - zmm2
+; X86-NEXT: vfmsub213ps {{.*#+}} zmm0 {%k1} {z} = (zmm1 * zmm0) - zmm2
; X86-NEXT: retl
;
; X64-LABEL: test_mm512_maskz_fmsub_ps:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfmsub213ps {{.*#+}} zmm0 = (zmm1 * zmm0) - zmm2
+; X64-NEXT: vfmsub213ps {{.*#+}} zmm0 {%k1} {z} = (zmm1 * zmm0) - zmm2
; X64-NEXT: retq
entry:
%sub.i = fsub <16 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %__C
@@ -3893,14 +3893,14 @@ define <16 x float> @test_mm512_mask3_fnmadd_ps(<16 x float> %__A, <16 x float>
; X86: # %bb.0: # %entry
; X86-NEXT: movzwl {{[0-9]+}}(%esp), %eax
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfnmadd231ps {{.*#+}} zmm2 = -(zmm0 * zmm1) + zmm2
+; X86-NEXT: vfnmadd231ps {{.*#+}} zmm2 {%k1} = -(zmm0 * zmm1) + zmm2
; X86-NEXT: vmovaps %zmm2, %zmm0
; X86-NEXT: retl
;
; X64-LABEL: test_mm512_mask3_fnmadd_ps:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfnmadd231ps {{.*#+}} zmm2 = -(zmm0 * zmm1) + zmm2
+; X64-NEXT: vfnmadd231ps {{.*#+}} zmm2 {%k1} = -(zmm0 * zmm1) + zmm2
; X64-NEXT: vmovaps %zmm2, %zmm0
; X64-NEXT: retq
entry:
@@ -3916,13 +3916,13 @@ define <16 x float> @test_mm512_maskz_fnmadd_ps(i16 zeroext %__U, <16 x float> %
; X86: # %bb.0: # %entry
; X86-NEXT: movzwl {{[0-9]+}}(%esp), %eax
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfnmadd213ps {{.*#+}} zmm0 = -(zmm1 * zmm0) + zmm2
+; X86-NEXT: vfnmadd213ps {{.*#+}} zmm0 {%k1} {z} = -(zmm1 * zmm0) + zmm2
; X86-NEXT: retl
;
; X64-LABEL: test_mm512_maskz_fnmadd_ps:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfnmadd213ps {{.*#+}} zmm0 = -(zmm1 * zmm0) + zmm2
+; X64-NEXT: vfnmadd213ps {{.*#+}} zmm0 {%k1} {z} = -(zmm1 * zmm0) + zmm2
; X64-NEXT: retq
entry:
%sub.i = fsub <16 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %__A
@@ -3952,13 +3952,13 @@ define <16 x float> @test_mm512_maskz_fnmsub_ps(i16 zeroext %__U, <16 x float> %
; X86: # %bb.0: # %entry
; X86-NEXT: movzwl {{[0-9]+}}(%esp), %eax
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfnmsub213ps {{.*#+}} zmm0 = -(zmm1 * zmm0) - zmm2
+; X86-NEXT: vfnmsub213ps {{.*#+}} zmm0 {%k1} {z} = -(zmm1 * zmm0) - zmm2
; X86-NEXT: retl
;
; X64-LABEL: test_mm512_maskz_fnmsub_ps:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfnmsub213ps {{.*#+}} zmm0 = -(zmm1 * zmm0) - zmm2
+; X64-NEXT: vfnmsub213ps {{.*#+}} zmm0 {%k1} {z} = -(zmm1 * zmm0) - zmm2
; X64-NEXT: retq
entry:
%sub.i = fsub <16 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %__A
@@ -4121,13 +4121,13 @@ define <8 x double> @test_mm512_mask_fmaddsub_pd(<8 x double> %__A, i8 zeroext %
; X86: # %bb.0: # %entry
; X86-NEXT: movb {{[0-9]+}}(%esp), %al
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfmaddsub132pd {{.*#+}} zmm0 = (zmm0 * zmm1) +/- zmm2
+; X86-NEXT: vfmaddsub132pd {{.*#+}} zmm0 {%k1} = (zmm0 * zmm1) +/- zmm2
; X86-NEXT: retl
;
; X64-LABEL: test_mm512_mask_fmaddsub_pd:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfmaddsub132pd {{.*#+}} zmm0 = (zmm0 * zmm1) +/- zmm2
+; X64-NEXT: vfmaddsub132pd {{.*#+}} zmm0 {%k1} = (zmm0 * zmm1) +/- zmm2
; X64-NEXT: retq
entry:
%0 = tail call <8 x double> @llvm.fma.v8f64(<8 x double> %__A, <8 x double> %__B, <8 x double> %__C) #10
@@ -4144,14 +4144,14 @@ define <8 x double> @test_mm512_mask3_fmaddsub_pd(<8 x double> %__A, <8 x double
; X86: # %bb.0: # %entry
; X86-NEXT: movb {{[0-9]+}}(%esp), %al
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfmaddsub231pd {{.*#+}} zmm2 = (zmm0 * zmm1) +/- zmm2
+; X86-NEXT: vfmaddsub231pd {{.*#+}} zmm2 {%k1} = (zmm0 * zmm1) +/- zmm2
; X86-NEXT: vmovapd %zmm2, %zmm0
; X86-NEXT: retl
;
; X64-LABEL: test_mm512_mask3_fmaddsub_pd:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfmaddsub231pd {{.*#+}} zmm2 = (zmm0 * zmm1) +/- zmm2
+; X64-NEXT: vfmaddsub231pd {{.*#+}} zmm2 {%k1} = (zmm0 * zmm1) +/- zmm2
; X64-NEXT: vmovapd %zmm2, %zmm0
; X64-NEXT: retq
entry:
@@ -4169,13 +4169,13 @@ define <8 x double> @test_mm512_maskz_fmaddsub_pd(i8 zeroext %__U, <8 x double>
; X86: # %bb.0: # %entry
; X86-NEXT: movb {{[0-9]+}}(%esp), %al
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfmaddsub213pd {{.*#+}} zmm0 = (zmm1 * zmm0) +/- zmm2
+; X86-NEXT: vfmaddsub213pd {{.*#+}} zmm0 {%k1} {z} = (zmm1 * zmm0) +/- zmm2
; X86-NEXT: retl
;
; X64-LABEL: test_mm512_maskz_fmaddsub_pd:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfmaddsub213pd {{.*#+}} zmm0 = (zmm1 * zmm0) +/- zmm2
+; X64-NEXT: vfmaddsub213pd {{.*#+}} zmm0 {%k1} {z} = (zmm1 * zmm0) +/- zmm2
; X64-NEXT: retq
entry:
%0 = tail call <8 x double> @llvm.fma.v8f64(<8 x double> %__A, <8 x double> %__B, <8 x double> %__C) #10
@@ -4205,13 +4205,13 @@ define <8 x double> @test_mm512_mask_fmsubadd_pd(<8 x double> %__A, i8 zeroext %
; X86: # %bb.0: # %entry
; X86-NEXT: movb {{[0-9]+}}(%esp), %al
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfmsubadd132pd {{.*#+}} zmm0 = (zmm0 * zmm1) -/+ zmm2
+; X86-NEXT: vfmsubadd132pd {{.*#+}} zmm0 {%k1} = (zmm0 * zmm1) -/+ zmm2
; X86-NEXT: retl
;
; X64-LABEL: test_mm512_mask_fmsubadd_pd:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfmsubadd132pd {{.*#+}} zmm0 = (zmm0 * zmm1) -/+ zmm2
+; X64-NEXT: vfmsubadd132pd {{.*#+}} zmm0 {%k1} = (zmm0 * zmm1) -/+ zmm2
; X64-NEXT: retq
entry:
%sub.i = fsub <8 x double> <double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00>, %__C
@@ -4228,13 +4228,13 @@ define <8 x double> @test_mm512_maskz_fmsubadd_pd(i8 zeroext %__U, <8 x double>
; X86: # %bb.0: # %entry
; X86-NEXT: movb {{[0-9]+}}(%esp), %al
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfmsubadd213pd {{.*#+}} zmm0 = (zmm1 * zmm0) -/+ zmm2
+; X86-NEXT: vfmsubadd213pd {{.*#+}} zmm0 {%k1} {z} = (zmm1 * zmm0) -/+ zmm2
; X86-NEXT: retl
;
; X64-LABEL: test_mm512_maskz_fmsubadd_pd:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfmsubadd213pd {{.*#+}} zmm0 = (zmm1 * zmm0) -/+ zmm2
+; X64-NEXT: vfmsubadd213pd {{.*#+}} zmm0 {%k1} {z} = (zmm1 * zmm0) -/+ zmm2
; X64-NEXT: retq
entry:
%sub.i = fsub <8 x double> <double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00>, %__C
@@ -4398,13 +4398,13 @@ define <16 x float> @test_mm512_mask_fmaddsub_ps(<16 x float> %__A, i16 zeroext
; X86: # %bb.0: # %entry
; X86-NEXT: movzwl {{[0-9]+}}(%esp), %eax
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfmaddsub132ps {{.*#+}} zmm0 = (zmm0 * zmm1) +/- zmm2
+; X86-NEXT: vfmaddsub132ps {{.*#+}} zmm0 {%k1} = (zmm0 * zmm1) +/- zmm2
; X86-NEXT: retl
;
; X64-LABEL: test_mm512_mask_fmaddsub_ps:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfmaddsub132ps {{.*#+}} zmm0 = (zmm0 * zmm1) +/- zmm2
+; X64-NEXT: vfmaddsub132ps {{.*#+}} zmm0 {%k1} = (zmm0 * zmm1) +/- zmm2
; X64-NEXT: retq
entry:
%0 = tail call <16 x float> @llvm.fma.v16f32(<16 x float> %__A, <16 x float> %__B, <16 x float> %__C) #10
@@ -4421,14 +4421,14 @@ define <16 x float> @test_mm512_mask3_fmaddsub_ps(<16 x float> %__A, <16 x float
; X86: # %bb.0: # %entry
; X86-NEXT: movzwl {{[0-9]+}}(%esp), %eax
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfmaddsub231ps {{.*#+}} zmm2 = (zmm0 * zmm1) +/- zmm2
+; X86-NEXT: vfmaddsub231ps {{.*#+}} zmm2 {%k1} = (zmm0 * zmm1) +/- zmm2
; X86-NEXT: vmovaps %zmm2, %zmm0
; X86-NEXT: retl
;
; X64-LABEL: test_mm512_mask3_fmaddsub_ps:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfmaddsub231ps {{.*#+}} zmm2 = (zmm0 * zmm1) +/- zmm2
+; X64-NEXT: vfmaddsub231ps {{.*#+}} zmm2 {%k1} = (zmm0 * zmm1) +/- zmm2
; X64-NEXT: vmovaps %zmm2, %zmm0
; X64-NEXT: retq
entry:
@@ -4446,13 +4446,13 @@ define <16 x float> @test_mm512_maskz_fmaddsub_ps(i16 zeroext %__U, <16 x float>
; X86: # %bb.0: # %entry
; X86-NEXT: movzwl {{[0-9]+}}(%esp), %eax
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfmaddsub213ps {{.*#+}} zmm0 = (zmm1 * zmm0) +/- zmm2
+; X86-NEXT: vfmaddsub213ps {{.*#+}} zmm0 {%k1} {z} = (zmm1 * zmm0) +/- zmm2
; X86-NEXT: retl
;
; X64-LABEL: test_mm512_maskz_fmaddsub_ps:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfmaddsub213ps {{.*#+}} zmm0 = (zmm1 * zmm0) +/- zmm2
+; X64-NEXT: vfmaddsub213ps {{.*#+}} zmm0 {%k1} {z} = (zmm1 * zmm0) +/- zmm2
; X64-NEXT: retq
entry:
%0 = tail call <16 x float> @llvm.fma.v16f32(<16 x float> %__A, <16 x float> %__B, <16 x float> %__C) #10
@@ -4482,13 +4482,13 @@ define <16 x float> @test_mm512_mask_fmsubadd_ps(<16 x float> %__A, i16 zeroext
; X86: # %bb.0: # %entry
; X86-NEXT: movzwl {{[0-9]+}}(%esp), %eax
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfmsubadd132ps {{.*#+}} zmm0 = (zmm0 * zmm1) -/+ zmm2
+; X86-NEXT: vfmsubadd132ps {{.*#+}} zmm0 {%k1} = (zmm0 * zmm1) -/+ zmm2
; X86-NEXT: retl
;
; X64-LABEL: test_mm512_mask_fmsubadd_ps:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfmsubadd132ps {{.*#+}} zmm0 = (zmm0 * zmm1) -/+ zmm2
+; X64-NEXT: vfmsubadd132ps {{.*#+}} zmm0 {%k1} = (zmm0 * zmm1) -/+ zmm2
; X64-NEXT: retq
entry:
%sub.i = fsub <16 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %__C
@@ -4505,13 +4505,13 @@ define <16 x float> @test_mm512_maskz_fmsubadd_ps(i16 zeroext %__U, <16 x float>
; X86: # %bb.0: # %entry
; X86-NEXT: movzwl {{[0-9]+}}(%esp), %eax
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfmsubadd213ps {{.*#+}} zmm0 = (zmm1 * zmm0) -/+ zmm2
+; X86-NEXT: vfmsubadd213ps {{.*#+}} zmm0 {%k1} {z} = (zmm1 * zmm0) -/+ zmm2
; X86-NEXT: retl
;
; X64-LABEL: test_mm512_maskz_fmsubadd_ps:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfmsubadd213ps {{.*#+}} zmm0 = (zmm1 * zmm0) -/+ zmm2
+; X64-NEXT: vfmsubadd213ps {{.*#+}} zmm0 {%k1} {z} = (zmm1 * zmm0) -/+ zmm2
; X64-NEXT: retq
entry:
%sub.i = fsub <16 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %__C
@@ -4551,14 +4551,14 @@ define <8 x double> @test_mm512_mask3_fmsub_pd(<8 x double> %__A, <8 x double> %
; X86: # %bb.0: # %entry
; X86-NEXT: movb {{[0-9]+}}(%esp), %al
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfmsub231pd {{.*#+}} zmm2 = (zmm0 * zmm1) - zmm2
+; X86-NEXT: vfmsub231pd {{.*#+}} zmm2 {%k1} = (zmm0 * zmm1) - zmm2
; X86-NEXT: vmovapd %zmm2, %zmm0
; X86-NEXT: retl
;
; X64-LABEL: test_mm512_mask3_fmsub_pd:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfmsub231pd {{.*#+}} zmm2 = (zmm0 * zmm1) - zmm2
+; X64-NEXT: vfmsub231pd {{.*#+}} zmm2 {%k1} = (zmm0 * zmm1) - zmm2
; X64-NEXT: vmovapd %zmm2, %zmm0
; X64-NEXT: retq
entry:
@@ -4597,14 +4597,14 @@ define <16 x float> @test_mm512_mask3_fmsub_ps(<16 x float> %__A, <16 x float> %
; X86: # %bb.0: # %entry
; X86-NEXT: movzwl {{[0-9]+}}(%esp), %eax
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfmsub231ps {{.*#+}} zmm2 = (zmm0 * zmm1) - zmm2
+; X86-NEXT: vfmsub231ps {{.*#+}} zmm2 {%k1} = (zmm0 * zmm1) - zmm2
; X86-NEXT: vmovaps %zmm2, %zmm0
; X86-NEXT: retl
;
; X64-LABEL: test_mm512_mask3_fmsub_ps:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfmsub231ps {{.*#+}} zmm2 = (zmm0 * zmm1) - zmm2
+; X64-NEXT: vfmsub231ps {{.*#+}} zmm2 {%k1} = (zmm0 * zmm1) - zmm2
; X64-NEXT: vmovaps %zmm2, %zmm0
; X64-NEXT: retq
entry:
@@ -4643,14 +4643,14 @@ define <8 x double> @test_mm512_mask3_fmsubadd_pd(<8 x double> %__A, <8 x double
; X86: # %bb.0: # %entry
; X86-NEXT: movb {{[0-9]+}}(%esp), %al
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfmsubadd231pd {{.*#+}} zmm2 = (zmm0 * zmm1) -/+ zmm2
+; X86-NEXT: vfmsubadd231pd {{.*#+}} zmm2 {%k1} = (zmm0 * zmm1) -/+ zmm2
; X86-NEXT: vmovapd %zmm2, %zmm0
; X86-NEXT: retl
;
; X64-LABEL: test_mm512_mask3_fmsubadd_pd:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfmsubadd231pd {{.*#+}} zmm2 = (zmm0 * zmm1) -/+ zmm2
+; X64-NEXT: vfmsubadd231pd {{.*#+}} zmm2 {%k1} = (zmm0 * zmm1) -/+ zmm2
; X64-NEXT: vmovapd %zmm2, %zmm0
; X64-NEXT: retq
entry:
@@ -4691,14 +4691,14 @@ define <16 x float> @test_mm512_mask3_fmsubadd_ps(<16 x float> %__A, <16 x float
; X86: # %bb.0: # %entry
; X86-NEXT: movzwl {{[0-9]+}}(%esp), %eax
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfmsubadd231ps {{.*#+}} zmm2 = (zmm0 * zmm1) -/+ zmm2
+; X86-NEXT: vfmsubadd231ps {{.*#+}} zmm2 {%k1} = (zmm0 * zmm1) -/+ zmm2
; X86-NEXT: vmovaps %zmm2, %zmm0
; X86-NEXT: retl
;
; X64-LABEL: test_mm512_mask3_fmsubadd_ps:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfmsubadd231ps {{.*#+}} zmm2 = (zmm0 * zmm1) -/+ zmm2
+; X64-NEXT: vfmsubadd231ps {{.*#+}} zmm2 {%k1} = (zmm0 * zmm1) -/+ zmm2
; X64-NEXT: vmovaps %zmm2, %zmm0
; X64-NEXT: retq
entry:
@@ -4737,13 +4737,13 @@ define <8 x double> @test_mm512_mask_fnmadd_pd(<8 x double> %__A, i8 zeroext %__
; X86: # %bb.0: # %entry
; X86-NEXT: movb {{[0-9]+}}(%esp), %al
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfnmadd132pd {{.*#+}} zmm0 = -(zmm0 * zmm1) + zmm2
+; X86-NEXT: vfnmadd132pd {{.*#+}} zmm0 {%k1} = -(zmm0 * zmm1) + zmm2
; X86-NEXT: retl
;
; X64-LABEL: test_mm512_mask_fnmadd_pd:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfnmadd132pd {{.*#+}} zmm0 = -(zmm0 * zmm1) + zmm2
+; X64-NEXT: vfnmadd132pd {{.*#+}} zmm0 {%k1} = -(zmm0 * zmm1) + zmm2
; X64-NEXT: retq
entry:
%sub.i = fsub <8 x double> <double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00>, %__A
@@ -4779,13 +4779,13 @@ define <16 x float> @test_mm512_mask_fnmadd_ps(<16 x float> %__A, i16 zeroext %_
; X86: # %bb.0: # %entry
; X86-NEXT: movzwl {{[0-9]+}}(%esp), %eax
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfnmadd132ps {{.*#+}} zmm0 = -(zmm0 * zmm1) + zmm2
+; X86-NEXT: vfnmadd132ps {{.*#+}} zmm0 {%k1} = -(zmm0 * zmm1) + zmm2
; X86-NEXT: retl
;
; X64-LABEL: test_mm512_mask_fnmadd_ps:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfnmadd132ps {{.*#+}} zmm0 = -(zmm0 * zmm1) + zmm2
+; X64-NEXT: vfnmadd132ps {{.*#+}} zmm0 {%k1} = -(zmm0 * zmm1) + zmm2
; X64-NEXT: retq
entry:
%sub.i = fsub <16 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %__A
@@ -4846,13 +4846,13 @@ define <8 x double> @test_mm512_mask_fnmsub_pd(<8 x double> %__A, i8 zeroext %__
; X86: # %bb.0: # %entry
; X86-NEXT: movb {{[0-9]+}}(%esp), %al
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfnmsub132pd {{.*#+}} zmm0 = -(zmm0 * zmm1) - zmm2
+; X86-NEXT: vfnmsub132pd {{.*#+}} zmm0 {%k1} = -(zmm0 * zmm1) - zmm2
; X86-NEXT: retl
;
; X64-LABEL: test_mm512_mask_fnmsub_pd:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfnmsub132pd {{.*#+}} zmm0 = -(zmm0 * zmm1) - zmm2
+; X64-NEXT: vfnmsub132pd {{.*#+}} zmm0 {%k1} = -(zmm0 * zmm1) - zmm2
; X64-NEXT: retq
entry:
%sub.i = fsub <8 x double> <double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00>, %__B
@@ -4868,14 +4868,14 @@ define <8 x double> @test_mm512_mask3_fnmsub_pd(<8 x double> %__A, <8 x double>
; X86: # %bb.0: # %entry
; X86-NEXT: movb {{[0-9]+}}(%esp), %al
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfnmsub231pd {{.*#+}} zmm2 = -(zmm0 * zmm1) - zmm2
+; X86-NEXT: vfnmsub231pd {{.*#+}} zmm2 {%k1} = -(zmm0 * zmm1) - zmm2
; X86-NEXT: vmovapd %zmm2, %zmm0
; X86-NEXT: retl
;
; X64-LABEL: test_mm512_mask3_fnmsub_pd:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfnmsub231pd {{.*#+}} zmm2 = -(zmm0 * zmm1) - zmm2
+; X64-NEXT: vfnmsub231pd {{.*#+}} zmm2 {%k1} = -(zmm0 * zmm1) - zmm2
; X64-NEXT: vmovapd %zmm2, %zmm0
; X64-NEXT: retq
entry:
@@ -4938,13 +4938,13 @@ define <16 x float> @test_mm512_mask_fnmsub_ps(<16 x float> %__A, i16 zeroext %_
; X86: # %bb.0: # %entry
; X86-NEXT: movzwl {{[0-9]+}}(%esp), %eax
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfnmsub132ps {{.*#+}} zmm0 = -(zmm0 * zmm1) - zmm2
+; X86-NEXT: vfnmsub132ps {{.*#+}} zmm0 {%k1} = -(zmm0 * zmm1) - zmm2
; X86-NEXT: retl
;
; X64-LABEL: test_mm512_mask_fnmsub_ps:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfnmsub132ps {{.*#+}} zmm0 = -(zmm0 * zmm1) - zmm2
+; X64-NEXT: vfnmsub132ps {{.*#+}} zmm0 {%k1} = -(zmm0 * zmm1) - zmm2
; X64-NEXT: retq
entry:
%sub.i = fsub <16 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %__B
@@ -4960,14 +4960,14 @@ define <16 x float> @test_mm512_mask3_fnmsub_ps(<16 x float> %__A, <16 x float>
; X86: # %bb.0: # %entry
; X86-NEXT: movzwl {{[0-9]+}}(%esp), %eax
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfnmsub231ps {{.*#+}} zmm2 = -(zmm0 * zmm1) - zmm2
+; X86-NEXT: vfnmsub231ps {{.*#+}} zmm2 {%k1} = -(zmm0 * zmm1) - zmm2
; X86-NEXT: vmovaps %zmm2, %zmm0
; X86-NEXT: retl
;
; X64-LABEL: test_mm512_mask3_fnmsub_ps:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfnmsub231ps {{.*#+}} zmm2 = -(zmm0 * zmm1) - zmm2
+; X64-NEXT: vfnmsub231ps {{.*#+}} zmm2 {%k1} = -(zmm0 * zmm1) - zmm2
; X64-NEXT: vmovaps %zmm2, %zmm0
; X64-NEXT: retq
entry:
@@ -4984,13 +4984,13 @@ define <4 x float> @test_mm_mask_fmadd_ss(<4 x float> %__W, i8 zeroext %__U, <4
; X86: # %bb.0: # %entry
; X86-NEXT: movb {{[0-9]+}}(%esp), %al
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfmadd213ss {{.*#+}} xmm0 = (xmm1 * xmm0) + xmm2
+; X86-NEXT: vfmadd213ss {{.*#+}} xmm0 {%k1} = (xmm1 * xmm0) + xmm2
; X86-NEXT: retl
;
; X64-LABEL: test_mm_mask_fmadd_ss:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfmadd213ss {{.*#+}} xmm0 = (xmm1 * xmm0) + xmm2
+; X64-NEXT: vfmadd213ss {{.*#+}} xmm0 {%k1} = (xmm1 * xmm0) + xmm2
; X64-NEXT: retq
entry:
%0 = extractelement <4 x float> %__W, i64 0
@@ -5037,13 +5037,13 @@ define <4 x float> @test_mm_maskz_fmadd_ss(i8 zeroext %__U, <4 x float> %__A, <4
; X86: # %bb.0: # %entry
; X86-NEXT: movb {{[0-9]+}}(%esp), %al
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfmadd213ss {{.*#+}} xmm0 = (xmm1 * xmm0) + xmm2
+; X86-NEXT: vfmadd213ss {{.*#+}} xmm0 {%k1} {z} = (xmm1 * xmm0) + xmm2
; X86-NEXT: retl
;
; X64-LABEL: test_mm_maskz_fmadd_ss:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfmadd213ss {{.*#+}} xmm0 = (xmm1 * xmm0) + xmm2
+; X64-NEXT: vfmadd213ss {{.*#+}} xmm0 {%k1} {z} = (xmm1 * xmm0) + xmm2
; X64-NEXT: retq
entry:
%0 = extractelement <4 x float> %__A, i64 0
@@ -5087,14 +5087,14 @@ define <4 x float> @test_mm_mask3_fmadd_ss(<4 x float> %__W, <4 x float> %__X, <
; X86: # %bb.0: # %entry
; X86-NEXT: movb {{[0-9]+}}(%esp), %al
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfmadd231ss {{.*#+}} xmm2 = (xmm0 * xmm1) + xmm2
+; X86-NEXT: vfmadd231ss {{.*#+}} xmm2 {%k1} = (xmm0 * xmm1) + xmm2
; X86-NEXT: vmovaps %xmm2, %xmm0
; X86-NEXT: retl
;
; X64-LABEL: test_mm_mask3_fmadd_ss:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfmadd231ss {{.*#+}} xmm2 = (xmm0 * xmm1) + xmm2
+; X64-NEXT: vfmadd231ss {{.*#+}} xmm2 {%k1} = (xmm0 * xmm1) + xmm2
; X64-NEXT: vmovaps %xmm2, %xmm0
; X64-NEXT: retq
entry:
@@ -5142,13 +5142,13 @@ define <4 x float> @test_mm_mask_fmsub_ss(<4 x float> %__W, i8 zeroext %__U, <4
; X86: # %bb.0: # %entry
; X86-NEXT: movb {{[0-9]+}}(%esp), %al
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfmsub213ss {{.*#+}} xmm0 = (xmm1 * xmm0) - xmm2
+; X86-NEXT: vfmsub213ss {{.*#+}} xmm0 {%k1} = (xmm1 * xmm0) - xmm2
; X86-NEXT: retl
;
; X64-LABEL: test_mm_mask_fmsub_ss:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfmsub213ss {{.*#+}} xmm0 = (xmm1 * xmm0) - xmm2
+; X64-NEXT: vfmsub213ss {{.*#+}} xmm0 {%k1} = (xmm1 * xmm0) - xmm2
; X64-NEXT: retq
entry:
%0 = extractelement <4 x float> %__W, i64 0
@@ -5195,13 +5195,13 @@ define <4 x float> @test_mm_maskz_fmsub_ss(i8 zeroext %__U, <4 x float> %__A, <4
; X86: # %bb.0: # %entry
; X86-NEXT: movb {{[0-9]+}}(%esp), %al
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfmsub213ss {{.*#+}} xmm0 = (xmm1 * xmm0) - xmm2
+; X86-NEXT: vfmsub213ss {{.*#+}} xmm0 {%k1} {z} = (xmm1 * xmm0) - xmm2
; X86-NEXT: retl
;
; X64-LABEL: test_mm_maskz_fmsub_ss:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfmsub213ss {{.*#+}} xmm0 = (xmm1 * xmm0) - xmm2
+; X64-NEXT: vfmsub213ss {{.*#+}} xmm0 {%k1} {z} = (xmm1 * xmm0) - xmm2
; X64-NEXT: retq
entry:
%0 = extractelement <4 x float> %__A, i64 0
@@ -5247,14 +5247,14 @@ define <4 x float> @test_mm_mask3_fmsub_ss(<4 x float> %__W, <4 x float> %__X, <
; X86: # %bb.0: # %entry
; X86-NEXT: movb {{[0-9]+}}(%esp), %al
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfmsub231ss {{.*#+}} xmm2 = (xmm0 * xmm1) - xmm2
+; X86-NEXT: vfmsub231ss {{.*#+}} xmm2 {%k1} = (xmm0 * xmm1) - xmm2
; X86-NEXT: vmovaps %xmm2, %xmm0
; X86-NEXT: retl
;
; X64-LABEL: test_mm_mask3_fmsub_ss:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfmsub231ss {{.*#+}} xmm2 = (xmm0 * xmm1) - xmm2
+; X64-NEXT: vfmsub231ss {{.*#+}} xmm2 {%k1} = (xmm0 * xmm1) - xmm2
; X64-NEXT: vmovaps %xmm2, %xmm0
; X64-NEXT: retq
entry:
@@ -5304,13 +5304,13 @@ define <4 x float> @test_mm_mask_fnmadd_ss(<4 x float> %__W, i8 zeroext %__U, <4
; X86: # %bb.0: # %entry
; X86-NEXT: movb {{[0-9]+}}(%esp), %al
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfnmadd213ss {{.*#+}} xmm0 = -(xmm1 * xmm0) + xmm2
+; X86-NEXT: vfnmadd213ss {{.*#+}} xmm0 {%k1} = -(xmm1 * xmm0) + xmm2
; X86-NEXT: retl
;
; X64-LABEL: test_mm_mask_fnmadd_ss:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfnmadd213ss {{.*#+}} xmm0 = -(xmm1 * xmm0) + xmm2
+; X64-NEXT: vfnmadd213ss {{.*#+}} xmm0 {%k1} = -(xmm1 * xmm0) + xmm2
; X64-NEXT: retq
entry:
%0 = extractelement <4 x float> %__W, i64 0
@@ -5357,13 +5357,13 @@ define <4 x float> @test_mm_maskz_fnmadd_ss(i8 zeroext %__U, <4 x float> %__A, <
; X86: # %bb.0: # %entry
; X86-NEXT: movb {{[0-9]+}}(%esp), %al
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfnmadd213ss {{.*#+}} xmm0 = -(xmm1 * xmm0) + xmm2
+; X86-NEXT: vfnmadd213ss {{.*#+}} xmm0 {%k1} {z} = -(xmm1 * xmm0) + xmm2
; X86-NEXT: retl
;
; X64-LABEL: test_mm_maskz_fnmadd_ss:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfnmadd213ss {{.*#+}} xmm0 = -(xmm1 * xmm0) + xmm2
+; X64-NEXT: vfnmadd213ss {{.*#+}} xmm0 {%k1} {z} = -(xmm1 * xmm0) + xmm2
; X64-NEXT: retq
entry:
%0 = extractelement <4 x float> %__A, i64 0
@@ -5409,14 +5409,14 @@ define <4 x float> @test_mm_mask3_fnmadd_ss(<4 x float> %__W, <4 x float> %__X,
; X86: # %bb.0: # %entry
; X86-NEXT: movb {{[0-9]+}}(%esp), %al
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfnmadd231ss {{.*#+}} xmm2 = -(xmm0 * xmm1) + xmm2
+; X86-NEXT: vfnmadd231ss {{.*#+}} xmm2 {%k1} = -(xmm0 * xmm1) + xmm2
; X86-NEXT: vmovaps %xmm2, %xmm0
; X86-NEXT: retl
;
; X64-LABEL: test_mm_mask3_fnmadd_ss:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfnmadd231ss {{.*#+}} xmm2 = -(xmm0 * xmm1) + xmm2
+; X64-NEXT: vfnmadd231ss {{.*#+}} xmm2 {%k1} = -(xmm0 * xmm1) + xmm2
; X64-NEXT: vmovaps %xmm2, %xmm0
; X64-NEXT: retq
entry:
@@ -5466,13 +5466,13 @@ define <4 x float> @test_mm_mask_fnmsub_ss(<4 x float> %__W, i8 zeroext %__U, <4
; X86: # %bb.0: # %entry
; X86-NEXT: movb {{[0-9]+}}(%esp), %al
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfnmsub213ss {{.*#+}} xmm0 = -(xmm1 * xmm0) - xmm2
+; X86-NEXT: vfnmsub213ss {{.*#+}} xmm0 {%k1} = -(xmm1 * xmm0) - xmm2
; X86-NEXT: retl
;
; X64-LABEL: test_mm_mask_fnmsub_ss:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfnmsub213ss {{.*#+}} xmm0 = -(xmm1 * xmm0) - xmm2
+; X64-NEXT: vfnmsub213ss {{.*#+}} xmm0 {%k1} = -(xmm1 * xmm0) - xmm2
; X64-NEXT: retq
entry:
%0 = extractelement <4 x float> %__W, i64 0
@@ -5521,13 +5521,13 @@ define <4 x float> @test_mm_maskz_fnmsub_ss(i8 zeroext %__U, <4 x float> %__A, <
; X86: # %bb.0: # %entry
; X86-NEXT: movb {{[0-9]+}}(%esp), %al
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfnmsub213ss {{.*#+}} xmm0 = -(xmm1 * xmm0) - xmm2
+; X86-NEXT: vfnmsub213ss {{.*#+}} xmm0 {%k1} {z} = -(xmm1 * xmm0) - xmm2
; X86-NEXT: retl
;
; X64-LABEL: test_mm_maskz_fnmsub_ss:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfnmsub213ss {{.*#+}} xmm0 = -(xmm1 * xmm0) - xmm2
+; X64-NEXT: vfnmsub213ss {{.*#+}} xmm0 {%k1} {z} = -(xmm1 * xmm0) - xmm2
; X64-NEXT: retq
entry:
%0 = extractelement <4 x float> %__A, i64 0
@@ -5575,14 +5575,14 @@ define <4 x float> @test_mm_mask3_fnmsub_ss(<4 x float> %__W, <4 x float> %__X,
; X86: # %bb.0: # %entry
; X86-NEXT: movb {{[0-9]+}}(%esp), %al
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfnmsub231ss {{.*#+}} xmm2 = -(xmm0 * xmm1) - xmm2
+; X86-NEXT: vfnmsub231ss {{.*#+}} xmm2 {%k1} = -(xmm0 * xmm1) - xmm2
; X86-NEXT: vmovaps %xmm2, %xmm0
; X86-NEXT: retl
;
; X64-LABEL: test_mm_mask3_fnmsub_ss:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfnmsub231ss {{.*#+}} xmm2 = -(xmm0 * xmm1) - xmm2
+; X64-NEXT: vfnmsub231ss {{.*#+}} xmm2 {%k1} = -(xmm0 * xmm1) - xmm2
; X64-NEXT: vmovaps %xmm2, %xmm0
; X64-NEXT: retq
entry:
@@ -5634,13 +5634,13 @@ define <2 x double> @test_mm_mask_fmadd_sd(<2 x double> %__W, i8 zeroext %__U, <
; X86: # %bb.0: # %entry
; X86-NEXT: movb {{[0-9]+}}(%esp), %al
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfmadd213sd {{.*#+}} xmm0 = (xmm1 * xmm0) + xmm2
+; X86-NEXT: vfmadd213sd {{.*#+}} xmm0 {%k1} = (xmm1 * xmm0) + xmm2
; X86-NEXT: retl
;
; X64-LABEL: test_mm_mask_fmadd_sd:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfmadd213sd {{.*#+}} xmm0 = (xmm1 * xmm0) + xmm2
+; X64-NEXT: vfmadd213sd {{.*#+}} xmm0 {%k1} = (xmm1 * xmm0) + xmm2
; X64-NEXT: retq
entry:
%0 = extractelement <2 x double> %__W, i64 0
@@ -5687,13 +5687,13 @@ define <2 x double> @test_mm_maskz_fmadd_sd(i8 zeroext %__U, <2 x double> %__A,
; X86: # %bb.0: # %entry
; X86-NEXT: movb {{[0-9]+}}(%esp), %al
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfmadd213sd {{.*#+}} xmm0 = (xmm1 * xmm0) + xmm2
+; X86-NEXT: vfmadd213sd {{.*#+}} xmm0 {%k1} {z} = (xmm1 * xmm0) + xmm2
; X86-NEXT: retl
;
; X64-LABEL: test_mm_maskz_fmadd_sd:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfmadd213sd {{.*#+}} xmm0 = (xmm1 * xmm0) + xmm2
+; X64-NEXT: vfmadd213sd {{.*#+}} xmm0 {%k1} {z} = (xmm1 * xmm0) + xmm2
; X64-NEXT: retq
entry:
%0 = extractelement <2 x double> %__A, i64 0
@@ -5737,14 +5737,14 @@ define <2 x double> @test_mm_mask3_fmadd_sd(<2 x double> %__W, <2 x double> %__X
; X86: # %bb.0: # %entry
; X86-NEXT: movb {{[0-9]+}}(%esp), %al
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfmadd231sd {{.*#+}} xmm2 = (xmm0 * xmm1) + xmm2
+; X86-NEXT: vfmadd231sd {{.*#+}} xmm2 {%k1} = (xmm0 * xmm1) + xmm2
; X86-NEXT: vmovapd %xmm2, %xmm0
; X86-NEXT: retl
;
; X64-LABEL: test_mm_mask3_fmadd_sd:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfmadd231sd {{.*#+}} xmm2 = (xmm0 * xmm1) + xmm2
+; X64-NEXT: vfmadd231sd {{.*#+}} xmm2 {%k1} = (xmm0 * xmm1) + xmm2
; X64-NEXT: vmovapd %xmm2, %xmm0
; X64-NEXT: retq
entry:
@@ -5792,13 +5792,13 @@ define <2 x double> @test_mm_mask_fmsub_sd(<2 x double> %__W, i8 zeroext %__U, <
; X86: # %bb.0: # %entry
; X86-NEXT: movb {{[0-9]+}}(%esp), %al
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfmsub213sd {{.*#+}} xmm0 = (xmm1 * xmm0) - xmm2
+; X86-NEXT: vfmsub213sd {{.*#+}} xmm0 {%k1} = (xmm1 * xmm0) - xmm2
; X86-NEXT: retl
;
; X64-LABEL: test_mm_mask_fmsub_sd:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfmsub213sd {{.*#+}} xmm0 = (xmm1 * xmm0) - xmm2
+; X64-NEXT: vfmsub213sd {{.*#+}} xmm0 {%k1} = (xmm1 * xmm0) - xmm2
; X64-NEXT: retq
entry:
%0 = extractelement <2 x double> %__W, i64 0
@@ -5845,13 +5845,13 @@ define <2 x double> @test_mm_maskz_fmsub_sd(i8 zeroext %__U, <2 x double> %__A,
; X86: # %bb.0: # %entry
; X86-NEXT: movb {{[0-9]+}}(%esp), %al
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfmsub213sd {{.*#+}} xmm0 = (xmm1 * xmm0) - xmm2
+; X86-NEXT: vfmsub213sd {{.*#+}} xmm0 {%k1} {z} = (xmm1 * xmm0) - xmm2
; X86-NEXT: retl
;
; X64-LABEL: test_mm_maskz_fmsub_sd:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfmsub213sd {{.*#+}} xmm0 = (xmm1 * xmm0) - xmm2
+; X64-NEXT: vfmsub213sd {{.*#+}} xmm0 {%k1} {z} = (xmm1 * xmm0) - xmm2
; X64-NEXT: retq
entry:
%0 = extractelement <2 x double> %__A, i64 0
@@ -5897,14 +5897,14 @@ define <2 x double> @test_mm_mask3_fmsub_sd(<2 x double> %__W, <2 x double> %__X
; X86: # %bb.0: # %entry
; X86-NEXT: movb {{[0-9]+}}(%esp), %al
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfmsub231sd {{.*#+}} xmm2 = (xmm0 * xmm1) - xmm2
+; X86-NEXT: vfmsub231sd {{.*#+}} xmm2 {%k1} = (xmm0 * xmm1) - xmm2
; X86-NEXT: vmovapd %xmm2, %xmm0
; X86-NEXT: retl
;
; X64-LABEL: test_mm_mask3_fmsub_sd:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfmsub231sd {{.*#+}} xmm2 = (xmm0 * xmm1) - xmm2
+; X64-NEXT: vfmsub231sd {{.*#+}} xmm2 {%k1} = (xmm0 * xmm1) - xmm2
; X64-NEXT: vmovapd %xmm2, %xmm0
; X64-NEXT: retq
entry:
@@ -5954,13 +5954,13 @@ define <2 x double> @test_mm_mask_fnmadd_sd(<2 x double> %__W, i8 zeroext %__U,
; X86: # %bb.0: # %entry
; X86-NEXT: movb {{[0-9]+}}(%esp), %al
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfnmadd213sd {{.*#+}} xmm0 = -(xmm1 * xmm0) + xmm2
+; X86-NEXT: vfnmadd213sd {{.*#+}} xmm0 {%k1} = -(xmm1 * xmm0) + xmm2
; X86-NEXT: retl
;
; X64-LABEL: test_mm_mask_fnmadd_sd:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfnmadd213sd {{.*#+}} xmm0 = -(xmm1 * xmm0) + xmm2
+; X64-NEXT: vfnmadd213sd {{.*#+}} xmm0 {%k1} = -(xmm1 * xmm0) + xmm2
; X64-NEXT: retq
entry:
%0 = extractelement <2 x double> %__W, i64 0
@@ -6007,13 +6007,13 @@ define <2 x double> @test_mm_maskz_fnmadd_sd(i8 zeroext %__U, <2 x double> %__A,
; X86: # %bb.0: # %entry
; X86-NEXT: movb {{[0-9]+}}(%esp), %al
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfnmadd213sd {{.*#+}} xmm0 = -(xmm1 * xmm0) + xmm2
+; X86-NEXT: vfnmadd213sd {{.*#+}} xmm0 {%k1} {z} = -(xmm1 * xmm0) + xmm2
; X86-NEXT: retl
;
; X64-LABEL: test_mm_maskz_fnmadd_sd:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfnmadd213sd {{.*#+}} xmm0 = -(xmm1 * xmm0) + xmm2
+; X64-NEXT: vfnmadd213sd {{.*#+}} xmm0 {%k1} {z} = -(xmm1 * xmm0) + xmm2
; X64-NEXT: retq
entry:
%0 = extractelement <2 x double> %__A, i64 0
@@ -6059,14 +6059,14 @@ define <2 x double> @test_mm_mask3_fnmadd_sd(<2 x double> %__W, <2 x double> %__
; X86: # %bb.0: # %entry
; X86-NEXT: movb {{[0-9]+}}(%esp), %al
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfnmadd231sd {{.*#+}} xmm2 = -(xmm0 * xmm1) + xmm2
+; X86-NEXT: vfnmadd231sd {{.*#+}} xmm2 {%k1} = -(xmm0 * xmm1) + xmm2
; X86-NEXT: vmovapd %xmm2, %xmm0
; X86-NEXT: retl
;
; X64-LABEL: test_mm_mask3_fnmadd_sd:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfnmadd231sd {{.*#+}} xmm2 = -(xmm0 * xmm1) + xmm2
+; X64-NEXT: vfnmadd231sd {{.*#+}} xmm2 {%k1} = -(xmm0 * xmm1) + xmm2
; X64-NEXT: vmovapd %xmm2, %xmm0
; X64-NEXT: retq
entry:
@@ -6116,13 +6116,13 @@ define <2 x double> @test_mm_mask_fnmsub_sd(<2 x double> %__W, i8 zeroext %__U,
; X86: # %bb.0: # %entry
; X86-NEXT: movb {{[0-9]+}}(%esp), %al
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfnmsub213sd {{.*#+}} xmm0 = -(xmm1 * xmm0) - xmm2
+; X86-NEXT: vfnmsub213sd {{.*#+}} xmm0 {%k1} = -(xmm1 * xmm0) - xmm2
; X86-NEXT: retl
;
; X64-LABEL: test_mm_mask_fnmsub_sd:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfnmsub213sd {{.*#+}} xmm0 = -(xmm1 * xmm0) - xmm2
+; X64-NEXT: vfnmsub213sd {{.*#+}} xmm0 {%k1} = -(xmm1 * xmm0) - xmm2
; X64-NEXT: retq
entry:
%0 = extractelement <2 x double> %__W, i64 0
@@ -6171,13 +6171,13 @@ define <2 x double> @test_mm_maskz_fnmsub_sd(i8 zeroext %__U, <2 x double> %__A,
; X86: # %bb.0: # %entry
; X86-NEXT: movb {{[0-9]+}}(%esp), %al
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfnmsub213sd {{.*#+}} xmm0 = -(xmm1 * xmm0) - xmm2
+; X86-NEXT: vfnmsub213sd {{.*#+}} xmm0 {%k1} {z} = -(xmm1 * xmm0) - xmm2
; X86-NEXT: retl
;
; X64-LABEL: test_mm_maskz_fnmsub_sd:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfnmsub213sd {{.*#+}} xmm0 = -(xmm1 * xmm0) - xmm2
+; X64-NEXT: vfnmsub213sd {{.*#+}} xmm0 {%k1} {z} = -(xmm1 * xmm0) - xmm2
; X64-NEXT: retq
entry:
%0 = extractelement <2 x double> %__A, i64 0
@@ -6225,14 +6225,14 @@ define <2 x double> @test_mm_mask3_fnmsub_sd(<2 x double> %__W, <2 x double> %__
; X86: # %bb.0: # %entry
; X86-NEXT: movb {{[0-9]+}}(%esp), %al
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfnmsub231sd {{.*#+}} xmm2 = -(xmm0 * xmm1) - xmm2
+; X86-NEXT: vfnmsub231sd {{.*#+}} xmm2 {%k1} = -(xmm0 * xmm1) - xmm2
; X86-NEXT: vmovapd %xmm2, %xmm0
; X86-NEXT: retl
;
; X64-LABEL: test_mm_mask3_fnmsub_sd:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfnmsub231sd {{.*#+}} xmm2 = -(xmm0 * xmm1) - xmm2
+; X64-NEXT: vfnmsub231sd {{.*#+}} xmm2 {%k1} = -(xmm0 * xmm1) - xmm2
; X64-NEXT: vmovapd %xmm2, %xmm0
; X64-NEXT: retq
entry:
diff --git a/llvm/test/CodeGen/X86/avx512-intrinsics-upgrade.ll b/llvm/test/CodeGen/X86/avx512-intrinsics-upgrade.ll
index a0631f1e8d5e..b6f2ae9690a6 100644
--- a/llvm/test/CodeGen/X86/avx512-intrinsics-upgrade.ll
+++ b/llvm/test/CodeGen/X86/avx512-intrinsics-upgrade.ll
@@ -9892,7 +9892,7 @@ define <2 x double>@test_int_x86_avx512_mask_vfmadd_sd(<2 x double> %x0, <2 x do
; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vmovapd %xmm0, %xmm3 ## encoding: [0xc5,0xf9,0x28,0xd8]
; X86-NEXT: vfmadd213sd %xmm2, %xmm1, %xmm3 {%k1} ## encoding: [0x62,0xf2,0xf5,0x09,0xa9,0xda]
-; X86-NEXT: ## xmm3 = (xmm1 * xmm3) + xmm2
+; X86-NEXT: ## xmm3 {%k1} = (xmm1 * xmm3) + xmm2
; X86-NEXT: vmovapd %xmm0, %xmm4 ## encoding: [0xc5,0xf9,0x28,0xe0]
; X86-NEXT: vfmadd213sd {rz-sae}, %xmm2, %xmm1, %xmm4 ## encoding: [0x62,0xf2,0xf5,0x78,0xa9,0xe2]
; X86-NEXT: vaddpd %xmm4, %xmm3, %xmm3 ## encoding: [0xc5,0xe1,0x58,0xdc]
@@ -9905,7 +9905,7 @@ define <2 x double>@test_int_x86_avx512_mask_vfmadd_sd(<2 x double> %x0, <2 x do
; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vmovapd %xmm0, %xmm3 ## encoding: [0xc5,0xf9,0x28,0xd8]
; X64-NEXT: vfmadd213sd %xmm2, %xmm1, %xmm3 {%k1} ## encoding: [0x62,0xf2,0xf5,0x09,0xa9,0xda]
-; X64-NEXT: ## xmm3 = (xmm1 * xmm3) + xmm2
+; X64-NEXT: ## xmm3 {%k1} = (xmm1 * xmm3) + xmm2
; X64-NEXT: vmovapd %xmm0, %xmm4 ## encoding: [0xc5,0xf9,0x28,0xe0]
; X64-NEXT: vfmadd213sd {rz-sae}, %xmm2, %xmm1, %xmm4 ## encoding: [0x62,0xf2,0xf5,0x78,0xa9,0xe2]
; X64-NEXT: vaddpd %xmm4, %xmm3, %xmm3 ## encoding: [0xc5,0xe1,0x58,0xdc]
@@ -9929,7 +9929,7 @@ define <4 x float>@test_int_x86_avx512_mask_vfmadd_ss(<4 x float> %x0, <4 x floa
; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vmovaps %xmm0, %xmm3 ## encoding: [0xc5,0xf8,0x28,0xd8]
; X86-NEXT: vfmadd213ss %xmm2, %xmm1, %xmm3 {%k1} ## encoding: [0x62,0xf2,0x75,0x09,0xa9,0xda]
-; X86-NEXT: ## xmm3 = (xmm1 * xmm3) + xmm2
+; X86-NEXT: ## xmm3 {%k1} = (xmm1 * xmm3) + xmm2
; X86-NEXT: vmovaps %xmm0, %xmm4 ## encoding: [0xc5,0xf8,0x28,0xe0]
; X86-NEXT: vfmadd213ss {rz-sae}, %xmm2, %xmm1, %xmm4 ## encoding: [0x62,0xf2,0x75,0x78,0xa9,0xe2]
; X86-NEXT: vaddps %xmm4, %xmm3, %xmm3 ## encoding: [0xc5,0xe0,0x58,0xdc]
@@ -9942,7 +9942,7 @@ define <4 x float>@test_int_x86_avx512_mask_vfmadd_ss(<4 x float> %x0, <4 x floa
; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vmovaps %xmm0, %xmm3 ## encoding: [0xc5,0xf8,0x28,0xd8]
; X64-NEXT: vfmadd213ss %xmm2, %xmm1, %xmm3 {%k1} ## encoding: [0x62,0xf2,0x75,0x09,0xa9,0xda]
-; X64-NEXT: ## xmm3 = (xmm1 * xmm3) + xmm2
+; X64-NEXT: ## xmm3 {%k1} = (xmm1 * xmm3) + xmm2
; X64-NEXT: vmovaps %xmm0, %xmm4 ## encoding: [0xc5,0xf8,0x28,0xe0]
; X64-NEXT: vfmadd213ss {rz-sae}, %xmm2, %xmm1, %xmm4 ## encoding: [0x62,0xf2,0x75,0x78,0xa9,0xe2]
; X64-NEXT: vaddps %xmm4, %xmm3, %xmm3 ## encoding: [0xc5,0xe0,0x58,0xdc]
@@ -9966,7 +9966,7 @@ define <2 x double>@test_int_x86_avx512_maskz_vfmadd_sd(<2 x double> %x0, <2 x d
; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vmovapd %xmm0, %xmm3 ## encoding: [0xc5,0xf9,0x28,0xd8]
