[llvm] r276613 - [AVX512] Cleanup FMA operand order in patterns to match the VEX versions and to really be 213, 231, and 132.
Craig Topper via llvm-commits
llvm-commits at lists.llvm.org
Mon Jul 25 00:20:28 PDT 2016
Author: ctopper
Date: Mon Jul 25 02:20:28 2016
New Revision: 276613
URL: http://llvm.org/viewvc/llvm-project?rev=276613&view=rev
Log:
[AVX512] Cleanup FMA operand order in patterns to match the VEX versions and to really be 213, 231, and 132.
Modified:
llvm/trunk/lib/Target/X86/X86InstrAVX512.td
llvm/trunk/test/CodeGen/X86/avx512-fma-intrinsics.ll
llvm/trunk/test/CodeGen/X86/avx512-fma.ll
llvm/trunk/test/CodeGen/X86/avx512bwvl-intrinsics.ll
Modified: llvm/trunk/lib/Target/X86/X86InstrAVX512.td
URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/lib/Target/X86/X86InstrAVX512.td?rev=276613&r1=276612&r2=276613&view=diff
==============================================================================
--- llvm/trunk/lib/Target/X86/X86InstrAVX512.td (original)
+++ llvm/trunk/lib/Target/X86/X86InstrAVX512.td Mon Jul 25 02:20:28 2016
@@ -4763,21 +4763,21 @@ multiclass avx512_fma3p_213_rm<bits<8> o
defm r: AVX512_maskable_3src<opc, MRMSrcReg, _, (outs _.RC:$dst),
(ins _.RC:$src2, _.RC:$src3),
OpcodeStr, "$src3, $src2", "$src2, $src3",
- (_.VT (OpNode _.RC:$src1, _.RC:$src2, _.RC:$src3))>,
+ (_.VT (OpNode _.RC:$src2, _.RC:$src1, _.RC:$src3))>,
AVX512FMA3Base;
defm m: AVX512_maskable_3src<opc, MRMSrcMem, _, (outs _.RC:$dst),
(ins _.RC:$src2, _.MemOp:$src3),
OpcodeStr, "$src3, $src2", "$src2, $src3",
- (_.VT (OpNode _.RC:$src1, _.RC:$src2, (_.LdFrag addr:$src3)))>,
+ (_.VT (OpNode _.RC:$src2, _.RC:$src1, (_.LdFrag addr:$src3)))>,
AVX512FMA3Base;
defm mb: AVX512_maskable_3src<opc, MRMSrcMem, _, (outs _.RC:$dst),
(ins _.RC:$src2, _.ScalarMemOp:$src3),
OpcodeStr, !strconcat("${src3}", _.BroadcastStr,", $src2"),
!strconcat("$src2, ${src3}", _.BroadcastStr ),
- (OpNode _.RC:$src1,
- _.RC:$src2,(_.VT (X86VBroadcast (_.ScalarLdFrag addr:$src3))))>,
+ (OpNode _.RC:$src2,
+ _.RC:$src1,(_.VT (X86VBroadcast (_.ScalarLdFrag addr:$src3))))>,
AVX512FMA3Base, EVEX_B;
}
}
@@ -4788,7 +4788,7 @@ multiclass avx512_fma3_213_round<bits<8>
defm rb: AVX512_maskable_3src<opc, MRMSrcReg, _, (outs _.RC:$dst),
(ins _.RC:$src2, _.RC:$src3, AVX512RC:$rc),
OpcodeStr, "$rc, $src3, $src2", "$src2, $src3, $rc",
- (_.VT ( OpNode _.RC:$src1, _.RC:$src2, _.RC:$src3, (i32 imm:$rc)))>,
+ (_.VT ( OpNode _.RC:$src2, _.RC:$src1, _.RC:$src3, (i32 imm:$rc)))>,
AVX512FMA3Base, EVEX_B, EVEX_RC;
}
} // Constraints = "$src1 = $dst"
@@ -4896,24 +4896,24 @@ multiclass avx512_fma3p_132_rm<bits<8> o
X86VectorVTInfo _> {
let ExeDomain = _.ExeDomain in {
defm r: AVX512_maskable_3src<opc, MRMSrcReg, _, (outs _.RC:$dst),
- (ins _.RC:$src3, _.RC:$src2),
- OpcodeStr, "$src2, $src3", "$src3, $src2",
- (_.VT (OpNode _.RC:$src1, _.RC:$src2, _.RC:$src3))>,
+ (ins _.RC:$src2, _.RC:$src3),
+ OpcodeStr, "$src3, $src2", "$src2, $src3",
+ (_.VT (OpNode _.RC:$src1, _.RC:$src3, _.RC:$src2))>,
AVX512FMA3Base;
defm m: AVX512_maskable_3src<opc, MRMSrcMem, _, (outs _.RC:$dst),
- (ins _.RC:$src3, _.MemOp:$src2),
- OpcodeStr, "$src2, $src3", "$src3, $src2",
- (_.VT (OpNode _.RC:$src1, (_.LdFrag addr:$src2), _.RC:$src3))>,
+ (ins _.RC:$src2, _.MemOp:$src3),
+ OpcodeStr, "$src3, $src2", "$src2, $src3",
+ (_.VT (OpNode _.RC:$src1, (_.LdFrag addr:$src3), _.RC:$src2))>,
AVX512FMA3Base;
defm mb: AVX512_maskable_3src<opc, MRMSrcMem, _, (outs _.RC:$dst),
- (ins _.RC:$src3, _.ScalarMemOp:$src2),
- OpcodeStr, "${src2}"##_.BroadcastStr##", $src3",
- "$src3, ${src2}"##_.BroadcastStr,
+ (ins _.RC:$src2, _.ScalarMemOp:$src3),
+ OpcodeStr, "${src3}"##_.BroadcastStr##", $src2",
+ "$src2, ${src3}"##_.BroadcastStr,
(_.VT (OpNode _.RC:$src1,
- (_.VT (X86VBroadcast(_.ScalarLdFrag addr:$src2))),
- _.RC:$src3))>, AVX512FMA3Base, EVEX_B;
+ (_.VT (X86VBroadcast(_.ScalarLdFrag addr:$src3))),
+ _.RC:$src2))>, AVX512FMA3Base, EVEX_B;
}
}
@@ -4921,9 +4921,9 @@ multiclass avx512_fma3_132_round<bits<8>
X86VectorVTInfo _> {
let ExeDomain = _.ExeDomain in
defm rb: AVX512_maskable_3src<opc, MRMSrcReg, _, (outs _.RC:$dst),
- (ins _.RC:$src3, _.RC:$src2, AVX512RC:$rc),
- OpcodeStr, "$rc, $src2, $src3", "$src3, $src2, $rc",
- (_.VT ( OpNode _.RC:$src1, _.RC:$src2, _.RC:$src3, (i32 imm:$rc)))>,
+ (ins _.RC:$src2, _.RC:$src3, AVX512RC:$rc),
+ OpcodeStr, "$rc, $src3, $src2", "$src2, $src3, $rc",
+ (_.VT ( OpNode _.RC:$src1, _.RC:$src3, _.RC:$src2, (i32 imm:$rc)))>,
AVX512FMA3Base, EVEX_B, EVEX_RC;
}
} // Constraints = "$src1 = $dst"
Modified: llvm/trunk/test/CodeGen/X86/avx512-fma-intrinsics.ll
URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/test/CodeGen/X86/avx512-fma-intrinsics.ll?rev=276613&r1=276612&r2=276613&view=diff
==============================================================================
--- llvm/trunk/test/CodeGen/X86/avx512-fma-intrinsics.ll (original)
+++ llvm/trunk/test/CodeGen/X86/avx512-fma-intrinsics.ll Mon Jul 25 02:20:28 2016
@@ -18,7 +18,7 @@ define <16 x float> @test_mask_vfnmadd_p
; CHECK-LABEL: test_mask_vfnmadd_ps:
; CHECK: ## BB#0:
; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vfnmadd213ps %zmm2, %zmm1, %zmm0 {%k1}
+; CHECK-NEXT: vfnmadd132ps %zmm1, %zmm2, %zmm0 {%k1}
; 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_p
; CHECK-LABEL: test_mask_vfnmadd_pd:
; CHECK: ## BB#0:
; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vfnmadd213pd %zmm2, %zmm1, %zmm0 {%k1}
+; CHECK-NEXT: vfnmadd132pd %zmm1, %zmm2, %zmm0 {%k1}
; 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_p
; CHECK-LABEL: test_mask_vfnmsub_ps:
; CHECK: ## BB#0:
; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vfnmsub213ps %zmm2, %zmm1, %zmm0 {%k1}
+; CHECK-NEXT: vfnmsub132ps %zmm1, %zmm2, %zmm0 {%k1}
; 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_p
; CHECK-LABEL: test_mask_vfnmsub_pd:
; CHECK: ## BB#0:
; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vfnmsub213pd %zmm2, %zmm1, %zmm0 {%k1}
+; CHECK-NEXT: vfnmsub132pd %zmm1, %zmm2, %zmm0 {%k1}