; X86-NEXT: vfmadd213sd %xmm2, %xmm1, %xmm3 {%k1} {z} ## encoding: [0x62,0xf2,0xf5,0x89,0xa9,0xda]
-; X86-NEXT: ## xmm3 = (xmm1 * xmm3) + xmm2
+; X86-NEXT: ## xmm3 {%k1} {z} = (xmm1 * xmm3) + xmm2
; X86-NEXT: vfmadd213sd {rz-sae}, %xmm2, %xmm1, %xmm0 {%k1} {z} ## encoding: [0x62,0xf2,0xf5,0xf9,0xa9,0xc2]
; X86-NEXT: vaddpd %xmm0, %xmm3, %xmm0 ## encoding: [0xc5,0xe1,0x58,0xc0]
; X86-NEXT: retl ## encoding: [0xc3]
@@ -9976,7 +9976,7 @@ define <2 x double>@test_int_x86_avx512_maskz_vfmadd_sd(<2 x double> %x0, <2 x d
; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vmovapd %xmm0, %xmm3 ## encoding: [0xc5,0xf9,0x28,0xd8]
; X64-NEXT: vfmadd213sd %xmm2, %xmm1, %xmm3 {%k1} {z} ## encoding: [0x62,0xf2,0xf5,0x89,0xa9,0xda]
-; X64-NEXT: ## xmm3 = (xmm1 * xmm3) + xmm2
+; X64-NEXT: ## xmm3 {%k1} {z} = (xmm1 * xmm3) + xmm2
; X64-NEXT: vfmadd213sd {rz-sae}, %xmm2, %xmm1, %xmm0 {%k1} {z} ## encoding: [0x62,0xf2,0xf5,0xf9,0xa9,0xc2]
; X64-NEXT: vaddpd %xmm0, %xmm3, %xmm0 ## encoding: [0xc5,0xe1,0x58,0xc0]
; X64-NEXT: retq ## encoding: [0xc3]
@@ -9994,14 +9994,14 @@ define <4 x float>@test_int_x86_avx512_maskz_vfmadd_ss(<4 x float> %x0, <4 x flo
; X86-NEXT: movb {{[0-9]+}}(%esp), %al ## encoding: [0x8a,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmadd213ss %xmm2, %xmm1, %xmm0 {%k1} {z} ## encoding: [0x62,0xf2,0x75,0x89,0xa9,0xc2]
-; X86-NEXT: ## xmm0 = (xmm1 * xmm0) + xmm2
+; X86-NEXT: ## xmm0 {%k1} {z} = (xmm1 * xmm0) + xmm2
; X86-NEXT: retl ## encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_maskz_vfmadd_ss:
; X64: ## %bb.0:
; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmadd213ss %xmm2, %xmm1, %xmm0 {%k1} {z} ## encoding: [0x62,0xf2,0x75,0x89,0xa9,0xc2]
-; X64-NEXT: ## xmm0 = (xmm1 * xmm0) + xmm2
+; X64-NEXT: ## xmm0 {%k1} {z} = (xmm1 * xmm0) + xmm2
; X64-NEXT: retq ## encoding: [0xc3]
%res = call <4 x float> @llvm.x86.avx512.maskz.vfmadd.ss(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2, i8 %x3, i32 4)
%res1 = call <4 x float> @llvm.x86.avx512.maskz.vfmadd.ss(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2, i8 %x3, i32 11)
@@ -10017,7 +10017,7 @@ define <2 x double>@test_int_x86_avx512_mask3_vfmadd_sd(<2 x double> %x0, <2 x d
; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vmovapd %xmm2, %xmm3 ## encoding: [0xc5,0xf9,0x28,0xda]
; X86-NEXT: vfmadd231sd %xmm1, %xmm0, %xmm3 {%k1} ## encoding: [0x62,0xf2,0xfd,0x09,0xb9,0xd9]
-; X86-NEXT: ## xmm3 = (xmm0 * xmm1) + xmm3
+; X86-NEXT: ## xmm3 {%k1} = (xmm0 * xmm1) + xmm3
; X86-NEXT: vmovapd %xmm2, %xmm4 ## encoding: [0xc5,0xf9,0x28,0xe2]
; X86-NEXT: vfmadd231sd {rz-sae}, %xmm1, %xmm0, %xmm4 ## encoding: [0x62,0xf2,0xfd,0x78,0xb9,0xe1]
; X86-NEXT: vaddpd %xmm4, %xmm3, %xmm3 ## encoding: [0xc5,0xe1,0x58,0xdc]
@@ -10030,7 +10030,7 @@ define <2 x double>@test_int_x86_avx512_mask3_vfmadd_sd(<2 x double> %x0, <2 x d
; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vmovapd %xmm2, %xmm3 ## encoding: [0xc5,0xf9,0x28,0xda]
; X64-NEXT: vfmadd231sd %xmm1, %xmm0, %xmm3 {%k1} ## encoding: [0x62,0xf2,0xfd,0x09,0xb9,0xd9]
-; X64-NEXT: ## xmm3 = (xmm0 * xmm1) + xmm3
+; X64-NEXT: ## xmm3 {%k1} = (xmm0 * xmm1) + xmm3
; X64-NEXT: vmovapd %xmm2, %xmm4 ## encoding: [0xc5,0xf9,0x28,0xe2]
; X64-NEXT: vfmadd231sd {rz-sae}, %xmm1, %xmm0, %xmm4 ## encoding: [0x62,0xf2,0xfd,0x78,0xb9,0xe1]
; X64-NEXT: vaddpd %xmm4, %xmm3, %xmm3 ## encoding: [0xc5,0xe1,0x58,0xdc]
@@ -10054,7 +10054,7 @@ define <4 x float>@test_int_x86_avx512_mask3_vfmadd_ss(<4 x float> %x0, <4 x flo
; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vmovaps %xmm2, %xmm3 ## encoding: [0xc5,0xf8,0x28,0xda]
; X86-NEXT: vfmadd231ss %xmm1, %xmm0, %xmm3 {%k1} ## encoding: [0x62,0xf2,0x7d,0x09,0xb9,0xd9]
-; X86-NEXT: ## xmm3 = (xmm0 * xmm1) + xmm3
+; X86-NEXT: ## xmm3 {%k1} = (xmm0 * xmm1) + xmm3
; X86-NEXT: vmovaps %xmm2, %xmm4 ## encoding: [0xc5,0xf8,0x28,0xe2]
; X86-NEXT: vfmadd231ss {rz-sae}, %xmm1, %xmm0, %xmm4 ## encoding: [0x62,0xf2,0x7d,0x78,0xb9,0xe1]
; X86-NEXT: vaddps %xmm4, %xmm3, %xmm3 ## encoding: [0xc5,0xe0,0x58,0xdc]
@@ -10067,7 +10067,7 @@ define <4 x float>@test_int_x86_avx512_mask3_vfmadd_ss(<4 x float> %x0, <4 x flo
; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vmovaps %xmm2, %xmm3 ## encoding: [0xc5,0xf8,0x28,0xda]
; X64-NEXT: vfmadd231ss %xmm1, %xmm0, %xmm3 {%k1} ## encoding: [0x62,0xf2,0x7d,0x09,0xb9,0xd9]
-; X64-NEXT: ## xmm3 = (xmm0 * xmm1) + xmm3
+; X64-NEXT: ## xmm3 {%k1} = (xmm0 * xmm1) + xmm3
; X64-NEXT: vmovaps %xmm2, %xmm4 ## encoding: [0xc5,0xf8,0x28,0xe2]
; X64-NEXT: vfmadd231ss {rz-sae}, %xmm1, %xmm0, %xmm4 ## encoding: [0x62,0xf2,0x7d,0x78,0xb9,0xe1]
; X64-NEXT: vaddps %xmm4, %xmm3, %xmm3 ## encoding: [0xc5,0xe0,0x58,0xdc]
@@ -10267,7 +10267,7 @@ define <2 x double>@test_int_x86_avx512_mask3_vfmsub_sd(<2 x double> %x0, <2 x d
; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vmovapd %xmm2, %xmm3 ## encoding: [0xc5,0xf9,0x28,0xda]
; X86-NEXT: vfmsub231sd %xmm1, %xmm0, %xmm3 {%k1} ## encoding: [0x62,0xf2,0xfd,0x09,0xbb,0xd9]
-; X86-NEXT: ## xmm3 = (xmm0 * xmm1) - xmm3
+; X86-NEXT: ## xmm3 {%k1} = (xmm0 * xmm1) - xmm3
; X86-NEXT: vmovapd %xmm2, %xmm4 ## encoding: [0xc5,0xf9,0x28,0xe2]
; X86-NEXT: vfmsub231sd {rz-sae}, %xmm1, %xmm0, %xmm4 ## encoding: [0x62,0xf2,0xfd,0x78,0xbb,0xe1]
; X86-NEXT: vaddpd %xmm4, %xmm3, %xmm3 ## encoding: [0xc5,0xe1,0x58,0xdc]
@@ -10280,7 +10280,7 @@ define <2 x double>@test_int_x86_avx512_mask3_vfmsub_sd(<2 x double> %x0, <2 x d
; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vmovapd %xmm2, %xmm3 ## encoding: [0xc5,0xf9,0x28,0xda]
; X64-NEXT: vfmsub231sd %xmm1, %xmm0, %xmm3 {%k1} ## encoding: [0x62,0xf2,0xfd,0x09,0xbb,0xd9]
-; X64-NEXT: ## xmm3 = (xmm0 * xmm1) - xmm3
+; X64-NEXT: ## xmm3 {%k1} = (xmm0 * xmm1) - xmm3
; X64-NEXT: vmovapd %xmm2, %xmm4 ## encoding: [0xc5,0xf9,0x28,0xe2]
; X64-NEXT: vfmsub231sd {rz-sae}, %xmm1, %xmm0, %xmm4 ## encoding: [0x62,0xf2,0xfd,0x78,0xbb,0xe1]
; X64-NEXT: vaddpd %xmm4, %xmm3, %xmm3 ## encoding: [0xc5,0xe1,0x58,0xdc]
@@ -10304,7 +10304,7 @@ define <4 x float>@test_int_x86_avx512_mask3_vfmsub_ss(<4 x float> %x0, <4 x flo
; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vmovaps %xmm2, %xmm3 ## encoding: [0xc5,0xf8,0x28,0xda]
; X86-NEXT: vfmsub231ss %xmm1, %xmm0, %xmm3 {%k1} ## encoding: [0x62,0xf2,0x7d,0x09,0xbb,0xd9]
-; X86-NEXT: ## xmm3 = (xmm0 * xmm1) - xmm3
+; X86-NEXT: ## xmm3 {%k1} = (xmm0 * xmm1) - xmm3
; X86-NEXT: vmovaps %xmm2, %xmm4 ## encoding: [0xc5,0xf8,0x28,0xe2]
; X86-NEXT: vfmsub231ss {rz-sae}, %xmm1, %xmm0, %xmm4 ## encoding: [0x62,0xf2,0x7d,0x78,0xbb,0xe1]
; X86-NEXT: vaddps %xmm4, %xmm3, %xmm3 ## encoding: [0xc5,0xe0,0x58,0xdc]
@@ -10317,7 +10317,7 @@ define <4 x float>@test_int_x86_avx512_mask3_vfmsub_ss(<4 x float> %x0, <4 x flo
; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vmovaps %xmm2, %xmm3 ## encoding: [0xc5,0xf8,0x28,0xda]
; X64-NEXT: vfmsub231ss %xmm1, %xmm0, %xmm3 {%k1} ## encoding: [0x62,0xf2,0x7d,0x09,0xbb,0xd9]
-; X64-NEXT: ## xmm3 = (xmm0 * xmm1) - xmm3
+; X64-NEXT: ## xmm3 {%k1} = (xmm0 * xmm1) - xmm3
; X64-NEXT: vmovaps %xmm2, %xmm4 ## encoding: [0xc5,0xf8,0x28,0xe2]
; X64-NEXT: vfmsub231ss {rz-sae}, %xmm1, %xmm0, %xmm4 ## encoding: [0x62,0xf2,0x7d,0x78,0xbb,0xe1]
; X64-NEXT: vaddps %xmm4, %xmm3, %xmm3 ## encoding: [0xc5,0xe0,0x58,0xdc]
@@ -10341,7 +10341,7 @@ define <2 x double>@test_int_x86_avx512_mask3_vfnmsub_sd(<2 x double> %x0, <2 x
; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vmovapd %xmm2, %xmm3 ## encoding: [0xc5,0xf9,0x28,0xda]
; X86-NEXT: vfnmsub231sd %xmm1, %xmm0, %xmm3 {%k1} ## encoding: [0x62,0xf2,0xfd,0x09,0xbf,0xd9]
-; X86-NEXT: ## xmm3 = -(xmm0 * xmm1) - xmm3
+; X86-NEXT: ## xmm3 {%k1} = -(xmm0 * xmm1) - xmm3
; X86-NEXT: vmovapd %xmm2, %xmm4 ## encoding: [0xc5,0xf9,0x28,0xe2]
; X86-NEXT: vfnmsub231sd {rz-sae}, %xmm1, %xmm0, %xmm4 ## encoding: [0x62,0xf2,0xfd,0x78,0xbf,0xe1]
; X86-NEXT: vaddpd %xmm4, %xmm3, %xmm3 ## encoding: [0xc5,0xe1,0x58,0xdc]
@@ -10354,7 +10354,7 @@ define <2 x double>@test_int_x86_avx512_mask3_vfnmsub_sd(<2 x double> %x0, <2 x
; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vmovapd %xmm2, %xmm3 ## encoding: [0xc5,0xf9,0x28,0xda]
; X64-NEXT: vfnmsub231sd %xmm1, %xmm0, %xmm3 {%k1} ## encoding: [0x62,0xf2,0xfd,0x09,0xbf,0xd9]
-; X64-NEXT: ## xmm3 = -(xmm0 * xmm1) - xmm3
+; X64-NEXT: ## xmm3 {%k1} = -(xmm0 * xmm1) - xmm3
; X64-NEXT: vmovapd %xmm2, %xmm4 ## encoding: [0xc5,0xf9,0x28,0xe2]
; X64-NEXT: vfnmsub231sd {rz-sae}, %xmm1, %xmm0, %xmm4 ## encoding: [0x62,0xf2,0xfd,0x78,0xbf,0xe1]
; X64-NEXT: vaddpd %xmm4, %xmm3, %xmm3 ## encoding: [0xc5,0xe1,0x58,0xdc]
@@ -10378,7 +10378,7 @@ define <4 x float>@test_int_x86_avx512_mask3_vfnmsub_ss(<4 x float> %x0, <4 x fl
; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vmovaps %xmm2, %xmm3 ## encoding: [0xc5,0xf8,0x28,0xda]
; X86-NEXT: vfnmsub231ss %xmm1, %xmm0, %xmm3 {%k1} ## encoding: [0x62,0xf2,0x7d,0x09,0xbf,0xd9]
-; X86-NEXT: ## xmm3 = -(xmm0 * xmm1) - xmm3
+; X86-NEXT: ## xmm3 {%k1} = -(xmm0 * xmm1) - xmm3
; X86-NEXT: vmovaps %xmm2, %xmm4 ## encoding: [0xc5,0xf8,0x28,0xe2]
; X86-NEXT: vfnmsub231ss {rz-sae}, %xmm1, %xmm0, %xmm4 ## encoding: [0x62,0xf2,0x7d,0x78,0xbf,0xe1]
; X86-NEXT: vaddps %xmm4, %xmm3, %xmm3 ## encoding: [0xc5,0xe0,0x58,0xdc]
@@ -10391,7 +10391,7 @@ define <4 x float>@test_int_x86_avx512_mask3_vfnmsub_ss(<4 x float> %x0, <4 x fl
; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vmovaps %xmm2, %xmm3 ## encoding: [0xc5,0xf8,0x28,0xda]
; X64-NEXT: vfnmsub231ss %xmm1, %xmm0, %xmm3 {%k1} ## encoding: [0x62,0xf2,0x7d,0x09,0xbf,0xd9]
-; X64-NEXT: ## xmm3 = -(xmm0 * xmm1) - xmm3
+; X64-NEXT: ## xmm3 {%k1} = -(xmm0 * xmm1) - xmm3
; X64-NEXT: vmovaps %xmm2, %xmm4 ## encoding: [0xc5,0xf8,0x28,0xe2]
; X64-NEXT: vfnmsub231ss {rz-sae}, %xmm1, %xmm0, %xmm4 ## encoding: [0x62,0xf2,0x7d,0x78,0xbf,0xe1]
; X64-NEXT: vaddps %xmm4, %xmm3, %xmm3 ## encoding: [0xc5,0xe0,0x58,0xdc]
@@ -10413,7 +10413,7 @@ define <4 x float>@test_int_x86_avx512_mask3_vfmadd_ss_rm(<4 x float> %x0, <4 x
; X86-NEXT: movb {{[0-9]+}}(%esp), %cl ## encoding: [0x8a,0x4c,0x24,0x08]
; X86-NEXT: kmovw %ecx, %k1 ## encoding: [0xc5,0xf8,0x92,0xc9]
; X86-NEXT: vfmadd231ss (%eax), %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf2,0x7d,0x09,0xb9,0x08]
-; X86-NEXT: ## xmm1 = (xmm0 * mem) + xmm1
+; X86-NEXT: ## xmm1 {%k1} = (xmm0 * mem) + xmm1
; X86-NEXT: vmovaps %xmm1, %xmm0 ## encoding: [0xc5,0xf8,0x28,0xc1]
; X86-NEXT: retl ## encoding: [0xc3]
;
@@ -10421,7 +10421,7 @@ define <4 x float>@test_int_x86_avx512_mask3_vfmadd_ss_rm(<4 x float> %x0, <4 x
; X64: ## %bb.0:
; X64-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; X64-NEXT: vfmadd231ss (%rdi), %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf2,0x7d,0x09,0xb9,0x0f]
-; X64-NEXT: ## xmm1 = (xmm0 * mem) + xmm1
+; X64-NEXT: ## xmm1 {%k1} = (xmm0 * mem) + xmm1
; X64-NEXT: vmovaps %xmm1, %xmm0 ## encoding: [0xc5,0xf8,0x28,0xc1]
; X64-NEXT: retq ## encoding: [0xc3]
%q = load float, float* %ptr_b
@@ -10437,14 +10437,14 @@ define <4 x float>@test_int_x86_avx512_mask_vfmadd_ss_rm(<4 x float> %x0, <4 x f
; X86-NEXT: movb {{[0-9]+}}(%esp), %cl ## encoding: [0x8a,0x4c,0x24,0x08]
; X86-NEXT: kmovw %ecx, %k1 ## encoding: [0xc5,0xf8,0x92,0xc9]
; X86-NEXT: vfmadd132ss (%eax), %xmm1, %xmm0 {%k1} ## encoding: [0x62,0xf2,0x75,0x09,0x99,0x00]
-; X86-NEXT: ## xmm0 = (xmm0 * mem) + xmm1
+; X86-NEXT: ## xmm0 {%k1} = (xmm0 * mem) + xmm1
; X86-NEXT: retl ## encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_vfmadd_ss_rm:
; X64: ## %bb.0:
; X64-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; X64-NEXT: vfmadd132ss (%rdi), %xmm1, %xmm0 {%k1} ## encoding: [0x62,0xf2,0x75,0x09,0x99,0x07]
-; X64-NEXT: ## xmm0 = (xmm0 * mem) + xmm1
+; X64-NEXT: ## xmm0 {%k1} = (xmm0 * mem) + xmm1
; X64-NEXT: retq ## encoding: [0xc3]
%q = load float, float* %ptr_b
%vecinit.i = insertelement <4 x float> undef, float %q, i32 0
diff --git a/llvm/test/CodeGen/X86/avx512-intrinsics.ll b/llvm/test/CodeGen/X86/avx512-intrinsics.ll
index 762a0909fb2c..2e5dc1e69c8a 100644
--- a/llvm/test/CodeGen/X86/avx512-intrinsics.ll
+++ b/llvm/test/CodeGen/X86/avx512-intrinsics.ll
@@ -5726,7 +5726,7 @@ define <2 x double> @test_int_x86_avx512_mask_vfmadd_sd(<2 x double> %x0, <2 x d
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1
; X64-NEXT: vmovapd %xmm0, %xmm3
-; X64-NEXT: vfmadd213sd {{.*#+}} xmm3 = (xmm1 * xmm3) + xmm2
+; X64-NEXT: vfmadd213sd {{.*#+}} xmm3 {%k1} = (xmm1 * xmm3) + xmm2
; X64-NEXT: vmovapd %xmm0, %xmm4
; X64-NEXT: vfmadd213sd {rz-sae}, %xmm2, %xmm1, %xmm4
; X64-NEXT: vaddpd %xmm4, %xmm3, %xmm3
@@ -5739,7 +5739,7 @@ define <2 x double> @test_int_x86_avx512_mask_vfmadd_sd(<2 x double> %x0, <2 x d
; X86-NEXT: movb {{[0-9]+}}(%esp), %al
; X86-NEXT: kmovw %eax, %k1
; X86-NEXT: vmovapd %xmm0, %xmm3
-; X86-NEXT: vfmadd213sd {{.*#+}} xmm3 = (xmm1 * xmm3) + xmm2
+; X86-NEXT: vfmadd213sd {{.*#+}} xmm3 {%k1} = (xmm1 * xmm3) + xmm2
; X86-NEXT: vmovapd %xmm0, %xmm4
; X86-NEXT: vfmadd213sd {rz-sae}, %xmm2, %xmm1, %xmm4
; X86-NEXT: vaddpd %xmm4, %xmm3, %xmm3
@@ -5777,7 +5777,7 @@ define <4 x float> @test_int_x86_avx512_mask_vfmadd_ss(<4 x float> %x0, <4 x flo
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1
; X64-NEXT: vmovaps %xmm0, %xmm3
-; X64-NEXT: vfmadd213ss {{.*#+}} xmm3 = (xmm1 * xmm3) + xmm2
+; X64-NEXT: vfmadd213ss {{.*#+}} xmm3 {%k1} = (xmm1 * xmm3) + xmm2
; X64-NEXT: vmovaps %xmm0, %xmm4
; X64-NEXT: vfmadd213ss {rz-sae}, %xmm2, %xmm1, %xmm4
; X64-NEXT: vaddps %xmm4, %xmm3, %xmm3
@@ -5790,7 +5790,7 @@ define <4 x float> @test_int_x86_avx512_mask_vfmadd_ss(<4 x float> %x0, <4 x flo
; X86-NEXT: movb {{[0-9]+}}(%esp), %al
; X86-NEXT: kmovw %eax, %k1
; X86-NEXT: vmovaps %xmm0, %xmm3
-; X86-NEXT: vfmadd213ss {{.*#+}} xmm3 = (xmm1 * xmm3) + xmm2
+; X86-NEXT: vfmadd213ss {{.*#+}} xmm3 {%k1} = (xmm1 * xmm3) + xmm2
; X86-NEXT: vmovaps %xmm0, %xmm4
; X86-NEXT: vfmadd213ss {rz-sae}, %xmm2, %xmm1, %xmm4
; X86-NEXT: vaddps %xmm4, %xmm3, %xmm3
@@ -5828,7 +5828,7 @@ define <2 x double>@test_int_x86_avx512_maskz_vfmadd_sd(<2 x double> %x0, <2 x d
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1
; X64-NEXT: vmovapd %xmm0, %xmm3
-; X64-NEXT: vfmadd213sd {{.*#+}} xmm3 = (xmm1 * xmm3) + xmm2
+; X64-NEXT: vfmadd213sd {{.*#+}} xmm3 {%k1} {z} = (xmm1 * xmm3) + xmm2
; X64-NEXT: vfmadd213sd {rz-sae}, %xmm2, %xmm1, %xmm0 {%k1} {z}
; X64-NEXT: vaddpd %xmm0, %xmm3, %xmm0
; X64-NEXT: retq
@@ -5838,7 +5838,7 @@ define <2 x double>@test_int_x86_avx512_maskz_vfmadd_sd(<2 x double> %x0, <2 x d
; X86-NEXT: movb {{[0-9]+}}(%esp), %al
; X86-NEXT: kmovw %eax, %k1
; X86-NEXT: vmovapd %xmm0, %xmm3
-; X86-NEXT: vfmadd213sd {{.*#+}} xmm3 = (xmm1 * xmm3) + xmm2
+; X86-NEXT: vfmadd213sd {{.*#+}} xmm3 {%k1} {z} = (xmm1 * xmm3) + xmm2
; X86-NEXT: vfmadd213sd {rz-sae}, %xmm2, %xmm1, %xmm0 {%k1} {z}
; X86-NEXT: vaddpd %xmm0, %xmm3, %xmm0
; X86-NEXT: retl
@@ -5870,7 +5870,7 @@ define <4 x float>@test_int_x86_avx512_maskz_vfmadd_ss(<4 x float> %x0, <4 x flo
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1
; X64-NEXT: vmovaps %xmm0, %xmm3
-; X64-NEXT: vfmadd213ss {{.*#+}} xmm3 = (xmm1 * xmm3) + xmm2
+; X64-NEXT: vfmadd213ss {{.*#+}} xmm3 {%k1} {z} = (xmm1 * xmm3) + xmm2
; X64-NEXT: vfmadd213ss {rz-sae}, %xmm2, %xmm1, %xmm0 {%k1} {z}
; X64-NEXT: vaddps %xmm0, %xmm3, %xmm0
; X64-NEXT: retq
@@ -5880,7 +5880,7 @@ define <4 x float>@test_int_x86_avx512_maskz_vfmadd_ss(<4 x float> %x0, <4 x flo
; X86-NEXT: movb {{[0-9]+}}(%esp), %al
; X86-NEXT: kmovw %eax, %k1
; X86-NEXT: vmovaps %xmm0, %xmm3
-; X86-NEXT: vfmadd213ss {{.*#+}} xmm3 = (xmm1 * xmm3) + xmm2
+; X86-NEXT: vfmadd213ss {{.*#+}} xmm3 {%k1} {z} = (xmm1 * xmm3) + xmm2
; X86-NEXT: vfmadd213ss {rz-sae}, %xmm2, %xmm1, %xmm0 {%k1} {z}
; X86-NEXT: vaddps %xmm0, %xmm3, %xmm0
; X86-NEXT: retl
@@ -5910,7 +5910,7 @@ define <4 x float> @test_int_x86_avx512_maskz_vfmadd_ss_load0(i8 zeroext %0, <4
; X64: # %bb.0:
; X64-NEXT: vmovaps (%rsi), %xmm2
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfmadd213ss {{.*#+}} xmm2 = (xmm0 * xmm2) + xmm1
+; X64-NEXT: vfmadd213ss {{.*#+}} xmm2 {%k1} {z} = (xmm0 * xmm2) + xmm1
; X64-NEXT: vmovaps %xmm2, %xmm0
; X64-NEXT: retq
;
@@ -5921,7 +5921,7 @@ define <4 x float> @test_int_x86_avx512_maskz_vfmadd_ss_load0(i8 zeroext %0, <4
; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx
; X86-NEXT: vmovaps (%ecx), %xmm0
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfmadd132ss {{.*#+}} xmm0 = (xmm0 * mem) + xmm1
+; X86-NEXT: vfmadd132ss {{.*#+}} xmm0 {%k1} {z} = (xmm0 * mem) + xmm1
; X86-NEXT: retl
%5 = load <4 x float>, <4 x float>* %1, align 16
%6 = extractelement <4 x float> %5, i64 0
@@ -5938,7 +5938,7 @@ define <2 x double> @test_int_x86_avx512_mask3_vfmadd_sd(<2 x double> %x0, <2 x
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1
; X64-NEXT: vmovapd %xmm2, %xmm3
-; X64-NEXT: vfmadd231sd {{.*#+}} xmm3 = (xmm0 * xmm1) + xmm3
+; X64-NEXT: vfmadd231sd {{.*#+}} xmm3 {%k1} = (xmm0 * xmm1) + xmm3
; X64-NEXT: vmovapd %xmm2, %xmm4
; X64-NEXT: vfmadd231sd {rz-sae}, %xmm1, %xmm0, %xmm4
; X64-NEXT: vaddpd %xmm4, %xmm3, %xmm3
@@ -5951,7 +5951,7 @@ define <2 x double> @test_int_x86_avx512_mask3_vfmadd_sd(<2 x double> %x0, <2 x
; X86-NEXT: movb {{[0-9]+}}(%esp), %al
; X86-NEXT: kmovw %eax, %k1
; X86-NEXT: vmovapd %xmm2, %xmm3
-; X86-NEXT: vfmadd231sd {{.*#+}} xmm3 = (xmm0 * xmm1) + xmm3
+; X86-NEXT: vfmadd231sd {{.*#+}} xmm3 {%k1} = (xmm0 * xmm1) + xmm3
; X86-NEXT: vmovapd %xmm2, %xmm4
; X86-NEXT: vfmadd231sd {rz-sae}, %xmm1, %xmm0, %xmm4
; X86-NEXT: vaddpd %xmm4, %xmm3, %xmm3
@@ -5989,7 +5989,7 @@ define <4 x float> @test_int_x86_avx512_mask3_vfmadd_ss(<4 x float> %x0, <4 x fl
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1
; X64-NEXT: vmovaps %xmm2, %xmm3
-; X64-NEXT: vfmadd231ss {{.*#+}} xmm3 = (xmm0 * xmm1) + xmm3
+; X64-NEXT: vfmadd231ss {{.*#+}} xmm3 {%k1} = (xmm0 * xmm1) + xmm3
; X64-NEXT: vmovaps %xmm2, %xmm4
; X64-NEXT: vfmadd231ss {rz-sae}, %xmm1, %xmm0, %xmm4
; X64-NEXT: vaddps %xmm4, %xmm3, %xmm3
@@ -6002,7 +6002,7 @@ define <4 x float> @test_int_x86_avx512_mask3_vfmadd_ss(<4 x float> %x0, <4 x fl
; X86-NEXT: movb {{[0-9]+}}(%esp), %al
; X86-NEXT: kmovw %eax, %k1
; X86-NEXT: vmovaps %xmm2, %xmm3
-; X86-NEXT: vfmadd231ss {{.*#+}} xmm3 = (xmm0 * xmm1) + xmm3
+; X86-NEXT: vfmadd231ss {{.*#+}} xmm3 {%k1} = (xmm0 * xmm1) + xmm3
; X86-NEXT: vmovaps %xmm2, %xmm4
; X86-NEXT: vfmadd231ss {rz-sae}, %xmm1, %xmm0, %xmm4
; X86-NEXT: vaddps %xmm4, %xmm3, %xmm3
@@ -6216,7 +6216,7 @@ define <2 x double> @test_int_x86_avx512_mask3_vfmsub_sd(<2 x double> %x0, <2 x
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1
; X64-NEXT: vmovapd %xmm2, %xmm3
-; X64-NEXT: vfmsub231sd {{.*#+}} xmm3 = (xmm0 * xmm1) - xmm3
+; X64-NEXT: vfmsub231sd {{.*#+}} xmm3 {%k1} = (xmm0 * xmm1) - xmm3
; X64-NEXT: vmovapd %xmm2, %xmm4
; X64-NEXT: vfmsub231sd {rz-sae}, %xmm1, %xmm0, %xmm4
; X64-NEXT: vaddpd %xmm4, %xmm3, %xmm3
@@ -6229,7 +6229,7 @@ define <2 x double> @test_int_x86_avx512_mask3_vfmsub_sd(<2 x double> %x0, <2 x
; X86-NEXT: movb {{[0-9]+}}(%esp), %al
; X86-NEXT: kmovw %eax, %k1
; X86-NEXT: vmovapd %xmm2, %xmm3
-; X86-NEXT: vfmsub231sd {{.*#+}} xmm3 = (xmm0 * xmm1) - xmm3
+; X86-NEXT: vfmsub231sd {{.*#+}} xmm3 {%k1} = (xmm0 * xmm1) - xmm3
; X86-NEXT: vmovapd %xmm2, %xmm4
; X86-NEXT: vfmsub231sd {rz-sae}, %xmm1, %xmm0, %xmm4
; X86-NEXT: vaddpd %xmm4, %xmm3, %xmm3
@@ -6273,7 +6273,7 @@ define <4 x float> @test_int_x86_avx512_mask3_vfmsub_ss(<4 x float> %x0, <4 x fl
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1
; X64-NEXT: vmovaps %xmm2, %xmm3
-; X64-NEXT: vfmsub231ss {{.*#+}} xmm3 = (xmm0 * xmm1) - xmm3
+; X64-NEXT: vfmsub231ss {{.*#+}} xmm3 {%k1} = (xmm0 * xmm1) - xmm3
; X64-NEXT: vmovaps %xmm2, %xmm4
; X64-NEXT: vfmsub231ss {rz-sae}, %xmm1, %xmm0, %xmm4
; X64-NEXT: vaddps %xmm4, %xmm3, %xmm3
@@ -6286,7 +6286,7 @@ define <4 x float> @test_int_x86_avx512_mask3_vfmsub_ss(<4 x float> %x0, <4 x fl
; X86-NEXT: movb {{[0-9]+}}(%esp), %al
; X86-NEXT: kmovw %eax, %k1
; X86-NEXT: vmovaps %xmm2, %xmm3
-; X86-NEXT: vfmsub231ss {{.*#+}} xmm3 = (xmm0 * xmm1) - xmm3
+; X86-NEXT: vfmsub231ss {{.*#+}} xmm3 {%k1} = (xmm0 * xmm1) - xmm3
; X86-NEXT: vmovaps %xmm2, %xmm4
; X86-NEXT: vfmsub231ss {rz-sae}, %xmm1, %xmm0, %xmm4
; X86-NEXT: vaddps %xmm4, %xmm3, %xmm3
@@ -6330,7 +6330,7 @@ define <2 x double> @test_int_x86_avx512_mask3_vfnmsub_sd(<2 x double> %x0, <2 x
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1
; X64-NEXT: vmovapd %xmm2, %xmm3
-; X64-NEXT: vfnmsub231sd {{.*#+}} xmm3 = -(xmm0 * xmm1) - xmm3
+; X64-NEXT: vfnmsub231sd {{.*#+}} xmm3 {%k1} = -(xmm0 * xmm1) - xmm3
; X64-NEXT: vmovapd %xmm2, %xmm4
; X64-NEXT: vfnmsub231sd {rz-sae}, %xmm1, %xmm0, %xmm4
; X64-NEXT: vaddpd %xmm4, %xmm3, %xmm3
@@ -6343,7 +6343,7 @@ define <2 x double> @test_int_x86_avx512_mask3_vfnmsub_sd(<2 x double> %x0, <2 x
; X86-NEXT: movb {{[0-9]+}}(%esp), %al
; X86-NEXT: kmovw %eax, %k1
; X86-NEXT: vmovapd %xmm2, %xmm3
-; X86-NEXT: vfnmsub231sd {{.*#+}} xmm3 = -(xmm0 * xmm1) - xmm3
+; X86-NEXT: vfnmsub231sd {{.*#+}} xmm3 {%k1} = -(xmm0 * xmm1) - xmm3
; X86-NEXT: vmovapd %xmm2, %xmm4
; X86-NEXT: vfnmsub231sd {rz-sae}, %xmm1, %xmm0, %xmm4
; X86-NEXT: vaddpd %xmm4, %xmm3, %xmm3
@@ -6390,7 +6390,7 @@ define <4 x float> @test_int_x86_avx512_mask3_vfnmsub_ss(<4 x float> %x0, <4 x f
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1
; X64-NEXT: vmovaps %xmm2, %xmm3
-; X64-NEXT: vfnmsub231ss {{.*#+}} xmm3 = -(xmm0 * xmm1) - xmm3
+; X64-NEXT: vfnmsub231ss {{.*#+}} xmm3 {%k1} = -(xmm0 * xmm1) - xmm3
; X64-NEXT: vmovaps %xmm2, %xmm4
; X64-NEXT: vfnmsub231ss {rz-sae}, %xmm1, %xmm0, %xmm4
; X64-NEXT: vaddps %xmm4, %xmm3, %xmm3
@@ -6403,7 +6403,7 @@ define <4 x float> @test_int_x86_avx512_mask3_vfnmsub_ss(<4 x float> %x0, <4 x f
; X86-NEXT: movb {{[0-9]+}}(%esp), %al
; X86-NEXT: kmovw %eax, %k1
; X86-NEXT: vmovaps %xmm2, %xmm3
-; X86-NEXT: vfnmsub231ss {{.*#+}} xmm3 = -(xmm0 * xmm1) - xmm3
+; X86-NEXT: vfnmsub231ss {{.*#+}} xmm3 {%k1} = -(xmm0 * xmm1) - xmm3
; X86-NEXT: vmovaps %xmm2, %xmm4
; X86-NEXT: vfnmsub231ss {rz-sae}, %xmm1, %xmm0, %xmm4
; X86-NEXT: vaddps %xmm4, %xmm3, %xmm3
@@ -6449,7 +6449,7 @@ define <4 x float>@test_int_x86_avx512_mask3_vfmadd_ss_rm(<4 x float> %x0, <4 x
; X64-LABEL: test_int_x86_avx512_mask3_vfmadd_ss_rm:
; X64: # %bb.0:
; X64-NEXT: kmovw %esi, %k1
-; X64-NEXT: vfmadd231ss {{.*#+}} xmm1 = (xmm0 * mem) + xmm1
+; X64-NEXT: vfmadd231ss {{.*#+}} xmm1 {%k1} = (xmm0 * mem) + xmm1
; X64-NEXT: vmovaps %xmm1, %xmm0
; X64-NEXT: retq
;
@@ -6458,7 +6458,7 @@ define <4 x float>@test_int_x86_avx512_mask3_vfmadd_ss_rm(<4 x float> %x0, <4 x
; X86-NEXT: movl {{[0-9]+}}(%esp), %eax
; X86-NEXT: movb {{[0-9]+}}(%esp), %cl
; X86-NEXT: kmovw %ecx, %k1
-; X86-NEXT: vfmadd231ss {{.*#+}} xmm1 = (xmm0 * mem) + xmm1
+; X86-NEXT: vfmadd231ss {{.*#+}} xmm1 {%k1} = (xmm0 * mem) + xmm1
; X86-NEXT: vmovaps %xmm1, %xmm0
; X86-NEXT: retl
%q = load float, float* %ptr_b
@@ -6478,7 +6478,7 @@ define <4 x float>@test_int_x86_avx512_mask_vfmadd_ss_rm(<4 x float> %x0, <4 x f
; X64-LABEL: test_int_x86_avx512_mask_vfmadd_ss_rm:
; X64: # %bb.0:
; X64-NEXT: kmovw %esi, %k1
-; X64-NEXT: vfmadd132ss {{.*#+}} xmm0 = (xmm0 * mem) + xmm1
+; X64-NEXT: vfmadd132ss {{.*#+}} xmm0 {%k1} = (xmm0 * mem) + xmm1
; X64-NEXT: retq
;
; X86-LABEL: test_int_x86_avx512_mask_vfmadd_ss_rm:
@@ -6486,7 +6486,7 @@ define <4 x float>@test_int_x86_avx512_mask_vfmadd_ss_rm(<4 x float> %x0, <4 x f
; X86-NEXT: movl {{[0-9]+}}(%esp), %eax
; X86-NEXT: movb {{[0-9]+}}(%esp), %cl
; X86-NEXT: kmovw %ecx, %k1
-; X86-NEXT: vfmadd132ss {{.*#+}} xmm0 = (xmm0 * mem) + xmm1
+; X86-NEXT: vfmadd132ss {{.*#+}} xmm0 {%k1} = (xmm0 * mem) + xmm1
; X86-NEXT: retl
%q = load float, float* %ptr_b
%vecinit.i = insertelement <4 x float> undef, float %q, i32 0
diff --git a/llvm/test/CodeGen/X86/avx512-memfold.ll b/llvm/test/CodeGen/X86/avx512-memfold.ll
index 3b1fe7cec002..e0db40c8fee8 100644
--- a/llvm/test/CodeGen/X86/avx512-memfold.ll
+++ b/llvm/test/CodeGen/X86/avx512-memfold.ll
@@ -58,7 +58,7 @@ define <2 x double> @test_int_x86_avx512_mask_vfmadd_sd(<2 x double> %a, <2 x do
; CHECK-LABEL: test_int_x86_avx512_mask_vfmadd_sd:
; CHECK: ## %bb.0:
; CHECK-NEXT: kmovw %esi, %k1
-; CHECK-NEXT: vfmadd213sd {{.*#+}} xmm0 = (xmm1 * xmm0) + mem
+; CHECK-NEXT: vfmadd213sd {{.*#+}} xmm0 {%k1} = (xmm1 * xmm0) + mem
; CHECK-NEXT: retq
%c.val = load double, double* %c
%cv0 = insertelement <2 x double> undef, double %c.val, i32 0
diff --git a/llvm/test/CodeGen/X86/avx512-scalar_mask.ll b/llvm/test/CodeGen/X86/avx512-scalar_mask.ll
index 27468fecfde7..9e9fc57cf7b6 100644
--- a/llvm/test/CodeGen/X86/avx512-scalar_mask.ll
+++ b/llvm/test/CodeGen/X86/avx512-scalar_mask.ll
@@ -8,7 +8,7 @@ define <4 x float>@test_var_mask(<4 x float> %v0, <4 x float> %v1, <4 x float> %
; CHECK-LABEL: test_var_mask:
; CHECK: ## %bb.0:
; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vfmadd213ss {{.*#+}} xmm0 = (xmm1 * xmm0) + xmm2
+; CHECK-NEXT: vfmadd213ss {{.*#+}} xmm0 {%k1} = (xmm1 * xmm0) + xmm2
; CHECK-NEXT: retq
%res = call <4 x float> @llvm.x86.avx512.mask.vfmadd.ss(<4 x float> %v0,<4 x float> %v1, <4 x float> %v2, i8 %mask, i32 4)
ret < 4 x float> %res
@@ -18,7 +18,7 @@ define <4 x float>@test_var_maskz(<4 x float> %v0, <4 x float> %v1, <4 x float>
; CHECK-LABEL: test_var_maskz:
; CHECK: ## %bb.0:
; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vfmadd213ss {{.*#+}} xmm0 = (xmm1 * xmm0) + xmm2
+; CHECK-NEXT: vfmadd213ss {{.*#+}} xmm0 {%k1} {z} = (xmm1 * xmm0) + xmm2
; CHECK-NEXT: retq
%res = call <4 x float> @llvm.x86.avx512.maskz.vfmadd.ss(<4 x float> %v0,<4 x float> %v1, <4 x float> %v2, i8 %mask, i32 4)
ret < 4 x float> %res
diff --git a/llvm/test/CodeGen/X86/avx512vl-intrinsics-canonical.ll b/llvm/test/CodeGen/X86/avx512vl-intrinsics-canonical.ll
index b6f61e04158f..429cd0d8c38e 100644
--- a/llvm/test/CodeGen/X86/avx512vl-intrinsics-canonical.ll
+++ b/llvm/test/CodeGen/X86/avx512vl-intrinsics-canonical.ll
@@ -10,14 +10,14 @@ define <2 x double> @test_mm_mask_fmadd_pd(<2 x double> %__A, i8 zeroext %__U, <
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmadd132pd %xmm1, %xmm2, %xmm0 {%k1} # encoding: [0x62,0xf2,0xed,0x09,0x98,0xc1]
-; X86-NEXT: # xmm0 = (xmm0 * xmm1) + xmm2
+; X86-NEXT: # xmm0 {%k1} = (xmm0 * xmm1) + xmm2
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_mm_mask_fmadd_pd:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmadd132pd %xmm1, %xmm2, %xmm0 {%k1} # encoding: [0x62,0xf2,0xed,0x09,0x98,0xc1]
-; X64-NEXT: # xmm0 = (xmm0 * xmm1) + xmm2
+; X64-NEXT: # xmm0 {%k1} = (xmm0 * xmm1) + xmm2
; X64-NEXT: retq # encoding: [0xc3]
entry:
%0 = tail call <2 x double> @llvm.fma.v2f64(<2 x double> %__A, <2 x double> %__B, <2 x double> %__C) #9
@@ -33,14 +33,14 @@ define <2 x double> @test_mm_mask_fmsub_pd(<2 x double> %__A, i8 zeroext %__U, <
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmsub132pd %xmm1, %xmm2, %xmm0 {%k1} # encoding: [0x62,0xf2,0xed,0x09,0x9a,0xc1]
-; X86-NEXT: # xmm0 = (xmm0 * xmm1) - xmm2
+; X86-NEXT: # xmm0 {%k1} = (xmm0 * xmm1) - xmm2
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_mm_mask_fmsub_pd:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmsub132pd %xmm1, %xmm2, %xmm0 {%k1} # encoding: [0x62,0xf2,0xed,0x09,0x9a,0xc1]
-; X64-NEXT: # xmm0 = (xmm0 * xmm1) - xmm2
+; X64-NEXT: # xmm0 {%k1} = (xmm0 * xmm1) - xmm2
; X64-NEXT: retq # encoding: [0xc3]
entry:
%sub.i = fsub <2 x double> <double -0.000000e+00, double -0.000000e+00>, %__C
@@ -57,7 +57,7 @@ define <2 x double> @test_mm_mask3_fmadd_pd(<2 x double> %__A, <2 x double> %__B
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmadd231pd %xmm1, %xmm0, %xmm2 {%k1} # encoding: [0x62,0xf2,0xfd,0x09,0xb8,0xd1]
-; X86-NEXT: # xmm2 = (xmm0 * xmm1) + xmm2
+; X86-NEXT: # xmm2 {%k1} = (xmm0 * xmm1) + xmm2
; X86-NEXT: vmovapd %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x28,0xc2]
; X86-NEXT: retl # encoding: [0xc3]
;
@@ -65,7 +65,7 @@ define <2 x double> @test_mm_mask3_fmadd_pd(<2 x double> %__A, <2 x double> %__B
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmadd231pd %xmm1, %xmm0, %xmm2 {%k1} # encoding: [0x62,0xf2,0xfd,0x09,0xb8,0xd1]
-; X64-NEXT: # xmm2 = (xmm0 * xmm1) + xmm2
+; X64-NEXT: # xmm2 {%k1} = (xmm0 * xmm1) + xmm2
; X64-NEXT: vmovapd %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x28,0xc2]
; X64-NEXT: retq # encoding: [0xc3]
entry:
@@ -82,7 +82,7 @@ define <2 x double> @test_mm_mask3_fnmadd_pd(<2 x double> %__A, <2 x double> %__
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfnmadd231pd %xmm1, %xmm0, %xmm2 {%k1} # encoding: [0x62,0xf2,0xfd,0x09,0xbc,0xd1]
-; X86-NEXT: # xmm2 = -(xmm0 * xmm1) + xmm2
+; X86-NEXT: # xmm2 {%k1} = -(xmm0 * xmm1) + xmm2
; X86-NEXT: vmovapd %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x28,0xc2]
; X86-NEXT: retl # encoding: [0xc3]
;
@@ -90,7 +90,7 @@ define <2 x double> @test_mm_mask3_fnmadd_pd(<2 x double> %__A, <2 x double> %__
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfnmadd231pd %xmm1, %xmm0, %xmm2 {%k1} # encoding: [0x62,0xf2,0xfd,0x09,0xbc,0xd1]
-; X64-NEXT: # xmm2 = -(xmm0 * xmm1) + xmm2
+; X64-NEXT: # xmm2 {%k1} = -(xmm0 * xmm1) + xmm2
; X64-NEXT: vmovapd %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x28,0xc2]
; X64-NEXT: retq # encoding: [0xc3]
entry:
@@ -108,14 +108,14 @@ define <2 x double> @test_mm_maskz_fmadd_pd(i8 zeroext %__U, <2 x double> %__A,
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmadd213pd %xmm2, %xmm1, %xmm0 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0x89,0xa8,0xc2]
-; X86-NEXT: # xmm0 = (xmm1 * xmm0) + xmm2
+; X86-NEXT: # xmm0 {%k1} {z} = (xmm1 * xmm0) + xmm2
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_mm_maskz_fmadd_pd:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmadd213pd %xmm2, %xmm1, %xmm0 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0x89,0xa8,0xc2]
-; X64-NEXT: # xmm0 = (xmm1 * xmm0) + xmm2
+; X64-NEXT: # xmm0 {%k1} {z} = (xmm1 * xmm0) + xmm2
; X64-NEXT: retq # encoding: [0xc3]
entry:
%0 = tail call <2 x double> @llvm.fma.v2f64(<2 x double> %__A, <2 x double> %__B, <2 x double> %__C) #9
@@ -131,14 +131,14 @@ define <2 x double> @test_mm_maskz_fmsub_pd(i8 zeroext %__U, <2 x double> %__A,
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmsub213pd %xmm2, %xmm1, %xmm0 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0x89,0xaa,0xc2]
-; X86-NEXT: # xmm0 = (xmm1 * xmm0) - xmm2
+; X86-NEXT: # xmm0 {%k1} {z} = (xmm1 * xmm0) - xmm2
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_mm_maskz_fmsub_pd:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmsub213pd %xmm2, %xmm1, %xmm0 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0x89,0xaa,0xc2]
-; X64-NEXT: # xmm0 = (xmm1 * xmm0) - xmm2
+; X64-NEXT: # xmm0 {%k1} {z} = (xmm1 * xmm0) - xmm2
; X64-NEXT: retq # encoding: [0xc3]
entry:
%sub.i = fsub <2 x double> <double -0.000000e+00, double -0.000000e+00>, %__C
@@ -155,14 +155,14 @@ define <2 x double> @test_mm_maskz_fnmadd_pd(i8 zeroext %__U, <2 x double> %__A,
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfnmadd213pd %xmm2, %xmm1, %xmm0 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0x89,0xac,0xc2]
-; X86-NEXT: # xmm0 = -(xmm1 * xmm0) + xmm2
+; X86-NEXT: # xmm0 {%k1} {z} = -(xmm1 * xmm0) + xmm2
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_mm_maskz_fnmadd_pd:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfnmadd213pd %xmm2, %xmm1, %xmm0 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0x89,0xac,0xc2]
-; X64-NEXT: # xmm0 = -(xmm1 * xmm0) + xmm2
+; X64-NEXT: # xmm0 {%k1} {z} = -(xmm1 * xmm0) + xmm2
; X64-NEXT: retq # encoding: [0xc3]
entry:
%sub.i = fsub <2 x double> <double -0.000000e+00, double -0.000000e+00>, %__A
@@ -179,14 +179,14 @@ define <2 x double> @test_mm_maskz_fnmsub_pd(i8 zeroext %__U, <2 x double> %__A,
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfnmsub213pd %xmm2, %xmm1, %xmm0 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0x89,0xae,0xc2]
-; X86-NEXT: # xmm0 = -(xmm1 * xmm0) - xmm2
+; X86-NEXT: # xmm0 {%k1} {z} = -(xmm1 * xmm0) - xmm2
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_mm_maskz_fnmsub_pd:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfnmsub213pd %xmm2, %xmm1, %xmm0 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0x89,0xae,0xc2]
-; X64-NEXT: # xmm0 = -(xmm1 * xmm0) - xmm2
+; X64-NEXT: # xmm0 {%k1} {z} = -(xmm1 * xmm0) - xmm2
; X64-NEXT: retq # encoding: [0xc3]
entry:
%sub.i = fsub <2 x double> <double -0.000000e+00, double -0.000000e+00>, %__A
@@ -204,14 +204,14 @@ define <4 x double> @test_mm256_mask_fmadd_pd(<4 x double> %__A, i8 zeroext %__U
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmadd132pd %ymm1, %ymm2, %ymm0 {%k1} # encoding: [0x62,0xf2,0xed,0x29,0x98,0xc1]
-; X86-NEXT: # ymm0 = (ymm0 * ymm1) + ymm2
+; X86-NEXT: # ymm0 {%k1} = (ymm0 * ymm1) + ymm2
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_mm256_mask_fmadd_pd:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmadd132pd %ymm1, %ymm2, %ymm0 {%k1} # encoding: [0x62,0xf2,0xed,0x29,0x98,0xc1]
-; X64-NEXT: # ymm0 = (ymm0 * ymm1) + ymm2
+; X64-NEXT: # ymm0 {%k1} = (ymm0 * ymm1) + ymm2
; X64-NEXT: retq # encoding: [0xc3]
entry:
%0 = tail call <4 x double> @llvm.fma.v4f64(<4 x double> %__A, <4 x double> %__B, <4 x double> %__C) #9
@@ -227,14 +227,14 @@ define <4 x double> @test_mm256_mask_fmsub_pd(<4 x double> %__A, i8 zeroext %__U
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmsub132pd %ymm1, %ymm2, %ymm0 {%k1} # encoding: [0x62,0xf2,0xed,0x29,0x9a,0xc1]
-; X86-NEXT: # ymm0 = (ymm0 * ymm1) - ymm2
+; X86-NEXT: # ymm0 {%k1} = (ymm0 * ymm1) - ymm2
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_mm256_mask_fmsub_pd:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmsub132pd %ymm1, %ymm2, %ymm0 {%k1} # encoding: [0x62,0xf2,0xed,0x29,0x9a,0xc1]
-; X64-NEXT: # ymm0 = (ymm0 * ymm1) - ymm2
+; X64-NEXT: # ymm0 {%k1} = (ymm0 * ymm1) - ymm2
; X64-NEXT: retq # encoding: [0xc3]
entry:
%sub.i = fsub <4 x double> <double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00>, %__C
@@ -251,7 +251,7 @@ define <4 x double> @test_mm256_mask3_fmadd_pd(<4 x double> %__A, <4 x double> %
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmadd231pd %ymm1, %ymm0, %ymm2 {%k1} # encoding: [0x62,0xf2,0xfd,0x29,0xb8,0xd1]
-; X86-NEXT: # ymm2 = (ymm0 * ymm1) + ymm2
+; X86-NEXT: # ymm2 {%k1} = (ymm0 * ymm1) + ymm2
; X86-NEXT: vmovapd %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x28,0xc2]
; X86-NEXT: retl # encoding: [0xc3]
;
@@ -259,7 +259,7 @@ define <4 x double> @test_mm256_mask3_fmadd_pd(<4 x double> %__A, <4 x double> %
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmadd231pd %ymm1, %ymm0, %ymm2 {%k1} # encoding: [0x62,0xf2,0xfd,0x29,0xb8,0xd1]
-; X64-NEXT: # ymm2 = (ymm0 * ymm1) + ymm2
+; X64-NEXT: # ymm2 {%k1} = (ymm0 * ymm1) + ymm2
; X64-NEXT: vmovapd %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x28,0xc2]
; X64-NEXT: retq # encoding: [0xc3]
entry:
@@ -276,7 +276,7 @@ define <4 x double> @test_mm256_mask3_fnmadd_pd(<4 x double> %__A, <4 x double>
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfnmadd231pd %ymm1, %ymm0, %ymm2 {%k1} # encoding: [0x62,0xf2,0xfd,0x29,0xbc,0xd1]
-; X86-NEXT: # ymm2 = -(ymm0 * ymm1) + ymm2
+; X86-NEXT: # ymm2 {%k1} = -(ymm0 * ymm1) + ymm2
; X86-NEXT: vmovapd %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x28,0xc2]
; X86-NEXT: retl # encoding: [0xc3]
;
@@ -284,7 +284,7 @@ define <4 x double> @test_mm256_mask3_fnmadd_pd(<4 x double> %__A, <4 x double>
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfnmadd231pd %ymm1, %ymm0, %ymm2 {%k1} # encoding: [0x62,0xf2,0xfd,0x29,0xbc,0xd1]
-; X64-NEXT: # ymm2 = -(ymm0 * ymm1) + ymm2
+; X64-NEXT: # ymm2 {%k1} = -(ymm0 * ymm1) + ymm2
; X64-NEXT: vmovapd %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x28,0xc2]
; X64-NEXT: retq # encoding: [0xc3]
entry:
@@ -302,14 +302,14 @@ define <4 x double> @test_mm256_maskz_fmadd_pd(i8 zeroext %__U, <4 x double> %__
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmadd213pd %ymm2, %ymm1, %ymm0 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0xa9,0xa8,0xc2]
-; X86-NEXT: # ymm0 = (ymm1 * ymm0) + ymm2
+; X86-NEXT: # ymm0 {%k1} {z} = (ymm1 * ymm0) + ymm2
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_mm256_maskz_fmadd_pd:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmadd213pd %ymm2, %ymm1, %ymm0 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0xa9,0xa8,0xc2]
-; X64-NEXT: # ymm0 = (ymm1 * ymm0) + ymm2
+; X64-NEXT: # ymm0 {%k1} {z} = (ymm1 * ymm0) + ymm2
; X64-NEXT: retq # encoding: [0xc3]
entry:
%0 = tail call <4 x double> @llvm.fma.v4f64(<4 x double> %__A, <4 x double> %__B, <4 x double> %__C) #9
@@ -325,14 +325,14 @@ define <4 x double> @test_mm256_maskz_fmsub_pd(i8 zeroext %__U, <4 x double> %__
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmsub213pd %ymm2, %ymm1, %ymm0 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0xa9,0xaa,0xc2]
-; X86-NEXT: # ymm0 = (ymm1 * ymm0) - ymm2
+; X86-NEXT: # ymm0 {%k1} {z} = (ymm1 * ymm0) - ymm2
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_mm256_maskz_fmsub_pd:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmsub213pd %ymm2, %ymm1, %ymm0 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0xa9,0xaa,0xc2]
-; X64-NEXT: # ymm0 = (ymm1 * ymm0) - ymm2
+; X64-NEXT: # ymm0 {%k1} {z} = (ymm1 * ymm0) - ymm2
; X64-NEXT: retq # encoding: [0xc3]
entry:
%sub.i = fsub <4 x double> <double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00>, %__C
@@ -349,14 +349,14 @@ define <4 x double> @test_mm256_maskz_fnmadd_pd(i8 zeroext %__U, <4 x double> %_
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfnmadd213pd %ymm2, %ymm1, %ymm0 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0xa9,0xac,0xc2]
-; X86-NEXT: # ymm0 = -(ymm1 * ymm0) + ymm2
+; X86-NEXT: # ymm0 {%k1} {z} = -(ymm1 * ymm0) + ymm2
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_mm256_maskz_fnmadd_pd:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfnmadd213pd %ymm2, %ymm1, %ymm0 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0xa9,0xac,0xc2]
-; X64-NEXT: # ymm0 = -(ymm1 * ymm0) + ymm2
+; X64-NEXT: # ymm0 {%k1} {z} = -(ymm1 * ymm0) + ymm2
; X64-NEXT: retq # encoding: [0xc3]
entry:
%sub.i = fsub <4 x double> <double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00>, %__A
@@ -373,14 +373,14 @@ define <4 x double> @test_mm256_maskz_fnmsub_pd(i8 zeroext %__U, <4 x double> %_
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfnmsub213pd %ymm2, %ymm1, %ymm0 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0xa9,0xae,0xc2]
-; X86-NEXT: # ymm0 = -(ymm1 * ymm0) - ymm2
+; X86-NEXT: # ymm0 {%k1} {z} = -(ymm1 * ymm0) - ymm2
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_mm256_maskz_fnmsub_pd:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfnmsub213pd %ymm2, %ymm1, %ymm0 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0xa9,0xae,0xc2]
-; X64-NEXT: # ymm0 = -(ymm1 * ymm0) - ymm2
+; X64-NEXT: # ymm0 {%k1} {z} = -(ymm1 * ymm0) - ymm2
; X64-NEXT: retq # encoding: [0xc3]
entry:
%sub.i = fsub <4 x double> <double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00>, %__A
@@ -398,14 +398,14 @@ define <4 x float> @test_mm_mask_fmadd_ps(<4 x float> %__A, i8 zeroext %__U, <4
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmadd132ps %xmm1, %xmm2, %xmm0 {%k1} # encoding: [0x62,0xf2,0x6d,0x09,0x98,0xc1]
-; X86-NEXT: # xmm0 = (xmm0 * xmm1) + xmm2
+; X86-NEXT: # xmm0 {%k1} = (xmm0 * xmm1) + xmm2
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_mm_mask_fmadd_ps:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmadd132ps %xmm1, %xmm2, %xmm0 {%k1} # encoding: [0x62,0xf2,0x6d,0x09,0x98,0xc1]
-; X64-NEXT: # xmm0 = (xmm0 * xmm1) + xmm2
+; X64-NEXT: # xmm0 {%k1} = (xmm0 * xmm1) + xmm2
; X64-NEXT: retq # encoding: [0xc3]
entry:
%0 = tail call <4 x float> @llvm.fma.v4f32(<4 x float> %__A, <4 x float> %__B, <4 x float> %__C) #9
@@ -421,14 +421,14 @@ define <4 x float> @test_mm_mask_fmsub_ps(<4 x float> %__A, i8 zeroext %__U, <4
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmsub132ps %xmm1, %xmm2, %xmm0 {%k1} # encoding: [0x62,0xf2,0x6d,0x09,0x9a,0xc1]
-; X86-NEXT: # xmm0 = (xmm0 * xmm1) - xmm2
+; X86-NEXT: # xmm0 {%k1} = (xmm0 * xmm1) - xmm2
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_mm_mask_fmsub_ps:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmsub132ps %xmm1, %xmm2, %xmm0 {%k1} # encoding: [0x62,0xf2,0x6d,0x09,0x9a,0xc1]
-; X64-NEXT: # xmm0 = (xmm0 * xmm1) - xmm2
+; X64-NEXT: # xmm0 {%k1} = (xmm0 * xmm1) - xmm2
; X64-NEXT: retq # encoding: [0xc3]
entry:
%sub.i = fsub <4 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %__C
@@ -445,7 +445,7 @@ define <4 x float> @test_mm_mask3_fmadd_ps(<4 x float> %__A, <4 x float> %__B, <
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmadd231ps %xmm1, %xmm0, %xmm2 {%k1} # encoding: [0x62,0xf2,0x7d,0x09,0xb8,0xd1]
-; X86-NEXT: # xmm2 = (xmm0 * xmm1) + xmm2
+; X86-NEXT: # xmm2 {%k1} = (xmm0 * xmm1) + xmm2
; X86-NEXT: vmovaps %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf8,0x28,0xc2]
; X86-NEXT: retl # encoding: [0xc3]
;
@@ -453,7 +453,7 @@ define <4 x float> @test_mm_mask3_fmadd_ps(<4 x float> %__A, <4 x float> %__B, <
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmadd231ps %xmm1, %xmm0, %xmm2 {%k1} # encoding: [0x62,0xf2,0x7d,0x09,0xb8,0xd1]
-; X64-NEXT: # xmm2 = (xmm0 * xmm1) + xmm2
+; X64-NEXT: # xmm2 {%k1} = (xmm0 * xmm1) + xmm2
; X64-NEXT: vmovaps %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf8,0x28,0xc2]
; X64-NEXT: retq # encoding: [0xc3]
entry:
@@ -470,7 +470,7 @@ define <4 x float> @test_mm_mask3_fnmadd_ps(<4 x float> %__A, <4 x float> %__B,
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfnmadd231ps %xmm1, %xmm0, %xmm2 {%k1} # encoding: [0x62,0xf2,0x7d,0x09,0xbc,0xd1]
-; X86-NEXT: # xmm2 = -(xmm0 * xmm1) + xmm2
+; X86-NEXT: # xmm2 {%k1} = -(xmm0 * xmm1) + xmm2
; X86-NEXT: vmovaps %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf8,0x28,0xc2]
; X86-NEXT: retl # encoding: [0xc3]
;
@@ -478,7 +478,7 @@ define <4 x float> @test_mm_mask3_fnmadd_ps(<4 x float> %__A, <4 x float> %__B,
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfnmadd231ps %xmm1, %xmm0, %xmm2 {%k1} # encoding: [0x62,0xf2,0x7d,0x09,0xbc,0xd1]
-; X64-NEXT: # xmm2 = -(xmm0 * xmm1) + xmm2
+; X64-NEXT: # xmm2 {%k1} = -(xmm0 * xmm1) + xmm2
; X64-NEXT: vmovaps %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf8,0x28,0xc2]
; X64-NEXT: retq # encoding: [0xc3]
entry:
@@ -496,14 +496,14 @@ define <4 x float> @test_mm_maskz_fmadd_ps(i8 zeroext %__U, <4 x float> %__A, <4
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmadd213ps %xmm2, %xmm1, %xmm0 {%k1} {z} # encoding: [0x62,0xf2,0x75,0x89,0xa8,0xc2]
-; X86-NEXT: # xmm0 = (xmm1 * xmm0) + xmm2
+; X86-NEXT: # xmm0 {%k1} {z} = (xmm1 * xmm0) + xmm2
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_mm_maskz_fmadd_ps:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmadd213ps %xmm2, %xmm1, %xmm0 {%k1} {z} # encoding: [0x62,0xf2,0x75,0x89,0xa8,0xc2]
-; X64-NEXT: # xmm0 = (xmm1 * xmm0) + xmm2
+; X64-NEXT: # xmm0 {%k1} {z} = (xmm1 * xmm0) + xmm2
; X64-NEXT: retq # encoding: [0xc3]
entry:
%0 = tail call <4 x float> @llvm.fma.v4f32(<4 x float> %__A, <4 x float> %__B, <4 x float> %__C) #9
@@ -519,14 +519,14 @@ define <4 x float> @test_mm_maskz_fmsub_ps(i8 zeroext %__U, <4 x float> %__A, <4
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmsub213ps %xmm2, %xmm1, %xmm0 {%k1} {z} # encoding: [0x62,0xf2,0x75,0x89,0xaa,0xc2]
-; X86-NEXT: # xmm0 = (xmm1 * xmm0) - xmm2
+; X86-NEXT: # xmm0 {%k1} {z} = (xmm1 * xmm0) - xmm2
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_mm_maskz_fmsub_ps:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmsub213ps %xmm2, %xmm1, %xmm0 {%k1} {z} # encoding: [0x62,0xf2,0x75,0x89,0xaa,0xc2]
-; X64-NEXT: # xmm0 = (xmm1 * xmm0) - xmm2
+; X64-NEXT: # xmm0 {%k1} {z} = (xmm1 * xmm0) - xmm2
; X64-NEXT: retq # encoding: [0xc3]
entry:
%sub.i = fsub <4 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %__C
@@ -543,14 +543,14 @@ define <4 x float> @test_mm_maskz_fnmadd_ps(i8 zeroext %__U, <4 x float> %__A, <
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfnmadd213ps %xmm2, %xmm1, %xmm0 {%k1} {z} # encoding: [0x62,0xf2,0x75,0x89,0xac,0xc2]
-; X86-NEXT: # xmm0 = -(xmm1 * xmm0) + xmm2
+; X86-NEXT: # xmm0 {%k1} {z} = -(xmm1 * xmm0) + xmm2
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_mm_maskz_fnmadd_ps:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfnmadd213ps %xmm2, %xmm1, %xmm0 {%k1} {z} # encoding: [0x62,0xf2,0x75,0x89,0xac,0xc2]
-; X64-NEXT: # xmm0 = -(xmm1 * xmm0) + xmm2
+; X64-NEXT: # xmm0 {%k1} {z} = -(xmm1 * xmm0) + xmm2
; X64-NEXT: retq # encoding: [0xc3]
entry:
%sub.i = fsub <4 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %__A
@@ -567,14 +567,14 @@ define <4 x float> @test_mm_maskz_fnmsub_ps(i8 zeroext %__U, <4 x float> %__A, <
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfnmsub213ps %xmm2, %xmm1, %xmm0 {%k1} {z} # encoding: [0x62,0xf2,0x75,0x89,0xae,0xc2]
-; X86-NEXT: # xmm0 = -(xmm1 * xmm0) - xmm2
+; X86-NEXT: # xmm0 {%k1} {z} = -(xmm1 * xmm0) - xmm2
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_mm_maskz_fnmsub_ps:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfnmsub213ps %xmm2, %xmm1, %xmm0 {%k1} {z} # encoding: [0x62,0xf2,0x75,0x89,0xae,0xc2]
-; X64-NEXT: # xmm0 = -(xmm1 * xmm0) - xmm2
+; X64-NEXT: # xmm0 {%k1} {z} = -(xmm1 * xmm0) - xmm2
; X64-NEXT: retq # encoding: [0xc3]
entry:
%sub.i = fsub <4 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %__A
@@ -592,14 +592,14 @@ define <8 x float> @test_mm256_mask_fmadd_ps(<8 x float> %__A, i8 zeroext %__U,
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmadd132ps %ymm1, %ymm2, %ymm0 {%k1} # encoding: [0x62,0xf2,0x6d,0x29,0x98,0xc1]
-; X86-NEXT: # ymm0 = (ymm0 * ymm1) + ymm2
+; X86-NEXT: # ymm0 {%k1} = (ymm0 * ymm1) + ymm2
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_mm256_mask_fmadd_ps:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmadd132ps %ymm1, %ymm2, %ymm0 {%k1} # encoding: [0x62,0xf2,0x6d,0x29,0x98,0xc1]
-; X64-NEXT: # ymm0 = (ymm0 * ymm1) + ymm2
+; X64-NEXT: # ymm0 {%k1} = (ymm0 * ymm1) + ymm2
; X64-NEXT: retq # encoding: [0xc3]
entry:
%0 = tail call <8 x float> @llvm.fma.v8f32(<8 x float> %__A, <8 x float> %__B, <8 x float> %__C) #9
@@ -614,14 +614,14 @@ define <8 x float> @test_mm256_mask_fmsub_ps(<8 x float> %__A, i8 zeroext %__U,
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmsub132ps %ymm1, %ymm2, %ymm0 {%k1} # encoding: [0x62,0xf2,0x6d,0x29,0x9a,0xc1]
-; X86-NEXT: # ymm0 = (ymm0 * ymm1) - ymm2
+; X86-NEXT: # ymm0 {%k1} = (ymm0 * ymm1) - ymm2
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_mm256_mask_fmsub_ps:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmsub132ps %ymm1, %ymm2, %ymm0 {%k1} # encoding: [0x62,0xf2,0x6d,0x29,0x9a,0xc1]
-; X64-NEXT: # ymm0 = (ymm0 * ymm1) - ymm2
+; X64-NEXT: # ymm0 {%k1} = (ymm0 * ymm1) - ymm2
; X64-NEXT: retq # encoding: [0xc3]
entry:
%sub.i = fsub <8 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %__C
@@ -637,7 +637,7 @@ define <8 x float> @test_mm256_mask3_fmadd_ps(<8 x float> %__A, <8 x float> %__B
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmadd231ps %ymm1, %ymm0, %ymm2 {%k1} # encoding: [0x62,0xf2,0x7d,0x29,0xb8,0xd1]
-; X86-NEXT: # ymm2 = (ymm0 * ymm1) + ymm2
+; X86-NEXT: # ymm2 {%k1} = (ymm0 * ymm1) + ymm2
; X86-NEXT: vmovaps %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xfc,0x28,0xc2]
; X86-NEXT: retl # encoding: [0xc3]
;
@@ -645,7 +645,7 @@ define <8 x float> @test_mm256_mask3_fmadd_ps(<8 x float> %__A, <8 x float> %__B
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmadd231ps %ymm1, %ymm0, %ymm2 {%k1} # encoding: [0x62,0xf2,0x7d,0x29,0xb8,0xd1]
-; X64-NEXT: # ymm2 = (ymm0 * ymm1) + ymm2
+; X64-NEXT: # ymm2 {%k1} = (ymm0 * ymm1) + ymm2
; X64-NEXT: vmovaps %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xfc,0x28,0xc2]
; X64-NEXT: retq # encoding: [0xc3]
entry:
@@ -661,7 +661,7 @@ define <8 x float> @test_mm256_mask3_fnmadd_ps(<8 x float> %__A, <8 x float> %__
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfnmadd231ps %ymm1, %ymm0, %ymm2 {%k1} # encoding: [0x62,0xf2,0x7d,0x29,0xbc,0xd1]
-; X86-NEXT: # ymm2 = -(ymm0 * ymm1) + ymm2
+; X86-NEXT: # ymm2 {%k1} = -(ymm0 * ymm1) + ymm2
; X86-NEXT: vmovaps %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xfc,0x28,0xc2]
; X86-NEXT: retl # encoding: [0xc3]
;
@@ -669,7 +669,7 @@ define <8 x float> @test_mm256_mask3_fnmadd_ps(<8 x float> %__A, <8 x float> %__
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfnmadd231ps %ymm1, %ymm0, %ymm2 {%k1} # encoding: [0x62,0xf2,0x7d,0x29,0xbc,0xd1]
-; X64-NEXT: # ymm2 = -(ymm0 * ymm1) + ymm2
+; X64-NEXT: # ymm2 {%k1} = -(ymm0 * ymm1) + ymm2
; X64-NEXT: vmovaps %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xfc,0x28,0xc2]
; X64-NEXT: retq # encoding: [0xc3]
entry:
@@ -686,14 +686,14 @@ define <8 x float> @test_mm256_maskz_fmadd_ps(i8 zeroext %__U, <8 x float> %__A,
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmadd213ps %ymm2, %ymm1, %ymm0 {%k1} {z} # encoding: [0x62,0xf2,0x75,0xa9,0xa8,0xc2]
-; X86-NEXT: # ymm0 = (ymm1 * ymm0) + ymm2
+; X86-NEXT: # ymm0 {%k1} {z} = (ymm1 * ymm0) + ymm2
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_mm256_maskz_fmadd_ps:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmadd213ps %ymm2, %ymm1, %ymm0 {%k1} {z} # encoding: [0x62,0xf2,0x75,0xa9,0xa8,0xc2]
-; X64-NEXT: # ymm0 = (ymm1 * ymm0) + ymm2
+; X64-NEXT: # ymm0 {%k1} {z} = (ymm1 * ymm0) + ymm2
; X64-NEXT: retq # encoding: [0xc3]
entry:
%0 = tail call <8 x float> @llvm.fma.v8f32(<8 x float> %__A, <8 x float> %__B, <8 x float> %__C) #9
@@ -708,14 +708,14 @@ define <8 x float> @test_mm256_maskz_fmsub_ps(i8 zeroext %__U, <8 x float> %__A,
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmsub213ps %ymm2, %ymm1, %ymm0 {%k1} {z} # encoding: [0x62,0xf2,0x75,0xa9,0xaa,0xc2]
-; X86-NEXT: # ymm0 = (ymm1 * ymm0) - ymm2
+; X86-NEXT: # ymm0 {%k1} {z} = (ymm1 * ymm0) - ymm2
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_mm256_maskz_fmsub_ps:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmsub213ps %ymm2, %ymm1, %ymm0 {%k1} {z} # encoding: [0x62,0xf2,0x75,0xa9,0xaa,0xc2]
-; X64-NEXT: # ymm0 = (ymm1 * ymm0) - ymm2
+; X64-NEXT: # ymm0 {%k1} {z} = (ymm1 * ymm0) - ymm2
; X64-NEXT: retq # encoding: [0xc3]
entry:
%sub.i = fsub <8 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %__C
@@ -731,14 +731,14 @@ define <8 x float> @test_mm256_maskz_fnmadd_ps(i8 zeroext %__U, <8 x float> %__A
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfnmadd213ps %ymm2, %ymm1, %ymm0 {%k1} {z} # encoding: [0x62,0xf2,0x75,0xa9,0xac,0xc2]
-; X86-NEXT: # ymm0 = -(ymm1 * ymm0) + ymm2
+; X86-NEXT: # ymm0 {%k1} {z} = -(ymm1 * ymm0) + ymm2
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_mm256_maskz_fnmadd_ps:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfnmadd213ps %ymm2, %ymm1, %ymm0 {%k1} {z} # encoding: [0x62,0xf2,0x75,0xa9,0xac,0xc2]
-; X64-NEXT: # ymm0 = -(ymm1 * ymm0) + ymm2
+; X64-NEXT: # ymm0 {%k1} {z} = -(ymm1 * ymm0) + ymm2
; X64-NEXT: retq # encoding: [0xc3]
entry:
%sub.i = fsub <8 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %__A
@@ -754,14 +754,14 @@ define <8 x float> @test_mm256_maskz_fnmsub_ps(i8 zeroext %__U, <8 x float> %__A
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfnmsub213ps %ymm2, %ymm1, %ymm0 {%k1} {z} # encoding: [0x62,0xf2,0x75,0xa9,0xae,0xc2]
-; X86-NEXT: # ymm0 = -(ymm1 * ymm0) - ymm2
+; X86-NEXT: # ymm0 {%k1} {z} = -(ymm1 * ymm0) - ymm2
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_mm256_maskz_fnmsub_ps:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfnmsub213ps %ymm2, %ymm1, %ymm0 {%k1} {z} # encoding: [0x62,0xf2,0x75,0xa9,0xae,0xc2]
-; X64-NEXT: # ymm0 = -(ymm1 * ymm0) - ymm2
+; X64-NEXT: # ymm0 {%k1} {z} = -(ymm1 * ymm0) - ymm2
; X64-NEXT: retq # encoding: [0xc3]
entry:
%sub.i = fsub <8 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %__A
@@ -778,14 +778,14 @@ define <2 x double> @test_mm_mask_fmaddsub_pd(<2 x double> %__A, i8 zeroext %__U
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmaddsub132pd %xmm1, %xmm2, %xmm0 {%k1} # encoding: [0x62,0xf2,0xed,0x09,0x96,0xc1]
-; X86-NEXT: # xmm0 = (xmm0 * xmm1) +/- xmm2
+; X86-NEXT: # xmm0 {%k1} = (xmm0 * xmm1) +/- xmm2
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_mm_mask_fmaddsub_pd:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmaddsub132pd %xmm1, %xmm2, %xmm0 {%k1} # encoding: [0x62,0xf2,0xed,0x09,0x96,0xc1]
-; X64-NEXT: # xmm0 = (xmm0 * xmm1) +/- xmm2
+; X64-NEXT: # xmm0 {%k1} = (xmm0 * xmm1) +/- xmm2
; X64-NEXT: retq # encoding: [0xc3]
entry:
%0 = tail call <2 x double> @llvm.fma.v2f64(<2 x double> %__A, <2 x double> %__B, <2 x double> %__C) #9
@@ -804,14 +804,14 @@ define <2 x double> @test_mm_mask_fmsubadd_pd(<2 x double> %__A, i8 zeroext %__U
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmsubadd132pd %xmm1, %xmm2, %xmm0 {%k1} # encoding: [0x62,0xf2,0xed,0x09,0x97,0xc1]
-; X86-NEXT: # xmm0 = (xmm0 * xmm1) -/+ xmm2
+; X86-NEXT: # xmm0 {%k1} = (xmm0 * xmm1) -/+ xmm2
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_mm_mask_fmsubadd_pd:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmsubadd132pd %xmm1, %xmm2, %xmm0 {%k1} # encoding: [0x62,0xf2,0xed,0x09,0x97,0xc1]
-; X64-NEXT: # xmm0 = (xmm0 * xmm1) -/+ xmm2
+; X64-NEXT: # xmm0 {%k1} = (xmm0 * xmm1) -/+ xmm2
; X64-NEXT: retq # encoding: [0xc3]
entry:
%sub.i = fsub <2 x double> <double -0.000000e+00, double -0.000000e+00>, %__C
@@ -830,7 +830,7 @@ define <2 x double> @test_mm_mask3_fmaddsub_pd(<2 x double> %__A, <2 x double> %
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmaddsub231pd %xmm1, %xmm0, %xmm2 {%k1} # encoding: [0x62,0xf2,0xfd,0x09,0xb6,0xd1]
-; X86-NEXT: # xmm2 = (xmm0 * xmm1) +/- xmm2
+; X86-NEXT: # xmm2 {%k1} = (xmm0 * xmm1) +/- xmm2
; X86-NEXT: vmovapd %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x28,0xc2]
; X86-NEXT: retl # encoding: [0xc3]
;
@@ -838,7 +838,7 @@ define <2 x double> @test_mm_mask3_fmaddsub_pd(<2 x double> %__A, <2 x double> %
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmaddsub231pd %xmm1, %xmm0, %xmm2 {%k1} # encoding: [0x62,0xf2,0xfd,0x09,0xb6,0xd1]
-; X64-NEXT: # xmm2 = (xmm0 * xmm1) +/- xmm2
+; X64-NEXT: # xmm2 {%k1} = (xmm0 * xmm1) +/- xmm2
; X64-NEXT: vmovapd %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x28,0xc2]
; X64-NEXT: retq # encoding: [0xc3]
entry:
@@ -858,14 +858,14 @@ define <2 x double> @test_mm_maskz_fmaddsub_pd(i8 zeroext %__U, <2 x double> %__
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmaddsub213pd %xmm2, %xmm1, %xmm0 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0x89,0xa6,0xc2]
-; X86-NEXT: # xmm0 = (xmm1 * xmm0) +/- xmm2
+; X86-NEXT: # xmm0 {%k1} {z} = (xmm1 * xmm0) +/- xmm2
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_mm_maskz_fmaddsub_pd:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmaddsub213pd %xmm2, %xmm1, %xmm0 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0x89,0xa6,0xc2]
-; X64-NEXT: # xmm0 = (xmm1 * xmm0) +/- xmm2
+; X64-NEXT: # xmm0 {%k1} {z} = (xmm1 * xmm0) +/- xmm2
; X64-NEXT: retq # encoding: [0xc3]
entry:
%0 = tail call <2 x double> @llvm.fma.v2f64(<2 x double> %__A, <2 x double> %__B, <2 x double> %__C) #9
@@ -884,14 +884,14 @@ define <2 x double> @test_mm_maskz_fmsubadd_pd(i8 zeroext %__U, <2 x double> %__
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmsubadd213pd %xmm2, %xmm1, %xmm0 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0x89,0xa7,0xc2]
-; X86-NEXT: # xmm0 = (xmm1 * xmm0) -/+ xmm2
+; X86-NEXT: # xmm0 {%k1} {z} = (xmm1 * xmm0) -/+ xmm2
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_mm_maskz_fmsubadd_pd:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmsubadd213pd %xmm2, %xmm1, %xmm0 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0x89,0xa7,0xc2]
-; X64-NEXT: # xmm0 = (xmm1 * xmm0) -/+ xmm2
+; X64-NEXT: # xmm0 {%k1} {z} = (xmm1 * xmm0) -/+ xmm2
; X64-NEXT: retq # encoding: [0xc3]
entry:
%sub.i = fsub <2 x double> <double -0.000000e+00, double -0.000000e+00>, %__C
@@ -910,14 +910,14 @@ define <4 x double> @test_mm256_mask_fmaddsub_pd(<4 x double> %__A, i8 zeroext %
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmaddsub132pd %ymm1, %ymm2, %ymm0 {%k1} # encoding: [0x62,0xf2,0xed,0x29,0x96,0xc1]
-; X86-NEXT: # ymm0 = (ymm0 * ymm1) +/- ymm2
+; X86-NEXT: # ymm0 {%k1} = (ymm0 * ymm1) +/- ymm2
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_mm256_mask_fmaddsub_pd:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmaddsub132pd %ymm1, %ymm2, %ymm0 {%k1} # encoding: [0x62,0xf2,0xed,0x29,0x96,0xc1]
-; X64-NEXT: # ymm0 = (ymm0 * ymm1) +/- ymm2
+; X64-NEXT: # ymm0 {%k1} = (ymm0 * ymm1) +/- ymm2
; X64-NEXT: retq # encoding: [0xc3]
entry:
%0 = tail call <4 x double> @llvm.fma.v4f64(<4 x double> %__A, <4 x double> %__B, <4 x double> %__C) #9
@@ -936,14 +936,14 @@ define <4 x double> @test_mm256_mask_fmsubadd_pd(<4 x double> %__A, i8 zeroext %
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmsubadd132pd %ymm1, %ymm2, %ymm0 {%k1} # encoding: [0x62,0xf2,0xed,0x29,0x97,0xc1]
-; X86-NEXT: # ymm0 = (ymm0 * ymm1) -/+ ymm2
+; X86-NEXT: # ymm0 {%k1} = (ymm0 * ymm1) -/+ ymm2
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_mm256_mask_fmsubadd_pd:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmsubadd132pd %ymm1, %ymm2, %ymm0 {%k1} # encoding: [0x62,0xf2,0xed,0x29,0x97,0xc1]
-; X64-NEXT: # ymm0 = (ymm0 * ymm1) -/+ ymm2
+; X64-NEXT: # ymm0 {%k1} = (ymm0 * ymm1) -/+ ymm2
; X64-NEXT: retq # encoding: [0xc3]
entry:
%sub.i = fsub <4 x double> <double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00>, %__C
@@ -962,7 +962,7 @@ define <4 x double> @test_mm256_mask3_fmaddsub_pd(<4 x double> %__A, <4 x double
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmaddsub231pd %ymm1, %ymm0, %ymm2 {%k1} # encoding: [0x62,0xf2,0xfd,0x29,0xb6,0xd1]
-; X86-NEXT: # ymm2 = (ymm0 * ymm1) +/- ymm2
+; X86-NEXT: # ymm2 {%k1} = (ymm0 * ymm1) +/- ymm2
; X86-NEXT: vmovapd %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x28,0xc2]
; X86-NEXT: retl # encoding: [0xc3]
;
@@ -970,7 +970,7 @@ define <4 x double> @test_mm256_mask3_fmaddsub_pd(<4 x double> %__A, <4 x double
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmaddsub231pd %ymm1, %ymm0, %ymm2 {%k1} # encoding: [0x62,0xf2,0xfd,0x29,0xb6,0xd1]
-; X64-NEXT: # ymm2 = (ymm0 * ymm1) +/- ymm2
+; X64-NEXT: # ymm2 {%k1} = (ymm0 * ymm1) +/- ymm2
; X64-NEXT: vmovapd %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x28,0xc2]
; X64-NEXT: retq # encoding: [0xc3]
entry:
@@ -990,14 +990,14 @@ define <4 x double> @test_mm256_maskz_fmaddsub_pd(i8 zeroext %__U, <4 x double>
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmaddsub213pd %ymm2, %ymm1, %ymm0 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0xa9,0xa6,0xc2]
-; X86-NEXT: # ymm0 = (ymm1 * ymm0) +/- ymm2
+; X86-NEXT: # ymm0 {%k1} {z} = (ymm1 * ymm0) +/- ymm2
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_mm256_maskz_fmaddsub_pd:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmaddsub213pd %ymm2, %ymm1, %ymm0 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0xa9,0xa6,0xc2]
-; X64-NEXT: # ymm0 = (ymm1 * ymm0) +/- ymm2
+; X64-NEXT: # ymm0 {%k1} {z} = (ymm1 * ymm0) +/- ymm2
; X64-NEXT: retq # encoding: [0xc3]
entry:
%0 = tail call <4 x double> @llvm.fma.v4f64(<4 x double> %__A, <4 x double> %__B, <4 x double> %__C) #9
@@ -1016,14 +1016,14 @@ define <4 x double> @test_mm256_maskz_fmsubadd_pd(i8 zeroext %__U, <4 x double>
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmsubadd213pd %ymm2, %ymm1, %ymm0 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0xa9,0xa7,0xc2]
-; X86-NEXT: # ymm0 = (ymm1 * ymm0) -/+ ymm2
+; X86-NEXT: # ymm0 {%k1} {z} = (ymm1 * ymm0) -/+ ymm2
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_mm256_maskz_fmsubadd_pd:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmsubadd213pd %ymm2, %ymm1, %ymm0 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0xa9,0xa7,0xc2]
-; X64-NEXT: # ymm0 = (ymm1 * ymm0) -/+ ymm2
+; X64-NEXT: # ymm0 {%k1} {z} = (ymm1 * ymm0) -/+ ymm2
; X64-NEXT: retq # encoding: [0xc3]
entry:
%sub.i = fsub <4 x double> <double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00>, %__C
@@ -1042,14 +1042,14 @@ define <4 x float> @test_mm_mask_fmaddsub_ps(<4 x float> %__A, i8 zeroext %__U,
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmaddsub132ps %xmm1, %xmm2, %xmm0 {%k1} # encoding: [0x62,0xf2,0x6d,0x09,0x96,0xc1]
-; X86-NEXT: # xmm0 = (xmm0 * xmm1) +/- xmm2
+; X86-NEXT: # xmm0 {%k1} = (xmm0 * xmm1) +/- xmm2
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_mm_mask_fmaddsub_ps:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmaddsub132ps %xmm1, %xmm2, %xmm0 {%k1} # encoding: [0x62,0xf2,0x6d,0x09,0x96,0xc1]
-; X64-NEXT: # xmm0 = (xmm0 * xmm1) +/- xmm2
+; X64-NEXT: # xmm0 {%k1} = (xmm0 * xmm1) +/- xmm2
; X64-NEXT: retq # encoding: [0xc3]
entry:
%0 = tail call <4 x float> @llvm.fma.v4f32(<4 x float> %__A, <4 x float> %__B, <4 x float> %__C) #9
@@ -1068,14 +1068,14 @@ define <4 x float> @test_mm_mask_fmsubadd_ps(<4 x float> %__A, i8 zeroext %__U,
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmsubadd132ps %xmm1, %xmm2, %xmm0 {%k1} # encoding: [0x62,0xf2,0x6d,0x09,0x97,0xc1]
-; X86-NEXT: # xmm0 = (xmm0 * xmm1) -/+ xmm2
+; X86-NEXT: # xmm0 {%k1} = (xmm0 * xmm1) -/+ xmm2
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_mm_mask_fmsubadd_ps:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmsubadd132ps %xmm1, %xmm2, %xmm0 {%k1} # encoding: [0x62,0xf2,0x6d,0x09,0x97,0xc1]
-; X64-NEXT: # xmm0 = (xmm0 * xmm1) -/+ xmm2
+; X64-NEXT: # xmm0 {%k1} = (xmm0 * xmm1) -/+ xmm2
; X64-NEXT: retq # encoding: [0xc3]
entry:
%sub.i = fsub <4 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %__C
@@ -1094,7 +1094,7 @@ define <4 x float> @test_mm_mask3_fmaddsub_ps(<4 x float> %__A, <4 x float> %__B
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmaddsub231ps %xmm1, %xmm0, %xmm2 {%k1} # encoding: [0x62,0xf2,0x7d,0x09,0xb6,0xd1]
-; X86-NEXT: # xmm2 = (xmm0 * xmm1) +/- xmm2
+; X86-NEXT: # xmm2 {%k1} = (xmm0 * xmm1) +/- xmm2
; X86-NEXT: vmovaps %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf8,0x28,0xc2]
; X86-NEXT: retl # encoding: [0xc3]
;
@@ -1102,7 +1102,7 @@ define <4 x float> @test_mm_mask3_fmaddsub_ps(<4 x float> %__A, <4 x float> %__B
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmaddsub231ps %xmm1, %xmm0, %xmm2 {%k1} # encoding: [0x62,0xf2,0x7d,0x09,0xb6,0xd1]
-; X64-NEXT: # xmm2 = (xmm0 * xmm1) +/- xmm2
+; X64-NEXT: # xmm2 {%k1} = (xmm0 * xmm1) +/- xmm2
; X64-NEXT: vmovaps %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf8,0x28,0xc2]
; X64-NEXT: retq # encoding: [0xc3]
entry:
@@ -1122,14 +1122,14 @@ define <4 x float> @test_mm_maskz_fmaddsub_ps(i8 zeroext %__U, <4 x float> %__A,
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmaddsub213ps %xmm2, %xmm1, %xmm0 {%k1} {z} # encoding: [0x62,0xf2,0x75,0x89,0xa6,0xc2]
-; X86-NEXT: # xmm0 = (xmm1 * xmm0) +/- xmm2
+; X86-NEXT: # xmm0 {%k1} {z} = (xmm1 * xmm0) +/- xmm2
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_mm_maskz_fmaddsub_ps:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmaddsub213ps %xmm2, %xmm1, %xmm0 {%k1} {z} # encoding: [0x62,0xf2,0x75,0x89,0xa6,0xc2]
-; X64-NEXT: # xmm0 = (xmm1 * xmm0) +/- xmm2
+; X64-NEXT: # xmm0 {%k1} {z} = (xmm1 * xmm0) +/- xmm2
; X64-NEXT: retq # encoding: [0xc3]
entry:
%0 = tail call <4 x float> @llvm.fma.v4f32(<4 x float> %__A, <4 x float> %__B, <4 x float> %__C) #9
@@ -1148,14 +1148,14 @@ define <4 x float> @test_mm_maskz_fmsubadd_ps(i8 zeroext %__U, <4 x float> %__A,
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmsubadd213ps %xmm2, %xmm1, %xmm0 {%k1} {z} # encoding: [0x62,0xf2,0x75,0x89,0xa7,0xc2]
-; X86-NEXT: # xmm0 = (xmm1 * xmm0) -/+ xmm2
+; X86-NEXT: # xmm0 {%k1} {z} = (xmm1 * xmm0) -/+ xmm2
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_mm_maskz_fmsubadd_ps:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmsubadd213ps %xmm2, %xmm1, %xmm0 {%k1} {z} # encoding: [0x62,0xf2,0x75,0x89,0xa7,0xc2]
-; X64-NEXT: # xmm0 = (xmm1 * xmm0) -/+ xmm2
+; X64-NEXT: # xmm0 {%k1} {z} = (xmm1 * xmm0) -/+ xmm2
; X64-NEXT: retq # encoding: [0xc3]
entry:
%sub.i = fsub <4 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %__C
@@ -1174,14 +1174,14 @@ define <8 x float> @test_mm256_mask_fmaddsub_ps(<8 x float> %__A, i8 zeroext %__
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmaddsub132ps %ymm1, %ymm2, %ymm0 {%k1} # encoding: [0x62,0xf2,0x6d,0x29,0x96,0xc1]
-; X86-NEXT: # ymm0 = (ymm0 * ymm1) +/- ymm2
+; X86-NEXT: # ymm0 {%k1} = (ymm0 * ymm1) +/- ymm2
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_mm256_mask_fmaddsub_ps:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmaddsub132ps %ymm1, %ymm2, %ymm0 {%k1} # encoding: [0x62,0xf2,0x6d,0x29,0x96,0xc1]
-; X64-NEXT: # ymm0 = (ymm0 * ymm1) +/- ymm2
+; X64-NEXT: # ymm0 {%k1} = (ymm0 * ymm1) +/- ymm2
; X64-NEXT: retq # encoding: [0xc3]
entry:
%0 = tail call <8 x float> @llvm.fma.v8f32(<8 x float> %__A, <8 x float> %__B, <8 x float> %__C) #9
@@ -1199,14 +1199,14 @@ define <8 x float> @test_mm256_mask_fmsubadd_ps(<8 x float> %__A, i8 zeroext %__
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmsubadd132ps %ymm1, %ymm2, %ymm0 {%k1} # encoding: [0x62,0xf2,0x6d,0x29,0x97,0xc1]
-; X86-NEXT: # ymm0 = (ymm0 * ymm1) -/+ ymm2
+; X86-NEXT: # ymm0 {%k1} = (ymm0 * ymm1) -/+ ymm2
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_mm256_mask_fmsubadd_ps:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmsubadd132ps %ymm1, %ymm2, %ymm0 {%k1} # encoding: [0x62,0xf2,0x6d,0x29,0x97,0xc1]
-; X64-NEXT: # ymm0 = (ymm0 * ymm1) -/+ ymm2
+; X64-NEXT: # ymm0 {%k1} = (ymm0 * ymm1) -/+ ymm2
; X64-NEXT: retq # encoding: [0xc3]
entry:
%sub.i = fsub <8 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %__C
@@ -1224,7 +1224,7 @@ define <8 x float> @test_mm256_mask3_fmaddsub_ps(<8 x float> %__A, <8 x float> %
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmaddsub231ps %ymm1, %ymm0, %ymm2 {%k1} # encoding: [0x62,0xf2,0x7d,0x29,0xb6,0xd1]
-; X86-NEXT: # ymm2 = (ymm0 * ymm1) +/- ymm2
+; X86-NEXT: # ymm2 {%k1} = (ymm0 * ymm1) +/- ymm2
; X86-NEXT: vmovaps %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xfc,0x28,0xc2]
; X86-NEXT: retl # encoding: [0xc3]
;
@@ -1232,7 +1232,7 @@ define <8 x float> @test_mm256_mask3_fmaddsub_ps(<8 x float> %__A, <8 x float> %
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmaddsub231ps %ymm1, %ymm0, %ymm2 {%k1} # encoding: [0x62,0xf2,0x7d,0x29,0xb6,0xd1]
-; X64-NEXT: # ymm2 = (ymm0 * ymm1) +/- ymm2
+; X64-NEXT: # ymm2 {%k1} = (ymm0 * ymm1) +/- ymm2
; X64-NEXT: vmovaps %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xfc,0x28,0xc2]
; X64-NEXT: retq # encoding: [0xc3]
entry:
@@ -1251,14 +1251,14 @@ define <8 x float> @test_mm256_maskz_fmaddsub_ps(i8 zeroext %__U, <8 x float> %_
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmaddsub213ps %ymm2, %ymm1, %ymm0 {%k1} {z} # encoding: [0x62,0xf2,0x75,0xa9,0xa6,0xc2]
-; X86-NEXT: # ymm0 = (ymm1 * ymm0) +/- ymm2
+; X86-NEXT: # ymm0 {%k1} {z} = (ymm1 * ymm0) +/- ymm2
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_mm256_maskz_fmaddsub_ps:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmaddsub213ps %ymm2, %ymm1, %ymm0 {%k1} {z} # encoding: [0x62,0xf2,0x75,0xa9,0xa6,0xc2]
-; X64-NEXT: # ymm0 = (ymm1 * ymm0) +/- ymm2
+; X64-NEXT: # ymm0 {%k1} {z} = (ymm1 * ymm0) +/- ymm2
; X64-NEXT: retq # encoding: [0xc3]
entry:
%0 = tail call <8 x float> @llvm.fma.v8f32(<8 x float> %__A, <8 x float> %__B, <8 x float> %__C) #9
@@ -1276,14 +1276,14 @@ define <8 x float> @test_mm256_maskz_fmsubadd_ps(i8 zeroext %__U, <8 x float> %_
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmsubadd213ps %ymm2, %ymm1, %ymm0 {%k1} {z} # encoding: [0x62,0xf2,0x75,0xa9,0xa7,0xc2]
-; X86-NEXT: # ymm0 = (ymm1 * ymm0) -/+ ymm2
+; X86-NEXT: # ymm0 {%k1} {z} = (ymm1 * ymm0) -/+ ymm2
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_mm256_maskz_fmsubadd_ps:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmsubadd213ps %ymm2, %ymm1, %ymm0 {%k1} {z} # encoding: [0x62,0xf2,0x75,0xa9,0xa7,0xc2]
-; X64-NEXT: # ymm0 = (ymm1 * ymm0) -/+ ymm2
+; X64-NEXT: # ymm0 {%k1} {z} = (ymm1 * ymm0) -/+ ymm2
; X64-NEXT: retq # encoding: [0xc3]
entry:
%sub.i = fsub <8 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %__C
@@ -1301,7 +1301,7 @@ define <2 x double> @test_mm_mask3_fmsub_pd(<2 x double> %__A, <2 x double> %__B
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmsub231pd %xmm1, %xmm0, %xmm2 {%k1} # encoding: [0x62,0xf2,0xfd,0x09,0xba,0xd1]
-; X86-NEXT: # xmm2 = (xmm0 * xmm1) - xmm2
+; X86-NEXT: # xmm2 {%k1} = (xmm0 * xmm1) - xmm2
; X86-NEXT: vmovapd %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x28,0xc2]
; X86-NEXT: retl # encoding: [0xc3]
;
@@ -1309,7 +1309,7 @@ define <2 x double> @test_mm_mask3_fmsub_pd(<2 x double> %__A, <2 x double> %__B
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmsub231pd %xmm1, %xmm0, %xmm2 {%k1} # encoding: [0x62,0xf2,0xfd,0x09,0xba,0xd1]
-; X64-NEXT: # xmm2 = (xmm0 * xmm1) - xmm2
+; X64-NEXT: # xmm2 {%k1} = (xmm0 * xmm1) - xmm2
; X64-NEXT: vmovapd %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x28,0xc2]
; X64-NEXT: retq # encoding: [0xc3]
entry:
@@ -1327,7 +1327,7 @@ define <4 x double> @test_mm256_mask3_fmsub_pd(<4 x double> %__A, <4 x double> %
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmsub231pd %ymm1, %ymm0, %ymm2 {%k1} # encoding: [0x62,0xf2,0xfd,0x29,0xba,0xd1]
-; X86-NEXT: # ymm2 = (ymm0 * ymm1) - ymm2
+; X86-NEXT: # ymm2 {%k1} = (ymm0 * ymm1) - ymm2
; X86-NEXT: vmovapd %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x28,0xc2]