; 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_
; CHECK-LABEL: test_mask_fmaddsub_ps:
; CHECK: ## BB#0:
; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vfmaddsub213ps %zmm2, %zmm1, %zmm0 {%k1}
+; CHECK-NEXT: vfmaddsub132ps %zmm1, %zmm2, %zmm0 {%k1}
; 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
; CHECK-LABEL: test_mask_vfmaddsub_pd:
; CHECK: ## BB#0:
; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vfmaddsub213pd %zmm2, %zmm1, %zmm0 {%k1}
+; CHECK-NEXT: vfmaddsub132pd %zmm1, %zmm2, %zmm0 {%k1}
; 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
@@ -130,9 +130,9 @@ define <8 x double>@test_int_x86_avx512_
; CHECK: ## BB#0:
; CHECK-NEXT: kmovw %edi, %k1
; CHECK-NEXT: vmovapd %zmm0, %zmm3
-; CHECK-NEXT: vfmaddsub213pd %zmm2, %zmm1, %zmm3 {%k1}
-; CHECK-NEXT: vfmaddsub213pd {rn-sae}, %zmm2, %zmm1, %zmm0
-; CHECK-NEXT: vaddpd %zmm0, %zmm3, %zmm0
+; CHECK-NEXT: vfmaddsub132pd %zmm1, %zmm2, %zmm3 {%k1}
+; CHECK-NEXT: vfmaddsub213pd {rn-sae}, %zmm2, %zmm0, %zmm1
+; CHECK-NEXT: vaddpd %zmm1, %zmm3, %zmm0
; 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)
%res1 = call <8 x double> @llvm.x86.avx512.mask.vfmaddsub.pd.512(<8 x double> %x0, <8 x double> %x1, <8 x double> %x2, i8 -1, i32 0)
@@ -148,8 +148,8 @@ define <8 x double>@test_int_x86_avx512_
; CHECK-NEXT: kmovw %edi, %k1
; CHECK-NEXT: vmovapd %zmm2, %zmm3
; CHECK-NEXT: vfmaddsub231pd %zmm1, %zmm0, %zmm3 {%k1}
-; CHECK-NEXT: vfmaddsub213pd {rn-sae}, %zmm2, %zmm1, %zmm0
-; CHECK-NEXT: vaddpd %zmm0, %zmm3, %zmm0
+; CHECK-NEXT: vfmaddsub213pd {rn-sae}, %zmm2, %zmm0, %zmm1
+; CHECK-NEXT: vaddpd %zmm1, %zmm3, %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)
%res1 = call <8 x double> @llvm.x86.avx512.mask3.vfmaddsub.pd.512(<8 x double> %x0, <8 x double> %x1, <8 x double> %x2, i8 -1, i32 0)
@@ -163,10 +163,10 @@ define <8 x double>@test_int_x86_avx512_
; CHECK-LABEL: test_int_x86_avx512_maskz_vfmaddsub_pd_512:
; CHECK: ## BB#0:
; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vmovapd %zmm0, %zmm3
-; CHECK-NEXT: vfmaddsub213pd %zmm2, %zmm1, %zmm3 {%k1} {z}
-; CHECK-NEXT: vfmaddsub213pd {rn-sae}, %zmm2, %zmm1, %zmm0
-; CHECK-NEXT: vaddpd %zmm0, %zmm3, %zmm0
+; CHECK-NEXT: vmovapd %zmm1, %zmm3
+; CHECK-NEXT: vfmaddsub213pd %zmm2, %zmm0, %zmm3 {%k1} {z}
+; CHECK-NEXT: vfmaddsub213pd {rn-sae}, %zmm2, %zmm0, %zmm1
+; CHECK-NEXT: vaddpd %zmm1, %zmm3, %zmm0
; 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)
%res1 = call <8 x double> @llvm.x86.avx512.maskz.vfmaddsub.pd.512(<8 x double> %x0, <8 x double> %x1, <8 x double> %x2, i8 -1, i32 0)
@@ -179,9 +179,9 @@ define <16 x float>@test_int_x86_avx512_
; CHECK: ## BB#0:
; CHECK-NEXT: kmovw %edi, %k1
; CHECK-NEXT: vmovaps %zmm0, %zmm3
-; CHECK-NEXT: vfmaddsub213ps %zmm2, %zmm1, %zmm3 {%k1}
-; CHECK-NEXT: vfmaddsub213ps {rn-sae}, %zmm2, %zmm1, %zmm0
-; CHECK-NEXT: vaddps %zmm0, %zmm3, %zmm0
+; CHECK-NEXT: vfmaddsub132ps %zmm1, %zmm2, %zmm3 {%k1}
+; CHECK-NEXT: vfmaddsub213ps {rn-sae}, %zmm2, %zmm0, %zmm1
+; CHECK-NEXT: vaddps %zmm1, %zmm3, %zmm0
; 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)
%res1 = call <16 x float> @llvm.x86.avx512.mask.vfmaddsub.ps.512(<16 x float> %x0, <16 x float> %x1, <16 x float> %x2, i16 -1, i32 0)
@@ -197,8 +197,8 @@ define <16 x float>@test_int_x86_avx512_
; CHECK-NEXT: kmovw %edi, %k1
; CHECK-NEXT: vmovaps %zmm2, %zmm3
; CHECK-NEXT: vfmaddsub231ps %zmm1, %zmm0, %zmm3 {%k1}
-; CHECK-NEXT: vfmaddsub213ps {rn-sae}, %zmm2, %zmm1, %zmm0
-; CHECK-NEXT: vaddps %zmm0, %zmm3, %zmm0
+; CHECK-NEXT: vfmaddsub213ps {rn-sae}, %zmm2, %zmm0, %zmm1
+; CHECK-NEXT: vaddps %zmm1, %zmm3, %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)
%res1 = call <16 x float> @llvm.x86.avx512.mask3.vfmaddsub.ps.512(<16 x float> %x0, <16 x float> %x1, <16 x float> %x2, i16 -1, i32 0)
@@ -212,10 +212,10 @@ define <16 x float>@test_int_x86_avx512_
; CHECK-LABEL: test_int_x86_avx512_maskz_vfmaddsub_ps_512:
; CHECK: ## BB#0:
; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vmovaps %zmm0, %zmm3
-; CHECK-NEXT: vfmaddsub213ps %zmm2, %zmm1, %zmm3 {%k1} {z}
-; CHECK-NEXT: vfmaddsub213ps {rn-sae}, %zmm2, %zmm1, %zmm0
-; CHECK-NEXT: vaddps %zmm0, %zmm3, %zmm0
+; CHECK-NEXT: vmovaps %zmm1, %zmm3
+; CHECK-NEXT: vfmaddsub213ps %zmm2, %zmm0, %zmm3 {%k1} {z}
+; CHECK-NEXT: vfmaddsub213ps {rn-sae}, %zmm2, %zmm0, %zmm1
+; CHECK-NEXT: vaddps %zmm1, %zmm3, %zmm0
; 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)
%res1 = call <16 x float> @llvm.x86.avx512.maskz.vfmaddsub.ps.512(<16 x float> %x0, <16 x float> %x1, <16 x float> %x2, i16 -1, i32 0)
@@ -231,8 +231,8 @@ define <8 x double>@test_int_x86_avx512_
; CHECK-NEXT: kmovw %edi, %k1
; CHECK-NEXT: vmovapd %zmm2, %zmm3
; CHECK-NEXT: vfmsubadd231pd %zmm1, %zmm0, %zmm3 {%k1}
-; CHECK-NEXT: vfmsubadd213pd {rn-sae}, %zmm2, %zmm1, %zmm0
-; CHECK-NEXT: vaddpd %zmm0, %zmm3, %zmm0
+; CHECK-NEXT: vfmsubadd213pd {rn-sae}, %zmm2, %zmm0, %zmm1
+; CHECK-NEXT: vaddpd %zmm1, %zmm3, %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)
%res1 = call <8 x double> @llvm.x86.avx512.mask3.vfmsubadd.pd.512(<8 x double> %x0, <8 x double> %x1, <8 x double> %x2, i8 -1, i32 0)
@@ -248,8 +248,8 @@ define <16 x float>@test_int_x86_avx512_
; CHECK-NEXT: kmovw %edi, %k1
; CHECK-NEXT: vmovaps %zmm2, %zmm3
; CHECK-NEXT: vfmsubadd231ps %zmm1, %zmm0, %zmm3 {%k1}
-; CHECK-NEXT: vfmsubadd213ps {rn-sae}, %zmm2, %zmm1, %zmm0
-; CHECK-NEXT: vaddps %zmm0, %zmm3, %zmm0
+; CHECK-NEXT: vfmsubadd213ps {rn-sae}, %zmm2, %zmm0, %zmm1
+; CHECK-NEXT: vaddps %zmm1, %zmm3, %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)
%res1 = call <16 x float> @llvm.x86.avx512.mask3.vfmsubadd.ps.512(<16 x float> %x0, <16 x float> %x1, <16 x float> %x2, i16 -1, i32 0)
@@ -261,7 +261,7 @@ define <16 x float> @test_mask_round_vfm
; CHECK-LABEL: test_mask_round_vfmadd512_ps_rrb_rne:
; CHECK: ## BB#0:
; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vfmadd213ps {rn-sae}, %zmm2, %zmm1, %zmm0 {%k1}
+; CHECK-NEXT: vfmadd132ps {rn-sae}, %zmm1, %zmm2, %zmm0 {%k1}