; X86-NEXT: retl # encoding: [0xc3]
;
@@ -1335,7 +1335,7 @@ define <4 x double> @test_mm256_mask3_fmsub_pd(<4 x double> %__A, <4 x double> %
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmsub231pd %ymm1, %ymm0, %ymm2 {%k1} # encoding: [0x62,0xf2,0xfd,0x29,0xba,0xd1]
-; X64-NEXT: # ymm2 = (ymm0 * ymm1) - ymm2
+; X64-NEXT: # ymm2 {%k1} = (ymm0 * ymm1) - ymm2
; X64-NEXT: vmovapd %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x28,0xc2]
; X64-NEXT: retq # encoding: [0xc3]
entry:
@@ -1353,7 +1353,7 @@ define <4 x float> @test_mm_mask3_fmsub_ps(<4 x float> %__A, <4 x float> %__B, <
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmsub231ps %xmm1, %xmm0, %xmm2 {%k1} # encoding: [0x62,0xf2,0x7d,0x09,0xba,0xd1]
-; X86-NEXT: # xmm2 = (xmm0 * xmm1) - xmm2
+; X86-NEXT: # xmm2 {%k1} = (xmm0 * xmm1) - xmm2
; X86-NEXT: vmovaps %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf8,0x28,0xc2]
; X86-NEXT: retl # encoding: [0xc3]
;
@@ -1361,7 +1361,7 @@ define <4 x float> @test_mm_mask3_fmsub_ps(<4 x float> %__A, <4 x float> %__B, <
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmsub231ps %xmm1, %xmm0, %xmm2 {%k1} # encoding: [0x62,0xf2,0x7d,0x09,0xba,0xd1]
-; X64-NEXT: # xmm2 = (xmm0 * xmm1) - xmm2
+; X64-NEXT: # xmm2 {%k1} = (xmm0 * xmm1) - xmm2
; X64-NEXT: vmovaps %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf8,0x28,0xc2]
; X64-NEXT: retq # encoding: [0xc3]
entry:
@@ -1379,7 +1379,7 @@ define <8 x float> @test_mm256_mask3_fmsub_ps(<8 x float> %__A, <8 x float> %__B
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmsub231ps %ymm1, %ymm0, %ymm2 {%k1} # encoding: [0x62,0xf2,0x7d,0x29,0xba,0xd1]
-; X86-NEXT: # ymm2 = (ymm0 * ymm1) - ymm2
+; X86-NEXT: # ymm2 {%k1} = (ymm0 * ymm1) - ymm2
; X86-NEXT: vmovaps %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xfc,0x28,0xc2]
; X86-NEXT: retl # encoding: [0xc3]
;
@@ -1387,7 +1387,7 @@ define <8 x float> @test_mm256_mask3_fmsub_ps(<8 x float> %__A, <8 x float> %__B
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmsub231ps %ymm1, %ymm0, %ymm2 {%k1} # encoding: [0x62,0xf2,0x7d,0x29,0xba,0xd1]
-; X64-NEXT: # ymm2 = (ymm0 * ymm1) - ymm2
+; X64-NEXT: # ymm2 {%k1} = (ymm0 * ymm1) - ymm2
; X64-NEXT: vmovaps %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xfc,0x28,0xc2]
; X64-NEXT: retq # encoding: [0xc3]
entry:
@@ -1404,7 +1404,7 @@ define <2 x double> @test_mm_mask3_fmsubadd_pd(<2 x double> %__A, <2 x double> %
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmsubadd231pd %xmm1, %xmm0, %xmm2 {%k1} # encoding: [0x62,0xf2,0xfd,0x09,0xb7,0xd1]
-; X86-NEXT: # xmm2 = (xmm0 * xmm1) -/+ xmm2
+; X86-NEXT: # xmm2 {%k1} = (xmm0 * xmm1) -/+ xmm2
; X86-NEXT: vmovapd %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x28,0xc2]
; X86-NEXT: retl # encoding: [0xc3]
;
@@ -1412,7 +1412,7 @@ define <2 x double> @test_mm_mask3_fmsubadd_pd(<2 x double> %__A, <2 x double> %
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmsubadd231pd %xmm1, %xmm0, %xmm2 {%k1} # encoding: [0x62,0xf2,0xfd,0x09,0xb7,0xd1]
-; X64-NEXT: # xmm2 = (xmm0 * xmm1) -/+ xmm2
+; X64-NEXT: # xmm2 {%k1} = (xmm0 * xmm1) -/+ xmm2
; X64-NEXT: vmovapd %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x28,0xc2]
; X64-NEXT: retq # encoding: [0xc3]
entry:
@@ -1432,7 +1432,7 @@ define <4 x double> @test_mm256_mask3_fmsubadd_pd(<4 x double> %__A, <4 x double
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmsubadd231pd %ymm1, %ymm0, %ymm2 {%k1} # encoding: [0x62,0xf2,0xfd,0x29,0xb7,0xd1]
-; X86-NEXT: # ymm2 = (ymm0 * ymm1) -/+ ymm2
+; X86-NEXT: # ymm2 {%k1} = (ymm0 * ymm1) -/+ ymm2
; X86-NEXT: vmovapd %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x28,0xc2]
; X86-NEXT: retl # encoding: [0xc3]
;
@@ -1440,7 +1440,7 @@ define <4 x double> @test_mm256_mask3_fmsubadd_pd(<4 x double> %__A, <4 x double
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmsubadd231pd %ymm1, %ymm0, %ymm2 {%k1} # encoding: [0x62,0xf2,0xfd,0x29,0xb7,0xd1]
-; X64-NEXT: # ymm2 = (ymm0 * ymm1) -/+ ymm2
+; X64-NEXT: # ymm2 {%k1} = (ymm0 * ymm1) -/+ ymm2
; X64-NEXT: vmovapd %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x28,0xc2]
; X64-NEXT: retq # encoding: [0xc3]
entry:
@@ -1460,7 +1460,7 @@ define <4 x float> @test_mm_mask3_fmsubadd_ps(<4 x float> %__A, <4 x float> %__B
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmsubadd231ps %xmm1, %xmm0, %xmm2 {%k1} # encoding: [0x62,0xf2,0x7d,0x09,0xb7,0xd1]
-; X86-NEXT: # xmm2 = (xmm0 * xmm1) -/+ xmm2
+; X86-NEXT: # xmm2 {%k1} = (xmm0 * xmm1) -/+ xmm2
; X86-NEXT: vmovaps %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf8,0x28,0xc2]
; X86-NEXT: retl # encoding: [0xc3]
;
@@ -1468,7 +1468,7 @@ define <4 x float> @test_mm_mask3_fmsubadd_ps(<4 x float> %__A, <4 x float> %__B
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmsubadd231ps %xmm1, %xmm0, %xmm2 {%k1} # encoding: [0x62,0xf2,0x7d,0x09,0xb7,0xd1]
-; X64-NEXT: # xmm2 = (xmm0 * xmm1) -/+ xmm2
+; X64-NEXT: # xmm2 {%k1} = (xmm0 * xmm1) -/+ xmm2
; X64-NEXT: vmovaps %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf8,0x28,0xc2]
; X64-NEXT: retq # encoding: [0xc3]
entry:
@@ -1488,7 +1488,7 @@ define <8 x float> @test_mm256_mask3_fmsubadd_ps(<8 x float> %__A, <8 x float> %
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmsubadd231ps %ymm1, %ymm0, %ymm2 {%k1} # encoding: [0x62,0xf2,0x7d,0x29,0xb7,0xd1]
-; X86-NEXT: # ymm2 = (ymm0 * ymm1) -/+ ymm2
+; X86-NEXT: # ymm2 {%k1} = (ymm0 * ymm1) -/+ ymm2
; X86-NEXT: vmovaps %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xfc,0x28,0xc2]
; X86-NEXT: retl # encoding: [0xc3]
;
@@ -1496,7 +1496,7 @@ define <8 x float> @test_mm256_mask3_fmsubadd_ps(<8 x float> %__A, <8 x float> %
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmsubadd231ps %ymm1, %ymm0, %ymm2 {%k1} # encoding: [0x62,0xf2,0x7d,0x29,0xb7,0xd1]
-; X64-NEXT: # ymm2 = (ymm0 * ymm1) -/+ ymm2
+; X64-NEXT: # ymm2 {%k1} = (ymm0 * ymm1) -/+ ymm2
; X64-NEXT: vmovaps %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xfc,0x28,0xc2]
; X64-NEXT: retq # encoding: [0xc3]
entry:
@@ -1515,14 +1515,14 @@ define <2 x double> @test_mm_mask_fnmadd_pd(<2 x double> %__A, i8 zeroext %__U,
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfnmadd132pd %xmm1, %xmm2, %xmm0 {%k1} # encoding: [0x62,0xf2,0xed,0x09,0x9c,0xc1]
-; X86-NEXT: # xmm0 = -(xmm0 * xmm1) + xmm2
+; X86-NEXT: # xmm0 {%k1} = -(xmm0 * xmm1) + xmm2
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_mm_mask_fnmadd_pd:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfnmadd132pd %xmm1, %xmm2, %xmm0 {%k1} # encoding: [0x62,0xf2,0xed,0x09,0x9c,0xc1]
-; X64-NEXT: # xmm0 = -(xmm0 * xmm1) + xmm2
+; X64-NEXT: # xmm0 {%k1} = -(xmm0 * xmm1) + xmm2
; X64-NEXT: retq # encoding: [0xc3]
entry:
%sub.i = fsub <2 x double> <double -0.000000e+00, double -0.000000e+00>, %__B
@@ -1539,14 +1539,14 @@ define <4 x double> @test_mm256_mask_fnmadd_pd(<4 x double> %__A, i8 zeroext %__
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfnmadd132pd %ymm1, %ymm2, %ymm0 {%k1} # encoding: [0x62,0xf2,0xed,0x29,0x9c,0xc1]
-; X86-NEXT: # ymm0 = -(ymm0 * ymm1) + ymm2
+; X86-NEXT: # ymm0 {%k1} = -(ymm0 * ymm1) + ymm2
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_mm256_mask_fnmadd_pd:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfnmadd132pd %ymm1, %ymm2, %ymm0 {%k1} # encoding: [0x62,0xf2,0xed,0x29,0x9c,0xc1]
-; X64-NEXT: # ymm0 = -(ymm0 * ymm1) + ymm2
+; X64-NEXT: # ymm0 {%k1} = -(ymm0 * ymm1) + ymm2
; X64-NEXT: retq # encoding: [0xc3]
entry:
%sub.i = fsub <4 x double> <double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00>, %__B
@@ -1563,14 +1563,14 @@ define <4 x float> @test_mm_mask_fnmadd_ps(<4 x float> %__A, i8 zeroext %__U, <4
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfnmadd132ps %xmm1, %xmm2, %xmm0 {%k1} # encoding: [0x62,0xf2,0x6d,0x09,0x9c,0xc1]
-; X86-NEXT: # xmm0 = -(xmm0 * xmm1) + xmm2
+; X86-NEXT: # xmm0 {%k1} = -(xmm0 * xmm1) + xmm2
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_mm_mask_fnmadd_ps:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfnmadd132ps %xmm1, %xmm2, %xmm0 {%k1} # encoding: [0x62,0xf2,0x6d,0x09,0x9c,0xc1]
-; X64-NEXT: # xmm0 = -(xmm0 * xmm1) + xmm2
+; X64-NEXT: # xmm0 {%k1} = -(xmm0 * xmm1) + xmm2
; X64-NEXT: retq # encoding: [0xc3]
entry:
%sub.i = fsub <4 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %__B
@@ -1587,14 +1587,14 @@ define <8 x float> @test_mm256_mask_fnmadd_ps(<8 x float> %__A, i8 zeroext %__U,
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfnmadd132ps %ymm1, %ymm2, %ymm0 {%k1} # encoding: [0x62,0xf2,0x6d,0x29,0x9c,0xc1]
-; X86-NEXT: # ymm0 = -(ymm0 * ymm1) + ymm2
+; X86-NEXT: # ymm0 {%k1} = -(ymm0 * ymm1) + ymm2
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_mm256_mask_fnmadd_ps:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfnmadd132ps %ymm1, %ymm2, %ymm0 {%k1} # encoding: [0x62,0xf2,0x6d,0x29,0x9c,0xc1]
-; X64-NEXT: # ymm0 = -(ymm0 * ymm1) + ymm2
+; X64-NEXT: # ymm0 {%k1} = -(ymm0 * ymm1) + ymm2
; X64-NEXT: retq # encoding: [0xc3]
entry:
%sub.i = fsub <8 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %__B
@@ -1610,14 +1610,14 @@ define <2 x double> @test_mm_mask_fnmsub_pd(<2 x double> %__A, i8 zeroext %__U,
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfnmsub132pd %xmm1, %xmm2, %xmm0 {%k1} # encoding: [0x62,0xf2,0xed,0x09,0x9e,0xc1]
-; X86-NEXT: # xmm0 = -(xmm0 * xmm1) - xmm2
+; X86-NEXT: # xmm0 {%k1} = -(xmm0 * xmm1) - xmm2
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_mm_mask_fnmsub_pd:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfnmsub132pd %xmm1, %xmm2, %xmm0 {%k1} # encoding: [0x62,0xf2,0xed,0x09,0x9e,0xc1]
-; X64-NEXT: # xmm0 = -(xmm0 * xmm1) - xmm2
+; X64-NEXT: # xmm0 {%k1} = -(xmm0 * xmm1) - xmm2
; X64-NEXT: retq # encoding: [0xc3]
entry:
%sub.i = fsub <2 x double> <double -0.000000e+00, double -0.000000e+00>, %__B
@@ -1635,7 +1635,7 @@ define <2 x double> @test_mm_mask3_fnmsub_pd(<2 x double> %__A, <2 x double> %__
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfnmsub231pd %xmm1, %xmm0, %xmm2 {%k1} # encoding: [0x62,0xf2,0xfd,0x09,0xbe,0xd1]
-; X86-NEXT: # xmm2 = -(xmm0 * xmm1) - xmm2
+; X86-NEXT: # xmm2 {%k1} = -(xmm0 * xmm1) - xmm2
; X86-NEXT: vmovapd %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x28,0xc2]
; X86-NEXT: retl # encoding: [0xc3]
;
@@ -1643,7 +1643,7 @@ define <2 x double> @test_mm_mask3_fnmsub_pd(<2 x double> %__A, <2 x double> %__
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfnmsub231pd %xmm1, %xmm0, %xmm2 {%k1} # encoding: [0x62,0xf2,0xfd,0x09,0xbe,0xd1]
-; X64-NEXT: # xmm2 = -(xmm0 * xmm1) - xmm2
+; X64-NEXT: # xmm2 {%k1} = -(xmm0 * xmm1) - xmm2
; X64-NEXT: vmovapd %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x28,0xc2]
; X64-NEXT: retq # encoding: [0xc3]
entry:
@@ -1662,14 +1662,14 @@ define <4 x double> @test_mm256_mask_fnmsub_pd(<4 x double> %__A, i8 zeroext %__
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfnmsub132pd %ymm1, %ymm2, %ymm0 {%k1} # encoding: [0x62,0xf2,0xed,0x29,0x9e,0xc1]
-; X86-NEXT: # ymm0 = -(ymm0 * ymm1) - ymm2
+; X86-NEXT: # ymm0 {%k1} = -(ymm0 * ymm1) - ymm2
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_mm256_mask_fnmsub_pd:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfnmsub132pd %ymm1, %ymm2, %ymm0 {%k1} # encoding: [0x62,0xf2,0xed,0x29,0x9e,0xc1]
-; X64-NEXT: # ymm0 = -(ymm0 * ymm1) - ymm2
+; X64-NEXT: # ymm0 {%k1} = -(ymm0 * ymm1) - ymm2
; X64-NEXT: retq # encoding: [0xc3]
entry:
%sub.i = fsub <4 x double> <double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00>, %__B
@@ -1687,7 +1687,7 @@ define <4 x double> @test_mm256_mask3_fnmsub_pd(<4 x double> %__A, <4 x double>
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfnmsub231pd %ymm1, %ymm0, %ymm2 {%k1} # encoding: [0x62,0xf2,0xfd,0x29,0xbe,0xd1]
-; X86-NEXT: # ymm2 = -(ymm0 * ymm1) - ymm2
+; X86-NEXT: # ymm2 {%k1} = -(ymm0 * ymm1) - ymm2
; X86-NEXT: vmovapd %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x28,0xc2]
; X86-NEXT: retl # encoding: [0xc3]
;
@@ -1695,7 +1695,7 @@ define <4 x double> @test_mm256_mask3_fnmsub_pd(<4 x double> %__A, <4 x double>
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfnmsub231pd %ymm1, %ymm0, %ymm2 {%k1} # encoding: [0x62,0xf2,0xfd,0x29,0xbe,0xd1]
-; X64-NEXT: # ymm2 = -(ymm0 * ymm1) - ymm2
+; X64-NEXT: # ymm2 {%k1} = -(ymm0 * ymm1) - ymm2
; X64-NEXT: vmovapd %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x28,0xc2]
; X64-NEXT: retq # encoding: [0xc3]
entry:
@@ -1714,14 +1714,14 @@ define <4 x float> @test_mm_mask_fnmsub_ps(<4 x float> %__A, i8 zeroext %__U, <4
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfnmsub132ps %xmm1, %xmm2, %xmm0 {%k1} # encoding: [0x62,0xf2,0x6d,0x09,0x9e,0xc1]
-; X86-NEXT: # xmm0 = -(xmm0 * xmm1) - xmm2
+; X86-NEXT: # xmm0 {%k1} = -(xmm0 * xmm1) - xmm2
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_mm_mask_fnmsub_ps:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfnmsub132ps %xmm1, %xmm2, %xmm0 {%k1} # encoding: [0x62,0xf2,0x6d,0x09,0x9e,0xc1]
-; X64-NEXT: # xmm0 = -(xmm0 * xmm1) - xmm2
+; X64-NEXT: # xmm0 {%k1} = -(xmm0 * xmm1) - xmm2
; X64-NEXT: retq # encoding: [0xc3]
entry:
%sub.i = fsub <4 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %__B
@@ -1739,7 +1739,7 @@ define <4 x float> @test_mm_mask3_fnmsub_ps(<4 x float> %__A, <4 x float> %__B,
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfnmsub231ps %xmm1, %xmm0, %xmm2 {%k1} # encoding: [0x62,0xf2,0x7d,0x09,0xbe,0xd1]
-; X86-NEXT: # xmm2 = -(xmm0 * xmm1) - xmm2
+; X86-NEXT: # xmm2 {%k1} = -(xmm0 * xmm1) - xmm2
; X86-NEXT: vmovaps %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf8,0x28,0xc2]
; X86-NEXT: retl # encoding: [0xc3]
;
@@ -1747,7 +1747,7 @@ define <4 x float> @test_mm_mask3_fnmsub_ps(<4 x float> %__A, <4 x float> %__B,
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfnmsub231ps %xmm1, %xmm0, %xmm2 {%k1} # encoding: [0x62,0xf2,0x7d,0x09,0xbe,0xd1]
-; X64-NEXT: # xmm2 = -(xmm0 * xmm1) - xmm2
+; X64-NEXT: # xmm2 {%k1} = -(xmm0 * xmm1) - xmm2
; X64-NEXT: vmovaps %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf8,0x28,0xc2]
; X64-NEXT: retq # encoding: [0xc3]
entry:
@@ -1766,14 +1766,14 @@ define <8 x float> @test_mm256_mask_fnmsub_ps(<8 x float> %__A, i8 zeroext %__U,
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfnmsub132ps %ymm1, %ymm2, %ymm0 {%k1} # encoding: [0x62,0xf2,0x6d,0x29,0x9e,0xc1]
-; X86-NEXT: # ymm0 = -(ymm0 * ymm1) - ymm2
+; X86-NEXT: # ymm0 {%k1} = -(ymm0 * ymm1) - ymm2
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_mm256_mask_fnmsub_ps:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfnmsub132ps %ymm1, %ymm2, %ymm0 {%k1} # encoding: [0x62,0xf2,0x6d,0x29,0x9e,0xc1]
-; X64-NEXT: # ymm0 = -(ymm0 * ymm1) - ymm2
+; X64-NEXT: # ymm0 {%k1} = -(ymm0 * ymm1) - ymm2
; X64-NEXT: retq # encoding: [0xc3]
entry:
%sub.i = fsub <8 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %__B
@@ -1790,7 +1790,7 @@ define <8 x float> @test_mm256_mask3_fnmsub_ps(<8 x float> %__A, <8 x float> %__
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfnmsub231ps %ymm1, %ymm0, %ymm2 {%k1} # encoding: [0x62,0xf2,0x7d,0x29,0xbe,0xd1]
-; X86-NEXT: # ymm2 = -(ymm0 * ymm1) - ymm2
+; X86-NEXT: # ymm2 {%k1} = -(ymm0 * ymm1) - ymm2
; X86-NEXT: vmovaps %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xfc,0x28,0xc2]
; X86-NEXT: retl # encoding: [0xc3]
;
@@ -1798,7 +1798,7 @@ define <8 x float> @test_mm256_mask3_fnmsub_ps(<8 x float> %__A, <8 x float> %__
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfnmsub231ps %ymm1, %ymm0, %ymm2 {%k1} # encoding: [0x62,0xf2,0x7d,0x29,0xbe,0xd1]
-; X64-NEXT: # ymm2 = -(ymm0 * ymm1) - ymm2
+; X64-NEXT: # ymm2 {%k1} = -(ymm0 * ymm1) - ymm2
; X64-NEXT: vmovaps %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xfc,0x28,0xc2]
; X64-NEXT: retq # encoding: [0xc3]
entry:
diff --git a/llvm/test/CodeGen/X86/avx512vl-intrinsics-fast-isel.ll b/llvm/test/CodeGen/X86/avx512vl-intrinsics-fast-isel.ll
index 0e48424a2d25..9cda6c97ae0a 100644
--- a/llvm/test/CodeGen/X86/avx512vl-intrinsics-fast-isel.ll
+++ b/llvm/test/CodeGen/X86/avx512vl-intrinsics-fast-isel.ll
@@ -4340,13 +4340,13 @@ define <2 x double> @test_mm_mask_fmadd_pd(<2 x double> %__A, i8 zeroext %__U, <
; X86: # %bb.0: # %entry
; X86-NEXT: movb {{[0-9]+}}(%esp), %al
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfmadd132pd {{.*#+}} xmm0 = (xmm0 * xmm1) + xmm2
+; X86-NEXT: vfmadd132pd {{.*#+}} xmm0 {%k1} = (xmm0 * xmm1) + xmm2
; X86-NEXT: retl
;
; X64-LABEL: test_mm_mask_fmadd_pd:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfmadd132pd {{.*#+}} xmm0 = (xmm0 * xmm1) + xmm2
+; X64-NEXT: vfmadd132pd {{.*#+}} xmm0 {%k1} = (xmm0 * xmm1) + xmm2
; X64-NEXT: retq
entry:
%0 = tail call <2 x double> @llvm.fma.v2f64(<2 x double> %__A, <2 x double> %__B, <2 x double> %__C) #9
@@ -4361,13 +4361,13 @@ define <2 x double> @test_mm_mask_fmsub_pd(<2 x double> %__A, i8 zeroext %__U, <
; X86: # %bb.0: # %entry
; X86-NEXT: movb {{[0-9]+}}(%esp), %al
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfmsub132pd {{.*#+}} xmm0 = (xmm0 * xmm1) - xmm2
+; X86-NEXT: vfmsub132pd {{.*#+}} xmm0 {%k1} = (xmm0 * xmm1) - xmm2
; X86-NEXT: retl
;
; X64-LABEL: test_mm_mask_fmsub_pd:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfmsub132pd {{.*#+}} xmm0 = (xmm0 * xmm1) - xmm2
+; X64-NEXT: vfmsub132pd {{.*#+}} xmm0 {%k1} = (xmm0 * xmm1) - xmm2
; X64-NEXT: retq
entry:
%sub.i = fsub <2 x double> <double -0.000000e+00, double -0.000000e+00>, %__C
@@ -4383,14 +4383,14 @@ define <2 x double> @test_mm_mask3_fmadd_pd(<2 x double> %__A, <2 x double> %__B
; X86: # %bb.0: # %entry
; X86-NEXT: movb {{[0-9]+}}(%esp), %al
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfmadd231pd {{.*#+}} xmm2 = (xmm0 * xmm1) + xmm2
+; X86-NEXT: vfmadd231pd {{.*#+}} xmm2 {%k1} = (xmm0 * xmm1) + xmm2
; X86-NEXT: vmovapd %xmm2, %xmm0
; X86-NEXT: retl
;
; X64-LABEL: test_mm_mask3_fmadd_pd:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfmadd231pd {{.*#+}} xmm2 = (xmm0 * xmm1) + xmm2
+; X64-NEXT: vfmadd231pd {{.*#+}} xmm2 {%k1} = (xmm0 * xmm1) + xmm2
; X64-NEXT: vmovapd %xmm2, %xmm0
; X64-NEXT: retq
entry:
@@ -4406,14 +4406,14 @@ define <2 x double> @test_mm_mask3_fnmadd_pd(<2 x double> %__A, <2 x double> %__
; X86: # %bb.0: # %entry
; X86-NEXT: movb {{[0-9]+}}(%esp), %al
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfnmadd231pd {{.*#+}} xmm2 = -(xmm0 * xmm1) + xmm2
+; X86-NEXT: vfnmadd231pd {{.*#+}} xmm2 {%k1} = -(xmm0 * xmm1) + xmm2
; X86-NEXT: vmovapd %xmm2, %xmm0
; X86-NEXT: retl
;
; X64-LABEL: test_mm_mask3_fnmadd_pd:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfnmadd231pd {{.*#+}} xmm2 = -(xmm0 * xmm1) + xmm2
+; X64-NEXT: vfnmadd231pd {{.*#+}} xmm2 {%k1} = -(xmm0 * xmm1) + xmm2
; X64-NEXT: vmovapd %xmm2, %xmm0
; X64-NEXT: retq
entry:
@@ -4430,13 +4430,13 @@ define <2 x double> @test_mm_maskz_fmadd_pd(i8 zeroext %__U, <2 x double> %__A,
; X86: # %bb.0: # %entry
; X86-NEXT: movb {{[0-9]+}}(%esp), %al
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfmadd213pd {{.*#+}} xmm0 = (xmm1 * xmm0) + xmm2
+; X86-NEXT: vfmadd213pd {{.*#+}} xmm0 {%k1} {z} = (xmm1 * xmm0) + xmm2
; X86-NEXT: retl
;
; X64-LABEL: test_mm_maskz_fmadd_pd:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfmadd213pd {{.*#+}} xmm0 = (xmm1 * xmm0) + xmm2
+; X64-NEXT: vfmadd213pd {{.*#+}} xmm0 {%k1} {z} = (xmm1 * xmm0) + xmm2
; X64-NEXT: retq
entry:
%0 = tail call <2 x double> @llvm.fma.v2f64(<2 x double> %__A, <2 x double> %__B, <2 x double> %__C) #9
@@ -4451,13 +4451,13 @@ define <2 x double> @test_mm_maskz_fmsub_pd(i8 zeroext %__U, <2 x double> %__A,
; X86: # %bb.0: # %entry
; X86-NEXT: movb {{[0-9]+}}(%esp), %al
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfmsub213pd {{.*#+}} xmm0 = (xmm1 * xmm0) - xmm2
+; X86-NEXT: vfmsub213pd {{.*#+}} xmm0 {%k1} {z} = (xmm1 * xmm0) - xmm2
; X86-NEXT: retl
;
; X64-LABEL: test_mm_maskz_fmsub_pd:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfmsub213pd {{.*#+}} xmm0 = (xmm1 * xmm0) - xmm2
+; X64-NEXT: vfmsub213pd {{.*#+}} xmm0 {%k1} {z} = (xmm1 * xmm0) - xmm2
; X64-NEXT: retq
entry:
%sub.i = fsub <2 x double> <double -0.000000e+00, double -0.000000e+00>, %__C
@@ -4473,13 +4473,13 @@ define <2 x double> @test_mm_maskz_fnmadd_pd(i8 zeroext %__U, <2 x double> %__A,
; X86: # %bb.0: # %entry
; X86-NEXT: movb {{[0-9]+}}(%esp), %al
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfnmadd213pd {{.*#+}} xmm0 = -(xmm1 * xmm0) + xmm2
+; X86-NEXT: vfnmadd213pd {{.*#+}} xmm0 {%k1} {z} = -(xmm1 * xmm0) + xmm2
; X86-NEXT: retl
;
; X64-LABEL: test_mm_maskz_fnmadd_pd:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfnmadd213pd {{.*#+}} xmm0 = -(xmm1 * xmm0) + xmm2
+; X64-NEXT: vfnmadd213pd {{.*#+}} xmm0 {%k1} {z} = -(xmm1 * xmm0) + xmm2
; X64-NEXT: retq
entry:
%sub.i = fsub <2 x double> <double -0.000000e+00, double -0.000000e+00>, %__A
@@ -4495,13 +4495,13 @@ define <2 x double> @test_mm_maskz_fnmsub_pd(i8 zeroext %__U, <2 x double> %__A,
; X86: # %bb.0: # %entry
; X86-NEXT: movb {{[0-9]+}}(%esp), %al
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfnmsub213pd {{.*#+}} xmm0 = -(xmm1 * xmm0) - xmm2
+; X86-NEXT: vfnmsub213pd {{.*#+}} xmm0 {%k1} {z} = -(xmm1 * xmm0) - xmm2
; X86-NEXT: retl
;
; X64-LABEL: test_mm_maskz_fnmsub_pd:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfnmsub213pd {{.*#+}} xmm0 = -(xmm1 * xmm0) - xmm2
+; X64-NEXT: vfnmsub213pd {{.*#+}} xmm0 {%k1} {z} = -(xmm1 * xmm0) - xmm2
; X64-NEXT: retq
entry:
%sub.i = fsub <2 x double> <double -0.000000e+00, double -0.000000e+00>, %__A
@@ -4518,13 +4518,13 @@ define <4 x double> @test_mm256_mask_fmadd_pd(<4 x double> %__A, i8 zeroext %__U
; X86: # %bb.0: # %entry
; X86-NEXT: movb {{[0-9]+}}(%esp), %al
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfmadd132pd {{.*#+}} ymm0 = (ymm0 * ymm1) + ymm2
+; X86-NEXT: vfmadd132pd {{.*#+}} ymm0 {%k1} = (ymm0 * ymm1) + ymm2
; X86-NEXT: retl
;
; X64-LABEL: test_mm256_mask_fmadd_pd:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfmadd132pd {{.*#+}} ymm0 = (ymm0 * ymm1) + ymm2
+; X64-NEXT: vfmadd132pd {{.*#+}} ymm0 {%k1} = (ymm0 * ymm1) + ymm2
; X64-NEXT: retq
entry:
%0 = tail call <4 x double> @llvm.fma.v4f64(<4 x double> %__A, <4 x double> %__B, <4 x double> %__C) #9
@@ -4539,13 +4539,13 @@ define <4 x double> @test_mm256_mask_fmsub_pd(<4 x double> %__A, i8 zeroext %__U
; X86: # %bb.0: # %entry
; X86-NEXT: movb {{[0-9]+}}(%esp), %al
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfmsub132pd {{.*#+}} ymm0 = (ymm0 * ymm1) - ymm2
+; X86-NEXT: vfmsub132pd {{.*#+}} ymm0 {%k1} = (ymm0 * ymm1) - ymm2
; X86-NEXT: retl
;
; X64-LABEL: test_mm256_mask_fmsub_pd:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfmsub132pd {{.*#+}} ymm0 = (ymm0 * ymm1) - ymm2
+; X64-NEXT: vfmsub132pd {{.*#+}} ymm0 {%k1} = (ymm0 * ymm1) - ymm2
; X64-NEXT: retq
entry:
%sub.i = fsub <4 x double> <double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00>, %__C
@@ -4561,14 +4561,14 @@ define <4 x double> @test_mm256_mask3_fmadd_pd(<4 x double> %__A, <4 x double> %
; X86: # %bb.0: # %entry
; X86-NEXT: movb {{[0-9]+}}(%esp), %al
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfmadd231pd {{.*#+}} ymm2 = (ymm0 * ymm1) + ymm2
+; X86-NEXT: vfmadd231pd {{.*#+}} ymm2 {%k1} = (ymm0 * ymm1) + ymm2
; X86-NEXT: vmovapd %ymm2, %ymm0
; X86-NEXT: retl
;
; X64-LABEL: test_mm256_mask3_fmadd_pd:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfmadd231pd {{.*#+}} ymm2 = (ymm0 * ymm1) + ymm2
+; X64-NEXT: vfmadd231pd {{.*#+}} ymm2 {%k1} = (ymm0 * ymm1) + ymm2
; X64-NEXT: vmovapd %ymm2, %ymm0
; X64-NEXT: retq
entry:
@@ -4584,14 +4584,14 @@ define <4 x double> @test_mm256_mask3_fnmadd_pd(<4 x double> %__A, <4 x double>
; X86: # %bb.0: # %entry
; X86-NEXT: movb {{[0-9]+}}(%esp), %al
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfnmadd231pd {{.*#+}} ymm2 = -(ymm0 * ymm1) + ymm2
+; X86-NEXT: vfnmadd231pd {{.*#+}} ymm2 {%k1} = -(ymm0 * ymm1) + ymm2
; X86-NEXT: vmovapd %ymm2, %ymm0
; X86-NEXT: retl
;
; X64-LABEL: test_mm256_mask3_fnmadd_pd:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfnmadd231pd {{.*#+}} ymm2 = -(ymm0 * ymm1) + ymm2
+; X64-NEXT: vfnmadd231pd {{.*#+}} ymm2 {%k1} = -(ymm0 * ymm1) + ymm2
; X64-NEXT: vmovapd %ymm2, %ymm0
; X64-NEXT: retq
entry:
@@ -4608,13 +4608,13 @@ define <4 x double> @test_mm256_maskz_fmadd_pd(i8 zeroext %__U, <4 x double> %__
; X86: # %bb.0: # %entry
; X86-NEXT: movb {{[0-9]+}}(%esp), %al
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfmadd213pd {{.*#+}} ymm0 = (ymm1 * ymm0) + ymm2
+; X86-NEXT: vfmadd213pd {{.*#+}} ymm0 {%k1} {z} = (ymm1 * ymm0) + ymm2
; X86-NEXT: retl
;
; X64-LABEL: test_mm256_maskz_fmadd_pd:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfmadd213pd {{.*#+}} ymm0 = (ymm1 * ymm0) + ymm2
+; X64-NEXT: vfmadd213pd {{.*#+}} ymm0 {%k1} {z} = (ymm1 * ymm0) + ymm2
; X64-NEXT: retq
entry:
%0 = tail call <4 x double> @llvm.fma.v4f64(<4 x double> %__A, <4 x double> %__B, <4 x double> %__C) #9
@@ -4629,13 +4629,13 @@ define <4 x double> @test_mm256_maskz_fmsub_pd(i8 zeroext %__U, <4 x double> %__
; X86: # %bb.0: # %entry
; X86-NEXT: movb {{[0-9]+}}(%esp), %al
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfmsub213pd {{.*#+}} ymm0 = (ymm1 * ymm0) - ymm2
+; X86-NEXT: vfmsub213pd {{.*#+}} ymm0 {%k1} {z} = (ymm1 * ymm0) - ymm2
; X86-NEXT: retl
;
; X64-LABEL: test_mm256_maskz_fmsub_pd:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfmsub213pd {{.*#+}} ymm0 = (ymm1 * ymm0) - ymm2
+; X64-NEXT: vfmsub213pd {{.*#+}} ymm0 {%k1} {z} = (ymm1 * ymm0) - ymm2
; X64-NEXT: retq
entry:
%sub.i = fsub <4 x double> <double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00>, %__C
@@ -4651,13 +4651,13 @@ define <4 x double> @test_mm256_maskz_fnmadd_pd(i8 zeroext %__U, <4 x double> %_
; X86: # %bb.0: # %entry
; X86-NEXT: movb {{[0-9]+}}(%esp), %al
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfnmadd213pd {{.*#+}} ymm0 = -(ymm1 * ymm0) + ymm2
+; X86-NEXT: vfnmadd213pd {{.*#+}} ymm0 {%k1} {z} = -(ymm1 * ymm0) + ymm2
; X86-NEXT: retl
;
; X64-LABEL: test_mm256_maskz_fnmadd_pd:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfnmadd213pd {{.*#+}} ymm0 = -(ymm1 * ymm0) + ymm2
+; X64-NEXT: vfnmadd213pd {{.*#+}} ymm0 {%k1} {z} = -(ymm1 * ymm0) + ymm2
; X64-NEXT: retq
entry:
%sub.i = fsub <4 x double> <double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00>, %__A
@@ -4673,13 +4673,13 @@ define <4 x double> @test_mm256_maskz_fnmsub_pd(i8 zeroext %__U, <4 x double> %_
; X86: # %bb.0: # %entry
; X86-NEXT: movb {{[0-9]+}}(%esp), %al
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfnmsub213pd {{.*#+}} ymm0 = -(ymm1 * ymm0) - ymm2
+; X86-NEXT: vfnmsub213pd {{.*#+}} ymm0 {%k1} {z} = -(ymm1 * ymm0) - ymm2
; X86-NEXT: retl
;
; X64-LABEL: test_mm256_maskz_fnmsub_pd:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfnmsub213pd {{.*#+}} ymm0 = -(ymm1 * ymm0) - ymm2
+; X64-NEXT: vfnmsub213pd {{.*#+}} ymm0 {%k1} {z} = -(ymm1 * ymm0) - ymm2
; X64-NEXT: retq
entry:
%sub.i = fsub <4 x double> <double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00>, %__A
@@ -4696,13 +4696,13 @@ define <4 x float> @test_mm_mask_fmadd_ps(<4 x float> %__A, i8 zeroext %__U, <4
; X86: # %bb.0: # %entry
; X86-NEXT: movb {{[0-9]+}}(%esp), %al
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfmadd132ps {{.*#+}} xmm0 = (xmm0 * xmm1) + xmm2
+; X86-NEXT: vfmadd132ps {{.*#+}} xmm0 {%k1} = (xmm0 * xmm1) + xmm2
; X86-NEXT: retl
;
; X64-LABEL: test_mm_mask_fmadd_ps:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfmadd132ps {{.*#+}} xmm0 = (xmm0 * xmm1) + xmm2
+; X64-NEXT: vfmadd132ps {{.*#+}} xmm0 {%k1} = (xmm0 * xmm1) + xmm2
; X64-NEXT: retq
entry:
%0 = tail call <4 x float> @llvm.fma.v4f32(<4 x float> %__A, <4 x float> %__B, <4 x float> %__C) #9
@@ -4717,13 +4717,13 @@ define <4 x float> @test_mm_mask_fmsub_ps(<4 x float> %__A, i8 zeroext %__U, <4
; X86: # %bb.0: # %entry
; X86-NEXT: movb {{[0-9]+}}(%esp), %al
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfmsub132ps {{.*#+}} xmm0 = (xmm0 * xmm1) - xmm2
+; X86-NEXT: vfmsub132ps {{.*#+}} xmm0 {%k1} = (xmm0 * xmm1) - xmm2
; X86-NEXT: retl
;
; X64-LABEL: test_mm_mask_fmsub_ps:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfmsub132ps {{.*#+}} xmm0 = (xmm0 * xmm1) - xmm2
+; X64-NEXT: vfmsub132ps {{.*#+}} xmm0 {%k1} = (xmm0 * xmm1) - xmm2
; X64-NEXT: retq
entry:
%sub.i = fsub <4 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %__C
@@ -4739,14 +4739,14 @@ define <4 x float> @test_mm_mask3_fmadd_ps(<4 x float> %__A, <4 x float> %__B, <
; X86: # %bb.0: # %entry
; X86-NEXT: movb {{[0-9]+}}(%esp), %al
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfmadd231ps {{.*#+}} xmm2 = (xmm0 * xmm1) + xmm2
+; X86-NEXT: vfmadd231ps {{.*#+}} xmm2 {%k1} = (xmm0 * xmm1) + xmm2
; X86-NEXT: vmovaps %xmm2, %xmm0
; X86-NEXT: retl
;
; X64-LABEL: test_mm_mask3_fmadd_ps:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfmadd231ps {{.*#+}} xmm2 = (xmm0 * xmm1) + xmm2
+; X64-NEXT: vfmadd231ps {{.*#+}} xmm2 {%k1} = (xmm0 * xmm1) + xmm2
; X64-NEXT: vmovaps %xmm2, %xmm0
; X64-NEXT: retq
entry:
@@ -4762,14 +4762,14 @@ define <4 x float> @test_mm_mask3_fnmadd_ps(<4 x float> %__A, <4 x float> %__B,
; X86: # %bb.0: # %entry
; X86-NEXT: movb {{[0-9]+}}(%esp), %al
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfnmadd231ps {{.*#+}} xmm2 = -(xmm0 * xmm1) + xmm2
+; X86-NEXT: vfnmadd231ps {{.*#+}} xmm2 {%k1} = -(xmm0 * xmm1) + xmm2
; X86-NEXT: vmovaps %xmm2, %xmm0
; X86-NEXT: retl
;
; X64-LABEL: test_mm_mask3_fnmadd_ps:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfnmadd231ps {{.*#+}} xmm2 = -(xmm0 * xmm1) + xmm2
+; X64-NEXT: vfnmadd231ps {{.*#+}} xmm2 {%k1} = -(xmm0 * xmm1) + xmm2
; X64-NEXT: vmovaps %xmm2, %xmm0
; X64-NEXT: retq
entry:
@@ -4786,13 +4786,13 @@ define <4 x float> @test_mm_maskz_fmadd_ps(i8 zeroext %__U, <4 x float> %__A, <4
; X86: # %bb.0: # %entry
; X86-NEXT: movb {{[0-9]+}}(%esp), %al
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfmadd213ps {{.*#+}} xmm0 = (xmm1 * xmm0) + xmm2
+; X86-NEXT: vfmadd213ps {{.*#+}} xmm0 {%k1} {z} = (xmm1 * xmm0) + xmm2
; X86-NEXT: retl
;
; X64-LABEL: test_mm_maskz_fmadd_ps:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfmadd213ps {{.*#+}} xmm0 = (xmm1 * xmm0) + xmm2
+; X64-NEXT: vfmadd213ps {{.*#+}} xmm0 {%k1} {z} = (xmm1 * xmm0) + xmm2
; X64-NEXT: retq
entry:
%0 = tail call <4 x float> @llvm.fma.v4f32(<4 x float> %__A, <4 x float> %__B, <4 x float> %__C) #9
@@ -4807,13 +4807,13 @@ define <4 x float> @test_mm_maskz_fmsub_ps(i8 zeroext %__U, <4 x float> %__A, <4
; X86: # %bb.0: # %entry
; X86-NEXT: movb {{[0-9]+}}(%esp), %al
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfmsub213ps {{.*#+}} xmm0 = (xmm1 * xmm0) - xmm2
+; X86-NEXT: vfmsub213ps {{.*#+}} xmm0 {%k1} {z} = (xmm1 * xmm0) - xmm2
; X86-NEXT: retl
;
; X64-LABEL: test_mm_maskz_fmsub_ps:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfmsub213ps {{.*#+}} xmm0 = (xmm1 * xmm0) - xmm2
+; X64-NEXT: vfmsub213ps {{.*#+}} xmm0 {%k1} {z} = (xmm1 * xmm0) - xmm2
; X64-NEXT: retq
entry:
%sub.i = fsub <4 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %__C
@@ -4829,13 +4829,13 @@ define <4 x float> @test_mm_maskz_fnmadd_ps(i8 zeroext %__U, <4 x float> %__A, <
; X86: # %bb.0: # %entry
; X86-NEXT: movb {{[0-9]+}}(%esp), %al
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfnmadd213ps {{.*#+}} xmm0 = -(xmm1 * xmm0) + xmm2
+; X86-NEXT: vfnmadd213ps {{.*#+}} xmm0 {%k1} {z} = -(xmm1 * xmm0) + xmm2
; X86-NEXT: retl
;
; X64-LABEL: test_mm_maskz_fnmadd_ps:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfnmadd213ps {{.*#+}} xmm0 = -(xmm1 * xmm0) + xmm2
+; X64-NEXT: vfnmadd213ps {{.*#+}} xmm0 {%k1} {z} = -(xmm1 * xmm0) + xmm2
; X64-NEXT: retq
entry:
%sub.i = fsub <4 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %__A
@@ -4851,13 +4851,13 @@ define <4 x float> @test_mm_maskz_fnmsub_ps(i8 zeroext %__U, <4 x float> %__A, <
; X86: # %bb.0: # %entry
; X86-NEXT: movb {{[0-9]+}}(%esp), %al
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfnmsub213ps {{.*#+}} xmm0 = -(xmm1 * xmm0) - xmm2
+; X86-NEXT: vfnmsub213ps {{.*#+}} xmm0 {%k1} {z} = -(xmm1 * xmm0) - xmm2
; X86-NEXT: retl
;
; X64-LABEL: test_mm_maskz_fnmsub_ps:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfnmsub213ps {{.*#+}} xmm0 = -(xmm1 * xmm0) - xmm2
+; X64-NEXT: vfnmsub213ps {{.*#+}} xmm0 {%k1} {z} = -(xmm1 * xmm0) - xmm2
; X64-NEXT: retq
entry:
%sub.i = fsub <4 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %__A
@@ -4874,13 +4874,13 @@ define <8 x float> @test_mm256_mask_fmadd_ps(<8 x float> %__A, i8 zeroext %__U,
; X86: # %bb.0: # %entry
; X86-NEXT: movb {{[0-9]+}}(%esp), %al
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfmadd132ps {{.*#+}} ymm0 = (ymm0 * ymm1) + ymm2
+; X86-NEXT: vfmadd132ps {{.*#+}} ymm0 {%k1} = (ymm0 * ymm1) + ymm2
; X86-NEXT: retl
;
; X64-LABEL: test_mm256_mask_fmadd_ps:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfmadd132ps {{.*#+}} ymm0 = (ymm0 * ymm1) + ymm2
+; X64-NEXT: vfmadd132ps {{.*#+}} ymm0 {%k1} = (ymm0 * ymm1) + ymm2
; X64-NEXT: retq
entry:
%0 = tail call <8 x float> @llvm.fma.v8f32(<8 x float> %__A, <8 x float> %__B, <8 x float> %__C) #9
@@ -4894,13 +4894,13 @@ define <8 x float> @test_mm256_mask_fmsub_ps(<8 x float> %__A, i8 zeroext %__U,
; X86: # %bb.0: # %entry
; X86-NEXT: movb {{[0-9]+}}(%esp), %al
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfmsub132ps {{.*#+}} ymm0 = (ymm0 * ymm1) - ymm2
+; X86-NEXT: vfmsub132ps {{.*#+}} ymm0 {%k1} = (ymm0 * ymm1) - ymm2
; X86-NEXT: retl
;
; X64-LABEL: test_mm256_mask_fmsub_ps:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfmsub132ps {{.*#+}} ymm0 = (ymm0 * ymm1) - ymm2
+; X64-NEXT: vfmsub132ps {{.*#+}} ymm0 {%k1} = (ymm0 * ymm1) - ymm2
; X64-NEXT: retq
entry:
%sub.i = fsub <8 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %__C
@@ -4915,14 +4915,14 @@ define <8 x float> @test_mm256_mask3_fmadd_ps(<8 x float> %__A, <8 x float> %__B
; X86: # %bb.0: # %entry
; X86-NEXT: movb {{[0-9]+}}(%esp), %al
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfmadd231ps {{.*#+}} ymm2 = (ymm0 * ymm1) + ymm2
+; X86-NEXT: vfmadd231ps {{.*#+}} ymm2 {%k1} = (ymm0 * ymm1) + ymm2
; X86-NEXT: vmovaps %ymm2, %ymm0
; X86-NEXT: retl
;
; X64-LABEL: test_mm256_mask3_fmadd_ps:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfmadd231ps {{.*#+}} ymm2 = (ymm0 * ymm1) + ymm2
+; X64-NEXT: vfmadd231ps {{.*#+}} ymm2 {%k1} = (ymm0 * ymm1) + ymm2
; X64-NEXT: vmovaps %ymm2, %ymm0
; X64-NEXT: retq
entry:
@@ -4937,14 +4937,14 @@ define <8 x float> @test_mm256_mask3_fnmadd_ps(<8 x float> %__A, <8 x float> %__
; X86: # %bb.0: # %entry
; X86-NEXT: movb {{[0-9]+}}(%esp), %al
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfnmadd231ps {{.*#+}} ymm2 = -(ymm0 * ymm1) + ymm2
+; X86-NEXT: vfnmadd231ps {{.*#+}} ymm2 {%k1} = -(ymm0 * ymm1) + ymm2
; X86-NEXT: vmovaps %ymm2, %ymm0
; X86-NEXT: retl
;
; X64-LABEL: test_mm256_mask3_fnmadd_ps:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfnmadd231ps {{.*#+}} ymm2 = -(ymm0 * ymm1) + ymm2
+; X64-NEXT: vfnmadd231ps {{.*#+}} ymm2 {%k1} = -(ymm0 * ymm1) + ymm2
; X64-NEXT: vmovaps %ymm2, %ymm0
; X64-NEXT: retq
entry:
@@ -4960,13 +4960,13 @@ define <8 x float> @test_mm256_maskz_fmadd_ps(i8 zeroext %__U, <8 x float> %__A,
; X86: # %bb.0: # %entry
; X86-NEXT: movb {{[0-9]+}}(%esp), %al
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfmadd213ps {{.*#+}} ymm0 = (ymm1 * ymm0) + ymm2
+; X86-NEXT: vfmadd213ps {{.*#+}} ymm0 {%k1} {z} = (ymm1 * ymm0) + ymm2
; X86-NEXT: retl
;
; X64-LABEL: test_mm256_maskz_fmadd_ps:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfmadd213ps {{.*#+}} ymm0 = (ymm1 * ymm0) + ymm2
+; X64-NEXT: vfmadd213ps {{.*#+}} ymm0 {%k1} {z} = (ymm1 * ymm0) + ymm2
; X64-NEXT: retq
entry:
%0 = tail call <8 x float> @llvm.fma.v8f32(<8 x float> %__A, <8 x float> %__B, <8 x float> %__C) #9
@@ -4980,13 +4980,13 @@ define <8 x float> @test_mm256_maskz_fmsub_ps(i8 zeroext %__U, <8 x float> %__A,
; X86: # %bb.0: # %entry
; X86-NEXT: movb {{[0-9]+}}(%esp), %al
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfmsub213ps {{.*#+}} ymm0 = (ymm1 * ymm0) - ymm2
+; X86-NEXT: vfmsub213ps {{.*#+}} ymm0 {%k1} {z} = (ymm1 * ymm0) - ymm2
; X86-NEXT: retl
;
; X64-LABEL: test_mm256_maskz_fmsub_ps:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfmsub213ps {{.*#+}} ymm0 = (ymm1 * ymm0) - ymm2
+; X64-NEXT: vfmsub213ps {{.*#+}} ymm0 {%k1} {z} = (ymm1 * ymm0) - ymm2
; X64-NEXT: retq
entry:
%sub.i = fsub <8 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %__C
@@ -5001,13 +5001,13 @@ define <8 x float> @test_mm256_maskz_fnmadd_ps(i8 zeroext %__U, <8 x float> %__A
; X86: # %bb.0: # %entry
; X86-NEXT: movb {{[0-9]+}}(%esp), %al
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfnmadd213ps {{.*#+}} ymm0 = -(ymm1 * ymm0) + ymm2
+; X86-NEXT: vfnmadd213ps {{.*#+}} ymm0 {%k1} {z} = -(ymm1 * ymm0) + ymm2
; X86-NEXT: retl
;
; X64-LABEL: test_mm256_maskz_fnmadd_ps:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfnmadd213ps {{.*#+}} ymm0 = -(ymm1 * ymm0) + ymm2
+; X64-NEXT: vfnmadd213ps {{.*#+}} ymm0 {%k1} {z} = -(ymm1 * ymm0) + ymm2
; X64-NEXT: retq
entry:
%sub.i = fsub <8 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %__A
@@ -5022,13 +5022,13 @@ define <8 x float> @test_mm256_maskz_fnmsub_ps(i8 zeroext %__U, <8 x float> %__A
; X86: # %bb.0: # %entry
; X86-NEXT: movb {{[0-9]+}}(%esp), %al
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfnmsub213ps {{.*#+}} ymm0 = -(ymm1 * ymm0) - ymm2
+; X86-NEXT: vfnmsub213ps {{.*#+}} ymm0 {%k1} {z} = -(ymm1 * ymm0) - ymm2
; X86-NEXT: retl
;
; X64-LABEL: test_mm256_maskz_fnmsub_ps:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfnmsub213ps {{.*#+}} ymm0 = -(ymm1 * ymm0) - ymm2
+; X64-NEXT: vfnmsub213ps {{.*#+}} ymm0 {%k1} {z} = -(ymm1 * ymm0) - ymm2
; X64-NEXT: retq
entry:
%sub.i = fsub <8 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %__A
@@ -5044,13 +5044,13 @@ define <2 x double> @test_mm_mask_fmaddsub_pd(<2 x double> %__A, i8 zeroext %__U
; X86: # %bb.0: # %entry
; X86-NEXT: movb {{[0-9]+}}(%esp), %al
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfmaddsub132pd {{.*#+}} xmm0 = (xmm0 * xmm1) +/- xmm2
+; X86-NEXT: vfmaddsub132pd {{.*#+}} xmm0 {%k1} = (xmm0 * xmm1) +/- xmm2
; X86-NEXT: retl
;
; X64-LABEL: test_mm_mask_fmaddsub_pd:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfmaddsub132pd {{.*#+}} xmm0 = (xmm0 * xmm1) +/- xmm2
+; X64-NEXT: vfmaddsub132pd {{.*#+}} xmm0 {%k1} = (xmm0 * xmm1) +/- xmm2
; X64-NEXT: retq
entry:
%0 = tail call <2 x double> @llvm.fma.v2f64(<2 x double> %__A, <2 x double> %__B, <2 x double> %__C) #9
@@ -5068,13 +5068,13 @@ define <2 x double> @test_mm_mask_fmsubadd_pd(<2 x double> %__A, i8 zeroext %__U
; X86: # %bb.0: # %entry
; X86-NEXT: movb {{[0-9]+}}(%esp), %al
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfmsubadd132pd {{.*#+}} xmm0 = (xmm0 * xmm1) -/+ xmm2
+; X86-NEXT: vfmsubadd132pd {{.*#+}} xmm0 {%k1} = (xmm0 * xmm1) -/+ xmm2
; X86-NEXT: retl
;
; X64-LABEL: test_mm_mask_fmsubadd_pd:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfmsubadd132pd {{.*#+}} xmm0 = (xmm0 * xmm1) -/+ xmm2
+; X64-NEXT: vfmsubadd132pd {{.*#+}} xmm0 {%k1} = (xmm0 * xmm1) -/+ xmm2
; X64-NEXT: retq
entry:
%sub.i = fsub <2 x double> <double -0.000000e+00, double -0.000000e+00>, %__C
@@ -5092,14 +5092,14 @@ define <2 x double> @test_mm_mask3_fmaddsub_pd(<2 x double> %__A, <2 x double> %
; X86: # %bb.0: # %entry
; X86-NEXT: movb {{[0-9]+}}(%esp), %al
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfmaddsub231pd {{.*#+}} xmm2 = (xmm0 * xmm1) +/- xmm2
+; X86-NEXT: vfmaddsub231pd {{.*#+}} xmm2 {%k1} = (xmm0 * xmm1) +/- xmm2
; X86-NEXT: vmovapd %xmm2, %xmm0
; X86-NEXT: retl
;
; X64-LABEL: test_mm_mask3_fmaddsub_pd:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfmaddsub231pd {{.*#+}} xmm2 = (xmm0 * xmm1) +/- xmm2
+; X64-NEXT: vfmaddsub231pd {{.*#+}} xmm2 {%k1} = (xmm0 * xmm1) +/- xmm2
; X64-NEXT: vmovapd %xmm2, %xmm0
; X64-NEXT: retq
entry:
@@ -5118,13 +5118,13 @@ define <2 x double> @test_mm_maskz_fmaddsub_pd(i8 zeroext %__U, <2 x double> %__
; X86: # %bb.0: # %entry
; X86-NEXT: movb {{[0-9]+}}(%esp), %al
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfmaddsub213pd {{.*#+}} xmm0 = (xmm1 * xmm0) +/- xmm2
+; X86-NEXT: vfmaddsub213pd {{.*#+}} xmm0 {%k1} {z} = (xmm1 * xmm0) +/- xmm2
; X86-NEXT: retl
;
; X64-LABEL: test_mm_maskz_fmaddsub_pd:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfmaddsub213pd {{.*#+}} xmm0 = (xmm1 * xmm0) +/- xmm2
+; X64-NEXT: vfmaddsub213pd {{.*#+}} xmm0 {%k1} {z} = (xmm1 * xmm0) +/- xmm2
; X64-NEXT: retq
entry:
%0 = tail call <2 x double> @llvm.fma.v2f64(<2 x double> %__A, <2 x double> %__B, <2 x double> %__C) #9
@@ -5142,13 +5142,13 @@ define <2 x double> @test_mm_maskz_fmsubadd_pd(i8 zeroext %__U, <2 x double> %__
; X86: # %bb.0: # %entry
; X86-NEXT: movb {{[0-9]+}}(%esp), %al
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfmsubadd213pd {{.*#+}} xmm0 = (xmm1 * xmm0) -/+ xmm2
+; X86-NEXT: vfmsubadd213pd {{.*#+}} xmm0 {%k1} {z} = (xmm1 * xmm0) -/+ xmm2
; X86-NEXT: retl
;
; X64-LABEL: test_mm_maskz_fmsubadd_pd:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfmsubadd213pd {{.*#+}} xmm0 = (xmm1 * xmm0) -/+ xmm2
+; X64-NEXT: vfmsubadd213pd {{.*#+}} xmm0 {%k1} {z} = (xmm1 * xmm0) -/+ xmm2
; X64-NEXT: retq
entry:
%sub.i = fsub <2 x double> <double -0.000000e+00, double -0.000000e+00>, %__C
@@ -5166,13 +5166,13 @@ define <4 x double> @test_mm256_mask_fmaddsub_pd(<4 x double> %__A, i8 zeroext %
; X86: # %bb.0: # %entry
; X86-NEXT: movb {{[0-9]+}}(%esp), %al
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfmaddsub132pd {{.*#+}} ymm0 = (ymm0 * ymm1) +/- ymm2
+; X86-NEXT: vfmaddsub132pd {{.*#+}} ymm0 {%k1} = (ymm0 * ymm1) +/- ymm2
; X86-NEXT: retl
;
; X64-LABEL: test_mm256_mask_fmaddsub_pd:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfmaddsub132pd {{.*#+}} ymm0 = (ymm0 * ymm1) +/- ymm2
+; X64-NEXT: vfmaddsub132pd {{.*#+}} ymm0 {%k1} = (ymm0 * ymm1) +/- ymm2
; X64-NEXT: retq
entry:
%0 = tail call <4 x double> @llvm.fma.v4f64(<4 x double> %__A, <4 x double> %__B, <4 x double> %__C) #9
@@ -5190,13 +5190,13 @@ define <4 x double> @test_mm256_mask_fmsubadd_pd(<4 x double> %__A, i8 zeroext %
; X86: # %bb.0: # %entry
; X86-NEXT: movb {{[0-9]+}}(%esp), %al
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfmsubadd132pd {{.*#+}} ymm0 = (ymm0 * ymm1) -/+ ymm2
+; X86-NEXT: vfmsubadd132pd {{.*#+}} ymm0 {%k1} = (ymm0 * ymm1) -/+ ymm2
; X86-NEXT: retl
;
; X64-LABEL: test_mm256_mask_fmsubadd_pd:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfmsubadd132pd {{.*#+}} ymm0 = (ymm0 * ymm1) -/+ ymm2
+; X64-NEXT: vfmsubadd132pd {{.*#+}} ymm0 {%k1} = (ymm0 * ymm1) -/+ ymm2
; X64-NEXT: retq
entry:
%sub.i = fsub <4 x double> <double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00>, %__C
@@ -5214,14 +5214,14 @@ define <4 x double> @test_mm256_mask3_fmaddsub_pd(<4 x double> %__A, <4 x double
; X86: # %bb.0: # %entry
; X86-NEXT: movb {{[0-9]+}}(%esp), %al
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfmaddsub231pd {{.*#+}} ymm2 = (ymm0 * ymm1) +/- ymm2
+; X86-NEXT: vfmaddsub231pd {{.*#+}} ymm2 {%k1} = (ymm0 * ymm1) +/- ymm2
; X86-NEXT: vmovapd %ymm2, %ymm0
; X86-NEXT: retl
;
; X64-LABEL: test_mm256_mask3_fmaddsub_pd:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfmaddsub231pd {{.*#+}} ymm2 = (ymm0 * ymm1) +/- ymm2
+; X64-NEXT: vfmaddsub231pd {{.*#+}} ymm2 {%k1} = (ymm0 * ymm1) +/- ymm2
; X64-NEXT: vmovapd %ymm2, %ymm0
; X64-NEXT: retq
entry:
@@ -5240,13 +5240,13 @@ define <4 x double> @test_mm256_maskz_fmaddsub_pd(i8 zeroext %__U, <4 x double>
; X86: # %bb.0: # %entry
; X86-NEXT: movb {{[0-9]+}}(%esp), %al
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfmaddsub213pd {{.*#+}} ymm0 = (ymm1 * ymm0) +/- ymm2
+; X86-NEXT: vfmaddsub213pd {{.*#+}} ymm0 {%k1} {z} = (ymm1 * ymm0) +/- ymm2
; X86-NEXT: retl
;
; X64-LABEL: test_mm256_maskz_fmaddsub_pd:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfmaddsub213pd {{.*#+}} ymm0 = (ymm1 * ymm0) +/- ymm2
+; X64-NEXT: vfmaddsub213pd {{.*#+}} ymm0 {%k1} {z} = (ymm1 * ymm0) +/- ymm2
; X64-NEXT: retq
entry:
%0 = tail call <4 x double> @llvm.fma.v4f64(<4 x double> %__A, <4 x double> %__B, <4 x double> %__C) #9
@@ -5264,13 +5264,13 @@ define <4 x double> @test_mm256_maskz_fmsubadd_pd(i8 zeroext %__U, <4 x double>
; X86: # %bb.0: # %entry
; X86-NEXT: movb {{[0-9]+}}(%esp), %al
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfmsubadd213pd {{.*#+}} ymm0 = (ymm1 * ymm0) -/+ ymm2
+; X86-NEXT: vfmsubadd213pd {{.*#+}} ymm0 {%k1} {z} = (ymm1 * ymm0) -/+ ymm2
; X86-NEXT: retl
;
; X64-LABEL: test_mm256_maskz_fmsubadd_pd:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfmsubadd213pd {{.*#+}} ymm0 = (ymm1 * ymm0) -/+ ymm2
+; X64-NEXT: vfmsubadd213pd {{.*#+}} ymm0 {%k1} {z} = (ymm1 * ymm0) -/+ ymm2
; X64-NEXT: retq
entry:
%sub.i = fsub <4 x double> <double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00>, %__C
@@ -5288,13 +5288,13 @@ define <4 x float> @test_mm_mask_fmaddsub_ps(<4 x float> %__A, i8 zeroext %__U,
; X86: # %bb.0: # %entry
; X86-NEXT: movb {{[0-9]+}}(%esp), %al
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfmaddsub132ps {{.*#+}} xmm0 = (xmm0 * xmm1) +/- xmm2
+; X86-NEXT: vfmaddsub132ps {{.*#+}} xmm0 {%k1} = (xmm0 * xmm1) +/- xmm2
; X86-NEXT: retl
;
; X64-LABEL: test_mm_mask_fmaddsub_ps:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfmaddsub132ps {{.*#+}} xmm0 = (xmm0 * xmm1) +/- xmm2
+; X64-NEXT: vfmaddsub132ps {{.*#+}} xmm0 {%k1} = (xmm0 * xmm1) +/- xmm2
; X64-NEXT: retq
entry:
%0 = tail call <4 x float> @llvm.fma.v4f32(<4 x float> %__A, <4 x float> %__B, <4 x float> %__C) #9
@@ -5312,13 +5312,13 @@ define <4 x float> @test_mm_mask_fmsubadd_ps(<4 x float> %__A, i8 zeroext %__U,
; X86: # %bb.0: # %entry
; X86-NEXT: movb {{[0-9]+}}(%esp), %al
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfmsubadd132ps {{.*#+}} xmm0 = (xmm0 * xmm1) -/+ xmm2
+; X86-NEXT: vfmsubadd132ps {{.*#+}} xmm0 {%k1} = (xmm0 * xmm1) -/+ xmm2
; X86-NEXT: retl
;
; X64-LABEL: test_mm_mask_fmsubadd_ps:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfmsubadd132ps {{.*#+}} xmm0 = (xmm0 * xmm1) -/+ xmm2
+; X64-NEXT: vfmsubadd132ps {{.*#+}} xmm0 {%k1} = (xmm0 * xmm1) -/+ xmm2
; X64-NEXT: retq
entry:
%sub.i = fsub <4 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %__C
@@ -5336,14 +5336,14 @@ define <4 x float> @test_mm_mask3_fmaddsub_ps(<4 x float> %__A, <4 x float> %__B
; X86: # %bb.0: # %entry
; X86-NEXT: movb {{[0-9]+}}(%esp), %al
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfmaddsub231ps {{.*#+}} xmm2 = (xmm0 * xmm1) +/- xmm2
+; X86-NEXT: vfmaddsub231ps {{.*#+}} xmm2 {%k1} = (xmm0 * xmm1) +/- xmm2
; X86-NEXT: vmovaps %xmm2, %xmm0
; X86-NEXT: retl
;
; X64-LABEL: test_mm_mask3_fmaddsub_ps:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfmaddsub231ps {{.*#+}} xmm2 = (xmm0 * xmm1) +/- xmm2
+; X64-NEXT: vfmaddsub231ps {{.*#+}} xmm2 {%k1} = (xmm0 * xmm1) +/- xmm2
; X64-NEXT: vmovaps %xmm2, %xmm0
; X64-NEXT: retq
entry:
@@ -5362,13 +5362,13 @@ define <4 x float> @test_mm_maskz_fmaddsub_ps(i8 zeroext %__U, <4 x float> %__A,
; X86: # %bb.0: # %entry
; X86-NEXT: movb {{[0-9]+}}(%esp), %al
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfmaddsub213ps {{.*#+}} xmm0 = (xmm1 * xmm0) +/- xmm2
+; X86-NEXT: vfmaddsub213ps {{.*#+}} xmm0 {%k1} {z} = (xmm1 * xmm0) +/- xmm2
; X86-NEXT: retl
;
; X64-LABEL: test_mm_maskz_fmaddsub_ps:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfmaddsub213ps {{.*#+}} xmm0 = (xmm1 * xmm0) +/- xmm2
+; X64-NEXT: vfmaddsub213ps {{.*#+}} xmm0 {%k1} {z} = (xmm1 * xmm0) +/- xmm2
; X64-NEXT: retq
entry:
%0 = tail call <4 x float> @llvm.fma.v4f32(<4 x float> %__A, <4 x float> %__B, <4 x float> %__C) #9
@@ -5386,13 +5386,13 @@ define <4 x float> @test_mm_maskz_fmsubadd_ps(i8 zeroext %__U, <4 x float> %__A,
; X86: # %bb.0: # %entry
; X86-NEXT: movb {{[0-9]+}}(%esp), %al
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfmsubadd213ps {{.*#+}} xmm0 = (xmm1 * xmm0) -/+ xmm2
+; X86-NEXT: vfmsubadd213ps {{.*#+}} xmm0 {%k1} {z} = (xmm1 * xmm0) -/+ xmm2
; X86-NEXT: retl
;
; X64-LABEL: test_mm_maskz_fmsubadd_ps:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfmsubadd213ps {{.*#+}} xmm0 = (xmm1 * xmm0) -/+ xmm2
+; X64-NEXT: vfmsubadd213ps {{.*#+}} xmm0 {%k1} {z} = (xmm1 * xmm0) -/+ xmm2
; X64-NEXT: retq
entry:
%sub.i = fsub <4 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %__C
@@ -5410,13 +5410,13 @@ define <8 x float> @test_mm256_mask_fmaddsub_ps(<8 x float> %__A, i8 zeroext %__
; X86: # %bb.0: # %entry
; X86-NEXT: movb {{[0-9]+}}(%esp), %al
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfmaddsub132ps {{.*#+}} ymm0 = (ymm0 * ymm1) +/- ymm2
+; X86-NEXT: vfmaddsub132ps {{.*#+}} ymm0 {%k1} = (ymm0 * ymm1) +/- ymm2
; X86-NEXT: retl
;
; X64-LABEL: test_mm256_mask_fmaddsub_ps:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfmaddsub132ps {{.*#+}} ymm0 = (ymm0 * ymm1) +/- ymm2
+; X64-NEXT: vfmaddsub132ps {{.*#+}} ymm0 {%k1} = (ymm0 * ymm1) +/- ymm2
; X64-NEXT: retq
entry:
%0 = tail call <8 x float> @llvm.fma.v8f32(<8 x float> %__A, <8 x float> %__B, <8 x float> %__C) #9
@@ -5433,13 +5433,13 @@ define <8 x float> @test_mm256_mask_fmsubadd_ps(<8 x float> %__A, i8 zeroext %__
; X86: # %bb.0: # %entry
; X86-NEXT: movb {{[0-9]+}}(%esp), %al
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfmsubadd132ps {{.*#+}} ymm0 = (ymm0 * ymm1) -/+ ymm2
+; X86-NEXT: vfmsubadd132ps {{.*#+}} ymm0 {%k1} = (ymm0 * ymm1) -/+ ymm2
; X86-NEXT: retl
;
; X64-LABEL: test_mm256_mask_fmsubadd_ps:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfmsubadd132ps {{.*#+}} ymm0 = (ymm0 * ymm1) -/+ ymm2
+; X64-NEXT: vfmsubadd132ps {{.*#+}} ymm0 {%k1} = (ymm0 * ymm1) -/+ ymm2
; X64-NEXT: retq
entry:
%sub.i = fsub <8 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %__C
@@ -5456,14 +5456,14 @@ define <8 x float> @test_mm256_mask3_fmaddsub_ps(<8 x float> %__A, <8 x float> %
; X86: # %bb.0: # %entry
; X86-NEXT: movb {{[0-9]+}}(%esp), %al
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfmaddsub231ps {{.*#+}} ymm2 = (ymm0 * ymm1) +/- ymm2
+; X86-NEXT: vfmaddsub231ps {{.*#+}} ymm2 {%k1} = (ymm0 * ymm1) +/- ymm2
; X86-NEXT: vmovaps %ymm2, %ymm0
; X86-NEXT: retl
;
; X64-LABEL: test_mm256_mask3_fmaddsub_ps:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfmaddsub231ps {{.*#+}} ymm2 = (ymm0 * ymm1) +/- ymm2
+; X64-NEXT: vfmaddsub231ps {{.*#+}} ymm2 {%k1} = (ymm0 * ymm1) +/- ymm2
; X64-NEXT: vmovaps %ymm2, %ymm0
; X64-NEXT: retq
entry:
@@ -5481,13 +5481,13 @@ define <8 x float> @test_mm256_maskz_fmaddsub_ps(i8 zeroext %__U, <8 x float> %_
; X86: # %bb.0: # %entry
; X86-NEXT: movb {{[0-9]+}}(%esp), %al
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfmaddsub213ps {{.*#+}} ymm0 = (ymm1 * ymm0) +/- ymm2
+; X86-NEXT: vfmaddsub213ps {{.*#+}} ymm0 {%k1} {z} = (ymm1 * ymm0) +/- ymm2
; X86-NEXT: retl
;
; X64-LABEL: test_mm256_maskz_fmaddsub_ps:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfmaddsub213ps {{.*#+}} ymm0 = (ymm1 * ymm0) +/- ymm2
+; X64-NEXT: vfmaddsub213ps {{.*#+}} ymm0 {%k1} {z} = (ymm1 * ymm0) +/- ymm2
; X64-NEXT: retq
entry:
%0 = tail call <8 x float> @llvm.fma.v8f32(<8 x float> %__A, <8 x float> %__B, <8 x float> %__C) #9
@@ -5504,13 +5504,13 @@ define <8 x float> @test_mm256_maskz_fmsubadd_ps(i8 zeroext %__U, <8 x float> %_
; X86: # %bb.0: # %entry
; X86-NEXT: movb {{[0-9]+}}(%esp), %al
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfmsubadd213ps {{.*#+}} ymm0 = (ymm1 * ymm0) -/+ ymm2
+; X86-NEXT: vfmsubadd213ps {{.*#+}} ymm0 {%k1} {z} = (ymm1 * ymm0) -/+ ymm2
; X86-NEXT: retl
;
; X64-LABEL: test_mm256_maskz_fmsubadd_ps:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfmsubadd213ps {{.*#+}} ymm0 = (ymm1 * ymm0) -/+ ymm2
+; X64-NEXT: vfmsubadd213ps {{.*#+}} ymm0 {%k1} {z} = (ymm1 * ymm0) -/+ ymm2
; X64-NEXT: retq
entry:
%sub.i = fsub <8 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %__C
@@ -5527,14 +5527,14 @@ define <2 x double> @test_mm_mask3_fmsub_pd(<2 x double> %__A, <2 x double> %__B
; X86: # %bb.0: # %entry
; X86-NEXT: movb {{[0-9]+}}(%esp), %al
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfmsub231pd {{.*#+}} xmm2 = (xmm0 * xmm1) - xmm2
+; X86-NEXT: vfmsub231pd {{.*#+}} xmm2 {%k1} = (xmm0 * xmm1) - xmm2
; X86-NEXT: vmovapd %xmm2, %xmm0
; X86-NEXT: retl
;
; X64-LABEL: test_mm_mask3_fmsub_pd:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfmsub231pd {{.*#+}} xmm2 = (xmm0 * xmm1) - xmm2
+; X64-NEXT: vfmsub231pd {{.*#+}} xmm2 {%k1} = (xmm0 * xmm1) - xmm2
; X64-NEXT: vmovapd %xmm2, %xmm0
; X64-NEXT: retq
entry:
@@ -5551,14 +5551,14 @@ define <4 x double> @test_mm256_mask3_fmsub_pd(<4 x double> %__A, <4 x double> %
; X86: # %bb.0: # %entry
; X86-NEXT: movb {{[0-9]+}}(%esp), %al
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfmsub231pd {{.*#+}} ymm2 = (ymm0 * ymm1) - ymm2
+; X86-NEXT: vfmsub231pd {{.*#+}} ymm2 {%k1} = (ymm0 * ymm1) - ymm2
; X86-NEXT: vmovapd %ymm2, %ymm0
; X86-NEXT: retl
;
; X64-LABEL: test_mm256_mask3_fmsub_pd:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfmsub231pd {{.*#+}} ymm2 = (ymm0 * ymm1) - ymm2
+; X64-NEXT: vfmsub231pd {{.*#+}} ymm2 {%k1} = (ymm0 * ymm1) - ymm2
; X64-NEXT: vmovapd %ymm2, %ymm0
; X64-NEXT: retq
entry:
@@ -5575,14 +5575,14 @@ define <4 x float> @test_mm_mask3_fmsub_ps(<4 x float> %__A, <4 x float> %__B, <
; X86: # %bb.0: # %entry
; X86-NEXT: movb {{[0-9]+}}(%esp), %al
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfmsub231ps {{.*#+}} xmm2 = (xmm0 * xmm1) - xmm2
+; X86-NEXT: vfmsub231ps {{.*#+}} xmm2 {%k1} = (xmm0 * xmm1) - xmm2
; X86-NEXT: vmovaps %xmm2, %xmm0
; X86-NEXT: retl
;
; X64-LABEL: test_mm_mask3_fmsub_ps:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfmsub231ps {{.*#+}} xmm2 = (xmm0 * xmm1) - xmm2
+; X64-NEXT: vfmsub231ps {{.*#+}} xmm2 {%k1} = (xmm0 * xmm1) - xmm2
; X64-NEXT: vmovaps %xmm2, %xmm0
; X64-NEXT: retq
entry:
@@ -5599,14 +5599,14 @@ define <8 x float> @test_mm256_mask3_fmsub_ps(<8 x float> %__A, <8 x float> %__B
; X86: # %bb.0: # %entry
; X86-NEXT: movb {{[0-9]+}}(%esp), %al
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfmsub231ps {{.*#+}} ymm2 = (ymm0 * ymm1) - ymm2
+; X86-NEXT: vfmsub231ps {{.*#+}} ymm2 {%k1} = (ymm0 * ymm1) - ymm2
; X86-NEXT: vmovaps %ymm2, %ymm0
; X86-NEXT: retl
;
; X64-LABEL: test_mm256_mask3_fmsub_ps:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfmsub231ps {{.*#+}} ymm2 = (ymm0 * ymm1) - ymm2
+; X64-NEXT: vfmsub231ps {{.*#+}} ymm2 {%k1} = (ymm0 * ymm1) - ymm2
; X64-NEXT: vmovaps %ymm2, %ymm0
; X64-NEXT: retq
entry:
@@ -5622,14 +5622,14 @@ define <2 x double> @test_mm_mask3_fmsubadd_pd(<2 x double> %__A, <2 x double> %
; X86: # %bb.0: # %entry
; X86-NEXT: movb {{[0-9]+}}(%esp), %al
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfmsubadd231pd {{.*#+}} xmm2 = (xmm0 * xmm1) -/+ xmm2
+; X86-NEXT: vfmsubadd231pd {{.*#+}} xmm2 {%k1} = (xmm0 * xmm1) -/+ xmm2
; X86-NEXT: vmovapd %xmm2, %xmm0
; X86-NEXT: retl
;
; X64-LABEL: test_mm_mask3_fmsubadd_pd:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfmsubadd231pd {{.*#+}} xmm2 = (xmm0 * xmm1) -/+ xmm2
+; X64-NEXT: vfmsubadd231pd {{.*#+}} xmm2 {%k1} = (xmm0 * xmm1) -/+ xmm2
; X64-NEXT: vmovapd %xmm2, %xmm0
; X64-NEXT: retq
entry:
@@ -5648,14 +5648,14 @@ define <4 x double> @test_mm256_mask3_fmsubadd_pd(<4 x double> %__A, <4 x double
; X86: # %bb.0: # %entry
; X86-NEXT: movb {{[0-9]+}}(%esp), %al
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfmsubadd231pd {{.*#+}} ymm2 = (ymm0 * ymm1) -/+ ymm2
+; X86-NEXT: vfmsubadd231pd {{.*#+}} ymm2 {%k1} = (ymm0 * ymm1) -/+ ymm2
; X86-NEXT: vmovapd %ymm2, %ymm0
; X86-NEXT: retl
;
; X64-LABEL: test_mm256_mask3_fmsubadd_pd:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfmsubadd231pd {{.*#+}} ymm2 = (ymm0 * ymm1) -/+ ymm2
+; X64-NEXT: vfmsubadd231pd {{.*#+}} ymm2 {%k1} = (ymm0 * ymm1) -/+ ymm2
; X64-NEXT: vmovapd %ymm2, %ymm0
; X64-NEXT: retq
entry:
@@ -5674,14 +5674,14 @@ define <4 x float> @test_mm_mask3_fmsubadd_ps(<4 x float> %__A, <4 x float> %__B
; X86: # %bb.0: # %entry
; X86-NEXT: movb {{[0-9]+}}(%esp), %al
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfmsubadd231ps {{.*#+}} xmm2 = (xmm0 * xmm1) -/+ xmm2
+; X86-NEXT: vfmsubadd231ps {{.*#+}} xmm2 {%k1} = (xmm0 * xmm1) -/+ xmm2
; X86-NEXT: vmovaps %xmm2, %xmm0
; X86-NEXT: retl
;
; X64-LABEL: test_mm_mask3_fmsubadd_ps:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfmsubadd231ps {{.*#+}} xmm2 = (xmm0 * xmm1) -/+ xmm2
+; X64-NEXT: vfmsubadd231ps {{.*#+}} xmm2 {%k1} = (xmm0 * xmm1) -/+ xmm2
; X64-NEXT: vmovaps %xmm2, %xmm0
; X64-NEXT: retq
entry:
@@ -5700,14 +5700,14 @@ define <8 x float> @test_mm256_mask3_fmsubadd_ps(<8 x float> %__A, <8 x float> %
; X86: # %bb.0: # %entry
; X86-NEXT: movb {{[0-9]+}}(%esp), %al
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfmsubadd231ps {{.*#+}} ymm2 = (ymm0 * ymm1) -/+ ymm2
+; X86-NEXT: vfmsubadd231ps {{.*#+}} ymm2 {%k1} = (ymm0 * ymm1) -/+ ymm2
; X86-NEXT: vmovaps %ymm2, %ymm0
; X86-NEXT: retl
;
; X64-LABEL: test_mm256_mask3_fmsubadd_ps:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfmsubadd231ps {{.*#+}} ymm2 = (ymm0 * ymm1) -/+ ymm2
+; X64-NEXT: vfmsubadd231ps {{.*#+}} ymm2 {%k1} = (ymm0 * ymm1) -/+ ymm2
; X64-NEXT: vmovaps %ymm2, %ymm0
; X64-NEXT: retq
entry:
@@ -5725,13 +5725,13 @@ define <2 x double> @test_mm_mask_fnmadd_pd(<2 x double> %__A, i8 zeroext %__U,
; X86: # %bb.0: # %entry
; X86-NEXT: movb {{[0-9]+}}(%esp), %al
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfnmadd132pd {{.*#+}} xmm0 = -(xmm0 * xmm1) + xmm2
+; X86-NEXT: vfnmadd132pd {{.*#+}} xmm0 {%k1} = -(xmm0 * xmm1) + xmm2
; X86-NEXT: retl
;
; X64-LABEL: test_mm_mask_fnmadd_pd:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfnmadd132pd {{.*#+}} xmm0 = -(xmm0 * xmm1) + xmm2
+; X64-NEXT: vfnmadd132pd {{.*#+}} xmm0 {%k1} = -(xmm0 * xmm1) + xmm2
; X64-NEXT: retq
entry:
%sub.i = fsub <2 x double> <double -0.000000e+00, double -0.000000e+00>, %__B
@@ -5747,13 +5747,13 @@ define <4 x double> @test_mm256_mask_fnmadd_pd(<4 x double> %__A, i8 zeroext %__
; X86: # %bb.0: # %entry
; X86-NEXT: movb {{[0-9]+}}(%esp), %al
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfnmadd132pd {{.*#+}} ymm0 = -(ymm0 * ymm1) + ymm2
+; X86-NEXT: vfnmadd132pd {{.*#+}} ymm0 {%k1} = -(ymm0 * ymm1) + ymm2
; X86-NEXT: retl
;
; X64-LABEL: test_mm256_mask_fnmadd_pd:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfnmadd132pd {{.*#+}} ymm0 = -(ymm0 * ymm1) + ymm2
+; X64-NEXT: vfnmadd132pd {{.*#+}} ymm0 {%k1} = -(ymm0 * ymm1) + ymm2
; X64-NEXT: retq
entry:
%sub.i = fsub <4 x double> <double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00>, %__B
@@ -5769,13 +5769,13 @@ define <4 x float> @test_mm_mask_fnmadd_ps(<4 x float> %__A, i8 zeroext %__U, <4
; X86: # %bb.0: # %entry
; X86-NEXT: movb {{[0-9]+}}(%esp), %al
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfnmadd132ps {{.*#+}} xmm0 = -(xmm0 * xmm1) + xmm2
+; X86-NEXT: vfnmadd132ps {{.*#+}} xmm0 {%k1} = -(xmm0 * xmm1) + xmm2
; X86-NEXT: retl
;
; X64-LABEL: test_mm_mask_fnmadd_ps:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfnmadd132ps {{.*#+}} xmm0 = -(xmm0 * xmm1) + xmm2
+; X64-NEXT: vfnmadd132ps {{.*#+}} xmm0 {%k1} = -(xmm0 * xmm1) + xmm2
; X64-NEXT: retq
entry:
%sub.i = fsub <4 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %__B
@@ -5791,13 +5791,13 @@ define <8 x float> @test_mm256_mask_fnmadd_ps(<8 x float> %__A, i8 zeroext %__U,
; X86: # %bb.0: # %entry
; X86-NEXT: movb {{[0-9]+}}(%esp), %al
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfnmadd132ps {{.*#+}} ymm0 = -(ymm0 * ymm1) + ymm2
+; X86-NEXT: vfnmadd132ps {{.*#+}} ymm0 {%k1} = -(ymm0 * ymm1) + ymm2
; X86-NEXT: retl
;
; X64-LABEL: test_mm256_mask_fnmadd_ps:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfnmadd132ps {{.*#+}} ymm0 = -(ymm0 * ymm1) + ymm2
+; X64-NEXT: vfnmadd132ps {{.*#+}} ymm0 {%k1} = -(ymm0 * ymm1) + ymm2
; X64-NEXT: retq
entry:
%sub.i = fsub <8 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %__B
@@ -5812,13 +5812,13 @@ define <2 x double> @test_mm_mask_fnmsub_pd(<2 x double> %__A, i8 zeroext %__U,
; X86: # %bb.0: # %entry
; X86-NEXT: movb {{[0-9]+}}(%esp), %al
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfnmsub132pd {{.*#+}} xmm0 = -(xmm0 * xmm1) - xmm2
+; X86-NEXT: vfnmsub132pd {{.*#+}} xmm0 {%k1} = -(xmm0 * xmm1) - xmm2
; X86-NEXT: retl
;
; X64-LABEL: test_mm_mask_fnmsub_pd:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfnmsub132pd {{.*#+}} xmm0 = -(xmm0 * xmm1) - xmm2
+; X64-NEXT: vfnmsub132pd {{.*#+}} xmm0 {%k1} = -(xmm0 * xmm1) - xmm2
; X64-NEXT: retq
entry:
%sub.i = fsub <2 x double> <double -0.000000e+00, double -0.000000e+00>, %__B
@@ -5835,14 +5835,14 @@ define <2 x double> @test_mm_mask3_fnmsub_pd(<2 x double> %__A, <2 x double> %__
; X86: # %bb.0: # %entry
; X86-NEXT: movb {{[0-9]+}}(%esp), %al
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfnmsub231pd {{.*#+}} xmm2 = -(xmm0 * xmm1) - xmm2
+; X86-NEXT: vfnmsub231pd {{.*#+}} xmm2 {%k1} = -(xmm0 * xmm1) - xmm2
; X86-NEXT: vmovapd %xmm2, %xmm0
; X86-NEXT: retl
;
; X64-LABEL: test_mm_mask3_fnmsub_pd:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfnmsub231pd {{.*#+}} xmm2 = -(xmm0 * xmm1) - xmm2
+; X64-NEXT: vfnmsub231pd {{.*#+}} xmm2 {%k1} = -(xmm0 * xmm1) - xmm2
; X64-NEXT: vmovapd %xmm2, %xmm0
; X64-NEXT: retq
entry:
@@ -5860,13 +5860,13 @@ define <4 x double> @test_mm256_mask_fnmsub_pd(<4 x double> %__A, i8 zeroext %__
; X86: # %bb.0: # %entry
; X86-NEXT: movb {{[0-9]+}}(%esp), %al
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfnmsub132pd {{.*#+}} ymm0 = -(ymm0 * ymm1) - ymm2
+; X86-NEXT: vfnmsub132pd {{.*#+}} ymm0 {%k1} = -(ymm0 * ymm1) - ymm2
; X86-NEXT: retl
;
; X64-LABEL: test_mm256_mask_fnmsub_pd:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfnmsub132pd {{.*#+}} ymm0 = -(ymm0 * ymm1) - ymm2
+; X64-NEXT: vfnmsub132pd {{.*#+}} ymm0 {%k1} = -(ymm0 * ymm1) - ymm2
; X64-NEXT: retq
entry:
%sub.i = fsub <4 x double> <double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00>, %__B
@@ -5883,14 +5883,14 @@ define <4 x double> @test_mm256_mask3_fnmsub_pd(<4 x double> %__A, <4 x double>
; X86: # %bb.0: # %entry
; X86-NEXT: movb {{[0-9]+}}(%esp), %al
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfnmsub231pd {{.*#+}} ymm2 = -(ymm0 * ymm1) - ymm2
+; X86-NEXT: vfnmsub231pd {{.*#+}} ymm2 {%k1} = -(ymm0 * ymm1) - ymm2
; X86-NEXT: vmovapd %ymm2, %ymm0
; X86-NEXT: retl
;
; X64-LABEL: test_mm256_mask3_fnmsub_pd:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfnmsub231pd {{.*#+}} ymm2 = -(ymm0 * ymm1) - ymm2
+; X64-NEXT: vfnmsub231pd {{.*#+}} ymm2 {%k1} = -(ymm0 * ymm1) - ymm2
; X64-NEXT: vmovapd %ymm2, %ymm0
; X64-NEXT: retq
entry:
@@ -5908,13 +5908,13 @@ define <4 x float> @test_mm_mask_fnmsub_ps(<4 x float> %__A, i8 zeroext %__U, <4
; X86: # %bb.0: # %entry
; X86-NEXT: movb {{[0-9]+}}(%esp), %al
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfnmsub132ps {{.*#+}} xmm0 = -(xmm0 * xmm1) - xmm2
+; X86-NEXT: vfnmsub132ps {{.*#+}} xmm0 {%k1} = -(xmm0 * xmm1) - xmm2
; X86-NEXT: retl
;
; X64-LABEL: test_mm_mask_fnmsub_ps:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfnmsub132ps {{.*#+}} xmm0 = -(xmm0 * xmm1) - xmm2
+; X64-NEXT: vfnmsub132ps {{.*#+}} xmm0 {%k1} = -(xmm0 * xmm1) - xmm2
; X64-NEXT: retq
entry:
%sub.i = fsub <4 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %__B
@@ -5931,14 +5931,14 @@ define <4 x float> @test_mm_mask3_fnmsub_ps(<4 x float> %__A, <4 x float> %__B,
; X86: # %bb.0: # %entry
; X86-NEXT: movb {{[0-9]+}}(%esp), %al
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfnmsub231ps {{.*#+}} xmm2 = -(xmm0 * xmm1) - xmm2
+; X86-NEXT: vfnmsub231ps {{.*#+}} xmm2 {%k1} = -(xmm0 * xmm1) - xmm2
; X86-NEXT: vmovaps %xmm2, %xmm0
; X86-NEXT: retl
;
; X64-LABEL: test_mm_mask3_fnmsub_ps:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfnmsub231ps {{.*#+}} xmm2 = -(xmm0 * xmm1) - xmm2
+; X64-NEXT: vfnmsub231ps {{.*#+}} xmm2 {%k1} = -(xmm0 * xmm1) - xmm2
; X64-NEXT: vmovaps %xmm2, %xmm0
; X64-NEXT: retq
entry:
@@ -5956,13 +5956,13 @@ define <8 x float> @test_mm256_mask_fnmsub_ps(<8 x float> %__A, i8 zeroext %__U,
; X86: # %bb.0: # %entry
; X86-NEXT: movb {{[0-9]+}}(%esp), %al
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfnmsub132ps {{.*#+}} ymm0 = -(ymm0 * ymm1) - ymm2
+; X86-NEXT: vfnmsub132ps {{.*#+}} ymm0 {%k1} = -(ymm0 * ymm1) - ymm2
; X86-NEXT: retl
;
; X64-LABEL: test_mm256_mask_fnmsub_ps:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfnmsub132ps {{.*#+}} ymm0 = -(ymm0 * ymm1) - ymm2
+; X64-NEXT: vfnmsub132ps {{.*#+}} ymm0 {%k1} = -(ymm0 * ymm1) - ymm2
; X64-NEXT: retq
entry:
%sub.i = fsub <8 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %__B
@@ -5978,14 +5978,14 @@ define <8 x float> @test_mm256_mask3_fnmsub_ps(<8 x float> %__A, <8 x float> %__
; X86: # %bb.0: # %entry
; X86-NEXT: movb {{[0-9]+}}(%esp), %al
; X86-NEXT: kmovw %eax, %k1
-; X86-NEXT: vfnmsub231ps {{.*#+}} ymm2 = -(ymm0 * ymm1) - ymm2
+; X86-NEXT: vfnmsub231ps {{.*#+}} ymm2 {%k1} = -(ymm0 * ymm1) - ymm2
; X86-NEXT: vmovaps %ymm2, %ymm0
; X86-NEXT: retl
;
; X64-LABEL: test_mm256_mask3_fnmsub_ps:
; X64: # %bb.0: # %entry
; X64-NEXT: kmovw %edi, %k1
-; X64-NEXT: vfnmsub231ps {{.*#+}} ymm2 = -(ymm0 * ymm1) - ymm2
+; X64-NEXT: vfnmsub231ps {{.*#+}} ymm2 {%k1} = -(ymm0 * ymm1) - ymm2
; X64-NEXT: vmovaps %ymm2, %ymm0
; X64-NEXT: retq
entry:
diff --git a/llvm/test/CodeGen/X86/avx512vl-intrinsics-upgrade.ll b/llvm/test/CodeGen/X86/avx512vl-intrinsics-upgrade.ll
index 352a7d6722d1..6e1b554b90bd 100644
--- a/llvm/test/CodeGen/X86/avx512vl-intrinsics-upgrade.ll
+++ b/llvm/test/CodeGen/X86/avx512vl-intrinsics-upgrade.ll
@@ -15172,14 +15172,14 @@ define <8 x float> @test_mask_vfmadd256_ps(<8 x float> %a0, <8 x float> %a1, <8
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmadd132ps %ymm1, %ymm2, %ymm0 {%k1} # encoding: [0x62,0xf2,0x6d,0x29,0x98,0xc1]
-; X86-NEXT: # ymm0 = (ymm0 * ymm1) + ymm2
+; X86-NEXT: # ymm0 {%k1} = (ymm0 * ymm1) + ymm2
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_mask_vfmadd256_ps:
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmadd132ps %ymm1, %ymm2, %ymm0 {%k1} # encoding: [0x62,0xf2,0x6d,0x29,0x98,0xc1]
-; X64-NEXT: # ymm0 = (ymm0 * ymm1) + ymm2
+; X64-NEXT: # ymm0 {%k1} = (ymm0 * ymm1) + ymm2
; X64-NEXT: retq # encoding: [0xc3]
%res = call <8 x float> @llvm.x86.avx512.mask.vfmadd.ps.256(<8 x float> %a0, <8 x float> %a1, <8 x float> %a2, i8 %mask) nounwind
ret <8 x float> %res
@@ -15203,14 +15203,14 @@ define <4 x float> @test_mask_vfmadd128_ps(<4 x float> %a0, <4 x float> %a1, <4
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmadd132ps %xmm1, %xmm2, %xmm0 {%k1} # encoding: [0x62,0xf2,0x6d,0x09,0x98,0xc1]
-; X86-NEXT: # xmm0 = (xmm0 * xmm1) + xmm2
+; X86-NEXT: # xmm0 {%k1} = (xmm0 * xmm1) + xmm2
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_mask_vfmadd128_ps:
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmadd132ps %xmm1, %xmm2, %xmm0 {%k1} # encoding: [0x62,0xf2,0x6d,0x09,0x98,0xc1]
-; X64-NEXT: # xmm0 = (xmm0 * xmm1) + xmm2
+; X64-NEXT: # xmm0 {%k1} = (xmm0 * xmm1) + xmm2
; X64-NEXT: retq # encoding: [0xc3]
%res = call <4 x float> @llvm.x86.avx512.mask.vfmadd.ps.128(<4 x float> %a0, <4 x float> %a1, <4 x float> %a2, i8 %mask) nounwind
ret <4 x float> %res
@@ -15234,14 +15234,14 @@ define <4 x double> @test_mask_fmadd256_pd(<4 x double> %a, <4 x double> %b, <4
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmadd132pd %ymm1, %ymm2, %ymm0 {%k1} # encoding: [0x62,0xf2,0xed,0x29,0x98,0xc1]
-; X86-NEXT: # ymm0 = (ymm0 * ymm1) + ymm2
+; X86-NEXT: # ymm0 {%k1} = (ymm0 * ymm1) + ymm2
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_mask_fmadd256_pd:
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmadd132pd %ymm1, %ymm2, %ymm0 {%k1} # encoding: [0x62,0xf2,0xed,0x29,0x98,0xc1]
-; X64-NEXT: # ymm0 = (ymm0 * ymm1) + ymm2
+; X64-NEXT: # ymm0 {%k1} = (ymm0 * ymm1) + ymm2
; X64-NEXT: retq # encoding: [0xc3]
%res = call <4 x double> @llvm.x86.avx512.mask.vfmadd.pd.256(<4 x double> %a, <4 x double> %b, <4 x double> %c, i8 %mask)
ret <4 x double> %res
@@ -15265,14 +15265,14 @@ define <2 x double> @test_mask_fmadd128_pd(<2 x double> %a, <2 x double> %b, <2
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmadd132pd %xmm1, %xmm2, %xmm0 {%k1} # encoding: [0x62,0xf2,0xed,0x09,0x98,0xc1]
-; X86-NEXT: # xmm0 = (xmm0 * xmm1) + xmm2
+; X86-NEXT: # xmm0 {%k1} = (xmm0 * xmm1) + xmm2
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_mask_fmadd128_pd:
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmadd132pd %xmm1, %xmm2, %xmm0 {%k1} # encoding: [0x62,0xf2,0xed,0x09,0x98,0xc1]
-; X64-NEXT: # xmm0 = (xmm0 * xmm1) + xmm2
+; X64-NEXT: # xmm0 {%k1} = (xmm0 * xmm1) + xmm2
; X64-NEXT: retq # encoding: [0xc3]
%res = call <2 x double> @llvm.x86.avx512.mask.vfmadd.pd.128(<2 x double> %a, <2 x double> %b, <2 x double> %c, i8 %mask)
ret <2 x double> %res
@@ -15286,7 +15286,7 @@ define <2 x double>@test_int_x86_avx512_mask3_vfmadd_pd_128(<2 x double> %x0, <2
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmadd231pd %xmm1, %xmm0, %xmm2 {%k1} # encoding: [0x62,0xf2,0xfd,0x09,0xb8,0xd1]
-; X86-NEXT: # xmm2 = (xmm0 * xmm1) + xmm2
+; X86-NEXT: # xmm2 {%k1} = (xmm0 * xmm1) + xmm2
; X86-NEXT: vmovapd %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x28,0xc2]
; X86-NEXT: retl # encoding: [0xc3]
;
@@ -15294,7 +15294,7 @@ define <2 x double>@test_int_x86_avx512_mask3_vfmadd_pd_128(<2 x double> %x0, <2
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmadd231pd %xmm1, %xmm0, %xmm2 {%k1} # encoding: [0x62,0xf2,0xfd,0x09,0xb8,0xd1]
-; X64-NEXT: # xmm2 = (xmm0 * xmm1) + xmm2
+; X64-NEXT: # xmm2 {%k1} = (xmm0 * xmm1) + xmm2
; X64-NEXT: vmovapd %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x28,0xc2]
; X64-NEXT: retq # encoding: [0xc3]