; 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 0) nounwind
ret <16 x float> %res
@@ -271,7 +271,7 @@ define <16 x float> @test_mask_round_vfm
; CHECK-LABEL: test_mask_round_vfmadd512_ps_rrb_rtn:
; CHECK: ## BB#0:
; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vfmadd213ps {rd-sae}, %zmm2, %zmm1, %zmm0 {%k1}
+; CHECK-NEXT: vfmadd132ps {rd-sae}, %zmm1, %zmm2, %zmm0 {%k1}
; 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 1) nounwind
ret <16 x float> %res
@@ -281,7 +281,7 @@ define <16 x float> @test_mask_round_vfm
; CHECK-LABEL: test_mask_round_vfmadd512_ps_rrb_rtp:
; CHECK: ## BB#0:
; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vfmadd213ps {ru-sae}, %zmm2, %zmm1, %zmm0 {%k1}
+; CHECK-NEXT: vfmadd132ps {ru-sae}, %zmm1, %zmm2, %zmm0 {%k1}
; 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 2) nounwind
ret <16 x float> %res
@@ -291,7 +291,7 @@ define <16 x float> @test_mask_round_vfm
; CHECK-LABEL: test_mask_round_vfmadd512_ps_rrb_rtz:
; CHECK: ## BB#0:
; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vfmadd213ps {rz-sae}, %zmm2, %zmm1, %zmm0 {%k1}
+; CHECK-NEXT: vfmadd132ps {rz-sae}, %zmm1, %zmm2, %zmm0 {%k1}
; 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 3) nounwind
ret <16 x float> %res
@@ -301,7 +301,7 @@ define <16 x float> @test_mask_round_vfm
; CHECK-LABEL: test_mask_round_vfmadd512_ps_rrb_current:
; CHECK: ## BB#0:
; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vfmadd213ps %zmm2, %zmm1, %zmm0 {%k1}
+; CHECK-NEXT: vfmadd132ps %zmm1, %zmm2, %zmm0 {%k1}
; 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
@@ -310,7 +310,8 @@ define <16 x float> @test_mask_round_vfm
define <16 x float> @test_mask_round_vfmadd512_ps_rrbz_rne(<16 x float> %a0, <16 x float> %a1, <16 x float> %a2) {
; CHECK-LABEL: test_mask_round_vfmadd512_ps_rrbz_rne:
; CHECK: ## BB#0:
-; CHECK-NEXT: vfmadd213ps {rn-sae}, %zmm2, %zmm1, %zmm0
+; CHECK-NEXT: vfmadd213ps {rn-sae}, %zmm2, %zmm0, %zmm1
+; CHECK-NEXT: vmovaps %zmm1, %zmm0
; 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 -1, i32 0) nounwind
ret <16 x float> %res
@@ -319,7 +320,8 @@ define <16 x float> @test_mask_round_vfm
define <16 x float> @test_mask_round_vfmadd512_ps_rrbz_rtn(<16 x float> %a0, <16 x float> %a1, <16 x float> %a2) {
; CHECK-LABEL: test_mask_round_vfmadd512_ps_rrbz_rtn:
; CHECK: ## BB#0:
-; CHECK-NEXT: vfmadd213ps {rd-sae}, %zmm2, %zmm1, %zmm0
+; CHECK-NEXT: vfmadd213ps {rd-sae}, %zmm2, %zmm0, %zmm1
+; CHECK-NEXT: vmovaps %zmm1, %zmm0
; 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 -1, i32 1) nounwind
ret <16 x float> %res
@@ -328,7 +330,8 @@ define <16 x float> @test_mask_round_vfm
define <16 x float> @test_mask_round_vfmadd512_ps_rrbz_rtp(<16 x float> %a0, <16 x float> %a1, <16 x float> %a2) {
; CHECK-LABEL: test_mask_round_vfmadd512_ps_rrbz_rtp:
; CHECK: ## BB#0:
-; CHECK-NEXT: vfmadd213ps {ru-sae}, %zmm2, %zmm1, %zmm0
+; CHECK-NEXT: vfmadd213ps {ru-sae}, %zmm2, %zmm0, %zmm1
+; CHECK-NEXT: vmovaps %zmm1, %zmm0
; 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 -1, i32 2) nounwind
ret <16 x float> %res
@@ -337,7 +340,8 @@ define <16 x float> @test_mask_round_vfm
define <16 x float> @test_mask_round_vfmadd512_ps_rrbz_rtz(<16 x float> %a0, <16 x float> %a1, <16 x float> %a2) {
; CHECK-LABEL: test_mask_round_vfmadd512_ps_rrbz_rtz:
; CHECK: ## BB#0:
-; CHECK-NEXT: vfmadd213ps {rz-sae}, %zmm2, %zmm1, %zmm0
+; CHECK-NEXT: vfmadd213ps {rz-sae}, %zmm2, %zmm0, %zmm1
+; CHECK-NEXT: vmovaps %zmm1, %zmm0
; 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 -1, i32 3) nounwind
ret <16 x float> %res
@@ -360,8 +364,8 @@ define <8 x double>@test_int_x86_avx512_
; CHECK-NEXT: kmovw %edi, %k1
; CHECK-NEXT: vmovapd %zmm2, %zmm3
; CHECK-NEXT: vfmsub231pd %zmm1, %zmm0, %zmm3 {%k1}
-; CHECK-NEXT: vfmsub213pd {rn-sae}, %zmm2, %zmm1, %zmm0
-; CHECK-NEXT: vaddpd %zmm0, %zmm3, %zmm0
+; CHECK-NEXT: vfmsub213pd {rn-sae}, %zmm2, %zmm0, %zmm1
+; CHECK-NEXT: vaddpd %zmm1, %zmm3, %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)
%res1 = call <8 x double> @llvm.x86.avx512.mask3.vfmsub.pd.512(<8 x double> %x0, <8 x double> %x1, <8 x double> %x2, i8 -1, i32 0)
@@ -377,8 +381,8 @@ define <16 x float>@test_int_x86_avx512_
; CHECK-NEXT: kmovw %edi, %k1
; CHECK-NEXT: vmovaps %zmm2, %zmm3
; CHECK-NEXT: vfmsub231ps %zmm1, %zmm0, %zmm3 {%k1}
-; CHECK-NEXT: vfmsub213ps {rn-sae}, %zmm2, %zmm1, %zmm0
-; CHECK-NEXT: vaddps %zmm0, %zmm3, %zmm0
+; CHECK-NEXT: vfmsub213ps {rn-sae}, %zmm2, %zmm0, %zmm1
+; CHECK-NEXT: vaddps %zmm1, %zmm3, %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)
%res1 = call <16 x float> @llvm.x86.avx512.mask3.vfmsub.ps.512(<16 x float> %x0, <16 x float> %x1, <16 x float> %x2, i16 -1, i32 0)
@@ -390,7 +394,7 @@ define <8 x double> @test_mask_round_vfm
; CHECK-LABEL: test_mask_round_vfmadd512_pd_rrb_rne:
; CHECK: ## BB#0:
; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vfmadd213pd {rn-sae}, %zmm2, %zmm1, %zmm0 {%k1}
+; CHECK-NEXT: vfmadd132pd {rn-sae}, %zmm1, %zmm2, %zmm0 {%k1}
; 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 0) nounwind
ret <8 x double> %res
@@ -400,7 +404,7 @@ define <8 x double> @test_mask_round_vfm
; CHECK-LABEL: test_mask_round_vfmadd512_pd_rrb_rtn:
; CHECK: ## BB#0:
; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vfmadd213pd {rd-sae}, %zmm2, %zmm1, %zmm0 {%k1}
+; CHECK-NEXT: vfmadd132pd {rd-sae}, %zmm1, %zmm2, %zmm0 {%k1}
; 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 1) nounwind
ret <8 x double> %res
@@ -410,7 +414,7 @@ define <8 x double> @test_mask_round_vfm
; CHECK-LABEL: test_mask_round_vfmadd512_pd_rrb_rtp:
; CHECK: ## BB#0:
; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vfmadd213pd {ru-sae}, %zmm2, %zmm1, %zmm0 {%k1}
+; CHECK-NEXT: vfmadd132pd {ru-sae}, %zmm1, %zmm2, %zmm0 {%k1}
; 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 2) nounwind
ret <8 x double> %res
@@ -420,7 +424,7 @@ define <8 x double> @test_mask_round_vfm
; CHECK-LABEL: test_mask_round_vfmadd512_pd_rrb_rtz:
; CHECK: ## BB#0:
; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vfmadd213pd {rz-sae}, %zmm2, %zmm1, %zmm0 {%k1}
+; CHECK-NEXT: vfmadd132pd {rz-sae}, %zmm1, %zmm2, %zmm0 {%k1}
; 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 3) nounwind
ret <8 x double> %res
@@ -430,7 +434,7 @@ define <8 x double> @test_mask_round_vfm
; CHECK-LABEL: test_mask_round_vfmadd512_pd_rrb_current:
; CHECK: ## BB#0:
; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vfmadd213pd %zmm2, %zmm1, %zmm0 {%k1}
+; CHECK-NEXT: vfmadd132pd %zmm1, %zmm2, %zmm0 {%k1}
; 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
@@ -439,7 +443,8 @@ define <8 x double> @test_mask_round_vfm
define <8 x double> @test_mask_round_vfmadd512_pd_rrbz_rne(<8 x double> %a0, <8 x double> %a1, <8 x double> %a2) {
; CHECK-LABEL: test_mask_round_vfmadd512_pd_rrbz_rne:
; CHECK: ## BB#0:
-; CHECK-NEXT: vfmadd213pd {rn-sae}, %zmm2, %zmm1, %zmm0
+; CHECK-NEXT: vfmadd213pd {rn-sae}, %zmm2, %zmm0, %zmm1
+; CHECK-NEXT: vmovapd %zmm1, %zmm0
; 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 -1, i32 0) nounwind
ret <8 x double> %res
@@ -448,7 +453,8 @@ define <8 x double> @test_mask_round_vfm
define <8 x double> @test_mask_round_vfmadd512_pd_rrbz_rtn(<8 x double> %a0, <8 x double> %a1, <8 x double> %a2) {
; CHECK-LABEL: test_mask_round_vfmadd512_pd_rrbz_rtn:
; CHECK: ## BB#0:
-; CHECK-NEXT: vfmadd213pd {rd-sae}, %zmm2, %zmm1, %zmm0
+; CHECK-NEXT: vfmadd213pd {rd-sae}, %zmm2, %zmm0, %zmm1
+; CHECK-NEXT: vmovapd %zmm1, %zmm0
; 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 -1, i32 1) nounwind
ret <8 x double> %res
@@ -457,7 +463,8 @@ define <8 x double> @test_mask_round_vfm
define <8 x double> @test_mask_round_vfmadd512_pd_rrbz_rtp(<8 x double> %a0, <8 x double> %a1, <8 x double> %a2) {
; CHECK-LABEL: test_mask_round_vfmadd512_pd_rrbz_rtp:
; CHECK: ## BB#0:
-; CHECK-NEXT: vfmadd213pd {ru-sae}, %zmm2, %zmm1, %zmm0
+; CHECK-NEXT: vfmadd213pd {ru-sae}, %zmm2, %zmm0, %zmm1
+; CHECK-NEXT: vmovapd %zmm1, %zmm0
; 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 -1, i32 2) nounwind
ret <8 x double> %res
@@ -466,7 +473,8 @@ define <8 x double> @test_mask_round_vfm
define <8 x double> @test_mask_round_vfmadd512_pd_rrbz_rtz(<8 x double> %a0, <8 x double> %a1, <8 x double> %a2) {
; CHECK-LABEL: test_mask_round_vfmadd512_pd_rrbz_rtz:
; CHECK: ## BB#0:
-; CHECK-NEXT: vfmadd213pd {rz-sae}, %zmm2, %zmm1, %zmm0
+; CHECK-NEXT: vfmadd213pd {rz-sae}, %zmm2, %zmm0, %zmm1
+; CHECK-NEXT: vmovapd %zmm1, %zmm0
; 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 -1, i32 3) nounwind
ret <8 x double> %res
@@ -486,9 +494,9 @@ define <8 x double>@test_int_x86_avx512_
; CHECK: ## BB#0:
; CHECK-NEXT: kmovw %edi, %k1
; CHECK-NEXT: vmovapd %zmm0, %zmm3
-; CHECK-NEXT: vfmadd213pd %zmm2, %zmm1, %zmm3 {%k1}
-; CHECK-NEXT: vfmadd213pd {rn-sae}, %zmm2, %zmm1, %zmm0
-; CHECK-NEXT: vaddpd %zmm0, %zmm3, %zmm0
+; CHECK-NEXT: vfmadd132pd %zmm1, %zmm2, %zmm3 {%k1}
+; CHECK-NEXT: vfmadd213pd {rn-sae}, %zmm2, %zmm0, %zmm1
+; CHECK-NEXT: vaddpd %zmm1, %zmm3, %zmm0
; 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)
%res1 = call <8 x double> @llvm.x86.avx512.mask.vfmadd.pd.512(<8 x double> %x0, <8 x double> %x1, <8 x double> %x2, i8 -1, i32 0)
@@ -504,8 +512,8 @@ define <8 x double>@test_int_x86_avx512_
; CHECK-NEXT: kmovw %edi, %k1
; CHECK-NEXT: vmovapd %zmm2, %zmm3
; CHECK-NEXT: vfmadd231pd %zmm1, %zmm0, %zmm3 {%k1}
-; CHECK-NEXT: vfmadd213pd {rn-sae}, %zmm2, %zmm1, %zmm0
-; CHECK-NEXT: vaddpd %zmm0, %zmm3, %zmm0
+; CHECK-NEXT: vfmadd213pd {rn-sae}, %zmm2, %zmm0, %zmm1
+; CHECK-NEXT: vaddpd %zmm1, %zmm3, %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)
%res1 = call <8 x double> @llvm.x86.avx512.mask3.vfmadd.pd.512(<8 x double> %x0, <8 x double> %x1, <8 x double> %x2, i8 -1, i32 0)
@@ -519,10 +527,10 @@ define <8 x double>@test_int_x86_avx512_
; CHECK-LABEL: test_int_x86_avx512_maskz_vfmadd_pd_512:
; CHECK: ## BB#0:
; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vmovapd %zmm0, %zmm3
-; CHECK-NEXT: vfmadd213pd %zmm2, %zmm1, %zmm3 {%k1} {z}
-; CHECK-NEXT: vfmadd213pd {rn-sae}, %zmm2, %zmm1, %zmm0
-; CHECK-NEXT: vaddpd %zmm0, %zmm3, %zmm0
+; CHECK-NEXT: vmovapd %zmm1, %zmm3
+; CHECK-NEXT: vfmadd213pd %zmm2, %zmm0, %zmm3 {%k1} {z}
+; CHECK-NEXT: vfmadd213pd {rn-sae}, %zmm2, %zmm0, %zmm1
+; CHECK-NEXT: vaddpd %zmm1, %zmm3, %zmm0
; 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)
%res1 = call <8 x double> @llvm.x86.avx512.maskz.vfmadd.pd.512(<8 x double> %x0, <8 x double> %x1, <8 x double> %x2, i8 -1, i32 0)
@@ -535,9 +543,9 @@ define <16 x float>@test_int_x86_avx512_
; CHECK: ## BB#0:
; CHECK-NEXT: kmovw %edi, %k1
; CHECK-NEXT: vmovaps %zmm0, %zmm3
-; CHECK-NEXT: vfmadd213ps %zmm2, %zmm1, %zmm3 {%k1}
-; CHECK-NEXT: vfmadd213ps {rn-sae}, %zmm2, %zmm1, %zmm0
-; CHECK-NEXT: vaddps %zmm0, %zmm3, %zmm0
+; CHECK-NEXT: vfmadd132ps %zmm1, %zmm2, %zmm3 {%k1}
+; CHECK-NEXT: vfmadd213ps {rn-sae}, %zmm2, %zmm0, %zmm1
+; CHECK-NEXT: vaddps %zmm1, %zmm3, %zmm0
; 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)
%res1 = call <16 x float> @llvm.x86.avx512.mask.vfmadd.ps.512(<16 x float> %x0, <16 x float> %x1, <16 x float> %x2, i16 -1, i32 0)
@@ -553,8 +561,8 @@ define <16 x float>@test_int_x86_avx512_
; CHECK-NEXT: kmovw %edi, %k1
; CHECK-NEXT: vmovaps %zmm2, %zmm3
; CHECK-NEXT: vfmadd231ps %zmm1, %zmm0, %zmm3 {%k1}
-; CHECK-NEXT: vfmadd213ps {rn-sae}, %zmm2, %zmm1, %zmm0
-; CHECK-NEXT: vaddps %zmm0, %zmm3, %zmm0
+; CHECK-NEXT: vfmadd213ps {rn-sae}, %zmm2, %zmm0, %zmm1
+; CHECK-NEXT: vaddps %zmm1, %zmm3, %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)
%res1 = call <16 x float> @llvm.x86.avx512.mask3.vfmadd.ps.512(<16 x float> %x0, <16 x float> %x1, <16 x float> %x2, i16 -1, i32 0)
@@ -568,10 +576,10 @@ define <16 x float>@test_int_x86_avx512_
; CHECK-LABEL: test_int_x86_avx512_maskz_vfmadd_ps_512:
; CHECK: ## BB#0:
; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vmovaps %zmm0, %zmm3
-; CHECK-NEXT: vfmadd213ps %zmm2, %zmm1, %zmm3 {%k1} {z}
-; CHECK-NEXT: vfmadd213ps {rn-sae}, %zmm2, %zmm1, %zmm0