%res = call <2 x double> @llvm.x86.avx512.mask3.vfmadd.pd.128(<2 x double> %x0, <2 x double> %x1, <2 x double> %x2, i8 %x3)
@@ -15309,14 +15309,14 @@ define <2 x double>@test_int_x86_avx512_maskz_vfmadd_pd_128(<2 x double> %x0, <2
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmadd213pd %xmm2, %xmm1, %xmm0 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0x89,0xa8,0xc2]
-; X86-NEXT: # xmm0 = (xmm1 * xmm0) + xmm2
+; X86-NEXT: # xmm0 {%k1} {z} = (xmm1 * xmm0) + xmm2
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_maskz_vfmadd_pd_128:
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmadd213pd %xmm2, %xmm1, %xmm0 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0x89,0xa8,0xc2]
-; X64-NEXT: # xmm0 = (xmm1 * xmm0) + xmm2
+; X64-NEXT: # xmm0 {%k1} {z} = (xmm1 * xmm0) + xmm2
; X64-NEXT: retq # encoding: [0xc3]
%res = call <2 x double> @llvm.x86.avx512.maskz.vfmadd.pd.128(<2 x double> %x0, <2 x double> %x1, <2 x double> %x2, i8 %x3)
ret <2 x double> %res
@@ -15330,7 +15330,7 @@ define <4 x double>@test_int_x86_avx512_mask3_vfmadd_pd_256(<4 x double> %x0, <4
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmadd231pd %ymm1, %ymm0, %ymm2 {%k1} # encoding: [0x62,0xf2,0xfd,0x29,0xb8,0xd1]
-; X86-NEXT: # ymm2 = (ymm0 * ymm1) + ymm2
+; X86-NEXT: # ymm2 {%k1} = (ymm0 * ymm1) + ymm2
; X86-NEXT: vmovapd %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x28,0xc2]
; X86-NEXT: retl # encoding: [0xc3]
;
@@ -15338,7 +15338,7 @@ define <4 x double>@test_int_x86_avx512_mask3_vfmadd_pd_256(<4 x double> %x0, <4
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmadd231pd %ymm1, %ymm0, %ymm2 {%k1} # encoding: [0x62,0xf2,0xfd,0x29,0xb8,0xd1]
-; X64-NEXT: # ymm2 = (ymm0 * ymm1) + ymm2
+; X64-NEXT: # ymm2 {%k1} = (ymm0 * ymm1) + ymm2
; X64-NEXT: vmovapd %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x28,0xc2]
; X64-NEXT: retq # encoding: [0xc3]
%res = call <4 x double> @llvm.x86.avx512.mask3.vfmadd.pd.256(<4 x double> %x0, <4 x double> %x1, <4 x double> %x2, i8 %x3)
@@ -15353,14 +15353,14 @@ define <4 x double>@test_int_x86_avx512_maskz_vfmadd_pd_256(<4 x double> %x0, <4
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmadd213pd %ymm2, %ymm1, %ymm0 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0xa9,0xa8,0xc2]
-; X86-NEXT: # ymm0 = (ymm1 * ymm0) + ymm2
+; X86-NEXT: # ymm0 {%k1} {z} = (ymm1 * ymm0) + ymm2
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_maskz_vfmadd_pd_256:
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmadd213pd %ymm2, %ymm1, %ymm0 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0xa9,0xa8,0xc2]
-; X64-NEXT: # ymm0 = (ymm1 * ymm0) + ymm2
+; X64-NEXT: # ymm0 {%k1} {z} = (ymm1 * ymm0) + ymm2
; X64-NEXT: retq # encoding: [0xc3]
%res = call <4 x double> @llvm.x86.avx512.maskz.vfmadd.pd.256(<4 x double> %x0, <4 x double> %x1, <4 x double> %x2, i8 %x3)
ret <4 x double> %res
@@ -15374,7 +15374,7 @@ define <4 x float>@test_int_x86_avx512_mask3_vfmadd_ps_128(<4 x float> %x0, <4 x
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmadd231ps %xmm1, %xmm0, %xmm2 {%k1} # encoding: [0x62,0xf2,0x7d,0x09,0xb8,0xd1]
-; X86-NEXT: # xmm2 = (xmm0 * xmm1) + xmm2
+; X86-NEXT: # xmm2 {%k1} = (xmm0 * xmm1) + xmm2
; X86-NEXT: vmovaps %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf8,0x28,0xc2]
; X86-NEXT: retl # encoding: [0xc3]
;
@@ -15382,7 +15382,7 @@ define <4 x float>@test_int_x86_avx512_mask3_vfmadd_ps_128(<4 x float> %x0, <4 x
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmadd231ps %xmm1, %xmm0, %xmm2 {%k1} # encoding: [0x62,0xf2,0x7d,0x09,0xb8,0xd1]
-; X64-NEXT: # xmm2 = (xmm0 * xmm1) + xmm2
+; X64-NEXT: # xmm2 {%k1} = (xmm0 * xmm1) + xmm2
; X64-NEXT: vmovaps %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf8,0x28,0xc2]
; X64-NEXT: retq # encoding: [0xc3]
%res = call <4 x float> @llvm.x86.avx512.mask3.vfmadd.ps.128(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2, i8 %x3)
@@ -15397,14 +15397,14 @@ define <4 x float>@test_int_x86_avx512_maskz_vfmadd_ps_128(<4 x float> %x0, <4 x
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmadd213ps %xmm2, %xmm1, %xmm0 {%k1} {z} # encoding: [0x62,0xf2,0x75,0x89,0xa8,0xc2]
-; X86-NEXT: # xmm0 = (xmm1 * xmm0) + xmm2
+; X86-NEXT: # xmm0 {%k1} {z} = (xmm1 * xmm0) + xmm2
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_maskz_vfmadd_ps_128:
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmadd213ps %xmm2, %xmm1, %xmm0 {%k1} {z} # encoding: [0x62,0xf2,0x75,0x89,0xa8,0xc2]
-; X64-NEXT: # xmm0 = (xmm1 * xmm0) + xmm2
+; X64-NEXT: # xmm0 {%k1} {z} = (xmm1 * xmm0) + xmm2
; X64-NEXT: retq # encoding: [0xc3]
%res = call <4 x float> @llvm.x86.avx512.maskz.vfmadd.ps.128(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2, i8 %x3)
ret <4 x float> %res
@@ -15418,7 +15418,7 @@ define <8 x float>@test_int_x86_avx512_mask3_vfmadd_ps_256(<8 x float> %x0, <8 x
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmadd231ps %ymm1, %ymm0, %ymm2 {%k1} # encoding: [0x62,0xf2,0x7d,0x29,0xb8,0xd1]
-; X86-NEXT: # ymm2 = (ymm0 * ymm1) + ymm2
+; X86-NEXT: # ymm2 {%k1} = (ymm0 * ymm1) + ymm2
; X86-NEXT: vmovaps %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xfc,0x28,0xc2]
; X86-NEXT: retl # encoding: [0xc3]
;
@@ -15426,7 +15426,7 @@ define <8 x float>@test_int_x86_avx512_mask3_vfmadd_ps_256(<8 x float> %x0, <8 x
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmadd231ps %ymm1, %ymm0, %ymm2 {%k1} # encoding: [0x62,0xf2,0x7d,0x29,0xb8,0xd1]
-; X64-NEXT: # ymm2 = (ymm0 * ymm1) + ymm2
+; X64-NEXT: # ymm2 {%k1} = (ymm0 * ymm1) + ymm2
; X64-NEXT: vmovaps %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xfc,0x28,0xc2]
; X64-NEXT: retq # encoding: [0xc3]
%res = call <8 x float> @llvm.x86.avx512.mask3.vfmadd.ps.256(<8 x float> %x0, <8 x float> %x1, <8 x float> %x2, i8 %x3)
@@ -15441,14 +15441,14 @@ define <8 x float>@test_int_x86_avx512_maskz_vfmadd_ps_256(<8 x float> %x0, <8 x
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmadd213ps %ymm2, %ymm1, %ymm0 {%k1} {z} # encoding: [0x62,0xf2,0x75,0xa9,0xa8,0xc2]
-; X86-NEXT: # ymm0 = (ymm1 * ymm0) + ymm2
+; X86-NEXT: # ymm0 {%k1} {z} = (ymm1 * ymm0) + ymm2
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_maskz_vfmadd_ps_256:
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmadd213ps %ymm2, %ymm1, %ymm0 {%k1} {z} # encoding: [0x62,0xf2,0x75,0xa9,0xa8,0xc2]
-; X64-NEXT: # ymm0 = (ymm1 * ymm0) + ymm2
+; X64-NEXT: # ymm0 {%k1} {z} = (ymm1 * ymm0) + ymm2
; X64-NEXT: retq # encoding: [0xc3]
%res = call <8 x float> @llvm.x86.avx512.maskz.vfmadd.ps.256(<8 x float> %x0, <8 x float> %x1, <8 x float> %x2, i8 %x3)
ret <8 x float> %res
@@ -15463,7 +15463,7 @@ define <2 x double>@test_int_x86_avx512_mask3_vfmsub_pd_128(<2 x double> %x0, <2
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmsub231pd %xmm1, %xmm0, %xmm2 {%k1} # encoding: [0x62,0xf2,0xfd,0x09,0xba,0xd1]
-; X86-NEXT: # xmm2 = (xmm0 * xmm1) - xmm2
+; X86-NEXT: # xmm2 {%k1} = (xmm0 * xmm1) - xmm2
; X86-NEXT: vmovapd %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x28,0xc2]
; X86-NEXT: retl # encoding: [0xc3]
;
@@ -15471,7 +15471,7 @@ define <2 x double>@test_int_x86_avx512_mask3_vfmsub_pd_128(<2 x double> %x0, <2
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmsub231pd %xmm1, %xmm0, %xmm2 {%k1} # encoding: [0x62,0xf2,0xfd,0x09,0xba,0xd1]
-; X64-NEXT: # xmm2 = (xmm0 * xmm1) - xmm2
+; X64-NEXT: # xmm2 {%k1} = (xmm0 * xmm1) - xmm2
; X64-NEXT: vmovapd %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x28,0xc2]
; X64-NEXT: retq # encoding: [0xc3]
%res = call <2 x double> @llvm.x86.avx512.mask3.vfmsub.pd.128(<2 x double> %x0, <2 x double> %x1, <2 x double> %x2, i8 %x3)
@@ -15487,7 +15487,7 @@ define <4 x double>@test_int_x86_avx512_mask3_vfmsub_pd_256(<4 x double> %x0, <4
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmsub231pd %ymm1, %ymm0, %ymm2 {%k1} # encoding: [0x62,0xf2,0xfd,0x29,0xba,0xd1]
-; X86-NEXT: # ymm2 = (ymm0 * ymm1) - ymm2
+; X86-NEXT: # ymm2 {%k1} = (ymm0 * ymm1) - ymm2
; X86-NEXT: vmovapd %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x28,0xc2]
; X86-NEXT: retl # encoding: [0xc3]
;
@@ -15495,7 +15495,7 @@ define <4 x double>@test_int_x86_avx512_mask3_vfmsub_pd_256(<4 x double> %x0, <4
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmsub231pd %ymm1, %ymm0, %ymm2 {%k1} # encoding: [0x62,0xf2,0xfd,0x29,0xba,0xd1]
-; X64-NEXT: # ymm2 = (ymm0 * ymm1) - ymm2
+; X64-NEXT: # ymm2 {%k1} = (ymm0 * ymm1) - ymm2
; X64-NEXT: vmovapd %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x28,0xc2]
; X64-NEXT: retq # encoding: [0xc3]
%res = call <4 x double> @llvm.x86.avx512.mask3.vfmsub.pd.256(<4 x double> %x0, <4 x double> %x1, <4 x double> %x2, i8 %x3)
@@ -15510,7 +15510,7 @@ define <4 x float>@test_int_x86_avx512_mask3_vfmsub_ps_128(<4 x float> %x0, <4 x
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmsub231ps %xmm1, %xmm0, %xmm2 {%k1} # encoding: [0x62,0xf2,0x7d,0x09,0xba,0xd1]
-; X86-NEXT: # xmm2 = (xmm0 * xmm1) - xmm2
+; X86-NEXT: # xmm2 {%k1} = (xmm0 * xmm1) - xmm2
; X86-NEXT: vmovaps %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf8,0x28,0xc2]
; X86-NEXT: retl # encoding: [0xc3]
;
@@ -15518,7 +15518,7 @@ define <4 x float>@test_int_x86_avx512_mask3_vfmsub_ps_128(<4 x float> %x0, <4 x
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmsub231ps %xmm1, %xmm0, %xmm2 {%k1} # encoding: [0x62,0xf2,0x7d,0x09,0xba,0xd1]
-; X64-NEXT: # xmm2 = (xmm0 * xmm1) - xmm2
+; X64-NEXT: # xmm2 {%k1} = (xmm0 * xmm1) - xmm2
; X64-NEXT: vmovaps %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf8,0x28,0xc2]
; X64-NEXT: retq # encoding: [0xc3]
%res = call <4 x float> @llvm.x86.avx512.mask3.vfmsub.ps.128(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2, i8 %x3)
@@ -15533,7 +15533,7 @@ define <8 x float>@test_int_x86_avx512_mask3_vfmsub_ps_256(<8 x float> %x0, <8 x
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmsub231ps %ymm1, %ymm0, %ymm2 {%k1} # encoding: [0x62,0xf2,0x7d,0x29,0xba,0xd1]
-; X86-NEXT: # ymm2 = (ymm0 * ymm1) - ymm2
+; X86-NEXT: # ymm2 {%k1} = (ymm0 * ymm1) - ymm2
; X86-NEXT: vmovaps %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xfc,0x28,0xc2]
; X86-NEXT: retl # encoding: [0xc3]
;
@@ -15541,7 +15541,7 @@ define <8 x float>@test_int_x86_avx512_mask3_vfmsub_ps_256(<8 x float> %x0, <8 x
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmsub231ps %ymm1, %ymm0, %ymm2 {%k1} # encoding: [0x62,0xf2,0x7d,0x29,0xba,0xd1]
-; X64-NEXT: # ymm2 = (ymm0 * ymm1) - ymm2
+; X64-NEXT: # ymm2 {%k1} = (ymm0 * ymm1) - ymm2
; X64-NEXT: vmovaps %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xfc,0x28,0xc2]
; X64-NEXT: retq # encoding: [0xc3]
%res = call <8 x float> @llvm.x86.avx512.mask3.vfmsub.ps.256(<8 x float> %x0, <8 x float> %x1, <8 x float> %x2, i8 %x3)
@@ -15566,14 +15566,14 @@ define <8 x float> @test_mask_vfnmadd256_ps(<8 x float> %a0, <8 x float> %a1, <8
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfnmadd132ps %ymm1, %ymm2, %ymm0 {%k1} # encoding: [0x62,0xf2,0x6d,0x29,0x9c,0xc1]
-; X86-NEXT: # ymm0 = -(ymm0 * ymm1) + ymm2
+; X86-NEXT: # ymm0 {%k1} = -(ymm0 * ymm1) + ymm2
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_mask_vfnmadd256_ps:
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfnmadd132ps %ymm1, %ymm2, %ymm0 {%k1} # encoding: [0x62,0xf2,0x6d,0x29,0x9c,0xc1]
-; X64-NEXT: # ymm0 = -(ymm0 * ymm1) + ymm2
+; X64-NEXT: # ymm0 {%k1} = -(ymm0 * ymm1) + ymm2
; X64-NEXT: retq # encoding: [0xc3]
%res = call <8 x float> @llvm.x86.avx512.mask.vfnmadd.ps.256(<8 x float> %a0, <8 x float> %a1, <8 x float> %a2, i8 %mask) nounwind
ret <8 x float> %res
@@ -15597,14 +15597,14 @@ define <4 x float> @test_mask_vfnmadd128_ps(<4 x float> %a0, <4 x float> %a1, <4
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfnmadd132ps %xmm1, %xmm2, %xmm0 {%k1} # encoding: [0x62,0xf2,0x6d,0x09,0x9c,0xc1]
-; X86-NEXT: # xmm0 = -(xmm0 * xmm1) + xmm2
+; X86-NEXT: # xmm0 {%k1} = -(xmm0 * xmm1) + xmm2
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_mask_vfnmadd128_ps:
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfnmadd132ps %xmm1, %xmm2, %xmm0 {%k1} # encoding: [0x62,0xf2,0x6d,0x09,0x9c,0xc1]
-; X64-NEXT: # xmm0 = -(xmm0 * xmm1) + xmm2
+; X64-NEXT: # xmm0 {%k1} = -(xmm0 * xmm1) + xmm2
; X64-NEXT: retq # encoding: [0xc3]
%res = call <4 x float> @llvm.x86.avx512.mask.vfnmadd.ps.128(<4 x float> %a0, <4 x float> %a1, <4 x float> %a2, i8 %mask) nounwind
ret <4 x float> %res
@@ -15628,14 +15628,14 @@ define <4 x double> @test_mask_vfnmadd256_pd(<4 x double> %a0, <4 x double> %a1,
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfnmadd132pd %ymm1, %ymm2, %ymm0 {%k1} # encoding: [0x62,0xf2,0xed,0x29,0x9c,0xc1]
-; X86-NEXT: # ymm0 = -(ymm0 * ymm1) + ymm2
+; X86-NEXT: # ymm0 {%k1} = -(ymm0 * ymm1) + ymm2
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_mask_vfnmadd256_pd:
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfnmadd132pd %ymm1, %ymm2, %ymm0 {%k1} # encoding: [0x62,0xf2,0xed,0x29,0x9c,0xc1]
-; X64-NEXT: # ymm0 = -(ymm0 * ymm1) + ymm2
+; X64-NEXT: # ymm0 {%k1} = -(ymm0 * ymm1) + ymm2
; X64-NEXT: retq # encoding: [0xc3]
%res = call <4 x double> @llvm.x86.avx512.mask.vfnmadd.pd.256(<4 x double> %a0, <4 x double> %a1, <4 x double> %a2, i8 %mask) nounwind
ret <4 x double> %res
@@ -15659,14 +15659,14 @@ define <2 x double> @test_mask_vfnmadd128_pd(<2 x double> %a0, <2 x double> %a1,
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfnmadd132pd %xmm1, %xmm2, %xmm0 {%k1} # encoding: [0x62,0xf2,0xed,0x09,0x9c,0xc1]
-; X86-NEXT: # xmm0 = -(xmm0 * xmm1) + xmm2
+; X86-NEXT: # xmm0 {%k1} = -(xmm0 * xmm1) + xmm2
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_mask_vfnmadd128_pd:
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfnmadd132pd %xmm1, %xmm2, %xmm0 {%k1} # encoding: [0x62,0xf2,0xed,0x09,0x9c,0xc1]
-; X64-NEXT: # xmm0 = -(xmm0 * xmm1) + xmm2
+; X64-NEXT: # xmm0 {%k1} = -(xmm0 * xmm1) + xmm2
; X64-NEXT: retq # encoding: [0xc3]
%res = call <2 x double> @llvm.x86.avx512.mask.vfnmadd.pd.128(<2 x double> %a0, <2 x double> %a1, <2 x double> %a2, i8 %mask) nounwind
ret <2 x double> %res
@@ -15690,14 +15690,14 @@ define <8 x float> @test_mask_vfnmsub256_ps(<8 x float> %a0, <8 x float> %a1, <8
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfnmsub132ps %ymm1, %ymm2, %ymm0 {%k1} # encoding: [0x62,0xf2,0x6d,0x29,0x9e,0xc1]
-; X86-NEXT: # ymm0 = -(ymm0 * ymm1) - ymm2
+; X86-NEXT: # ymm0 {%k1} = -(ymm0 * ymm1) - ymm2
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_mask_vfnmsub256_ps:
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfnmsub132ps %ymm1, %ymm2, %ymm0 {%k1} # encoding: [0x62,0xf2,0x6d,0x29,0x9e,0xc1]
-; X64-NEXT: # ymm0 = -(ymm0 * ymm1) - ymm2
+; X64-NEXT: # ymm0 {%k1} = -(ymm0 * ymm1) - ymm2
; X64-NEXT: retq # encoding: [0xc3]
%res = call <8 x float> @llvm.x86.avx512.mask.vfnmsub.ps.256(<8 x float> %a0, <8 x float> %a1, <8 x float> %a2, i8 %mask) nounwind
ret <8 x float> %res
@@ -15721,14 +15721,14 @@ define <4 x float> @test_mask_vfnmsub128_ps(<4 x float> %a0, <4 x float> %a1, <4
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfnmsub132ps %xmm1, %xmm2, %xmm0 {%k1} # encoding: [0x62,0xf2,0x6d,0x09,0x9e,0xc1]
-; X86-NEXT: # xmm0 = -(xmm0 * xmm1) - xmm2
+; X86-NEXT: # xmm0 {%k1} = -(xmm0 * xmm1) - xmm2
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_mask_vfnmsub128_ps:
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfnmsub132ps %xmm1, %xmm2, %xmm0 {%k1} # encoding: [0x62,0xf2,0x6d,0x09,0x9e,0xc1]
-; X64-NEXT: # xmm0 = -(xmm0 * xmm1) - xmm2
+; X64-NEXT: # xmm0 {%k1} = -(xmm0 * xmm1) - xmm2
; X64-NEXT: retq # encoding: [0xc3]
%res = call <4 x float> @llvm.x86.avx512.mask.vfnmsub.ps.128(<4 x float> %a0, <4 x float> %a1, <4 x float> %a2, i8 %mask) nounwind
ret <4 x float> %res
@@ -15752,14 +15752,14 @@ define <4 x double> @test_mask_vfnmsub256_pd(<4 x double> %a0, <4 x double> %a1,
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfnmsub132pd %ymm1, %ymm2, %ymm0 {%k1} # encoding: [0x62,0xf2,0xed,0x29,0x9e,0xc1]
-; X86-NEXT: # ymm0 = -(ymm0 * ymm1) - ymm2
+; X86-NEXT: # ymm0 {%k1} = -(ymm0 * ymm1) - ymm2
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_mask_vfnmsub256_pd:
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfnmsub132pd %ymm1, %ymm2, %ymm0 {%k1} # encoding: [0x62,0xf2,0xed,0x29,0x9e,0xc1]
-; X64-NEXT: # ymm0 = -(ymm0 * ymm1) - ymm2
+; X64-NEXT: # ymm0 {%k1} = -(ymm0 * ymm1) - ymm2
; X64-NEXT: retq # encoding: [0xc3]
%res = call <4 x double> @llvm.x86.avx512.mask.vfnmsub.pd.256(<4 x double> %a0, <4 x double> %a1, <4 x double> %a2, i8 %mask) nounwind
ret <4 x double> %res
@@ -15783,14 +15783,14 @@ define <2 x double> @test_mask_vfnmsub128_pd(<2 x double> %a0, <2 x double> %a1,
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfnmsub132pd %xmm1, %xmm2, %xmm0 {%k1} # encoding: [0x62,0xf2,0xed,0x09,0x9e,0xc1]
-; X86-NEXT: # xmm0 = -(xmm0 * xmm1) - xmm2
+; X86-NEXT: # xmm0 {%k1} = -(xmm0 * xmm1) - xmm2
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_mask_vfnmsub128_pd:
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfnmsub132pd %xmm1, %xmm2, %xmm0 {%k1} # encoding: [0x62,0xf2,0xed,0x09,0x9e,0xc1]
-; X64-NEXT: # xmm0 = -(xmm0 * xmm1) - xmm2
+; X64-NEXT: # xmm0 {%k1} = -(xmm0 * xmm1) - xmm2
; X64-NEXT: retq # encoding: [0xc3]
%res = call <2 x double> @llvm.x86.avx512.mask.vfnmsub.pd.128(<2 x double> %a0, <2 x double> %a1, <2 x double> %a2, i8 %mask) nounwind
ret <2 x double> %res
@@ -15804,7 +15804,7 @@ define <2 x double>@test_int_x86_avx512_mask3_vfnmsub_pd_128(<2 x double> %x0, <
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfnmsub231pd %xmm1, %xmm0, %xmm2 {%k1} # encoding: [0x62,0xf2,0xfd,0x09,0xbe,0xd1]
-; X86-NEXT: # xmm2 = -(xmm0 * xmm1) - xmm2
+; X86-NEXT: # xmm2 {%k1} = -(xmm0 * xmm1) - xmm2
; X86-NEXT: vmovapd %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x28,0xc2]
; X86-NEXT: retl # encoding: [0xc3]
;
@@ -15812,7 +15812,7 @@ define <2 x double>@test_int_x86_avx512_mask3_vfnmsub_pd_128(<2 x double> %x0, <
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfnmsub231pd %xmm1, %xmm0, %xmm2 {%k1} # encoding: [0x62,0xf2,0xfd,0x09,0xbe,0xd1]
-; X64-NEXT: # xmm2 = -(xmm0 * xmm1) - xmm2
+; X64-NEXT: # xmm2 {%k1} = -(xmm0 * xmm1) - xmm2
; X64-NEXT: vmovapd %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x28,0xc2]
; X64-NEXT: retq # encoding: [0xc3]
%res = call <2 x double> @llvm.x86.avx512.mask3.vfnmsub.pd.128(<2 x double> %x0, <2 x double> %x1, <2 x double> %x2, i8 %x3)
@@ -15827,7 +15827,7 @@ define <4 x double>@test_int_x86_avx512_mask3_vfnmsub_pd_256(<4 x double> %x0, <
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfnmsub231pd %ymm1, %ymm0, %ymm2 {%k1} # encoding: [0x62,0xf2,0xfd,0x29,0xbe,0xd1]
-; X86-NEXT: # ymm2 = -(ymm0 * ymm1) - ymm2
+; X86-NEXT: # ymm2 {%k1} = -(ymm0 * ymm1) - ymm2
; X86-NEXT: vmovapd %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x28,0xc2]
; X86-NEXT: retl # encoding: [0xc3]
;
@@ -15835,7 +15835,7 @@ define <4 x double>@test_int_x86_avx512_mask3_vfnmsub_pd_256(<4 x double> %x0, <
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfnmsub231pd %ymm1, %ymm0, %ymm2 {%k1} # encoding: [0x62,0xf2,0xfd,0x29,0xbe,0xd1]
-; X64-NEXT: # ymm2 = -(ymm0 * ymm1) - ymm2
+; X64-NEXT: # ymm2 {%k1} = -(ymm0 * ymm1) - ymm2
; X64-NEXT: vmovapd %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x28,0xc2]
; X64-NEXT: retq # encoding: [0xc3]
%res = call <4 x double> @llvm.x86.avx512.mask3.vfnmsub.pd.256(<4 x double> %x0, <4 x double> %x1, <4 x double> %x2, i8 %x3)
@@ -15850,7 +15850,7 @@ define <4 x float>@test_int_x86_avx512_mask3_vfnmsub_ps_128(<4 x float> %x0, <4
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfnmsub231ps %xmm1, %xmm0, %xmm2 {%k1} # encoding: [0x62,0xf2,0x7d,0x09,0xbe,0xd1]
-; X86-NEXT: # xmm2 = -(xmm0 * xmm1) - xmm2
+; X86-NEXT: # xmm2 {%k1} = -(xmm0 * xmm1) - xmm2
; X86-NEXT: vmovaps %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf8,0x28,0xc2]
; X86-NEXT: retl # encoding: [0xc3]
;
@@ -15858,7 +15858,7 @@ define <4 x float>@test_int_x86_avx512_mask3_vfnmsub_ps_128(<4 x float> %x0, <4
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfnmsub231ps %xmm1, %xmm0, %xmm2 {%k1} # encoding: [0x62,0xf2,0x7d,0x09,0xbe,0xd1]
-; X64-NEXT: # xmm2 = -(xmm0 * xmm1) - xmm2
+; X64-NEXT: # xmm2 {%k1} = -(xmm0 * xmm1) - xmm2
; X64-NEXT: vmovaps %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf8,0x28,0xc2]
; X64-NEXT: retq # encoding: [0xc3]
%res = call <4 x float> @llvm.x86.avx512.mask3.vfnmsub.ps.128(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2, i8 %x3)
@@ -15873,7 +15873,7 @@ define <8 x float>@test_int_x86_avx512_mask3_vfnmsub_ps_256(<8 x float> %x0, <8
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfnmsub231ps %ymm1, %ymm0, %ymm2 {%k1} # encoding: [0x62,0xf2,0x7d,0x29,0xbe,0xd1]
-; X86-NEXT: # ymm2 = -(ymm0 * ymm1) - ymm2
+; X86-NEXT: # ymm2 {%k1} = -(ymm0 * ymm1) - ymm2
; X86-NEXT: vmovaps %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xfc,0x28,0xc2]
; X86-NEXT: retl # encoding: [0xc3]
;
@@ -15881,7 +15881,7 @@ define <8 x float>@test_int_x86_avx512_mask3_vfnmsub_ps_256(<8 x float> %x0, <8
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfnmsub231ps %ymm1, %ymm0, %ymm2 {%k1} # encoding: [0x62,0xf2,0x7d,0x29,0xbe,0xd1]
-; X64-NEXT: # ymm2 = -(ymm0 * ymm1) - ymm2
+; X64-NEXT: # ymm2 {%k1} = -(ymm0 * ymm1) - ymm2
; X64-NEXT: vmovaps %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xfc,0x28,0xc2]
; X64-NEXT: retq # encoding: [0xc3]
%res = call <8 x float> @llvm.x86.avx512.mask3.vfnmsub.ps.256(<8 x float> %x0, <8 x float> %x1, <8 x float> %x2, i8 %x3)
@@ -15906,14 +15906,14 @@ define <8 x float> @test_mask_fmaddsub256_ps(<8 x float> %a, <8 x float> %b, <8
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmaddsub132ps %ymm1, %ymm2, %ymm0 {%k1} # encoding: [0x62,0xf2,0x6d,0x29,0x96,0xc1]
-; X86-NEXT: # ymm0 = (ymm0 * ymm1) +/- ymm2
+; X86-NEXT: # ymm0 {%k1} = (ymm0 * ymm1) +/- ymm2
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_mask_fmaddsub256_ps:
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmaddsub132ps %ymm1, %ymm2, %ymm0 {%k1} # encoding: [0x62,0xf2,0x6d,0x29,0x96,0xc1]
-; X64-NEXT: # ymm0 = (ymm0 * ymm1) +/- ymm2
+; X64-NEXT: # ymm0 {%k1} = (ymm0 * ymm1) +/- ymm2
; X64-NEXT: retq # encoding: [0xc3]
%res = call <8 x float> @llvm.x86.avx512.mask.vfmaddsub.ps.256(<8 x float> %a, <8 x float> %b, <8 x float> %c, i8 %mask)
ret <8 x float> %res
@@ -15937,14 +15937,14 @@ define <4 x float> @test_mask_fmaddsub128_ps(<4 x float> %a, <4 x float> %b, <4
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmaddsub132ps %xmm1, %xmm2, %xmm0 {%k1} # encoding: [0x62,0xf2,0x6d,0x09,0x96,0xc1]
-; X86-NEXT: # xmm0 = (xmm0 * xmm1) +/- xmm2
+; X86-NEXT: # xmm0 {%k1} = (xmm0 * xmm1) +/- xmm2
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_mask_fmaddsub128_ps:
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmaddsub132ps %xmm1, %xmm2, %xmm0 {%k1} # encoding: [0x62,0xf2,0x6d,0x09,0x96,0xc1]
-; X64-NEXT: # xmm0 = (xmm0 * xmm1) +/- xmm2
+; X64-NEXT: # xmm0 {%k1} = (xmm0 * xmm1) +/- xmm2
; X64-NEXT: retq # encoding: [0xc3]
%res = call <4 x float> @llvm.x86.avx512.mask.vfmaddsub.ps.128(<4 x float> %a, <4 x float> %b, <4 x float> %c, i8 %mask)
ret <4 x float> %res
@@ -15968,14 +15968,14 @@ define <4 x double> @test_mask_vfmaddsub256_pd(<4 x double> %a0, <4 x double> %a
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmaddsub132pd %ymm1, %ymm2, %ymm0 {%k1} # encoding: [0x62,0xf2,0xed,0x29,0x96,0xc1]
-; X86-NEXT: # ymm0 = (ymm0 * ymm1) +/- ymm2
+; X86-NEXT: # ymm0 {%k1} = (ymm0 * ymm1) +/- ymm2
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_mask_vfmaddsub256_pd:
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmaddsub132pd %ymm1, %ymm2, %ymm0 {%k1} # encoding: [0x62,0xf2,0xed,0x29,0x96,0xc1]
-; X64-NEXT: # ymm0 = (ymm0 * ymm1) +/- ymm2
+; X64-NEXT: # ymm0 {%k1} = (ymm0 * ymm1) +/- ymm2
; X64-NEXT: retq # encoding: [0xc3]
%res = call <4 x double> @llvm.x86.avx512.mask.vfmaddsub.pd.256(<4 x double> %a0, <4 x double> %a1, <4 x double> %a2, i8 %mask) nounwind
ret <4 x double> %res
@@ -15999,14 +15999,14 @@ define <2 x double> @test_mask_vfmaddsub128_pd(<2 x double> %a0, <2 x double> %a
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmaddsub132pd %xmm1, %xmm2, %xmm0 {%k1} # encoding: [0x62,0xf2,0xed,0x09,0x96,0xc1]
-; X86-NEXT: # xmm0 = (xmm0 * xmm1) +/- xmm2
+; X86-NEXT: # xmm0 {%k1} = (xmm0 * xmm1) +/- xmm2
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_mask_vfmaddsub128_pd:
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmaddsub132pd %xmm1, %xmm2, %xmm0 {%k1} # encoding: [0x62,0xf2,0xed,0x09,0x96,0xc1]
-; X64-NEXT: # xmm0 = (xmm0 * xmm1) +/- xmm2
+; X64-NEXT: # xmm0 {%k1} = (xmm0 * xmm1) +/- xmm2
; X64-NEXT: retq # encoding: [0xc3]
%res = call <2 x double> @llvm.x86.avx512.mask.vfmaddsub.pd.128(<2 x double> %a0, <2 x double> %a1, <2 x double> %a2, i8 %mask) nounwind
ret <2 x double> %res
@@ -16020,7 +16020,7 @@ define <2 x double>@test_int_x86_avx512_mask3_vfmaddsub_pd_128(<2 x double> %x0,
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmaddsub231pd %xmm1, %xmm0, %xmm2 {%k1} # encoding: [0x62,0xf2,0xfd,0x09,0xb6,0xd1]
-; X86-NEXT: # xmm2 = (xmm0 * xmm1) +/- xmm2
+; X86-NEXT: # xmm2 {%k1} = (xmm0 * xmm1) +/- xmm2
; X86-NEXT: vmovapd %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x28,0xc2]
; X86-NEXT: retl # encoding: [0xc3]
;
@@ -16028,7 +16028,7 @@ define <2 x double>@test_int_x86_avx512_mask3_vfmaddsub_pd_128(<2 x double> %x0,
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmaddsub231pd %xmm1, %xmm0, %xmm2 {%k1} # encoding: [0x62,0xf2,0xfd,0x09,0xb6,0xd1]
-; X64-NEXT: # xmm2 = (xmm0 * xmm1) +/- xmm2
+; X64-NEXT: # xmm2 {%k1} = (xmm0 * xmm1) +/- xmm2
; X64-NEXT: vmovapd %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x28,0xc2]
; X64-NEXT: retq # encoding: [0xc3]
%res = call <2 x double> @llvm.x86.avx512.mask3.vfmaddsub.pd.128(<2 x double> %x0, <2 x double> %x1, <2 x double> %x2, i8 %x3)
@@ -16043,14 +16043,14 @@ define <2 x double>@test_int_x86_avx512_maskz_vfmaddsub_pd_128(<2 x double> %x0,
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmaddsub213pd %xmm2, %xmm1, %xmm0 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0x89,0xa6,0xc2]
-; X86-NEXT: # xmm0 = (xmm1 * xmm0) +/- xmm2
+; X86-NEXT: # xmm0 {%k1} {z} = (xmm1 * xmm0) +/- xmm2
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_maskz_vfmaddsub_pd_128:
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmaddsub213pd %xmm2, %xmm1, %xmm0 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0x89,0xa6,0xc2]
-; X64-NEXT: # xmm0 = (xmm1 * xmm0) +/- xmm2
+; X64-NEXT: # xmm0 {%k1} {z} = (xmm1 * xmm0) +/- xmm2
; X64-NEXT: retq # encoding: [0xc3]
%res = call <2 x double> @llvm.x86.avx512.maskz.vfmaddsub.pd.128(<2 x double> %x0, <2 x double> %x1, <2 x double> %x2, i8 %x3)
ret <2 x double> %res
@@ -16064,7 +16064,7 @@ define <4 x double>@test_int_x86_avx512_mask3_vfmaddsub_pd_256(<4 x double> %x0,
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmaddsub231pd %ymm1, %ymm0, %ymm2 {%k1} # encoding: [0x62,0xf2,0xfd,0x29,0xb6,0xd1]
-; X86-NEXT: # ymm2 = (ymm0 * ymm1) +/- ymm2
+; X86-NEXT: # ymm2 {%k1} = (ymm0 * ymm1) +/- ymm2
; X86-NEXT: vmovapd %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x28,0xc2]
; X86-NEXT: retl # encoding: [0xc3]
;
@@ -16072,7 +16072,7 @@ define <4 x double>@test_int_x86_avx512_mask3_vfmaddsub_pd_256(<4 x double> %x0,
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmaddsub231pd %ymm1, %ymm0, %ymm2 {%k1} # encoding: [0x62,0xf2,0xfd,0x29,0xb6,0xd1]
-; X64-NEXT: # ymm2 = (ymm0 * ymm1) +/- ymm2
+; X64-NEXT: # ymm2 {%k1} = (ymm0 * ymm1) +/- ymm2
; X64-NEXT: vmovapd %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x28,0xc2]
; X64-NEXT: retq # encoding: [0xc3]
%res = call <4 x double> @llvm.x86.avx512.mask3.vfmaddsub.pd.256(<4 x double> %x0, <4 x double> %x1, <4 x double> %x2, i8 %x3)
@@ -16087,14 +16087,14 @@ define <4 x double>@test_int_x86_avx512_maskz_vfmaddsub_pd_256(<4 x double> %x0,
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmaddsub213pd %ymm2, %ymm1, %ymm0 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0xa9,0xa6,0xc2]
-; X86-NEXT: # ymm0 = (ymm1 * ymm0) +/- ymm2
+; X86-NEXT: # ymm0 {%k1} {z} = (ymm1 * ymm0) +/- ymm2
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_maskz_vfmaddsub_pd_256:
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmaddsub213pd %ymm2, %ymm1, %ymm0 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0xa9,0xa6,0xc2]
-; X64-NEXT: # ymm0 = (ymm1 * ymm0) +/- ymm2
+; X64-NEXT: # ymm0 {%k1} {z} = (ymm1 * ymm0) +/- ymm2
; X64-NEXT: retq # encoding: [0xc3]
%res = call <4 x double> @llvm.x86.avx512.maskz.vfmaddsub.pd.256(<4 x double> %x0, <4 x double> %x1, <4 x double> %x2, i8 %x3)
ret <4 x double> %res
@@ -16108,7 +16108,7 @@ define <4 x float>@test_int_x86_avx512_mask3_vfmaddsub_ps_128(<4 x float> %x0, <
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmaddsub231ps %xmm1, %xmm0, %xmm2 {%k1} # encoding: [0x62,0xf2,0x7d,0x09,0xb6,0xd1]
-; X86-NEXT: # xmm2 = (xmm0 * xmm1) +/- xmm2
+; X86-NEXT: # xmm2 {%k1} = (xmm0 * xmm1) +/- xmm2
; X86-NEXT: vmovaps %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf8,0x28,0xc2]
; X86-NEXT: retl # encoding: [0xc3]
;
@@ -16116,7 +16116,7 @@ define <4 x float>@test_int_x86_avx512_mask3_vfmaddsub_ps_128(<4 x float> %x0, <
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmaddsub231ps %xmm1, %xmm0, %xmm2 {%k1} # encoding: [0x62,0xf2,0x7d,0x09,0xb6,0xd1]
-; X64-NEXT: # xmm2 = (xmm0 * xmm1) +/- xmm2
+; X64-NEXT: # xmm2 {%k1} = (xmm0 * xmm1) +/- xmm2
; X64-NEXT: vmovaps %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf8,0x28,0xc2]
; X64-NEXT: retq # encoding: [0xc3]
%res = call <4 x float> @llvm.x86.avx512.mask3.vfmaddsub.ps.128(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2, i8 %x3)
@@ -16131,14 +16131,14 @@ define <4 x float>@test_int_x86_avx512_maskz_vfmaddsub_ps_128(<4 x float> %x0, <
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmaddsub213ps %xmm2, %xmm1, %xmm0 {%k1} {z} # encoding: [0x62,0xf2,0x75,0x89,0xa6,0xc2]
-; X86-NEXT: # xmm0 = (xmm1 * xmm0) +/- xmm2
+; X86-NEXT: # xmm0 {%k1} {z} = (xmm1 * xmm0) +/- xmm2
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_maskz_vfmaddsub_ps_128:
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmaddsub213ps %xmm2, %xmm1, %xmm0 {%k1} {z} # encoding: [0x62,0xf2,0x75,0x89,0xa6,0xc2]
-; X64-NEXT: # xmm0 = (xmm1 * xmm0) +/- xmm2
+; X64-NEXT: # xmm0 {%k1} {z} = (xmm1 * xmm0) +/- xmm2
; X64-NEXT: retq # encoding: [0xc3]
%res = call <4 x float> @llvm.x86.avx512.maskz.vfmaddsub.ps.128(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2, i8 %x3)
ret <4 x float> %res
@@ -16152,7 +16152,7 @@ define <8 x float>@test_int_x86_avx512_mask3_vfmaddsub_ps_256(<8 x float> %x0, <
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmaddsub231ps %ymm1, %ymm0, %ymm2 {%k1} # encoding: [0x62,0xf2,0x7d,0x29,0xb6,0xd1]
-; X86-NEXT: # ymm2 = (ymm0 * ymm1) +/- ymm2
+; X86-NEXT: # ymm2 {%k1} = (ymm0 * ymm1) +/- ymm2
; X86-NEXT: vmovaps %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xfc,0x28,0xc2]
; X86-NEXT: retl # encoding: [0xc3]
;
@@ -16160,7 +16160,7 @@ define <8 x float>@test_int_x86_avx512_mask3_vfmaddsub_ps_256(<8 x float> %x0, <
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmaddsub231ps %ymm1, %ymm0, %ymm2 {%k1} # encoding: [0x62,0xf2,0x7d,0x29,0xb6,0xd1]
-; X64-NEXT: # ymm2 = (ymm0 * ymm1) +/- ymm2
+; X64-NEXT: # ymm2 {%k1} = (ymm0 * ymm1) +/- ymm2
; X64-NEXT: vmovaps %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xfc,0x28,0xc2]
; X64-NEXT: retq # encoding: [0xc3]
%res = call <8 x float> @llvm.x86.avx512.mask3.vfmaddsub.ps.256(<8 x float> %x0, <8 x float> %x1, <8 x float> %x2, i8 %x3)
@@ -16175,14 +16175,14 @@ define <8 x float>@test_int_x86_avx512_maskz_vfmaddsub_ps_256(<8 x float> %x0, <
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmaddsub213ps %ymm2, %ymm1, %ymm0 {%k1} {z} # encoding: [0x62,0xf2,0x75,0xa9,0xa6,0xc2]
-; X86-NEXT: # ymm0 = (ymm1 * ymm0) +/- ymm2
+; X86-NEXT: # ymm0 {%k1} {z} = (ymm1 * ymm0) +/- ymm2
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_maskz_vfmaddsub_ps_256:
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmaddsub213ps %ymm2, %ymm1, %ymm0 {%k1} {z} # encoding: [0x62,0xf2,0x75,0xa9,0xa6,0xc2]
-; X64-NEXT: # ymm0 = (ymm1 * ymm0) +/- ymm2
+; X64-NEXT: # ymm0 {%k1} {z} = (ymm1 * ymm0) +/- ymm2
; X64-NEXT: retq # encoding: [0xc3]
%res = call <8 x float> @llvm.x86.avx512.maskz.vfmaddsub.ps.256(<8 x float> %x0, <8 x float> %x1, <8 x float> %x2, i8 %x3)
ret <8 x float> %res
@@ -16196,7 +16196,7 @@ define <2 x double>@test_int_x86_avx512_mask3_vfmsubadd_pd_128(<2 x double> %x0,
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmsubadd231pd %xmm1, %xmm0, %xmm2 {%k1} # encoding: [0x62,0xf2,0xfd,0x09,0xb7,0xd1]
-; X86-NEXT: # xmm2 = (xmm0 * xmm1) -/+ xmm2
+; X86-NEXT: # xmm2 {%k1} = (xmm0 * xmm1) -/+ xmm2
; X86-NEXT: vmovapd %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x28,0xc2]
; X86-NEXT: retl # encoding: [0xc3]
;
@@ -16204,7 +16204,7 @@ define <2 x double>@test_int_x86_avx512_mask3_vfmsubadd_pd_128(<2 x double> %x0,
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmsubadd231pd %xmm1, %xmm0, %xmm2 {%k1} # encoding: [0x62,0xf2,0xfd,0x09,0xb7,0xd1]
-; X64-NEXT: # xmm2 = (xmm0 * xmm1) -/+ xmm2
+; X64-NEXT: # xmm2 {%k1} = (xmm0 * xmm1) -/+ xmm2
; X64-NEXT: vmovapd %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x28,0xc2]
; X64-NEXT: retq # encoding: [0xc3]
%res = call <2 x double> @llvm.x86.avx512.mask3.vfmsubadd.pd.128(<2 x double> %x0, <2 x double> %x1, <2 x double> %x2, i8 %x3)
@@ -16219,7 +16219,7 @@ define <4 x double>@test_int_x86_avx512_mask3_vfmsubadd_pd_256(<4 x double> %x0,
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmsubadd231pd %ymm1, %ymm0, %ymm2 {%k1} # encoding: [0x62,0xf2,0xfd,0x29,0xb7,0xd1]
-; X86-NEXT: # ymm2 = (ymm0 * ymm1) -/+ ymm2
+; X86-NEXT: # ymm2 {%k1} = (ymm0 * ymm1) -/+ ymm2
; X86-NEXT: vmovapd %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x28,0xc2]
; X86-NEXT: retl # encoding: [0xc3]
;
@@ -16227,7 +16227,7 @@ define <4 x double>@test_int_x86_avx512_mask3_vfmsubadd_pd_256(<4 x double> %x0,
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmsubadd231pd %ymm1, %ymm0, %ymm2 {%k1} # encoding: [0x62,0xf2,0xfd,0x29,0xb7,0xd1]
-; X64-NEXT: # ymm2 = (ymm0 * ymm1) -/+ ymm2
+; X64-NEXT: # ymm2 {%k1} = (ymm0 * ymm1) -/+ ymm2
; X64-NEXT: vmovapd %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x28,0xc2]
; X64-NEXT: retq # encoding: [0xc3]
%res = call <4 x double> @llvm.x86.avx512.mask3.vfmsubadd.pd.256(<4 x double> %x0, <4 x double> %x1, <4 x double> %x2, i8 %x3)
@@ -16242,7 +16242,7 @@ define <4 x float>@test_int_x86_avx512_mask3_vfmsubadd_ps_128(<4 x float> %x0, <
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmsubadd231ps %xmm1, %xmm0, %xmm2 {%k1} # encoding: [0x62,0xf2,0x7d,0x09,0xb7,0xd1]
-; X86-NEXT: # xmm2 = (xmm0 * xmm1) -/+ xmm2
+; X86-NEXT: # xmm2 {%k1} = (xmm0 * xmm1) -/+ xmm2
; X86-NEXT: vmovaps %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf8,0x28,0xc2]
; X86-NEXT: retl # encoding: [0xc3]
;
@@ -16250,7 +16250,7 @@ define <4 x float>@test_int_x86_avx512_mask3_vfmsubadd_ps_128(<4 x float> %x0, <
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmsubadd231ps %xmm1, %xmm0, %xmm2 {%k1} # encoding: [0x62,0xf2,0x7d,0x09,0xb7,0xd1]
-; X64-NEXT: # xmm2 = (xmm0 * xmm1) -/+ xmm2
+; X64-NEXT: # xmm2 {%k1} = (xmm0 * xmm1) -/+ xmm2
; X64-NEXT: vmovaps %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf8,0x28,0xc2]
; X64-NEXT: retq # encoding: [0xc3]
%res = call <4 x float> @llvm.x86.avx512.mask3.vfmsubadd.ps.128(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2, i8 %x3)
@@ -16265,7 +16265,7 @@ define <8 x float>@test_int_x86_avx512_mask3_vfmsubadd_ps_256(<8 x float> %x0, <
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmsubadd231ps %ymm1, %ymm0, %ymm2 {%k1} # encoding: [0x62,0xf2,0x7d,0x29,0xb7,0xd1]
-; X86-NEXT: # ymm2 = (ymm0 * ymm1) -/+ ymm2
+; X86-NEXT: # ymm2 {%k1} = (ymm0 * ymm1) -/+ ymm2
; X86-NEXT: vmovaps %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xfc,0x28,0xc2]
; X86-NEXT: retl # encoding: [0xc3]
;
@@ -16273,7 +16273,7 @@ define <8 x float>@test_int_x86_avx512_mask3_vfmsubadd_ps_256(<8 x float> %x0, <
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmsubadd231ps %ymm1, %ymm0, %ymm2 {%k1} # encoding: [0x62,0xf2,0x7d,0x29,0xb7,0xd1]
-; X64-NEXT: # ymm2 = (ymm0 * ymm1) -/+ ymm2
+; X64-NEXT: # ymm2 {%k1} = (ymm0 * ymm1) -/+ ymm2
; X64-NEXT: vmovaps %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xfc,0x28,0xc2]
; X64-NEXT: retq # encoding: [0xc3]
%res = call <8 x float> @llvm.x86.avx512.mask3.vfmsubadd.ps.256(<8 x float> %x0, <8 x float> %x1, <8 x float> %x2, i8 %x3)
@@ -16288,14 +16288,14 @@ define <4 x float> @test_mask_vfmadd128_ps_rmk(<4 x float> %a0, <4 x float> %a1,
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %ecx # encoding: [0x0f,0xb6,0x4c,0x24,0x08]
; X86-NEXT: kmovw %ecx, %k1 # encoding: [0xc5,0xf8,0x92,0xc9]
; X86-NEXT: vfmadd213ps (%eax), %xmm1, %xmm0 {%k1} # encoding: [0x62,0xf2,0x75,0x09,0xa8,0x00]
-; X86-NEXT: # xmm0 = (xmm1 * xmm0) + mem
+; X86-NEXT: # xmm0 {%k1} = (xmm1 * xmm0) + mem
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_mask_vfmadd128_ps_rmk:
; X64: # %bb.0:
; X64-NEXT: kmovw %esi, %k1 # encoding: [0xc5,0xf8,0x92,0xce]
; X64-NEXT: vfmadd213ps (%rdi), %xmm1, %xmm0 {%k1} # encoding: [0x62,0xf2,0x75,0x09,0xa8,0x07]
-; X64-NEXT: # xmm0 = (xmm1 * xmm0) + mem
+; X64-NEXT: # xmm0 {%k1} = (xmm1 * xmm0) + mem
; X64-NEXT: retq # encoding: [0xc3]
%a2 = load <4 x float>, <4 x float>* %ptr_a2
%res = call <4 x float> @llvm.x86.avx512.mask.vfmadd.ps.128(<4 x float> %a0, <4 x float> %a1, <4 x float> %a2, i8 %mask) nounwind
@@ -16309,14 +16309,14 @@ define <4 x float> @test_mask_vfmadd128_ps_rmka(<4 x float> %a0, <4 x float> %a1
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %ecx # encoding: [0x0f,0xb6,0x4c,0x24,0x08]
; X86-NEXT: kmovw %ecx, %k1 # encoding: [0xc5,0xf8,0x92,0xc9]
; X86-NEXT: vfmadd213ps (%eax), %xmm1, %xmm0 {%k1} # encoding: [0x62,0xf2,0x75,0x09,0xa8,0x00]
-; X86-NEXT: # xmm0 = (xmm1 * xmm0) + mem
+; X86-NEXT: # xmm0 {%k1} = (xmm1 * xmm0) + mem
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_mask_vfmadd128_ps_rmka:
; X64: # %bb.0:
; X64-NEXT: kmovw %esi, %k1 # encoding: [0xc5,0xf8,0x92,0xce]
; X64-NEXT: vfmadd213ps (%rdi), %xmm1, %xmm0 {%k1} # encoding: [0x62,0xf2,0x75,0x09,0xa8,0x07]
-; X64-NEXT: # xmm0 = (xmm1 * xmm0) + mem
+; X64-NEXT: # xmm0 {%k1} = (xmm1 * xmm0) + mem
; X64-NEXT: retq # encoding: [0xc3]
%a2 = load <4 x float>, <4 x float>* %ptr_a2, align 8
%res = call <4 x float> @llvm.x86.avx512.mask.vfmadd.ps.128(<4 x float> %a0, <4 x float> %a1, <4 x float> %a2, i8 %mask) nounwind
@@ -16366,14 +16366,14 @@ define <4 x float> @test_mask_vfmadd128_ps_rmb(<4 x float> %a0, <4 x float> %a1,
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %ecx # encoding: [0x0f,0xb6,0x4c,0x24,0x08]
; X86-NEXT: kmovw %ecx, %k1 # encoding: [0xc5,0xf8,0x92,0xc9]
; X86-NEXT: vfmadd213ps (%eax){1to4}, %xmm1, %xmm0 {%k1} # encoding: [0x62,0xf2,0x75,0x19,0xa8,0x00]
-; X86-NEXT: # xmm0 = (xmm1 * xmm0) + mem
+; X86-NEXT: # xmm0 {%k1} = (xmm1 * xmm0) + mem
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_mask_vfmadd128_ps_rmb:
; X64: # %bb.0:
; X64-NEXT: kmovw %esi, %k1 # encoding: [0xc5,0xf8,0x92,0xce]
; X64-NEXT: vfmadd213ps (%rdi){1to4}, %xmm1, %xmm0 {%k1} # encoding: [0x62,0xf2,0x75,0x19,0xa8,0x07]
-; X64-NEXT: # xmm0 = (xmm1 * xmm0) + mem
+; X64-NEXT: # xmm0 {%k1} = (xmm1 * xmm0) + mem
; X64-NEXT: retq # encoding: [0xc3]
%q = load float, float* %ptr_a2
%vecinit.i = insertelement <4 x float> undef, float %q, i32 0
@@ -16391,14 +16391,14 @@ define <4 x float> @test_mask_vfmadd128_ps_rmba(<4 x float> %a0, <4 x float> %a1
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %ecx # encoding: [0x0f,0xb6,0x4c,0x24,0x08]
; X86-NEXT: kmovw %ecx, %k1 # encoding: [0xc5,0xf8,0x92,0xc9]
; X86-NEXT: vfmadd213ps (%eax){1to4}, %xmm1, %xmm0 {%k1} # encoding: [0x62,0xf2,0x75,0x19,0xa8,0x00]
-; X86-NEXT: # xmm0 = (xmm1 * xmm0) + mem
+; X86-NEXT: # xmm0 {%k1} = (xmm1 * xmm0) + mem
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_mask_vfmadd128_ps_rmba:
; X64: # %bb.0:
; X64-NEXT: kmovw %esi, %k1 # encoding: [0xc5,0xf8,0x92,0xce]
; X64-NEXT: vfmadd213ps (%rdi){1to4}, %xmm1, %xmm0 {%k1} # encoding: [0x62,0xf2,0x75,0x19,0xa8,0x07]
-; X64-NEXT: # xmm0 = (xmm1 * xmm0) + mem
+; X64-NEXT: # xmm0 {%k1} = (xmm1 * xmm0) + mem
; X64-NEXT: retq # encoding: [0xc3]
%q = load float, float* %ptr_a2, align 4
%vecinit.i = insertelement <4 x float> undef, float %q, i32 0
@@ -16460,14 +16460,14 @@ define <2 x double> @test_mask_vfmadd128_pd_rmk(<2 x double> %a0, <2 x double> %
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %ecx # encoding: [0x0f,0xb6,0x4c,0x24,0x08]
; X86-NEXT: kmovw %ecx, %k1 # encoding: [0xc5,0xf8,0x92,0xc9]
; X86-NEXT: vfmadd213pd (%eax), %xmm1, %xmm0 {%k1} # encoding: [0x62,0xf2,0xf5,0x09,0xa8,0x00]
-; X86-NEXT: # xmm0 = (xmm1 * xmm0) + mem
+; X86-NEXT: # xmm0 {%k1} = (xmm1 * xmm0) + mem
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_mask_vfmadd128_pd_rmk:
; X64: # %bb.0:
; X64-NEXT: kmovw %esi, %k1 # encoding: [0xc5,0xf8,0x92,0xce]
; X64-NEXT: vfmadd213pd (%rdi), %xmm1, %xmm0 {%k1} # encoding: [0x62,0xf2,0xf5,0x09,0xa8,0x07]
-; X64-NEXT: # xmm0 = (xmm1 * xmm0) + mem
+; X64-NEXT: # xmm0 {%k1} = (xmm1 * xmm0) + mem
; X64-NEXT: retq # encoding: [0xc3]
%a2 = load <2 x double>, <2 x double>* %ptr_a2
%res = call <2 x double> @llvm.x86.avx512.mask.vfmadd.pd.128(<2 x double> %a0, <2 x double> %a1, <2 x double> %a2, i8 %mask) nounwind
@@ -16499,14 +16499,14 @@ define <4 x double> @test_mask_vfmadd256_pd_rmk(<4 x double> %a0, <4 x double> %
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %ecx # encoding: [0x0f,0xb6,0x4c,0x24,0x08]
; X86-NEXT: kmovw %ecx, %k1 # encoding: [0xc5,0xf8,0x92,0xc9]
; X86-NEXT: vfmadd213pd (%eax), %ymm1, %ymm0 {%k1} # encoding: [0x62,0xf2,0xf5,0x29,0xa8,0x00]
-; X86-NEXT: # ymm0 = (ymm1 * ymm0) + mem
+; X86-NEXT: # ymm0 {%k1} = (ymm1 * ymm0) + mem
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_mask_vfmadd256_pd_rmk:
; X64: # %bb.0:
; X64-NEXT: kmovw %esi, %k1 # encoding: [0xc5,0xf8,0x92,0xce]
; X64-NEXT: vfmadd213pd (%rdi), %ymm1, %ymm0 {%k1} # encoding: [0x62,0xf2,0xf5,0x29,0xa8,0x07]
-; X64-NEXT: # ymm0 = (ymm1 * ymm0) + mem
+; X64-NEXT: # ymm0 {%k1} = (ymm1 * ymm0) + mem
; X64-NEXT: retq # encoding: [0xc3]
%a2 = load <4 x double>, <4 x double>* %ptr_a2
%res = call <4 x double> @llvm.x86.avx512.mask.vfmadd.pd.256(<4 x double> %a0, <4 x double> %a1, <4 x double> %a2, i8 %mask) nounwind
diff --git a/llvm/test/CodeGen/X86/avx512vl-intrinsics.ll b/llvm/test/CodeGen/X86/avx512vl-intrinsics.ll
index 6fb460b5cc66..6905fafc3a7b 100644
--- a/llvm/test/CodeGen/X86/avx512vl-intrinsics.ll
+++ b/llvm/test/CodeGen/X86/avx512vl-intrinsics.ll
@@ -5503,14 +5503,14 @@ define <8 x float> @test_mask_vfmadd256_ps(<8 x float> %a0, <8 x float> %a1, <8
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmadd132ps %ymm1, %ymm2, %ymm0 {%k1} # encoding: [0x62,0xf2,0x6d,0x29,0x98,0xc1]
-; X86-NEXT: # ymm0 = (ymm0 * ymm1) + ymm2
+; X86-NEXT: # ymm0 {%k1} = (ymm0 * ymm1) + ymm2
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_mask_vfmadd256_ps:
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmadd132ps %ymm1, %ymm2, %ymm0 {%k1} # encoding: [0x62,0xf2,0x6d,0x29,0x98,0xc1]
-; X64-NEXT: # ymm0 = (ymm0 * ymm1) + ymm2
+; X64-NEXT: # ymm0 {%k1} = (ymm0 * ymm1) + ymm2
; X64-NEXT: retq # encoding: [0xc3]
%1 = call <8 x float> @llvm.fma.v8f32(<8 x float> %a0, <8 x float> %a1, <8 x float> %a2)
%2 = bitcast i8 %mask to <8 x i1>
@@ -5534,14 +5534,14 @@ define <4 x float> @test_mask_vfmadd128_ps(<4 x float> %a0, <4 x float> %a1, <4
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmadd132ps %xmm1, %xmm2, %xmm0 {%k1} # encoding: [0x62,0xf2,0x6d,0x09,0x98,0xc1]
-; X86-NEXT: # xmm0 = (xmm0 * xmm1) + xmm2
+; X86-NEXT: # xmm0 {%k1} = (xmm0 * xmm1) + xmm2
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_mask_vfmadd128_ps:
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmadd132ps %xmm1, %xmm2, %xmm0 {%k1} # encoding: [0x62,0xf2,0x6d,0x09,0x98,0xc1]
-; X64-NEXT: # xmm0 = (xmm0 * xmm1) + xmm2
+; X64-NEXT: # xmm0 {%k1} = (xmm0 * xmm1) + xmm2
; X64-NEXT: retq # encoding: [0xc3]
%1 = call <4 x float> @llvm.fma.v4f32(<4 x float> %a0, <4 x float> %a1, <4 x float> %a2)
%2 = bitcast i8 %mask to <8 x i1>
@@ -5566,14 +5566,14 @@ define <4 x double> @test_mask_fmadd256_pd(<4 x double> %a, <4 x double> %b, <4
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmadd132pd %ymm1, %ymm2, %ymm0 {%k1} # encoding: [0x62,0xf2,0xed,0x29,0x98,0xc1]
-; X86-NEXT: # ymm0 = (ymm0 * ymm1) + ymm2
+; X86-NEXT: # ymm0 {%k1} = (ymm0 * ymm1) + ymm2
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_mask_fmadd256_pd:
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmadd132pd %ymm1, %ymm2, %ymm0 {%k1} # encoding: [0x62,0xf2,0xed,0x29,0x98,0xc1]
-; X64-NEXT: # ymm0 = (ymm0 * ymm1) + ymm2
+; X64-NEXT: # ymm0 {%k1} = (ymm0 * ymm1) + ymm2
; X64-NEXT: retq # encoding: [0xc3]
%1 = call <4 x double> @llvm.fma.v4f64(<4 x double> %a, <4 x double> %b, <4 x double> %c)
%2 = bitcast i8 %mask to <8 x i1>
@@ -5598,14 +5598,14 @@ define <2 x double> @test_mask_fmadd128_pd(<2 x double> %a, <2 x double> %b, <2
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmadd132pd %xmm1, %xmm2, %xmm0 {%k1} # encoding: [0x62,0xf2,0xed,0x09,0x98,0xc1]
-; X86-NEXT: # xmm0 = (xmm0 * xmm1) + xmm2
+; X86-NEXT: # xmm0 {%k1} = (xmm0 * xmm1) + xmm2
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_mask_fmadd128_pd:
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmadd132pd %xmm1, %xmm2, %xmm0 {%k1} # encoding: [0x62,0xf2,0xed,0x09,0x98,0xc1]
-; X64-NEXT: # xmm0 = (xmm0 * xmm1) + xmm2
+; X64-NEXT: # xmm0 {%k1} = (xmm0 * xmm1) + xmm2
; X64-NEXT: retq # encoding: [0xc3]
%1 = call <2 x double> @llvm.fma.v2f64(<2 x double> %a, <2 x double> %b, <2 x double> %c)
%2 = bitcast i8 %mask to <8 x i1>
@@ -5620,7 +5620,7 @@ define <2 x double>@test_int_x86_avx512_mask3_vfmadd_pd_128(<2 x double> %x0, <2
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmadd231pd %xmm1, %xmm0, %xmm2 {%k1} # encoding: [0x62,0xf2,0xfd,0x09,0xb8,0xd1]
-; X86-NEXT: # xmm2 = (xmm0 * xmm1) + xmm2
+; X86-NEXT: # xmm2 {%k1} = (xmm0 * xmm1) + xmm2
; X86-NEXT: vmovapd %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x28,0xc2]
; X86-NEXT: retl # encoding: [0xc3]
;
@@ -5628,7 +5628,7 @@ define <2 x double>@test_int_x86_avx512_mask3_vfmadd_pd_128(<2 x double> %x0, <2
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmadd231pd %xmm1, %xmm0, %xmm2 {%k1} # encoding: [0x62,0xf2,0xfd,0x09,0xb8,0xd1]
-; X64-NEXT: # xmm2 = (xmm0 * xmm1) + xmm2
+; X64-NEXT: # xmm2 {%k1} = (xmm0 * xmm1) + xmm2
; X64-NEXT: vmovapd %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x28,0xc2]
; X64-NEXT: retq # encoding: [0xc3]
%1 = call <2 x double> @llvm.fma.v2f64(<2 x double> %x0, <2 x double> %x1, <2 x double> %x2)
@@ -5644,14 +5644,14 @@ define <2 x double>@test_int_x86_avx512_maskz_vfmadd_pd_128(<2 x double> %x0, <2
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmadd213pd %xmm2, %xmm1, %xmm0 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0x89,0xa8,0xc2]
-; X86-NEXT: # xmm0 = (xmm1 * xmm0) + xmm2
+; X86-NEXT: # xmm0 {%k1} {z} = (xmm1 * xmm0) + xmm2
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_maskz_vfmadd_pd_128:
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmadd213pd %xmm2, %xmm1, %xmm0 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0x89,0xa8,0xc2]
-; X64-NEXT: # xmm0 = (xmm1 * xmm0) + xmm2
+; X64-NEXT: # xmm0 {%k1} {z} = (xmm1 * xmm0) + xmm2
; X64-NEXT: retq # encoding: [0xc3]
%1 = call <2 x double> @llvm.fma.v2f64(<2 x double> %x0, <2 x double> %x1, <2 x double> %x2)
%2 = bitcast i8 %x3 to <8 x i1>
@@ -5666,7 +5666,7 @@ define <4 x double>@test_int_x86_avx512_mask3_vfmadd_pd_256(<4 x double> %x0, <4
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmadd231pd %ymm1, %ymm0, %ymm2 {%k1} # encoding: [0x62,0xf2,0xfd,0x29,0xb8,0xd1]
-; X86-NEXT: # ymm2 = (ymm0 * ymm1) + ymm2
+; X86-NEXT: # ymm2 {%k1} = (ymm0 * ymm1) + ymm2
; X86-NEXT: vmovapd %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x28,0xc2]
; X86-NEXT: retl # encoding: [0xc3]
;
@@ -5674,7 +5674,7 @@ define <4 x double>@test_int_x86_avx512_mask3_vfmadd_pd_256(<4 x double> %x0, <4
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmadd231pd %ymm1, %ymm0, %ymm2 {%k1} # encoding: [0x62,0xf2,0xfd,0x29,0xb8,0xd1]
-; X64-NEXT: # ymm2 = (ymm0 * ymm1) + ymm2
+; X64-NEXT: # ymm2 {%k1} = (ymm0 * ymm1) + ymm2
; X64-NEXT: vmovapd %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x28,0xc2]
; X64-NEXT: retq # encoding: [0xc3]
%1 = call <4 x double> @llvm.fma.v4f64(<4 x double> %x0, <4 x double> %x1, <4 x double> %x2)
@@ -5690,14 +5690,14 @@ define <4 x double>@test_int_x86_avx512_maskz_vfmadd_pd_256(<4 x double> %x0, <4
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmadd213pd %ymm2, %ymm1, %ymm0 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0xa9,0xa8,0xc2]
-; X86-NEXT: # ymm0 = (ymm1 * ymm0) + ymm2
+; X86-NEXT: # ymm0 {%k1} {z} = (ymm1 * ymm0) + ymm2
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_maskz_vfmadd_pd_256:
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmadd213pd %ymm2, %ymm1, %ymm0 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0xa9,0xa8,0xc2]
-; X64-NEXT: # ymm0 = (ymm1 * ymm0) + ymm2
+; X64-NEXT: # ymm0 {%k1} {z} = (ymm1 * ymm0) + ymm2
; X64-NEXT: retq # encoding: [0xc3]
%1 = call <4 x double> @llvm.fma.v4f64(<4 x double> %x0, <4 x double> %x1, <4 x double> %x2)
%2 = bitcast i8 %x3 to <8 x i1>
@@ -5712,7 +5712,7 @@ define <4 x float>@test_int_x86_avx512_mask3_vfmadd_ps_128(<4 x float> %x0, <4 x
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmadd231ps %xmm1, %xmm0, %xmm2 {%k1} # encoding: [0x62,0xf2,0x7d,0x09,0xb8,0xd1]
-; X86-NEXT: # xmm2 = (xmm0 * xmm1) + xmm2
+; X86-NEXT: # xmm2 {%k1} = (xmm0 * xmm1) + xmm2
; X86-NEXT: vmovaps %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf8,0x28,0xc2]
; X86-NEXT: retl # encoding: [0xc3]
;
@@ -5720,7 +5720,7 @@ define <4 x float>@test_int_x86_avx512_mask3_vfmadd_ps_128(<4 x float> %x0, <4 x
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmadd231ps %xmm1, %xmm0, %xmm2 {%k1} # encoding: [0x62,0xf2,0x7d,0x09,0xb8,0xd1]
-; X64-NEXT: # xmm2 = (xmm0 * xmm1) + xmm2
+; X64-NEXT: # xmm2 {%k1} = (xmm0 * xmm1) + xmm2
; X64-NEXT: vmovaps %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf8,0x28,0xc2]
; X64-NEXT: retq # encoding: [0xc3]
%1 = call <4 x float> @llvm.fma.v4f32(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2)
@@ -5736,14 +5736,14 @@ define <4 x float>@test_int_x86_avx512_maskz_vfmadd_ps_128(<4 x float> %x0, <4 x
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmadd213ps %xmm2, %xmm1, %xmm0 {%k1} {z} # encoding: [0x62,0xf2,0x75,0x89,0xa8,0xc2]
-; X86-NEXT: # xmm0 = (xmm1 * xmm0) + xmm2
+; X86-NEXT: # xmm0 {%k1} {z} = (xmm1 * xmm0) + xmm2
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_maskz_vfmadd_ps_128:
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmadd213ps %xmm2, %xmm1, %xmm0 {%k1} {z} # encoding: [0x62,0xf2,0x75,0x89,0xa8,0xc2]
-; X64-NEXT: # xmm0 = (xmm1 * xmm0) + xmm2
+; X64-NEXT: # xmm0 {%k1} {z} = (xmm1 * xmm0) + xmm2
; X64-NEXT: retq # encoding: [0xc3]
%1 = call <4 x float> @llvm.fma.v4f32(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2)
%2 = bitcast i8 %x3 to <8 x i1>
@@ -5758,7 +5758,7 @@ define <8 x float>@test_int_x86_avx512_mask3_vfmadd_ps_256(<8 x float> %x0, <8 x
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmadd231ps %ymm1, %ymm0, %ymm2 {%k1} # encoding: [0x62,0xf2,0x7d,0x29,0xb8,0xd1]
-; X86-NEXT: # ymm2 = (ymm0 * ymm1) + ymm2
+; X86-NEXT: # ymm2 {%k1} = (ymm0 * ymm1) + ymm2
; X86-NEXT: vmovaps %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xfc,0x28,0xc2]
; X86-NEXT: retl # encoding: [0xc3]
;
@@ -5766,7 +5766,7 @@ define <8 x float>@test_int_x86_avx512_mask3_vfmadd_ps_256(<8 x float> %x0, <8 x
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmadd231ps %ymm1, %ymm0, %ymm2 {%k1} # encoding: [0x62,0xf2,0x7d,0x29,0xb8,0xd1]
-; X64-NEXT: # ymm2 = (ymm0 * ymm1) + ymm2
+; X64-NEXT: # ymm2 {%k1} = (ymm0 * ymm1) + ymm2
; X64-NEXT: vmovaps %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xfc,0x28,0xc2]
; X64-NEXT: retq # encoding: [0xc3]
%1 = call <8 x float> @llvm.fma.v8f32(<8 x float> %x0, <8 x float> %x1, <8 x float> %x2)
@@ -5781,14 +5781,14 @@ define <8 x float>@test_int_x86_avx512_maskz_vfmadd_ps_256(<8 x float> %x0, <8 x
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmadd213ps %ymm2, %ymm1, %ymm0 {%k1} {z} # encoding: [0x62,0xf2,0x75,0xa9,0xa8,0xc2]
-; X86-NEXT: # ymm0 = (ymm1 * ymm0) + ymm2
+; X86-NEXT: # ymm0 {%k1} {z} = (ymm1 * ymm0) + ymm2
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_maskz_vfmadd_ps_256:
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmadd213ps %ymm2, %ymm1, %ymm0 {%k1} {z} # encoding: [0x62,0xf2,0x75,0xa9,0xa8,0xc2]
-; X64-NEXT: # ymm0 = (ymm1 * ymm0) + ymm2
+; X64-NEXT: # ymm0 {%k1} {z} = (ymm1 * ymm0) + ymm2
; X64-NEXT: retq # encoding: [0xc3]
%1 = call <8 x float> @llvm.fma.v8f32(<8 x float> %x0, <8 x float> %x1, <8 x float> %x2)
%2 = bitcast i8 %x3 to <8 x i1>
@@ -5802,7 +5802,7 @@ define <2 x double>@test_int_x86_avx512_mask3_vfmsub_pd_128(<2 x double> %x0, <2
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmsub231pd %xmm1, %xmm0, %xmm2 {%k1} # encoding: [0x62,0xf2,0xfd,0x09,0xba,0xd1]
-; X86-NEXT: # xmm2 = (xmm0 * xmm1) - xmm2
+; X86-NEXT: # xmm2 {%k1} = (xmm0 * xmm1) - xmm2
; X86-NEXT: vmovapd %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x28,0xc2]
; X86-NEXT: retl # encoding: [0xc3]
;
@@ -5810,7 +5810,7 @@ define <2 x double>@test_int_x86_avx512_mask3_vfmsub_pd_128(<2 x double> %x0, <2
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmsub231pd %xmm1, %xmm0, %xmm2 {%k1} # encoding: [0x62,0xf2,0xfd,0x09,0xba,0xd1]
-; X64-NEXT: # xmm2 = (xmm0 * xmm1) - xmm2
+; X64-NEXT: # xmm2 {%k1} = (xmm0 * xmm1) - xmm2
; X64-NEXT: vmovapd %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x28,0xc2]
; X64-NEXT: retq # encoding: [0xc3]
%1 = fsub <2 x double> <double -0.000000e+00, double -0.000000e+00>, %x2
@@ -5827,7 +5827,7 @@ define <4 x double>@test_int_x86_avx512_mask3_vfmsub_pd_256(<4 x double> %x0, <4
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmsub231pd %ymm1, %ymm0, %ymm2 {%k1} # encoding: [0x62,0xf2,0xfd,0x29,0xba,0xd1]
-; X86-NEXT: # ymm2 = (ymm0 * ymm1) - ymm2
+; X86-NEXT: # ymm2 {%k1} = (ymm0 * ymm1) - ymm2
; X86-NEXT: vmovapd %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x28,0xc2]
; X86-NEXT: retl # encoding: [0xc3]
;
@@ -5835,7 +5835,7 @@ define <4 x double>@test_int_x86_avx512_mask3_vfmsub_pd_256(<4 x double> %x0, <4
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmsub231pd %ymm1, %ymm0, %ymm2 {%k1} # encoding: [0x62,0xf2,0xfd,0x29,0xba,0xd1]
-; X64-NEXT: # ymm2 = (ymm0 * ymm1) - ymm2
+; X64-NEXT: # ymm2 {%k1} = (ymm0 * ymm1) - ymm2
; X64-NEXT: vmovapd %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x28,0xc2]
; X64-NEXT: retq # encoding: [0xc3]
%1 = fsub <4 x double> <double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00>, %x2
@@ -5852,7 +5852,7 @@ define <4 x float>@test_int_x86_avx512_mask3_vfmsub_ps_128(<4 x float> %x0, <4 x
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmsub231ps %xmm1, %xmm0, %xmm2 {%k1} # encoding: [0x62,0xf2,0x7d,0x09,0xba,0xd1]
-; X86-NEXT: # xmm2 = (xmm0 * xmm1) - xmm2
+; X86-NEXT: # xmm2 {%k1} = (xmm0 * xmm1) - xmm2
; X86-NEXT: vmovaps %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf8,0x28,0xc2]
; X86-NEXT: retl # encoding: [0xc3]
;
@@ -5860,7 +5860,7 @@ define <4 x float>@test_int_x86_avx512_mask3_vfmsub_ps_128(<4 x float> %x0, <4 x
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmsub231ps %xmm1, %xmm0, %xmm2 {%k1} # encoding: [0x62,0xf2,0x7d,0x09,0xba,0xd1]
-; X64-NEXT: # xmm2 = (xmm0 * xmm1) - xmm2
+; X64-NEXT: # xmm2 {%k1} = (xmm0 * xmm1) - xmm2
; X64-NEXT: vmovaps %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf8,0x28,0xc2]
; X64-NEXT: retq # encoding: [0xc3]
%1 = fsub <4 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %x2
@@ -5877,7 +5877,7 @@ define <8 x float>@test_int_x86_avx512_mask3_vfmsub_ps_256(<8 x float> %x0, <8 x
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmsub231ps %ymm1, %ymm0, %ymm2 {%k1} # encoding: [0x62,0xf2,0x7d,0x29,0xba,0xd1]
-; X86-NEXT: # ymm2 = (ymm0 * ymm1) - ymm2
+; X86-NEXT: # ymm2 {%k1} = (ymm0 * ymm1) - ymm2
; X86-NEXT: vmovaps %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xfc,0x28,0xc2]
; X86-NEXT: retl # encoding: [0xc3]
;
@@ -5885,7 +5885,7 @@ define <8 x float>@test_int_x86_avx512_mask3_vfmsub_ps_256(<8 x float> %x0, <8 x
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmsub231ps %ymm1, %ymm0, %ymm2 {%k1} # encoding: [0x62,0xf2,0x7d,0x29,0xba,0xd1]
-; X64-NEXT: # ymm2 = (ymm0 * ymm1) - ymm2
+; X64-NEXT: # ymm2 {%k1} = (ymm0 * ymm1) - ymm2
; X64-NEXT: vmovaps %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xfc,0x28,0xc2]
; X64-NEXT: retq # encoding: [0xc3]
%1 = fsub <8 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %x2
@@ -5912,14 +5912,14 @@ define <8 x float> @test_mask_vfnmadd256_ps(<8 x float> %a0, <8 x float> %a1, <8
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfnmadd132ps %ymm1, %ymm2, %ymm0 {%k1} # encoding: [0x62,0xf2,0x6d,0x29,0x9c,0xc1]
-; X86-NEXT: # ymm0 = -(ymm0 * ymm1) + ymm2
+; X86-NEXT: # ymm0 {%k1} = -(ymm0 * ymm1) + ymm2
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_mask_vfnmadd256_ps:
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfnmadd132ps %ymm1, %ymm2, %ymm0 {%k1} # encoding: [0x62,0xf2,0x6d,0x29,0x9c,0xc1]
-; X64-NEXT: # ymm0 = -(ymm0 * ymm1) + ymm2
+; X64-NEXT: # ymm0 {%k1} = -(ymm0 * ymm1) + ymm2
; X64-NEXT: retq # encoding: [0xc3]
%1 = fsub <8 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %a1
%2 = call <8 x float> @llvm.fma.v8f32(<8 x float> %a0, <8 x float> %1, <8 x float> %a2)
@@ -5945,14 +5945,14 @@ define <4 x float> @test_mask_vfnmadd128_ps(<4 x float> %a0, <4 x float> %a1, <4
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfnmadd132ps %xmm1, %xmm2, %xmm0 {%k1} # encoding: [0x62,0xf2,0x6d,0x09,0x9c,0xc1]
-; X86-NEXT: # xmm0 = -(xmm0 * xmm1) + xmm2
+; X86-NEXT: # xmm0 {%k1} = -(xmm0 * xmm1) + xmm2
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_mask_vfnmadd128_ps:
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfnmadd132ps %xmm1, %xmm2, %xmm0 {%k1} # encoding: [0x62,0xf2,0x6d,0x09,0x9c,0xc1]
-; X64-NEXT: # xmm0 = -(xmm0 * xmm1) + xmm2
+; X64-NEXT: # xmm0 {%k1} = -(xmm0 * xmm1) + xmm2
; X64-NEXT: retq # encoding: [0xc3]
%1 = fsub <4 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %a1
%2 = call <4 x float> @llvm.fma.v4f32(<4 x float> %a0, <4 x float> %1, <4 x float> %a2)
@@ -5979,14 +5979,14 @@ define <4 x double> @test_mask_vfnmadd256_pd(<4 x double> %a0, <4 x double> %a1,
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfnmadd132pd %ymm1, %ymm2, %ymm0 {%k1} # encoding: [0x62,0xf2,0xed,0x29,0x9c,0xc1]
-; X86-NEXT: # ymm0 = -(ymm0 * ymm1) + ymm2
+; X86-NEXT: # ymm0 {%k1} = -(ymm0 * ymm1) + ymm2
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_mask_vfnmadd256_pd:
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfnmadd132pd %ymm1, %ymm2, %ymm0 {%k1} # encoding: [0x62,0xf2,0xed,0x29,0x9c,0xc1]
-; X64-NEXT: # ymm0 = -(ymm0 * ymm1) + ymm2
+; X64-NEXT: # ymm0 {%k1} = -(ymm0 * ymm1) + ymm2
; X64-NEXT: retq # encoding: [0xc3]
%1 = fsub <4 x double> <double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00>, %a1
%2 = call <4 x double> @llvm.fma.v4f64(<4 x double> %a0, <4 x double> %1, <4 x double> %a2)
@@ -6013,14 +6013,14 @@ define <2 x double> @test_mask_vfnmadd128_pd(<2 x double> %a0, <2 x double> %a1,
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfnmadd132pd %xmm1, %xmm2, %xmm0 {%k1} # encoding: [0x62,0xf2,0xed,0x09,0x9c,0xc1]
-; X86-NEXT: # xmm0 = -(xmm0 * xmm1) + xmm2
+; X86-NEXT: # xmm0 {%k1} = -(xmm0 * xmm1) + xmm2
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_mask_vfnmadd128_pd:
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfnmadd132pd %xmm1, %xmm2, %xmm0 {%k1} # encoding: [0x62,0xf2,0xed,0x09,0x9c,0xc1]
-; X64-NEXT: # xmm0 = -(xmm0 * xmm1) + xmm2
+; X64-NEXT: # xmm0 {%k1} = -(xmm0 * xmm1) + xmm2
; X64-NEXT: retq # encoding: [0xc3]
%1 = fsub <2 x double> <double -0.000000e+00, double -0.000000e+00>, %a1
%2 = call <2 x double> @llvm.fma.v2f64(<2 x double> %a0, <2 x double> %1, <2 x double> %a2)
@@ -6048,14 +6048,14 @@ define <8 x float> @test_mask_vfnmsub256_ps(<8 x float> %a0, <8 x float> %a1, <8
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfnmsub132ps %ymm1, %ymm2, %ymm0 {%k1} # encoding: [0x62,0xf2,0x6d,0x29,0x9e,0xc1]
-; X86-NEXT: # ymm0 = -(ymm0 * ymm1) - ymm2
+; X86-NEXT: # ymm0 {%k1} = -(ymm0 * ymm1) - ymm2
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_mask_vfnmsub256_ps:
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfnmsub132ps %ymm1, %ymm2, %ymm0 {%k1} # encoding: [0x62,0xf2,0x6d,0x29,0x9e,0xc1]
-; X64-NEXT: # ymm0 = -(ymm0 * ymm1) - ymm2
+; X64-NEXT: # ymm0 {%k1} = -(ymm0 * ymm1) - ymm2
; X64-NEXT: retq # encoding: [0xc3]
%1 = fsub <8 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %a1
%2 = fsub <8 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %a2
@@ -6083,14 +6083,14 @@ define <4 x float> @test_mask_vfnmsub128_ps(<4 x float> %a0, <4 x float> %a1, <4
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfnmsub132ps %xmm1, %xmm2, %xmm0 {%k1} # encoding: [0x62,0xf2,0x6d,0x09,0x9e,0xc1]
-; X86-NEXT: # xmm0 = -(xmm0 * xmm1) - xmm2
+; X86-NEXT: # xmm0 {%k1} = -(xmm0 * xmm1) - xmm2
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_mask_vfnmsub128_ps:
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfnmsub132ps %xmm1, %xmm2, %xmm0 {%k1} # encoding: [0x62,0xf2,0x6d,0x09,0x9e,0xc1]
-; X64-NEXT: # xmm0 = -(xmm0 * xmm1) - xmm2
+; X64-NEXT: # xmm0 {%k1} = -(xmm0 * xmm1) - xmm2
; X64-NEXT: retq # encoding: [0xc3]
%1 = fsub <4 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %a1
%2 = fsub <4 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %a2
@@ -6119,14 +6119,14 @@ define <4 x double> @test_mask_vfnmsub256_pd(<4 x double> %a0, <4 x double> %a1,
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfnmsub132pd %ymm1, %ymm2, %ymm0 {%k1} # encoding: [0x62,0xf2,0xed,0x29,0x9e,0xc1]
-; X86-NEXT: # ymm0 = -(ymm0 * ymm1) - ymm2
+; X86-NEXT: # ymm0 {%k1} = -(ymm0 * ymm1) - ymm2
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_mask_vfnmsub256_pd:
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfnmsub132pd %ymm1, %ymm2, %ymm0 {%k1} # encoding: [0x62,0xf2,0xed,0x29,0x9e,0xc1]
-; X64-NEXT: # ymm0 = -(ymm0 * ymm1) - ymm2
+; X64-NEXT: # ymm0 {%k1} = -(ymm0 * ymm1) - ymm2
; X64-NEXT: retq # encoding: [0xc3]
%1 = fsub <4 x double> <double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00>, %a1
%2 = fsub <4 x double> <double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00>, %a2
@@ -6155,14 +6155,14 @@ define <2 x double> @test_mask_vfnmsub128_pd(<2 x double> %a0, <2 x double> %a1,
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfnmsub132pd %xmm1, %xmm2, %xmm0 {%k1} # encoding: [0x62,0xf2,0xed,0x09,0x9e,0xc1]
-; X86-NEXT: # xmm0 = -(xmm0 * xmm1) - xmm2
+; X86-NEXT: # xmm0 {%k1} = -(xmm0 * xmm1) - xmm2
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_mask_vfnmsub128_pd:
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfnmsub132pd %xmm1, %xmm2, %xmm0 {%k1} # encoding: [0x62,0xf2,0xed,0x09,0x9e,0xc1]
-; X64-NEXT: # xmm0 = -(xmm0 * xmm1) - xmm2
+; X64-NEXT: # xmm0 {%k1} = -(xmm0 * xmm1) - xmm2
; X64-NEXT: retq # encoding: [0xc3]
%1 = fsub <2 x double> <double -0.000000e+00, double -0.000000e+00>, %a1
%2 = fsub <2 x double> <double -0.000000e+00, double -0.000000e+00>, %a2
@@ -6179,7 +6179,7 @@ define <2 x double>@test_int_x86_avx512_mask3_vfnmsub_pd_128(<2 x double> %x0, <
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfnmsub231pd %xmm1, %xmm0, %xmm2 {%k1} # encoding: [0x62,0xf2,0xfd,0x09,0xbe,0xd1]
-; X86-NEXT: # xmm2 = -(xmm0 * xmm1) - xmm2
+; X86-NEXT: # xmm2 {%k1} = -(xmm0 * xmm1) - xmm2
; X86-NEXT: vmovapd %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x28,0xc2]
; X86-NEXT: retl # encoding: [0xc3]
;
@@ -6187,7 +6187,7 @@ define <2 x double>@test_int_x86_avx512_mask3_vfnmsub_pd_128(<2 x double> %x0, <
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfnmsub231pd %xmm1, %xmm0, %xmm2 {%k1} # encoding: [0x62,0xf2,0xfd,0x09,0xbe,0xd1]
-; X64-NEXT: # xmm2 = -(xmm0 * xmm1) - xmm2
+; X64-NEXT: # xmm2 {%k1} = -(xmm0 * xmm1) - xmm2
; X64-NEXT: vmovapd %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x28,0xc2]
; X64-NEXT: retq # encoding: [0xc3]
%1 = fsub <2 x double> <double -0.000000e+00, double -0.000000e+00>, %x0
@@ -6205,7 +6205,7 @@ define <4 x double>@test_int_x86_avx512_mask3_vfnmsub_pd_256(<4 x double> %x0, <
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfnmsub231pd %ymm1, %ymm0, %ymm2 {%k1} # encoding: [0x62,0xf2,0xfd,0x29,0xbe,0xd1]
-; X86-NEXT: # ymm2 = -(ymm0 * ymm1) - ymm2
+; X86-NEXT: # ymm2 {%k1} = -(ymm0 * ymm1) - ymm2
; X86-NEXT: vmovapd %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x28,0xc2]
; X86-NEXT: retl # encoding: [0xc3]
;
@@ -6213,7 +6213,7 @@ define <4 x double>@test_int_x86_avx512_mask3_vfnmsub_pd_256(<4 x double> %x0, <
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfnmsub231pd %ymm1, %ymm0, %ymm2 {%k1} # encoding: [0x62,0xf2,0xfd,0x29,0xbe,0xd1]
-; X64-NEXT: # ymm2 = -(ymm0 * ymm1) - ymm2
+; X64-NEXT: # ymm2 {%k1} = -(ymm0 * ymm1) - ymm2
; X64-NEXT: vmovapd %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x28,0xc2]
; X64-NEXT: retq # encoding: [0xc3]
%1 = fsub <4 x double> <double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00>, %x0
@@ -6231,7 +6231,7 @@ define <4 x float>@test_int_x86_avx512_mask3_vfnmsub_ps_128(<4 x float> %x0, <4
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfnmsub231ps %xmm1, %xmm0, %xmm2 {%k1} # encoding: [0x62,0xf2,0x7d,0x09,0xbe,0xd1]
-; X86-NEXT: # xmm2 = -(xmm0 * xmm1) - xmm2
+; X86-NEXT: # xmm2 {%k1} = -(xmm0 * xmm1) - xmm2
; X86-NEXT: vmovaps %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf8,0x28,0xc2]
; X86-NEXT: retl # encoding: [0xc3]
;
@@ -6239,7 +6239,7 @@ define <4 x float>@test_int_x86_avx512_mask3_vfnmsub_ps_128(<4 x float> %x0, <4
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfnmsub231ps %xmm1, %xmm0, %xmm2 {%k1} # encoding: [0x62,0xf2,0x7d,0x09,0xbe,0xd1]
-; X64-NEXT: # xmm2 = -(xmm0 * xmm1) - xmm2
+; X64-NEXT: # xmm2 {%k1} = -(xmm0 * xmm1) - xmm2
; X64-NEXT: vmovaps %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf8,0x28,0xc2]
; X64-NEXT: retq # encoding: [0xc3]
%1 = fsub <4 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %x0
@@ -6257,7 +6257,7 @@ define <8 x float>@test_int_x86_avx512_mask3_vfnmsub_ps_256(<8 x float> %x0, <8
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfnmsub231ps %ymm1, %ymm0, %ymm2 {%k1} # encoding: [0x62,0xf2,0x7d,0x29,0xbe,0xd1]
-; X86-NEXT: # ymm2 = -(ymm0 * ymm1) - ymm2
+; X86-NEXT: # ymm2 {%k1} = -(ymm0 * ymm1) - ymm2
; X86-NEXT: vmovaps %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xfc,0x28,0xc2]
; X86-NEXT: retl # encoding: [0xc3]
;
@@ -6265,7 +6265,7 @@ define <8 x float>@test_int_x86_avx512_mask3_vfnmsub_ps_256(<8 x float> %x0, <8
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfnmsub231ps %ymm1, %ymm0, %ymm2 {%k1} # encoding: [0x62,0xf2,0x7d,0x29,0xbe,0xd1]
-; X64-NEXT: # ymm2 = -(ymm0 * ymm1) - ymm2
+; X64-NEXT: # ymm2 {%k1} = -(ymm0 * ymm1) - ymm2
; X64-NEXT: vmovaps %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xfc,0x28,0xc2]
; X64-NEXT: retq # encoding: [0xc3]
%1 = fsub <8 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %x0
@@ -6295,14 +6295,14 @@ define <8 x float> @test_mask_fmaddsub256_ps(<8 x float> %a, <8 x float> %b, <8
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmaddsub132ps %ymm1, %ymm2, %ymm0 {%k1} # encoding: [0x62,0xf2,0x6d,0x29,0x96,0xc1]
-; X86-NEXT: # ymm0 = (ymm0 * ymm1) +/- ymm2
+; X86-NEXT: # ymm0 {%k1} = (ymm0 * ymm1) +/- ymm2
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_mask_fmaddsub256_ps:
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmaddsub132ps %ymm1, %ymm2, %ymm0 {%k1} # encoding: [0x62,0xf2,0x6d,0x29,0x96,0xc1]
-; X64-NEXT: # ymm0 = (ymm0 * ymm1) +/- ymm2
+; X64-NEXT: # ymm0 {%k1} = (ymm0 * ymm1) +/- ymm2
; X64-NEXT: retq # encoding: [0xc3]
%1 = call <8 x float> @llvm.fma.v8f32(<8 x float> %a, <8 x float> %b, <8 x float> %c)
%2 = fsub <8 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %c
@@ -6332,14 +6332,14 @@ define <4 x float> @test_mask_fmaddsub128_ps(<4 x float> %a, <4 x float> %b, <4
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmaddsub132ps %xmm1, %xmm2, %xmm0 {%k1} # encoding: [0x62,0xf2,0x6d,0x09,0x96,0xc1]
-; X86-NEXT: # xmm0 = (xmm0 * xmm1) +/- xmm2
+; X86-NEXT: # xmm0 {%k1} = (xmm0 * xmm1) +/- xmm2
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_mask_fmaddsub128_ps:
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmaddsub132ps %xmm1, %xmm2, %xmm0 {%k1} # encoding: [0x62,0xf2,0x6d,0x09,0x96,0xc1]
-; X64-NEXT: # xmm0 = (xmm0 * xmm1) +/- xmm2
+; X64-NEXT: # xmm0 {%k1} = (xmm0 * xmm1) +/- xmm2
; X64-NEXT: retq # encoding: [0xc3]
%1 = call <4 x float> @llvm.fma.v4f32(<4 x float> %a, <4 x float> %b, <4 x float> %c)
%2 = fsub <4 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %c
@@ -6370,14 +6370,14 @@ define <4 x double> @test_mask_vfmaddsub256_pd(<4 x double> %a0, <4 x double> %a