-; CHECK-NEXT: vaddps %zmm0, %zmm3, %zmm0
+; CHECK-NEXT: vmovaps %zmm1, %zmm3
+; CHECK-NEXT: vfmadd213ps %zmm2, %zmm0, %zmm3 {%k1} {z}
+; CHECK-NEXT: vfmadd213ps {rn-sae}, %zmm2, %zmm0, %zmm1
+; CHECK-NEXT: vaddps %zmm1, %zmm3, %zmm0
; 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)
%res1 = call <16 x float> @llvm.x86.avx512.maskz.vfmadd.ps.512(<16 x float> %x0, <16 x float> %x1, <16 x float> %x2, i16 -1, i32 0)
@@ -584,7 +592,7 @@ define <8 x double> @test_mask_round_vfn
; CHECK-LABEL: test_mask_round_vfnmsub512_pd_rrb_rne:
; CHECK: ## BB#0:
; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vfnmsub213pd {rn-sae}, %zmm2, %zmm1, %zmm0 {%k1}
+; CHECK-NEXT: vfnmsub132pd {rn-sae}, %zmm1, %zmm2, %zmm0 {%k1}
; 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 0) nounwind
ret <8 x double> %res
@@ -594,7 +602,7 @@ define <8 x double> @test_mask_round_vfn
; CHECK-LABEL: test_mask_round_vfnmsub512_pd_rrb_rtn:
; CHECK: ## BB#0:
; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vfnmsub213pd {rd-sae}, %zmm2, %zmm1, %zmm0 {%k1}
+; CHECK-NEXT: vfnmsub132pd {rd-sae}, %zmm1, %zmm2, %zmm0 {%k1}
; 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 1) nounwind
ret <8 x double> %res
@@ -604,7 +612,7 @@ define <8 x double> @test_mask_round_vfn
; CHECK-LABEL: test_mask_round_vfnmsub512_pd_rrb_rtp:
; CHECK: ## BB#0:
; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vfnmsub213pd {ru-sae}, %zmm2, %zmm1, %zmm0 {%k1}
+; CHECK-NEXT: vfnmsub132pd {ru-sae}, %zmm1, %zmm2, %zmm0 {%k1}
; 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 2) nounwind
ret <8 x double> %res
@@ -614,7 +622,7 @@ define <8 x double> @test_mask_round_vfn
; CHECK-LABEL: test_mask_round_vfnmsub512_pd_rrb_rtz:
; CHECK: ## BB#0:
; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vfnmsub213pd {rz-sae}, %zmm2, %zmm1, %zmm0 {%k1}
+; CHECK-NEXT: vfnmsub132pd {rz-sae}, %zmm1, %zmm2, %zmm0 {%k1}
; 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 3) nounwind
ret <8 x double> %res
@@ -624,7 +632,7 @@ define <8 x double> @test_mask_round_vfn
; CHECK-LABEL: test_mask_round_vfnmsub512_pd_rrb_current:
; CHECK: ## BB#0:
; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vfnmsub213pd %zmm2, %zmm1, %zmm0 {%k1}
+; CHECK-NEXT: vfnmsub132pd %zmm1, %zmm2, %zmm0 {%k1}
; 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
@@ -633,7 +641,8 @@ define <8 x double> @test_mask_round_vfn
define <8 x double> @test_mask_round_vfnmsub512_pd_rrbz_rne(<8 x double> %a0, <8 x double> %a1, <8 x double> %a2) {
; CHECK-LABEL: test_mask_round_vfnmsub512_pd_rrbz_rne:
; CHECK: ## BB#0:
-; CHECK-NEXT: vfnmsub213pd {rn-sae}, %zmm2, %zmm1, %zmm0
+; CHECK-NEXT: vfnmsub213pd {rn-sae}, %zmm2, %zmm0, %zmm1
+; CHECK-NEXT: vmovapd %zmm1, %zmm0
; 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 -1, i32 0) nounwind
ret <8 x double> %res
@@ -642,7 +651,8 @@ define <8 x double> @test_mask_round_vfn
define <8 x double> @test_mask_round_vfnmsub512_pd_rrbz_rtn(<8 x double> %a0, <8 x double> %a1, <8 x double> %a2) {
; CHECK-LABEL: test_mask_round_vfnmsub512_pd_rrbz_rtn:
; CHECK: ## BB#0:
-; CHECK-NEXT: vfnmsub213pd {rd-sae}, %zmm2, %zmm1, %zmm0
+; CHECK-NEXT: vfnmsub213pd {rd-sae}, %zmm2, %zmm0, %zmm1
+; CHECK-NEXT: vmovapd %zmm1, %zmm0
; 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 -1, i32 1) nounwind
ret <8 x double> %res
@@ -651,7 +661,8 @@ define <8 x double> @test_mask_round_vfn
define <8 x double> @test_mask_round_vfnmsub512_pd_rrbz_rtp(<8 x double> %a0, <8 x double> %a1, <8 x double> %a2) {
; CHECK-LABEL: test_mask_round_vfnmsub512_pd_rrbz_rtp:
; CHECK: ## BB#0:
-; CHECK-NEXT: vfnmsub213pd {ru-sae}, %zmm2, %zmm1, %zmm0
+; CHECK-NEXT: vfnmsub213pd {ru-sae}, %zmm2, %zmm0, %zmm1
+; CHECK-NEXT: vmovapd %zmm1, %zmm0
; 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 -1, i32 2) nounwind
ret <8 x double> %res
@@ -660,7 +671,8 @@ define <8 x double> @test_mask_round_vfn
define <8 x double> @test_mask_round_vfnmsub512_pd_rrbz_rtz(<8 x double> %a0, <8 x double> %a1, <8 x double> %a2) {
; CHECK-LABEL: test_mask_round_vfnmsub512_pd_rrbz_rtz:
; CHECK: ## BB#0:
-; CHECK-NEXT: vfnmsub213pd {rz-sae}, %zmm2, %zmm1, %zmm0
+; CHECK-NEXT: vfnmsub213pd {rz-sae}, %zmm2, %zmm0, %zmm1
+; CHECK-NEXT: vmovapd %zmm1, %zmm0
; 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 -1, i32 3) nounwind
ret <8 x double> %res
@@ -680,9 +692,9 @@ define <8 x double>@test_int_x86_avx512_
; CHECK: ## BB#0:
; CHECK-NEXT: kmovw %edi, %k1
; CHECK-NEXT: vmovapd %zmm0, %zmm3
-; CHECK-NEXT: vfnmsub213pd %zmm2, %zmm1, %zmm3 {%k1}
-; CHECK-NEXT: vfnmsub213pd {rn-sae}, %zmm2, %zmm1, %zmm0
-; CHECK-NEXT: vaddpd %zmm0, %zmm3, %zmm0
+; CHECK-NEXT: vfnmsub132pd %zmm1, %zmm2, %zmm3 {%k1}
+; CHECK-NEXT: vfnmsub213pd {rn-sae}, %zmm2, %zmm0, %zmm1
+; CHECK-NEXT: vaddpd %zmm1, %zmm3, %zmm0
; 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)
%res1 = call <8 x double> @llvm.x86.avx512.mask.vfnmsub.pd.512(<8 x double> %x0, <8 x double> %x1, <8 x double> %x2, i8 -1, i32 0)
@@ -698,8 +710,8 @@ define <8 x double>@test_int_x86_avx512_
; CHECK-NEXT: kmovw %edi, %k1
; CHECK-NEXT: vmovapd %zmm2, %zmm3
; CHECK-NEXT: vfnmsub231pd %zmm1, %zmm0, %zmm3 {%k1}
-; CHECK-NEXT: vfnmsub213pd {rn-sae}, %zmm2, %zmm1, %zmm0
-; CHECK-NEXT: vaddpd %zmm0, %zmm3, %zmm0
+; CHECK-NEXT: vfnmsub213pd {rn-sae}, %zmm2, %zmm0, %zmm1
+; CHECK-NEXT: vaddpd %zmm1, %zmm3, %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)
%res1 = call <8 x double> @llvm.x86.avx512.mask3.vfnmsub.pd.512(<8 x double> %x0, <8 x double> %x1, <8 x double> %x2, i8 -1, i32 0)
@@ -712,9 +724,9 @@ define <16 x float>@test_int_x86_avx512_
; CHECK: ## BB#0:
; CHECK-NEXT: kmovw %edi, %k1
; CHECK-NEXT: vmovaps %zmm0, %zmm3
-; CHECK-NEXT: vfnmsub213ps %zmm2, %zmm1, %zmm3 {%k1}
-; CHECK-NEXT: vfnmsub213ps {rn-sae}, %zmm2, %zmm1, %zmm0
-; CHECK-NEXT: vaddps %zmm0, %zmm3, %zmm0
+; CHECK-NEXT: vfnmsub132ps %zmm1, %zmm2, %zmm3 {%k1}
+; CHECK-NEXT: vfnmsub213ps {rn-sae}, %zmm2, %zmm0, %zmm1
+; CHECK-NEXT: vaddps %zmm1, %zmm3, %zmm0
; 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)
%res1 = call <16 x float> @llvm.x86.avx512.mask.vfnmsub.ps.512(<16 x float> %x0, <16 x float> %x1, <16 x float> %x2, i16 -1, i32 0)
@@ -730,8 +742,8 @@ define <16 x float>@test_int_x86_avx512_
; CHECK-NEXT: kmovw %edi, %k1
; CHECK-NEXT: vmovaps %zmm2, %zmm3
; CHECK-NEXT: vfnmsub231ps %zmm1, %zmm0, %zmm3 {%k1}
-; CHECK-NEXT: vfnmsub213ps {rn-sae}, %zmm2, %zmm1, %zmm0
-; CHECK-NEXT: vaddps %zmm0, %zmm3, %zmm0
+; CHECK-NEXT: vfnmsub213ps {rn-sae}, %zmm2, %zmm0, %zmm1
+; CHECK-NEXT: vaddps %zmm1, %zmm3, %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)
%res1 = call <16 x float> @llvm.x86.avx512.mask3.vfnmsub.ps.512(<16 x float> %x0, <16 x float> %x1, <16 x float> %x2, i16 -1, i32 0)
@@ -744,9 +756,9 @@ define <8 x double>@test_int_x86_avx512_
; CHECK: ## BB#0:
; CHECK-NEXT: kmovw %edi, %k1
; CHECK-NEXT: vmovapd %zmm0, %zmm3
-; CHECK-NEXT: vfnmadd213pd %zmm2, %zmm1, %zmm3 {%k1}
-; CHECK-NEXT: vfnmadd213pd {rn-sae}, %zmm2, %zmm1, %zmm0
-; CHECK-NEXT: vaddpd %zmm0, %zmm3, %zmm0
+; CHECK-NEXT: vfnmadd132pd %zmm1, %zmm2, %zmm3 {%k1}
+; CHECK-NEXT: vfnmadd213pd {rn-sae}, %zmm2, %zmm0, %zmm1
+; CHECK-NEXT: vaddpd %zmm1, %zmm3, %zmm0
; 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)
%res1 = call <8 x double> @llvm.x86.avx512.mask.vfnmadd.pd.512(<8 x double> %x0, <8 x double> %x1, <8 x double> %x2, i8 -1, i32 0)
@@ -759,9 +771,9 @@ define <16 x float>@test_int_x86_avx512_
; CHECK: ## BB#0:
; CHECK-NEXT: kmovw %edi, %k1
; CHECK-NEXT: vmovaps %zmm0, %zmm3
-; CHECK-NEXT: vfnmadd213ps %zmm2, %zmm1, %zmm3 {%k1}
-; CHECK-NEXT: vfnmadd213ps {rn-sae}, %zmm2, %zmm1, %zmm0
-; CHECK-NEXT: vaddps %zmm0, %zmm3, %zmm0
+; CHECK-NEXT: vfnmadd132ps %zmm1, %zmm2, %zmm3 {%k1}
+; CHECK-NEXT: vfnmadd213ps {rn-sae}, %zmm2, %zmm0, %zmm1
+; CHECK-NEXT: vaddps %zmm1, %zmm3, %zmm0
; 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)
%res1 = call <16 x float> @llvm.x86.avx512.mask.vfnmadd.ps.512(<16 x float> %x0, <16 x float> %x1, <16 x float> %x2, i16 -1, i32 0)
Modified: llvm/trunk/test/CodeGen/X86/avx512-fma.ll
URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/test/CodeGen/X86/avx512-fma.ll?rev=276613&r1=276612&r2=276613&view=diff
==============================================================================
--- llvm/trunk/test/CodeGen/X86/avx512-fma.ll (original)
+++ llvm/trunk/test/CodeGen/X86/avx512-fma.ll Mon Jul 25 02:20:28 2016
@@ -112,7 +112,8 @@ define <16 x float> @test231_br(<16 x fl
define <16 x float> @test213_br(<16 x float> %a1, <16 x float> %a2) nounwind {
; ALL-LABEL: test213_br:
; ALL: ## BB#0:
-; ALL-NEXT: vfmadd213ps {{.*}}(%rip){1to16}, %zmm1, %zmm0
+; ALL-NEXT: vfmadd213ps {{.*}}(%rip){1to16}, %zmm0, %zmm1
+; ALL-NEXT: vmovaps %zmm1, %zmm0
; ALL-NEXT: retq
%b1 = fmul <16 x float> %a1, %a2
%b2 = fadd <16 x float> %b1, <float 0x3FB99999A0000000, float 0x3FB99999A0000000, float 0x3FB99999A0000000, float 0x3FB99999A0000000, float 0x3FB99999A0000000, float 0x3FB99999A0000000, float 0x3FB99999A0000000, float 0x3FB99999A0000000, float 0x3FB99999A0000000, float 0x3FB99999A0000000, float 0x3FB99999A0000000, float 0x3FB99999A0000000, float 0x3FB99999A0000000, float 0x3FB99999A0000000, float 0x3FB99999A0000000, float 0x3FB99999A0000000>
@@ -174,7 +175,8 @@ define <16 x float> @test_x86_fmadd213_p
; KNL-NEXT: vpmovsxbd %xmm2, %zmm2
; KNL-NEXT: vpslld $31, %zmm2, %zmm2
; KNL-NEXT: vptestmd %zmm2, %zmm2, %k1
-; KNL-NEXT: vfmadd213ps (%rdi), %zmm0, %zmm1 {%k1}
+; KNL-NEXT: vmovups (%rdi), %zmm2
+; KNL-NEXT: vfmadd132ps %zmm0, %zmm2, %zmm1 {%k1}
; KNL-NEXT: vmovaps %zmm1, %zmm0
; KNL-NEXT: retq
;
@@ -182,7 +184,8 @@ define <16 x float> @test_x86_fmadd213_p
; SKX: ## BB#0:
; SKX-NEXT: vpsllw $7, %xmm2, %xmm2
; SKX-NEXT: vpmovb2m %xmm2, %k1
-; SKX-NEXT: vfmadd213ps (%rdi), %zmm0, %zmm1 {%k1}
+; SKX-NEXT: vmovups (%rdi), %zmm2
+; SKX-NEXT: vfmadd132ps %zmm0, %zmm2, %zmm1 {%k1}
; SKX-NEXT: vmovaps %zmm1, %zmm0
; SKX-NEXT: retq
%a2 = load <16 x float>,<16 x float> *%a2_ptrt,align 1
Modified: llvm/trunk/test/CodeGen/X86/avx512bwvl-intrinsics.ll
URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/test/CodeGen/X86/avx512bwvl-intrinsics.ll?rev=276613&r1=276612&r2=276613&view=diff
==============================================================================
--- llvm/trunk/test/CodeGen/X86/avx512bwvl-intrinsics.ll (original)
+++ llvm/trunk/test/CodeGen/X86/avx512bwvl-intrinsics.ll Mon Jul 25 02:20:28 2016
@@ -891,7 +891,7 @@ define <8 x float> @test_mask_vfmadd256_
; CHECK-LABEL: test_mask_vfmadd256_ps:
; CHECK: ## BB#0:
; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfmadd213ps %ymm2, %ymm1, %ymm0 {%k1} ## encoding: [0x62,0xf2,0x75,0x29,0xa8,0xc2]
+; CHECK-NEXT: vfmadd132ps %ymm1, %ymm2, %ymm0 {%k1} ## encoding: [0x62,0xf2,0x6d,0x29,0x98,0xc1]
; CHECK-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
@@ -903,7 +903,7 @@ define <4 x float> @test_mask_vfmadd128_
; CHECK-LABEL: test_mask_vfmadd128_ps:
; CHECK: ## BB#0:
; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfmadd213ps %xmm2, %xmm1, %xmm0 {%k1} ## encoding: [0x62,0xf2,0x75,0x09,0xa8,0xc2]
+; CHECK-NEXT: vfmadd132ps %xmm1, %xmm2, %xmm0 {%k1} ## encoding: [0x62,0xf2,0x6d,0x09,0x98,0xc1]
; CHECK-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
@@ -915,7 +915,7 @@ define <4 x double> @test_mask_fmadd256_
; CHECK-LABEL: test_mask_fmadd256_pd:
; CHECK: ## BB#0:
; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfmadd213pd %ymm2, %ymm1, %ymm0 {%k1} ## encoding: [0x62,0xf2,0xf5,0x29,0xa8,0xc2]
+; CHECK-NEXT: vfmadd132pd %ymm1, %ymm2, %ymm0 {%k1} ## encoding: [0x62,0xf2,0xed,0x29,0x98,0xc1]
; CHECK-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
@@ -927,7 +927,7 @@ define <2 x double> @test_mask_fmadd128_
; CHECK-LABEL: test_mask_fmadd128_pd:
; CHECK: ## BB#0:
; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfmadd213pd %xmm2, %xmm1, %xmm0 {%k1} ## encoding: [0x62,0xf2,0xf5,0x09,0xa8,0xc2]
+; CHECK-NEXT: vfmadd132pd %xmm1, %xmm2, %xmm0 {%k1} ## encoding: [0x62,0xf2,0xed,0x09,0x98,0xc1]