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmaddsub132pd %ymm1, %ymm2, %ymm0 {%k1} # encoding: [0x62,0xf2,0xed,0x29,0x96,0xc1]
-; X86-NEXT: # ymm0 = (ymm0 * ymm1) +/- ymm2
+; X86-NEXT: # ymm0 {%k1} = (ymm0 * ymm1) +/- ymm2
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_mask_vfmaddsub256_pd:
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmaddsub132pd %ymm1, %ymm2, %ymm0 {%k1} # encoding: [0x62,0xf2,0xed,0x29,0x96,0xc1]
-; X64-NEXT: # ymm0 = (ymm0 * ymm1) +/- ymm2
+; X64-NEXT: # ymm0 {%k1} = (ymm0 * ymm1) +/- ymm2
; X64-NEXT: retq # encoding: [0xc3]
%1 = call <4 x double> @llvm.fma.v4f64(<4 x double> %a0, <4 x double> %a1, <4 x double> %a2)
%2 = fsub <4 x double> <double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00>, %a2
@@ -6408,14 +6408,14 @@ define <2 x double> @test_mask_vfmaddsub128_pd(<2 x double> %a0, <2 x double> %a
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmaddsub132pd %xmm1, %xmm2, %xmm0 {%k1} # encoding: [0x62,0xf2,0xed,0x09,0x96,0xc1]
-; X86-NEXT: # xmm0 = (xmm0 * xmm1) +/- xmm2
+; X86-NEXT: # xmm0 {%k1} = (xmm0 * xmm1) +/- xmm2
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_mask_vfmaddsub128_pd:
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmaddsub132pd %xmm1, %xmm2, %xmm0 {%k1} # encoding: [0x62,0xf2,0xed,0x09,0x96,0xc1]
-; X64-NEXT: # xmm0 = (xmm0 * xmm1) +/- xmm2
+; X64-NEXT: # xmm0 {%k1} = (xmm0 * xmm1) +/- xmm2
; X64-NEXT: retq # encoding: [0xc3]
%1 = call <2 x double> @llvm.fma.v2f64(<2 x double> %a0, <2 x double> %a1, <2 x double> %a2)
%2 = fsub <2 x double> <double -0.000000e+00, double -0.000000e+00>, %a2
@@ -6433,7 +6433,7 @@ define <2 x double>@test_int_x86_avx512_mask3_vfmaddsub_pd_128(<2 x double> %x0,
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmaddsub231pd %xmm1, %xmm0, %xmm2 {%k1} # encoding: [0x62,0xf2,0xfd,0x09,0xb6,0xd1]
-; X86-NEXT: # xmm2 = (xmm0 * xmm1) +/- xmm2
+; X86-NEXT: # xmm2 {%k1} = (xmm0 * xmm1) +/- xmm2
; X86-NEXT: vmovapd %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x28,0xc2]
; X86-NEXT: retl # encoding: [0xc3]
;
@@ -6441,7 +6441,7 @@ define <2 x double>@test_int_x86_avx512_mask3_vfmaddsub_pd_128(<2 x double> %x0,
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmaddsub231pd %xmm1, %xmm0, %xmm2 {%k1} # encoding: [0x62,0xf2,0xfd,0x09,0xb6,0xd1]
-; X64-NEXT: # xmm2 = (xmm0 * xmm1) +/- xmm2
+; X64-NEXT: # xmm2 {%k1} = (xmm0 * xmm1) +/- xmm2
; X64-NEXT: vmovapd %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x28,0xc2]
; X64-NEXT: retq # encoding: [0xc3]
%1 = call <2 x double> @llvm.fma.v2f64(<2 x double> %x0, <2 x double> %x1, <2 x double> %x2)
@@ -6460,14 +6460,14 @@ define <2 x double>@test_int_x86_avx512_maskz_vfmaddsub_pd_128(<2 x double> %x0,
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmaddsub213pd %xmm2, %xmm1, %xmm0 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0x89,0xa6,0xc2]
-; X86-NEXT: # xmm0 = (xmm1 * xmm0) +/- xmm2
+; X86-NEXT: # xmm0 {%k1} {z} = (xmm1 * xmm0) +/- xmm2
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_maskz_vfmaddsub_pd_128:
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmaddsub213pd %xmm2, %xmm1, %xmm0 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0x89,0xa6,0xc2]
-; X64-NEXT: # xmm0 = (xmm1 * xmm0) +/- xmm2
+; X64-NEXT: # xmm0 {%k1} {z} = (xmm1 * xmm0) +/- xmm2
; X64-NEXT: retq # encoding: [0xc3]
%1 = call <2 x double> @llvm.fma.v2f64(<2 x double> %x0, <2 x double> %x1, <2 x double> %x2)
%2 = fsub <2 x double> <double -0.000000e+00, double -0.000000e+00>, %x2
@@ -6485,7 +6485,7 @@ define <4 x double>@test_int_x86_avx512_mask3_vfmaddsub_pd_256(<4 x double> %x0,
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmaddsub231pd %ymm1, %ymm0, %ymm2 {%k1} # encoding: [0x62,0xf2,0xfd,0x29,0xb6,0xd1]
-; X86-NEXT: # ymm2 = (ymm0 * ymm1) +/- ymm2
+; X86-NEXT: # ymm2 {%k1} = (ymm0 * ymm1) +/- ymm2
; X86-NEXT: vmovapd %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x28,0xc2]
; X86-NEXT: retl # encoding: [0xc3]
;
@@ -6493,7 +6493,7 @@ define <4 x double>@test_int_x86_avx512_mask3_vfmaddsub_pd_256(<4 x double> %x0,
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmaddsub231pd %ymm1, %ymm0, %ymm2 {%k1} # encoding: [0x62,0xf2,0xfd,0x29,0xb6,0xd1]
-; X64-NEXT: # ymm2 = (ymm0 * ymm1) +/- ymm2
+; X64-NEXT: # ymm2 {%k1} = (ymm0 * ymm1) +/- ymm2
; X64-NEXT: vmovapd %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x28,0xc2]
; X64-NEXT: retq # encoding: [0xc3]
%1 = call <4 x double> @llvm.fma.v4f64(<4 x double> %x0, <4 x double> %x1, <4 x double> %x2)
@@ -6512,14 +6512,14 @@ define <4 x double>@test_int_x86_avx512_maskz_vfmaddsub_pd_256(<4 x double> %x0,
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmaddsub213pd %ymm2, %ymm1, %ymm0 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0xa9,0xa6,0xc2]
-; X86-NEXT: # ymm0 = (ymm1 * ymm0) +/- ymm2
+; X86-NEXT: # ymm0 {%k1} {z} = (ymm1 * ymm0) +/- ymm2
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_maskz_vfmaddsub_pd_256:
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmaddsub213pd %ymm2, %ymm1, %ymm0 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0xa9,0xa6,0xc2]
-; X64-NEXT: # ymm0 = (ymm1 * ymm0) +/- ymm2
+; X64-NEXT: # ymm0 {%k1} {z} = (ymm1 * ymm0) +/- ymm2
; X64-NEXT: retq # encoding: [0xc3]
%1 = call <4 x double> @llvm.fma.v4f64(<4 x double> %x0, <4 x double> %x1, <4 x double> %x2)
%2 = fsub <4 x double> <double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00>, %x2
@@ -6537,7 +6537,7 @@ define <4 x float>@test_int_x86_avx512_mask3_vfmaddsub_ps_128(<4 x float> %x0, <
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmaddsub231ps %xmm1, %xmm0, %xmm2 {%k1} # encoding: [0x62,0xf2,0x7d,0x09,0xb6,0xd1]
-; X86-NEXT: # xmm2 = (xmm0 * xmm1) +/- xmm2
+; X86-NEXT: # xmm2 {%k1} = (xmm0 * xmm1) +/- xmm2
; X86-NEXT: vmovaps %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf8,0x28,0xc2]
; X86-NEXT: retl # encoding: [0xc3]
;
@@ -6545,7 +6545,7 @@ define <4 x float>@test_int_x86_avx512_mask3_vfmaddsub_ps_128(<4 x float> %x0, <
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmaddsub231ps %xmm1, %xmm0, %xmm2 {%k1} # encoding: [0x62,0xf2,0x7d,0x09,0xb6,0xd1]
-; X64-NEXT: # xmm2 = (xmm0 * xmm1) +/- xmm2
+; X64-NEXT: # xmm2 {%k1} = (xmm0 * xmm1) +/- xmm2
; X64-NEXT: vmovaps %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf8,0x28,0xc2]
; X64-NEXT: retq # encoding: [0xc3]
%1 = call <4 x float> @llvm.fma.v4f32(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2)
@@ -6564,14 +6564,14 @@ define <4 x float>@test_int_x86_avx512_maskz_vfmaddsub_ps_128(<4 x float> %x0, <
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmaddsub213ps %xmm2, %xmm1, %xmm0 {%k1} {z} # encoding: [0x62,0xf2,0x75,0x89,0xa6,0xc2]
-; X86-NEXT: # xmm0 = (xmm1 * xmm0) +/- xmm2
+; X86-NEXT: # xmm0 {%k1} {z} = (xmm1 * xmm0) +/- xmm2
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_maskz_vfmaddsub_ps_128:
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmaddsub213ps %xmm2, %xmm1, %xmm0 {%k1} {z} # encoding: [0x62,0xf2,0x75,0x89,0xa6,0xc2]
-; X64-NEXT: # xmm0 = (xmm1 * xmm0) +/- xmm2
+; X64-NEXT: # xmm0 {%k1} {z} = (xmm1 * xmm0) +/- xmm2
; X64-NEXT: retq # encoding: [0xc3]
%1 = call <4 x float> @llvm.fma.v4f32(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2)
%2 = fsub <4 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %x2
@@ -6589,7 +6589,7 @@ define <8 x float>@test_int_x86_avx512_mask3_vfmaddsub_ps_256(<8 x float> %x0, <
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmaddsub231ps %ymm1, %ymm0, %ymm2 {%k1} # encoding: [0x62,0xf2,0x7d,0x29,0xb6,0xd1]
-; X86-NEXT: # ymm2 = (ymm0 * ymm1) +/- ymm2
+; X86-NEXT: # ymm2 {%k1} = (ymm0 * ymm1) +/- ymm2
; X86-NEXT: vmovaps %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xfc,0x28,0xc2]
; X86-NEXT: retl # encoding: [0xc3]
;
@@ -6597,7 +6597,7 @@ define <8 x float>@test_int_x86_avx512_mask3_vfmaddsub_ps_256(<8 x float> %x0, <
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmaddsub231ps %ymm1, %ymm0, %ymm2 {%k1} # encoding: [0x62,0xf2,0x7d,0x29,0xb6,0xd1]
-; X64-NEXT: # ymm2 = (ymm0 * ymm1) +/- ymm2
+; X64-NEXT: # ymm2 {%k1} = (ymm0 * ymm1) +/- ymm2
; X64-NEXT: vmovaps %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xfc,0x28,0xc2]
; X64-NEXT: retq # encoding: [0xc3]
%1 = call <8 x float> @llvm.fma.v8f32(<8 x float> %x0, <8 x float> %x1, <8 x float> %x2)
@@ -6615,14 +6615,14 @@ define <8 x float>@test_int_x86_avx512_maskz_vfmaddsub_ps_256(<8 x float> %x0, <
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmaddsub213ps %ymm2, %ymm1, %ymm0 {%k1} {z} # encoding: [0x62,0xf2,0x75,0xa9,0xa6,0xc2]
-; X86-NEXT: # ymm0 = (ymm1 * ymm0) +/- ymm2
+; X86-NEXT: # ymm0 {%k1} {z} = (ymm1 * ymm0) +/- ymm2
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_maskz_vfmaddsub_ps_256:
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmaddsub213ps %ymm2, %ymm1, %ymm0 {%k1} {z} # encoding: [0x62,0xf2,0x75,0xa9,0xa6,0xc2]
-; X64-NEXT: # ymm0 = (ymm1 * ymm0) +/- ymm2
+; X64-NEXT: # ymm0 {%k1} {z} = (ymm1 * ymm0) +/- ymm2
; X64-NEXT: retq # encoding: [0xc3]
%1 = call <8 x float> @llvm.fma.v8f32(<8 x float> %x0, <8 x float> %x1, <8 x float> %x2)
%2 = fsub <8 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %x2
@@ -6639,7 +6639,7 @@ define <2 x double>@test_int_x86_avx512_mask3_vfmsubadd_pd_128(<2 x double> %x0,
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmsubadd231pd %xmm1, %xmm0, %xmm2 {%k1} # encoding: [0x62,0xf2,0xfd,0x09,0xb7,0xd1]
-; X86-NEXT: # xmm2 = (xmm0 * xmm1) -/+ xmm2
+; X86-NEXT: # xmm2 {%k1} = (xmm0 * xmm1) -/+ xmm2
; X86-NEXT: vmovapd %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x28,0xc2]
; X86-NEXT: retl # encoding: [0xc3]
;
@@ -6647,7 +6647,7 @@ define <2 x double>@test_int_x86_avx512_mask3_vfmsubadd_pd_128(<2 x double> %x0,
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmsubadd231pd %xmm1, %xmm0, %xmm2 {%k1} # encoding: [0x62,0xf2,0xfd,0x09,0xb7,0xd1]
-; X64-NEXT: # xmm2 = (xmm0 * xmm1) -/+ xmm2
+; X64-NEXT: # xmm2 {%k1} = (xmm0 * xmm1) -/+ xmm2
; X64-NEXT: vmovapd %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x28,0xc2]
; X64-NEXT: retq # encoding: [0xc3]
%1 = call <2 x double> @llvm.fma.v2f64(<2 x double> %x0, <2 x double> %x1, <2 x double> %x2)
@@ -6666,7 +6666,7 @@ define <4 x double>@test_int_x86_avx512_mask3_vfmsubadd_pd_256(<4 x double> %x0,
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmsubadd231pd %ymm1, %ymm0, %ymm2 {%k1} # encoding: [0x62,0xf2,0xfd,0x29,0xb7,0xd1]
-; X86-NEXT: # ymm2 = (ymm0 * ymm1) -/+ ymm2
+; X86-NEXT: # ymm2 {%k1} = (ymm0 * ymm1) -/+ ymm2
; X86-NEXT: vmovapd %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x28,0xc2]
; X86-NEXT: retl # encoding: [0xc3]
;
@@ -6674,7 +6674,7 @@ define <4 x double>@test_int_x86_avx512_mask3_vfmsubadd_pd_256(<4 x double> %x0,
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmsubadd231pd %ymm1, %ymm0, %ymm2 {%k1} # encoding: [0x62,0xf2,0xfd,0x29,0xb7,0xd1]
-; X64-NEXT: # ymm2 = (ymm0 * ymm1) -/+ ymm2
+; X64-NEXT: # ymm2 {%k1} = (ymm0 * ymm1) -/+ ymm2
; X64-NEXT: vmovapd %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x28,0xc2]
; X64-NEXT: retq # encoding: [0xc3]
%1 = call <4 x double> @llvm.fma.v4f64(<4 x double> %x0, <4 x double> %x1, <4 x double> %x2)
@@ -6693,7 +6693,7 @@ define <4 x float>@test_int_x86_avx512_mask3_vfmsubadd_ps_128(<4 x float> %x0, <
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmsubadd231ps %xmm1, %xmm0, %xmm2 {%k1} # encoding: [0x62,0xf2,0x7d,0x09,0xb7,0xd1]
-; X86-NEXT: # xmm2 = (xmm0 * xmm1) -/+ xmm2
+; X86-NEXT: # xmm2 {%k1} = (xmm0 * xmm1) -/+ xmm2
; X86-NEXT: vmovaps %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf8,0x28,0xc2]
; X86-NEXT: retl # encoding: [0xc3]
;
@@ -6701,7 +6701,7 @@ define <4 x float>@test_int_x86_avx512_mask3_vfmsubadd_ps_128(<4 x float> %x0, <
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmsubadd231ps %xmm1, %xmm0, %xmm2 {%k1} # encoding: [0x62,0xf2,0x7d,0x09,0xb7,0xd1]
-; X64-NEXT: # xmm2 = (xmm0 * xmm1) -/+ xmm2
+; X64-NEXT: # xmm2 {%k1} = (xmm0 * xmm1) -/+ xmm2
; X64-NEXT: vmovaps %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf8,0x28,0xc2]
; X64-NEXT: retq # encoding: [0xc3]
%1 = call <4 x float> @llvm.fma.v4f32(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2)
@@ -6720,7 +6720,7 @@ define <8 x float>@test_int_x86_avx512_mask3_vfmsubadd_ps_256(<8 x float> %x0, <
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vfmsubadd231ps %ymm1, %ymm0, %ymm2 {%k1} # encoding: [0x62,0xf2,0x7d,0x29,0xb7,0xd1]
-; X86-NEXT: # ymm2 = (ymm0 * ymm1) -/+ ymm2
+; X86-NEXT: # ymm2 {%k1} = (ymm0 * ymm1) -/+ ymm2
; X86-NEXT: vmovaps %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xfc,0x28,0xc2]
; X86-NEXT: retl # encoding: [0xc3]
;
@@ -6728,7 +6728,7 @@ define <8 x float>@test_int_x86_avx512_mask3_vfmsubadd_ps_256(<8 x float> %x0, <
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vfmsubadd231ps %ymm1, %ymm0, %ymm2 {%k1} # encoding: [0x62,0xf2,0x7d,0x29,0xb7,0xd1]
-; X64-NEXT: # ymm2 = (ymm0 * ymm1) -/+ ymm2
+; X64-NEXT: # ymm2 {%k1} = (ymm0 * ymm1) -/+ ymm2
; X64-NEXT: vmovaps %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xfc,0x28,0xc2]
; X64-NEXT: retq # encoding: [0xc3]
%1 = call <8 x float> @llvm.fma.v8f32(<8 x float> %x0, <8 x float> %x1, <8 x float> %x2)
@@ -6747,14 +6747,14 @@ define <4 x float> @test_mask_vfmadd128_ps_rmk(<4 x float> %a0, <4 x float> %a1,
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %ecx # encoding: [0x0f,0xb6,0x4c,0x24,0x08]
; X86-NEXT: kmovw %ecx, %k1 # encoding: [0xc5,0xf8,0x92,0xc9]
; X86-NEXT: vfmadd213ps (%eax), %xmm1, %xmm0 {%k1} # encoding: [0x62,0xf2,0x75,0x09,0xa8,0x00]
-; X86-NEXT: # xmm0 = (xmm1 * xmm0) + mem
+; X86-NEXT: # xmm0 {%k1} = (xmm1 * xmm0) + mem
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_mask_vfmadd128_ps_rmk:
; X64: # %bb.0:
; X64-NEXT: kmovw %esi, %k1 # encoding: [0xc5,0xf8,0x92,0xce]
; X64-NEXT: vfmadd213ps (%rdi), %xmm1, %xmm0 {%k1} # encoding: [0x62,0xf2,0x75,0x09,0xa8,0x07]
-; X64-NEXT: # xmm0 = (xmm1 * xmm0) + mem
+; X64-NEXT: # xmm0 {%k1} = (xmm1 * xmm0) + mem
; X64-NEXT: retq # encoding: [0xc3]
%a2 = load <4 x float>, <4 x float>* %ptr_a2
%1 = call <4 x float> @llvm.fma.v4f32(<4 x float> %a0, <4 x float> %a1, <4 x float> %a2)
@@ -6771,14 +6771,14 @@ define <4 x float> @test_mask_vfmadd128_ps_rmka(<4 x float> %a0, <4 x float> %a1
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %ecx # encoding: [0x0f,0xb6,0x4c,0x24,0x08]
; X86-NEXT: kmovw %ecx, %k1 # encoding: [0xc5,0xf8,0x92,0xc9]
; X86-NEXT: vfmadd213ps (%eax), %xmm1, %xmm0 {%k1} # encoding: [0x62,0xf2,0x75,0x09,0xa8,0x00]
-; X86-NEXT: # xmm0 = (xmm1 * xmm0) + mem
+; X86-NEXT: # xmm0 {%k1} = (xmm1 * xmm0) + mem
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_mask_vfmadd128_ps_rmka:
; X64: # %bb.0:
; X64-NEXT: kmovw %esi, %k1 # encoding: [0xc5,0xf8,0x92,0xce]
; X64-NEXT: vfmadd213ps (%rdi), %xmm1, %xmm0 {%k1} # encoding: [0x62,0xf2,0x75,0x09,0xa8,0x07]
-; X64-NEXT: # xmm0 = (xmm1 * xmm0) + mem
+; X64-NEXT: # xmm0 {%k1} = (xmm1 * xmm0) + mem
; X64-NEXT: retq # encoding: [0xc3]
%a2 = load <4 x float>, <4 x float>* %ptr_a2, align 8
%1 = call <4 x float> @llvm.fma.v4f32(<4 x float> %a0, <4 x float> %a1, <4 x float> %a2)
@@ -6831,14 +6831,14 @@ define <4 x float> @test_mask_vfmadd128_ps_rmb(<4 x float> %a0, <4 x float> %a1,
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %ecx # encoding: [0x0f,0xb6,0x4c,0x24,0x08]
; X86-NEXT: kmovw %ecx, %k1 # encoding: [0xc5,0xf8,0x92,0xc9]
; X86-NEXT: vfmadd213ps (%eax){1to4}, %xmm1, %xmm0 {%k1} # encoding: [0x62,0xf2,0x75,0x19,0xa8,0x00]
-; X86-NEXT: # xmm0 = (xmm1 * xmm0) + mem
+; X86-NEXT: # xmm0 {%k1} = (xmm1 * xmm0) + mem
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_mask_vfmadd128_ps_rmb:
; X64: # %bb.0:
; X64-NEXT: kmovw %esi, %k1 # encoding: [0xc5,0xf8,0x92,0xce]
; X64-NEXT: vfmadd213ps (%rdi){1to4}, %xmm1, %xmm0 {%k1} # encoding: [0x62,0xf2,0x75,0x19,0xa8,0x07]
-; X64-NEXT: # xmm0 = (xmm1 * xmm0) + mem
+; X64-NEXT: # xmm0 {%k1} = (xmm1 * xmm0) + mem
; X64-NEXT: retq # encoding: [0xc3]
%q = load float, float* %ptr_a2
%vecinit.i = insertelement <4 x float> undef, float %q, i32 0
@@ -6859,14 +6859,14 @@ define <4 x float> @test_mask_vfmadd128_ps_rmba(<4 x float> %a0, <4 x float> %a1
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %ecx # encoding: [0x0f,0xb6,0x4c,0x24,0x08]
; X86-NEXT: kmovw %ecx, %k1 # encoding: [0xc5,0xf8,0x92,0xc9]
; X86-NEXT: vfmadd213ps (%eax){1to4}, %xmm1, %xmm0 {%k1} # encoding: [0x62,0xf2,0x75,0x19,0xa8,0x00]
-; X86-NEXT: # xmm0 = (xmm1 * xmm0) + mem
+; X86-NEXT: # xmm0 {%k1} = (xmm1 * xmm0) + mem
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_mask_vfmadd128_ps_rmba:
; X64: # %bb.0:
; X64-NEXT: kmovw %esi, %k1 # encoding: [0xc5,0xf8,0x92,0xce]
; X64-NEXT: vfmadd213ps (%rdi){1to4}, %xmm1, %xmm0 {%k1} # encoding: [0x62,0xf2,0x75,0x19,0xa8,0x07]
-; X64-NEXT: # xmm0 = (xmm1 * xmm0) + mem
+; X64-NEXT: # xmm0 {%k1} = (xmm1 * xmm0) + mem
; X64-NEXT: retq # encoding: [0xc3]
%q = load float, float* %ptr_a2, align 4
%vecinit.i = insertelement <4 x float> undef, float %q, i32 0
@@ -6931,14 +6931,14 @@ define <2 x double> @test_mask_vfmadd128_pd_rmk(<2 x double> %a0, <2 x double> %
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %ecx # encoding: [0x0f,0xb6,0x4c,0x24,0x08]
; X86-NEXT: kmovw %ecx, %k1 # encoding: [0xc5,0xf8,0x92,0xc9]
; X86-NEXT: vfmadd213pd (%eax), %xmm1, %xmm0 {%k1} # encoding: [0x62,0xf2,0xf5,0x09,0xa8,0x00]
-; X86-NEXT: # xmm0 = (xmm1 * xmm0) + mem
+; X86-NEXT: # xmm0 {%k1} = (xmm1 * xmm0) + mem
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_mask_vfmadd128_pd_rmk:
; X64: # %bb.0:
; X64-NEXT: kmovw %esi, %k1 # encoding: [0xc5,0xf8,0x92,0xce]
; X64-NEXT: vfmadd213pd (%rdi), %xmm1, %xmm0 {%k1} # encoding: [0x62,0xf2,0xf5,0x09,0xa8,0x07]
-; X64-NEXT: # xmm0 = (xmm1 * xmm0) + mem
+; X64-NEXT: # xmm0 {%k1} = (xmm1 * xmm0) + mem
; X64-NEXT: retq # encoding: [0xc3]
%a2 = load <2 x double>, <2 x double>* %ptr_a2
%1 = call <2 x double> @llvm.fma.v2f64(<2 x double> %a0, <2 x double> %a1, <2 x double> %a2)
@@ -6973,14 +6973,14 @@ define <4 x double> @test_mask_vfmadd256_pd_rmk(<4 x double> %a0, <4 x double> %
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %ecx # encoding: [0x0f,0xb6,0x4c,0x24,0x08]
; X86-NEXT: kmovw %ecx, %k1 # encoding: [0xc5,0xf8,0x92,0xc9]
; X86-NEXT: vfmadd213pd (%eax), %ymm1, %ymm0 {%k1} # encoding: [0x62,0xf2,0xf5,0x29,0xa8,0x00]
-; X86-NEXT: # ymm0 = (ymm1 * ymm0) + mem
+; X86-NEXT: # ymm0 {%k1} = (ymm1 * ymm0) + mem
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_mask_vfmadd256_pd_rmk:
; X64: # %bb.0:
; X64-NEXT: kmovw %esi, %k1 # encoding: [0xc5,0xf8,0x92,0xce]
; X64-NEXT: vfmadd213pd (%rdi), %ymm1, %ymm0 {%k1} # encoding: [0x62,0xf2,0xf5,0x29,0xa8,0x07]
-; X64-NEXT: # ymm0 = (ymm1 * ymm0) + mem
+; X64-NEXT: # ymm0 {%k1} = (ymm1 * ymm0) + mem
; X64-NEXT: retq # encoding: [0xc3]
%a2 = load <4 x double>, <4 x double>* %ptr_a2
%1 = call <4 x double> @llvm.fma.v4f64(<4 x double> %a0, <4 x double> %a1, <4 x double> %a2)
diff --git a/llvm/test/CodeGen/X86/fma-fneg-combine.ll b/llvm/test/CodeGen/X86/fma-fneg-combine.ll
index d583c54a086b..a67595c1e95a 100644
--- a/llvm/test/CodeGen/X86/fma-fneg-combine.ll
+++ b/llvm/test/CodeGen/X86/fma-fneg-combine.ll
@@ -160,13 +160,13 @@ define <4 x float> @test11b(<4 x float> %a, <4 x float> %b, <4 x float> %c, i8 z
; SKX-LABEL: test11b:
; SKX: # %bb.0: # %entry
; SKX-NEXT: kmovd %edi, %k1
-; SKX-NEXT: vfmsub213ss {{.*#+}} xmm0 = (xmm1 * xmm0) - xmm2
+; SKX-NEXT: vfmsub213ss {{.*#+}} xmm0 {%k1} = (xmm1 * xmm0) - xmm2
; SKX-NEXT: retq
;
; KNL-LABEL: test11b:
; KNL: # %bb.0: # %entry
; KNL-NEXT: kmovw %edi, %k1
-; KNL-NEXT: vfmsub213ss {{.*#+}} xmm0 = (xmm1 * xmm0) - xmm2
+; KNL-NEXT: vfmsub213ss {{.*#+}} xmm0 {%k1} = (xmm1 * xmm0) - xmm2
; KNL-NEXT: retq
entry:
%sub.i = fsub <4 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %c
@@ -180,14 +180,14 @@ define <8 x double> @test12(<8 x double> %a, <8 x double> %b, <8 x double> %c, i
; SKX-LABEL: test12:
; SKX: # %bb.0: # %entry
; SKX-NEXT: kmovd %edi, %k1
-; SKX-NEXT: vfmadd132pd {{.*#+}} zmm0 = (zmm0 * zmm1) + zmm2
+; SKX-NEXT: vfmadd132pd {{.*#+}} zmm0 {%k1} = (zmm0 * zmm1) + zmm2
; SKX-NEXT: vxorpd {{.*}}(%rip){1to8}, %zmm0, %zmm0
; SKX-NEXT: retq
;
; KNL-LABEL: test12:
; KNL: # %bb.0: # %entry
; KNL-NEXT: kmovw %edi, %k1
-; KNL-NEXT: vfmadd132pd {{.*#+}} zmm0 = (zmm0 * zmm1) + zmm2
+; KNL-NEXT: vfmadd132pd {{.*#+}} zmm0 {%k1} = (zmm0 * zmm1) + zmm2
; KNL-NEXT: vpxorq {{.*}}(%rip){1to8}, %zmm0, %zmm0
; KNL-NEXT: retq
entry:
@@ -297,13 +297,13 @@ define <8 x double> @test17(<8 x double> %a, <8 x double> %b, <8 x double> %c, i
; SKX-LABEL: test17:
; SKX: # %bb.0:
; SKX-NEXT: kmovd %edi, %k1
-; SKX-NEXT: vfmsubadd132pd {{.*#+}} zmm0 = (zmm0 * zmm1) -/+ zmm2
+; SKX-NEXT: vfmsubadd132pd {{.*#+}} zmm0 {%k1} = (zmm0 * zmm1) -/+ zmm2
; SKX-NEXT: retq
;
; KNL-LABEL: test17:
; KNL: # %bb.0:
; KNL-NEXT: kmovw %edi, %k1
-; KNL-NEXT: vfmsubadd132pd {{.*#+}} zmm0 = (zmm0 * zmm1) -/+ zmm2
+; KNL-NEXT: vfmsubadd132pd {{.*#+}} zmm0 {%k1} = (zmm0 * zmm1) -/+ zmm2
; KNL-NEXT: retq
%sub.i = fsub <8 x double> <double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00>, %c
%res = call <8 x double> @llvm.x86.avx512.vfmaddsub.pd.512(<8 x double> %a, <8 x double> %b, <8 x double> %sub.i, i32 4)
@@ -317,13 +317,13 @@ define <4 x float> @test18(<4 x float> %a, <4 x float> %b, <4 x float> %c, i8 ze
; SKX-LABEL: test18:
; SKX: # %bb.0: # %entry
; SKX-NEXT: kmovd %edi, %k1
-; SKX-NEXT: vfnmadd213ss {{.*#+}} xmm0 = -(xmm1 * xmm0) + xmm2
+; SKX-NEXT: vfnmadd213ss {{.*#+}} xmm0 {%k1} = -(xmm1 * xmm0) + xmm2
; SKX-NEXT: retq
;
; KNL-LABEL: test18:
; KNL: # %bb.0: # %entry
; KNL-NEXT: kmovw %edi, %k1
-; KNL-NEXT: vfnmadd213ss {{.*#+}} xmm0 = -(xmm1 * xmm0) + xmm2
+; KNL-NEXT: vfnmadd213ss {{.*#+}} xmm0 {%k1} = -(xmm1 * xmm0) + xmm2
; KNL-NEXT: retq
entry:
%sub.i = fsub <4 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %b
@@ -335,13 +335,13 @@ define <4 x float> @test19(<4 x float> %a, <4 x float> %b, <4 x float> %c, i8 ze
; SKX-LABEL: test19:
; SKX: # %bb.0: # %entry
; SKX-NEXT: kmovd %edi, %k1
-; SKX-NEXT: vfnmsub213ss {{.*#+}} xmm0 = -(xmm1 * xmm0) - xmm2
+; SKX-NEXT: vfnmsub213ss {{.*#+}} xmm0 {%k1} = -(xmm1 * xmm0) - xmm2
; SKX-NEXT: retq
;
; KNL-LABEL: test19:
; KNL: # %bb.0: # %entry
; KNL-NEXT: kmovw %edi, %k1
-; KNL-NEXT: vfnmsub213ss {{.*#+}} xmm0 = -(xmm1 * xmm0) - xmm2
+; KNL-NEXT: vfnmsub213ss {{.*#+}} xmm0 {%k1} = -(xmm1 * xmm0) - xmm2
; KNL-NEXT: retq
entry:
%sub.i = fsub <4 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %b
@@ -354,14 +354,14 @@ define <4 x float> @test20(<4 x float> %a, <4 x float> %b, <4 x float> %c, i8 ze
; SKX-LABEL: test20:
; SKX: # %bb.0: # %entry
; SKX-NEXT: kmovd %edi, %k1
-; SKX-NEXT: vfnmadd231ss {{.*#+}} xmm2 = -(xmm0 * xmm1) + xmm2
+; SKX-NEXT: vfnmadd231ss {{.*#+}} xmm2 {%k1} = -(xmm0 * xmm1) + xmm2
; SKX-NEXT: vmovaps %xmm2, %xmm0
; SKX-NEXT: retq
;
; KNL-LABEL: test20:
; KNL: # %bb.0: # %entry
; KNL-NEXT: kmovw %edi, %k1
-; KNL-NEXT: vfnmadd231ss {{.*#+}} xmm2 = -(xmm0 * xmm1) + xmm2
+; KNL-NEXT: vfnmadd231ss {{.*#+}} xmm2 {%k1} = -(xmm0 * xmm1) + xmm2
; KNL-NEXT: vmovaps %xmm2, %xmm0
; KNL-NEXT: retq
entry:
diff --git a/llvm/test/CodeGen/X86/fma-scalar-combine.ll b/llvm/test/CodeGen/X86/fma-scalar-combine.ll
index f8e823525b85..aa7374747cf8 100644
--- a/llvm/test/CodeGen/X86/fma-scalar-combine.ll
+++ b/llvm/test/CodeGen/X86/fma-scalar-combine.ll
@@ -6,7 +6,7 @@ define <2 x double> @combine_scalar_mask_fmadd_f32(<2 x double> %a, i8 zeroext %
; CHECK: # %bb.0: # %entry
; CHECK-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT: vfmadd213ss %xmm2, %xmm1, %xmm0 {%k1} # encoding: [0x62,0xf2,0x75,0x09,0xa9,0xc2]
-; CHECK-NEXT: # xmm0 = (xmm1 * xmm0) + xmm2
+; CHECK-NEXT: # xmm0 {%k1} = (xmm1 * xmm0) + xmm2
; CHECK-NEXT: retq # encoding: [0xc3]
entry:
%0 = bitcast <2 x double> %a to <4 x float>
@@ -30,7 +30,7 @@ define <2 x double> @combine_scalar_mask_fmadd_f64(<2 x double> %a, i8 zeroext %
; CHECK: # %bb.0: # %entry
; CHECK-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT: vfmadd213sd %xmm2, %xmm1, %xmm0 {%k1} # encoding: [0x62,0xf2,0xf5,0x09,0xa9,0xc2]
-; CHECK-NEXT: # xmm0 = (xmm1 * xmm0) + xmm2
+; CHECK-NEXT: # xmm0 {%k1} = (xmm1 * xmm0) + xmm2
; CHECK-NEXT: retq # encoding: [0xc3]
entry:
%0 = extractelement <2 x double> %a, i64 0
@@ -50,7 +50,7 @@ define <2 x double> @combine_scalar_maskz_fmadd_32(i8 zeroext %k, <2 x double> %
; CHECK: # %bb.0: # %entry
; CHECK-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT: vfmadd213ss %xmm2, %xmm1, %xmm0 {%k1} {z} # encoding: [0x62,0xf2,0x75,0x89,0xa9,0xc2]
-; CHECK-NEXT: # xmm0 = (xmm1 * xmm0) + xmm2
+; CHECK-NEXT: # xmm0 {%k1} {z} = (xmm1 * xmm0) + xmm2
; CHECK-NEXT: retq # encoding: [0xc3]
entry:
%0 = bitcast <2 x double> %a to <4 x float>
@@ -74,7 +74,7 @@ define <2 x double> @combine_scalar_maskz_fmadd_64(i8 zeroext %k, <2 x double> %
; CHECK: # %bb.0: # %entry
; CHECK-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT: vfmadd213sd %xmm2, %xmm1, %xmm0 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0x89,0xa9,0xc2]
-; CHECK-NEXT: # xmm0 = (xmm1 * xmm0) + xmm2
+; CHECK-NEXT: # xmm0 {%k1} {z} = (xmm1 * xmm0) + xmm2
; CHECK-NEXT: retq # encoding: [0xc3]
entry:
%0 = extractelement <2 x double> %a, i64 0
@@ -94,7 +94,7 @@ define <2 x double> @combine_scalar_mask3_fmadd_32(<2 x double> %a, <2 x double>
; CHECK: # %bb.0: # %entry
; CHECK-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT: vfmadd231ss %xmm0, %xmm1, %xmm2 {%k1} # encoding: [0x62,0xf2,0x75,0x09,0xb9,0xd0]
-; CHECK-NEXT: # xmm2 = (xmm1 * xmm0) + xmm2
+; CHECK-NEXT: # xmm2 {%k1} = (xmm1 * xmm0) + xmm2
; CHECK-NEXT: vmovaps %xmm2, %xmm0 # encoding: [0xc5,0xf8,0x28,0xc2]
; CHECK-NEXT: retq # encoding: [0xc3]
entry:
@@ -119,7 +119,7 @@ define <2 x double> @combine_scalar_mask3_fmadd_64(<2 x double> %a, <2 x double>
; CHECK: # %bb.0: # %entry
; CHECK-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT: vfmadd231sd %xmm0, %xmm1, %xmm2 {%k1} # encoding: [0x62,0xf2,0xf5,0x09,0xb9,0xd0]
-; CHECK-NEXT: # xmm2 = (xmm1 * xmm0) + xmm2
+; CHECK-NEXT: # xmm2 {%k1} = (xmm1 * xmm0) + xmm2
; CHECK-NEXT: vmovapd %xmm2, %xmm0 # encoding: [0xc5,0xf9,0x28,0xc2]
; CHECK-NEXT: retq # encoding: [0xc3]
entry:
@@ -140,7 +140,7 @@ define <2 x double> @combine_scalar_mask_fmsub_f32(<2 x double> %a, i8 zeroext %
; CHECK: # %bb.0: # %entry
; CHECK-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT: vfmsub213ss %xmm2, %xmm1, %xmm0 {%k1} # encoding: [0x62,0xf2,0x75,0x09,0xab,0xc2]
-; CHECK-NEXT: # xmm0 = (xmm1 * xmm0) - xmm2
+; CHECK-NEXT: # xmm0 {%k1} = (xmm1 * xmm0) - xmm2
; CHECK-NEXT: retq # encoding: [0xc3]
entry:
%0 = bitcast <2 x double> %a to <4 x float>
@@ -164,7 +164,7 @@ define <2 x double> @combine_scalar_mask_fmsub_f64(<2 x double> %a, i8 zeroext %
; CHECK: # %bb.0: # %entry
; CHECK-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT: vfmsub213sd %xmm2, %xmm1, %xmm0 {%k1} # encoding: [0x62,0xf2,0xf5,0x09,0xab,0xc2]
-; CHECK-NEXT: # xmm0 = (xmm1 * xmm0) - xmm2
+; CHECK-NEXT: # xmm0 {%k1} = (xmm1 * xmm0) - xmm2
; CHECK-NEXT: retq # encoding: [0xc3]
entry:
%0 = extractelement <2 x double> %a, i64 0
@@ -184,7 +184,7 @@ define <2 x double> @combine_scalar_maskz_fmsub_32(i8 zeroext %k, <2 x double> %
; CHECK: # %bb.0: # %entry
; CHECK-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT: vfmsub213ss %xmm2, %xmm1, %xmm0 {%k1} {z} # encoding: [0x62,0xf2,0x75,0x89,0xab,0xc2]
-; CHECK-NEXT: # xmm0 = (xmm1 * xmm0) - xmm2
+; CHECK-NEXT: # xmm0 {%k1} {z} = (xmm1 * xmm0) - xmm2
; CHECK-NEXT: retq # encoding: [0xc3]
entry:
%0 = bitcast <2 x double> %a to <4 x float>
@@ -208,7 +208,7 @@ define <2 x double> @combine_scalar_maskz_fmsub_64(i8 zeroext %k, <2 x double> %
; CHECK: # %bb.0: # %entry
; CHECK-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT: vfmsub213sd %xmm2, %xmm1, %xmm0 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0x89,0xab,0xc2]
-; CHECK-NEXT: # xmm0 = (xmm1 * xmm0) - xmm2
+; CHECK-NEXT: # xmm0 {%k1} {z} = (xmm1 * xmm0) - xmm2
; CHECK-NEXT: retq # encoding: [0xc3]
entry:
%0 = extractelement <2 x double> %a, i64 0
@@ -228,7 +228,7 @@ define <2 x double> @combine_scalar_mask3_fmsub_32(<2 x double> %a, <2 x double>
; CHECK: # %bb.0: # %entry
; CHECK-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT: vfmsub231ss %xmm0, %xmm1, %xmm2 {%k1} # encoding: [0x62,0xf2,0x75,0x09,0xbb,0xd0]
-; CHECK-NEXT: # xmm2 = (xmm1 * xmm0) - xmm2
+; CHECK-NEXT: # xmm2 {%k1} = (xmm1 * xmm0) - xmm2
; CHECK-NEXT: vmovaps %xmm2, %xmm0 # encoding: [0xc5,0xf8,0x28,0xc2]
; CHECK-NEXT: retq # encoding: [0xc3]
entry:
@@ -253,7 +253,7 @@ define <2 x double> @combine_scalar_mask3_fmsub_64(<2 x double> %a, <2 x double>
; CHECK: # %bb.0: # %entry
; CHECK-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT: vfmsub231sd %xmm0, %xmm1, %xmm2 {%k1} # encoding: [0x62,0xf2,0xf5,0x09,0xbb,0xd0]
-; CHECK-NEXT: # xmm2 = (xmm1 * xmm0) - xmm2
+; CHECK-NEXT: # xmm2 {%k1} = (xmm1 * xmm0) - xmm2
; CHECK-NEXT: vmovapd %xmm2, %xmm0 # encoding: [0xc5,0xf9,0x28,0xc2]
; CHECK-NEXT: retq # encoding: [0xc3]
entry:
@@ -274,7 +274,7 @@ define <2 x double> @combine_scalar_mask_fnmadd_f32(<2 x double> %a, i8 zeroext
; CHECK: # %bb.0: # %entry
; CHECK-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT: vfnmadd213ss %xmm2, %xmm1, %xmm0 {%k1} # encoding: [0x62,0xf2,0x75,0x09,0xad,0xc2]
-; CHECK-NEXT: # xmm0 = -(xmm1 * xmm0) + xmm2
+; CHECK-NEXT: # xmm0 {%k1} = -(xmm1 * xmm0) + xmm2
; CHECK-NEXT: retq # encoding: [0xc3]
entry:
%0 = bitcast <2 x double> %a to <4 x float>
@@ -298,7 +298,7 @@ define <2 x double> @combine_scalar_mask_fnmadd_f64(<2 x double> %a, i8 zeroext
; CHECK: # %bb.0: # %entry
; CHECK-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT: vfnmadd213sd %xmm2, %xmm1, %xmm0 {%k1} # encoding: [0x62,0xf2,0xf5,0x09,0xad,0xc2]
-; CHECK-NEXT: # xmm0 = -(xmm1 * xmm0) + xmm2
+; CHECK-NEXT: # xmm0 {%k1} = -(xmm1 * xmm0) + xmm2
; CHECK-NEXT: retq # encoding: [0xc3]
entry:
%0 = extractelement <2 x double> %a, i64 0
@@ -318,7 +318,7 @@ define <2 x double> @combine_scalar_maskz_fnmadd_32(i8 zeroext %k, <2 x double>
; CHECK: # %bb.0: # %entry
; CHECK-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT: vfnmadd213ss %xmm2, %xmm1, %xmm0 {%k1} {z} # encoding: [0x62,0xf2,0x75,0x89,0xad,0xc2]
-; CHECK-NEXT: # xmm0 = -(xmm1 * xmm0) + xmm2
+; CHECK-NEXT: # xmm0 {%k1} {z} = -(xmm1 * xmm0) + xmm2
; CHECK-NEXT: retq # encoding: [0xc3]
entry:
%0 = bitcast <2 x double> %a to <4 x float>
@@ -342,7 +342,7 @@ define <2 x double> @combine_scalar_maskz_fnmadd_64(i8 zeroext %k, <2 x double>
; CHECK: # %bb.0: # %entry
; CHECK-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT: vfnmadd213sd %xmm2, %xmm1, %xmm0 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0x89,0xad,0xc2]
-; CHECK-NEXT: # xmm0 = -(xmm1 * xmm0) + xmm2
+; CHECK-NEXT: # xmm0 {%k1} {z} = -(xmm1 * xmm0) + xmm2
; CHECK-NEXT: retq # encoding: [0xc3]
entry:
%0 = extractelement <2 x double> %a, i64 0
@@ -362,7 +362,7 @@ define <2 x double> @combine_scalar_mask3_fnmadd_32(<2 x double> %a, <2 x double
; CHECK: # %bb.0: # %entry
; CHECK-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT: vfnmadd231ss %xmm0, %xmm1, %xmm2 {%k1} # encoding: [0x62,0xf2,0x75,0x09,0xbd,0xd0]
-; CHECK-NEXT: # xmm2 = -(xmm1 * xmm0) + xmm2
+; CHECK-NEXT: # xmm2 {%k1} = -(xmm1 * xmm0) + xmm2
; CHECK-NEXT: vmovaps %xmm2, %xmm0 # encoding: [0xc5,0xf8,0x28,0xc2]
; CHECK-NEXT: retq # encoding: [0xc3]
entry:
@@ -387,7 +387,7 @@ define <2 x double> @combine_scalar_mask3_fnmadd_64(<2 x double> %a, <2 x double
; CHECK: # %bb.0: # %entry
; CHECK-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT: vfnmadd231sd %xmm0, %xmm1, %xmm2 {%k1} # encoding: [0x62,0xf2,0xf5,0x09,0xbd,0xd0]
-; CHECK-NEXT: # xmm2 = -(xmm1 * xmm0) + xmm2
+; CHECK-NEXT: # xmm2 {%k1} = -(xmm1 * xmm0) + xmm2
; CHECK-NEXT: vmovapd %xmm2, %xmm0 # encoding: [0xc5,0xf9,0x28,0xc2]
; CHECK-NEXT: retq # encoding: [0xc3]
entry:
@@ -408,7 +408,7 @@ define <2 x double> @combine_scalar_mask_fnmsub_f32(<2 x double> %a, i8 zeroext
; CHECK: # %bb.0: # %entry
; CHECK-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT: vfnmsub213ss %xmm2, %xmm1, %xmm0 {%k1} # encoding: [0x62,0xf2,0x75,0x09,0xaf,0xc2]
-; CHECK-NEXT: # xmm0 = -(xmm1 * xmm0) - xmm2
+; CHECK-NEXT: # xmm0 {%k1} = -(xmm1 * xmm0) - xmm2
; CHECK-NEXT: retq # encoding: [0xc3]
entry:
%0 = bitcast <2 x double> %a to <4 x float>
@@ -433,7 +433,7 @@ define <2 x double> @combine_scalar_mask_fnmsub_f64(<2 x double> %a, i8 zeroext
; CHECK: # %bb.0: # %entry
; CHECK-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT: vfnmsub213sd %xmm2, %xmm1, %xmm0 {%k1} # encoding: [0x62,0xf2,0xf5,0x09,0xaf,0xc2]
-; CHECK-NEXT: # xmm0 = -(xmm1 * xmm0) - xmm2
+; CHECK-NEXT: # xmm0 {%k1} = -(xmm1 * xmm0) - xmm2
; CHECK-NEXT: retq # encoding: [0xc3]
entry:
%0 = extractelement <2 x double> %a, i64 0
@@ -454,7 +454,7 @@ define <2 x double> @combine_scalar_maskz_fnmsub_32(i8 zeroext %k, <2 x double>
; CHECK: # %bb.0: # %entry
; CHECK-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT: vfnmsub213ss %xmm2, %xmm1, %xmm0 {%k1} {z} # encoding: [0x62,0xf2,0x75,0x89,0xaf,0xc2]
-; CHECK-NEXT: # xmm0 = -(xmm1 * xmm0) - xmm2
+; CHECK-NEXT: # xmm0 {%k1} {z} = -(xmm1 * xmm0) - xmm2
; CHECK-NEXT: retq # encoding: [0xc3]
entry:
%0 = bitcast <2 x double> %a to <4 x float>
@@ -479,7 +479,7 @@ define <2 x double> @combine_scalar_maskz_fnmsub_64(i8 zeroext %k, <2 x double>
; CHECK: # %bb.0: # %entry
; CHECK-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT: vfnmsub213sd %xmm2, %xmm1, %xmm0 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0x89,0xaf,0xc2]
-; CHECK-NEXT: # xmm0 = -(xmm1 * xmm0) - xmm2
+; CHECK-NEXT: # xmm0 {%k1} {z} = -(xmm1 * xmm0) - xmm2
; CHECK-NEXT: retq # encoding: [0xc3]
entry:
%0 = extractelement <2 x double> %a, i64 0
@@ -500,7 +500,7 @@ define <2 x double> @combine_scalar_mask3_fnmsub_32(<2 x double> %a, <2 x double
; CHECK: # %bb.0: # %entry
; CHECK-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT: vfnmsub231ss %xmm0, %xmm1, %xmm2 {%k1} # encoding: [0x62,0xf2,0x75,0x09,0xbf,0xd0]
-; CHECK-NEXT: # xmm2 = -(xmm1 * xmm0) - xmm2
+; CHECK-NEXT: # xmm2 {%k1} = -(xmm1 * xmm0) - xmm2
; CHECK-NEXT: vmovaps %xmm2, %xmm0 # encoding: [0xc5,0xf8,0x28,0xc2]
; CHECK-NEXT: retq # encoding: [0xc3]
entry:
@@ -526,7 +526,7 @@ define <2 x double> @combine_scalar_mask3_fnmsub_64(<2 x double> %a, <2 x double
; CHECK: # %bb.0: # %entry
; CHECK-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT: vfnmsub231sd %xmm0, %xmm1, %xmm2 {%k1} # encoding: [0x62,0xf2,0xf5,0x09,0xbf,0xd0]
-; CHECK-NEXT: # xmm2 = -(xmm1 * xmm0) - xmm2
+; CHECK-NEXT: # xmm2 {%k1} = -(xmm1 * xmm0) - xmm2
; CHECK-NEXT: vmovapd %xmm2, %xmm0 # encoding: [0xc5,0xf9,0x28,0xc2]
; CHECK-NEXT: retq # encoding: [0xc3]
entry:
More information about the llvm-commits
mailing list