; CHECK-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
@@ -938,9 +938,9 @@ define <2 x double>@test_int_x86_avx512_
; CHECK: ## BB#0:
; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT: vmovaps %xmm0, %xmm3 ## encoding: [0x62,0xf1,0x7c,0x08,0x28,0xd8]
-; CHECK-NEXT: vfmadd213pd %xmm2, %xmm1, %xmm3 {%k1} ## encoding: [0x62,0xf2,0xf5,0x09,0xa8,0xda]
-; CHECK-NEXT: vfmadd213pd %xmm2, %xmm1, %xmm0 ## encoding: [0x62,0xf2,0xf5,0x08,0xa8,0xc2]
-; CHECK-NEXT: vaddpd %xmm0, %xmm3, %xmm0 ## encoding: [0x62,0xf1,0xe5,0x08,0x58,0xc0]
+; CHECK-NEXT: vfmadd132pd %xmm1, %xmm2, %xmm3 {%k1} ## encoding: [0x62,0xf2,0xed,0x09,0x98,0xd9]
+; CHECK-NEXT: vfmadd213pd %xmm2, %xmm0, %xmm1 ## encoding: [0x62,0xf2,0xfd,0x08,0xa8,0xca]
+; CHECK-NEXT: vaddpd %xmm1, %xmm3, %xmm0 ## encoding: [0x62,0xf1,0xe5,0x08,0x58,0xc1]
; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <2 x double> @llvm.x86.avx512.mask.vfmadd.pd.128(<2 x double> %x0, <2 x double> %x1, <2 x double> %x2, i8 %x3)
%res1 = call <2 x double> @llvm.x86.avx512.mask.vfmadd.pd.128(<2 x double> %x0, <2 x double> %x1, <2 x double> %x2, i8 -1)
@@ -956,8 +956,8 @@ define <2 x double>@test_int_x86_avx512_
; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT: vmovaps %xmm2, %xmm3 ## encoding: [0x62,0xf1,0x7c,0x08,0x28,0xda]
; CHECK-NEXT: vfmadd231pd %xmm1, %xmm0, %xmm3 {%k1} ## encoding: [0x62,0xf2,0xfd,0x09,0xb8,0xd9]
-; CHECK-NEXT: vfmadd213pd %xmm2, %xmm1, %xmm0 ## encoding: [0x62,0xf2,0xf5,0x08,0xa8,0xc2]
-; CHECK-NEXT: vaddpd %xmm0, %xmm3, %xmm0 ## encoding: [0x62,0xf1,0xe5,0x08,0x58,0xc0]
+; CHECK-NEXT: vfmadd213pd %xmm2, %xmm0, %xmm1 ## encoding: [0x62,0xf2,0xfd,0x08,0xa8,0xca]
+; CHECK-NEXT: vaddpd %xmm1, %xmm3, %xmm0 ## encoding: [0x62,0xf1,0xe5,0x08,0x58,0xc1]
; CHECK-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)
%res1 = call <2 x double> @llvm.x86.avx512.mask3.vfmadd.pd.128(<2 x double> %x0, <2 x double> %x1, <2 x double> %x2, i8 -1)
@@ -971,10 +971,10 @@ define <2 x double>@test_int_x86_avx512_
; CHECK-LABEL: test_int_x86_avx512_maskz_vfmadd_pd_128:
; CHECK: ## BB#0:
; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vmovaps %xmm0, %xmm3 ## encoding: [0x62,0xf1,0x7c,0x08,0x28,0xd8]
-; CHECK-NEXT: vfmadd213pd %xmm2, %xmm1, %xmm3 {%k1} {z} ## encoding: [0x62,0xf2,0xf5,0x89,0xa8,0xda]
-; CHECK-NEXT: vfmadd213pd %xmm2, %xmm1, %xmm0 ## encoding: [0x62,0xf2,0xf5,0x08,0xa8,0xc2]
-; CHECK-NEXT: vaddpd %xmm0, %xmm3, %xmm0 ## encoding: [0x62,0xf1,0xe5,0x08,0x58,0xc0]
+; CHECK-NEXT: vmovaps %xmm1, %xmm3 ## encoding: [0x62,0xf1,0x7c,0x08,0x28,0xd9]
+; CHECK-NEXT: vfmadd213pd %xmm2, %xmm0, %xmm3 {%k1} {z} ## encoding: [0x62,0xf2,0xfd,0x89,0xa8,0xda]
+; CHECK-NEXT: vfmadd213pd %xmm2, %xmm0, %xmm1 ## encoding: [0x62,0xf2,0xfd,0x08,0xa8,0xca]
+; CHECK-NEXT: vaddpd %xmm1, %xmm3, %xmm0 ## encoding: [0x62,0xf1,0xe5,0x08,0x58,0xc1]
; CHECK-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)
%res1 = call <2 x double> @llvm.x86.avx512.maskz.vfmadd.pd.128(<2 x double> %x0, <2 x double> %x1, <2 x double> %x2, i8 -1)
@@ -987,9 +987,9 @@ define <4 x double>@test_int_x86_avx512_
; CHECK: ## BB#0:
; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT: vmovaps %ymm0, %ymm3 ## encoding: [0x62,0xf1,0x7c,0x28,0x28,0xd8]
-; CHECK-NEXT: vfmadd213pd %ymm2, %ymm1, %ymm3 {%k1} ## encoding: [0x62,0xf2,0xf5,0x29,0xa8,0xda]
-; CHECK-NEXT: vfmadd213pd %ymm2, %ymm1, %ymm0 ## encoding: [0x62,0xf2,0xf5,0x28,0xa8,0xc2]
-; CHECK-NEXT: vaddpd %ymm0, %ymm3, %ymm0 ## encoding: [0x62,0xf1,0xe5,0x28,0x58,0xc0]
+; CHECK-NEXT: vfmadd132pd %ymm1, %ymm2, %ymm3 {%k1} ## encoding: [0x62,0xf2,0xed,0x29,0x98,0xd9]
+; CHECK-NEXT: vfmadd213pd %ymm2, %ymm0, %ymm1 ## encoding: [0x62,0xf2,0xfd,0x28,0xa8,0xca]
+; CHECK-NEXT: vaddpd %ymm1, %ymm3, %ymm0 ## encoding: [0x62,0xf1,0xe5,0x28,0x58,0xc1]
; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <4 x double> @llvm.x86.avx512.mask.vfmadd.pd.256(<4 x double> %x0, <4 x double> %x1, <4 x double> %x2, i8 %x3)
%res1 = call <4 x double> @llvm.x86.avx512.mask.vfmadd.pd.256(<4 x double> %x0, <4 x double> %x1, <4 x double> %x2, i8 -1)
@@ -1005,8 +1005,8 @@ define <4 x double>@test_int_x86_avx512_
; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT: vmovaps %ymm2, %ymm3 ## encoding: [0x62,0xf1,0x7c,0x28,0x28,0xda]
; CHECK-NEXT: vfmadd231pd %ymm1, %ymm0, %ymm3 {%k1} ## encoding: [0x62,0xf2,0xfd,0x29,0xb8,0xd9]
-; CHECK-NEXT: vfmadd213pd %ymm2, %ymm1, %ymm0 ## encoding: [0x62,0xf2,0xf5,0x28,0xa8,0xc2]
-; CHECK-NEXT: vaddpd %ymm0, %ymm3, %ymm0 ## encoding: [0x62,0xf1,0xe5,0x28,0x58,0xc0]
+; CHECK-NEXT: vfmadd213pd %ymm2, %ymm0, %ymm1 ## encoding: [0x62,0xf2,0xfd,0x28,0xa8,0xca]
+; CHECK-NEXT: vaddpd %ymm1, %ymm3, %ymm0 ## encoding: [0x62,0xf1,0xe5,0x28,0x58,0xc1]
; CHECK-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)
%res1 = call <4 x double> @llvm.x86.avx512.mask3.vfmadd.pd.256(<4 x double> %x0, <4 x double> %x1, <4 x double> %x2, i8 -1)
@@ -1020,10 +1020,10 @@ define <4 x double>@test_int_x86_avx512_
; CHECK-LABEL: test_int_x86_avx512_maskz_vfmadd_pd_256:
; CHECK: ## BB#0:
; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vmovaps %ymm0, %ymm3 ## encoding: [0x62,0xf1,0x7c,0x28,0x28,0xd8]
-; CHECK-NEXT: vfmadd213pd %ymm2, %ymm1, %ymm3 {%k1} {z} ## encoding: [0x62,0xf2,0xf5,0xa9,0xa8,0xda]
-; CHECK-NEXT: vfmadd213pd %ymm2, %ymm1, %ymm0 ## encoding: [0x62,0xf2,0xf5,0x28,0xa8,0xc2]
-; CHECK-NEXT: vaddpd %ymm0, %ymm3, %ymm0 ## encoding: [0x62,0xf1,0xe5,0x28,0x58,0xc0]
+; CHECK-NEXT: vmovaps %ymm1, %ymm3 ## encoding: [0x62,0xf1,0x7c,0x28,0x28,0xd9]
+; CHECK-NEXT: vfmadd213pd %ymm2, %ymm0, %ymm3 {%k1} {z} ## encoding: [0x62,0xf2,0xfd,0xa9,0xa8,0xda]
+; CHECK-NEXT: vfmadd213pd %ymm2, %ymm0, %ymm1 ## encoding: [0x62,0xf2,0xfd,0x28,0xa8,0xca]
+; CHECK-NEXT: vaddpd %ymm1, %ymm3, %ymm0 ## encoding: [0x62,0xf1,0xe5,0x28,0x58,0xc1]
; CHECK-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)
%res1 = call <4 x double> @llvm.x86.avx512.maskz.vfmadd.pd.256(<4 x double> %x0, <4 x double> %x1, <4 x double> %x2, i8 -1)
@@ -1036,9 +1036,9 @@ define <4 x float>@test_int_x86_avx512_m
; CHECK: ## BB#0:
; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT: vmovaps %xmm0, %xmm3 ## encoding: [0x62,0xf1,0x7c,0x08,0x28,0xd8]
-; CHECK-NEXT: vfmadd213ps %xmm2, %xmm1, %xmm3 {%k1} ## encoding: [0x62,0xf2,0x75,0x09,0xa8,0xda]
-; CHECK-NEXT: vfmadd213ps %xmm2, %xmm1, %xmm0 ## encoding: [0x62,0xf2,0x75,0x08,0xa8,0xc2]
-; CHECK-NEXT: vaddps %xmm0, %xmm3, %xmm0 ## encoding: [0x62,0xf1,0x64,0x08,0x58,0xc0]
+; CHECK-NEXT: vfmadd132ps %xmm1, %xmm2, %xmm3 {%k1} ## encoding: [0x62,0xf2,0x6d,0x09,0x98,0xd9]
+; CHECK-NEXT: vfmadd213ps %xmm2, %xmm0, %xmm1 ## encoding: [0x62,0xf2,0x7d,0x08,0xa8,0xca]
+; CHECK-NEXT: vaddps %xmm1, %xmm3, %xmm0 ## encoding: [0x62,0xf1,0x64,0x08,0x58,0xc1]
; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <4 x float> @llvm.x86.avx512.mask.vfmadd.ps.128(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2, i8 %x3)
%res1 = call <4 x float> @llvm.x86.avx512.mask.vfmadd.ps.128(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2, i8 -1)
@@ -1054,8 +1054,8 @@ define <4 x float>@test_int_x86_avx512_m
; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT: vmovaps %xmm2, %xmm3 ## encoding: [0x62,0xf1,0x7c,0x08,0x28,0xda]
; CHECK-NEXT: vfmadd231ps %xmm1, %xmm0, %xmm3 {%k1} ## encoding: [0x62,0xf2,0x7d,0x09,0xb8,0xd9]
-; CHECK-NEXT: vfmadd213ps %xmm2, %xmm1, %xmm0 ## encoding: [0x62,0xf2,0x75,0x08,0xa8,0xc2]
-; CHECK-NEXT: vaddps %xmm0, %xmm3, %xmm0 ## encoding: [0x62,0xf1,0x64,0x08,0x58,0xc0]
+; CHECK-NEXT: vfmadd213ps %xmm2, %xmm0, %xmm1 ## encoding: [0x62,0xf2,0x7d,0x08,0xa8,0xca]
+; CHECK-NEXT: vaddps %xmm1, %xmm3, %xmm0 ## encoding: [0x62,0xf1,0x64,0x08,0x58,0xc1]
; CHECK-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)
%res1 = call <4 x float> @llvm.x86.avx512.mask3.vfmadd.ps.128(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2, i8 -1)
@@ -1069,10 +1069,10 @@ define <4 x float>@test_int_x86_avx512_m
; CHECK-LABEL: test_int_x86_avx512_maskz_vfmadd_ps_128:
; CHECK: ## BB#0:
; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vmovaps %xmm0, %xmm3 ## encoding: [0x62,0xf1,0x7c,0x08,0x28,0xd8]
-; CHECK-NEXT: vfmadd213ps %xmm2, %xmm1, %xmm3 {%k1} {z} ## encoding: [0x62,0xf2,0x75,0x89,0xa8,0xda]
-; CHECK-NEXT: vfmadd213ps %xmm2, %xmm1, %xmm0 ## encoding: [0x62,0xf2,0x75,0x08,0xa8,0xc2]
-; CHECK-NEXT: vaddps %xmm0, %xmm3, %xmm0 ## encoding: [0x62,0xf1,0x64,0x08,0x58,0xc0]
+; CHECK-NEXT: vmovaps %xmm1, %xmm3 ## encoding: [0x62,0xf1,0x7c,0x08,0x28,0xd9]
+; CHECK-NEXT: vfmadd213ps %xmm2, %xmm0, %xmm3 {%k1} {z} ## encoding: [0x62,0xf2,0x7d,0x89,0xa8,0xda]
+; CHECK-NEXT: vfmadd213ps %xmm2, %xmm0, %xmm1 ## encoding: [0x62,0xf2,0x7d,0x08,0xa8,0xca]
+; CHECK-NEXT: vaddps %xmm1, %xmm3, %xmm0 ## encoding: [0x62,0xf1,0x64,0x08,0x58,0xc1]
; CHECK-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)
%res1 = call <4 x float> @llvm.x86.avx512.maskz.vfmadd.ps.128(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2, i8 -1)
@@ -1085,9 +1085,9 @@ define <8 x float>@test_int_x86_avx512_m
; CHECK: ## BB#0:
; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT: vmovaps %ymm0, %ymm3 ## encoding: [0x62,0xf1,0x7c,0x28,0x28,0xd8]
-; CHECK-NEXT: vfmadd213ps %ymm2, %ymm1, %ymm3 {%k1} ## encoding: [0x62,0xf2,0x75,0x29,0xa8,0xda]
-; CHECK-NEXT: vfmadd213ps %ymm2, %ymm1, %ymm0 ## encoding: [0x62,0xf2,0x75,0x28,0xa8,0xc2]
-; CHECK-NEXT: vaddps %ymm0, %ymm3, %ymm0 ## encoding: [0x62,0xf1,0x64,0x28,0x58,0xc0]
+; CHECK-NEXT: vfmadd132ps %ymm1, %ymm2, %ymm3 {%k1} ## encoding: [0x62,0xf2,0x6d,0x29,0x98,0xd9]
+; CHECK-NEXT: vfmadd213ps %ymm2, %ymm0, %ymm1 ## encoding: [0x62,0xf2,0x7d,0x28,0xa8,0xca]
+; CHECK-NEXT: vaddps %ymm1, %ymm3, %ymm0 ## encoding: [0x62,0xf1,0x64,0x28,0x58,0xc1]
; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <8 x float> @llvm.x86.avx512.mask.vfmadd.ps.256(<8 x float> %x0, <8 x float> %x1, <8 x float> %x2, i8 %x3)
%res1 = call <8 x float> @llvm.x86.avx512.mask.vfmadd.ps.256(<8 x float> %x0, <8 x float> %x1, <8 x float> %x2, i8 -1)
@@ -1103,8 +1103,8 @@ define <8 x float>@test_int_x86_avx512_m
; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT: vmovaps %ymm2, %ymm3 ## encoding: [0x62,0xf1,0x7c,0x28,0x28,0xda]
; CHECK-NEXT: vfmadd231ps %ymm1, %ymm0, %ymm3 {%k1} ## encoding: [0x62,0xf2,0x7d,0x29,0xb8,0xd9]
-; CHECK-NEXT: vfmadd213ps %ymm2, %ymm1, %ymm0 ## encoding: [0x62,0xf2,0x75,0x28,0xa8,0xc2]
-; CHECK-NEXT: vaddps %ymm0, %ymm3, %ymm0 ## encoding: [0x62,0xf1,0x64,0x28,0x58,0xc0]
+; CHECK-NEXT: vfmadd213ps %ymm2, %ymm0, %ymm1 ## encoding: [0x62,0xf2,0x7d,0x28,0xa8,0xca]
+; CHECK-NEXT: vaddps %ymm1, %ymm3, %ymm0 ## encoding: [0x62,0xf1,0x64,0x28,0x58,0xc1]
; CHECK-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)
%res1 = call <8 x float> @llvm.x86.avx512.mask3.vfmadd.ps.256(<8 x float> %x0, <8 x float> %x1, <8 x float> %x2, i8 -1)
@@ -1118,10 +1118,10 @@ define <8 x float>@test_int_x86_avx512_m
; CHECK-LABEL: test_int_x86_avx512_maskz_vfmadd_ps_256:
; CHECK: ## BB#0:
; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vmovaps %ymm0, %ymm3 ## encoding: [0x62,0xf1,0x7c,0x28,0x28,0xd8]
-; CHECK-NEXT: vfmadd213ps %ymm2, %ymm1, %ymm3 {%k1} {z} ## encoding: [0x62,0xf2,0x75,0xa9,0xa8,0xda]
-; CHECK-NEXT: vfmadd213ps %ymm2, %ymm1, %ymm0 ## encoding: [0x62,0xf2,0x75,0x28,0xa8,0xc2]
-; CHECK-NEXT: vaddps %ymm0, %ymm3, %ymm0 ## encoding: [0x62,0xf1,0x64,0x28,0x58,0xc0]
+; CHECK-NEXT: vmovaps %ymm1, %ymm3 ## encoding: [0x62,0xf1,0x7c,0x28,0x28,0xd9]
+; CHECK-NEXT: vfmadd213ps %ymm2, %ymm0, %ymm3 {%k1} {z} ## encoding: [0x62,0xf2,0x7d,0xa9,0xa8,0xda]
+; CHECK-NEXT: vfmadd213ps %ymm2, %ymm0, %ymm1 ## encoding: [0x62,0xf2,0x7d,0x28,0xa8,0xca]
+; CHECK-NEXT: vaddps %ymm1, %ymm3, %ymm0 ## encoding: [0x62,0xf1,0x64,0x28,0x58,0xc1]
; CHECK-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)
%res1 = call <8 x float> @llvm.x86.avx512.maskz.vfmadd.ps.256(<8 x float> %x0, <8 x float> %x1, <8 x float> %x2, i8 -1)
@@ -1138,8 +1138,8 @@ define <2 x double>@test_int_x86_avx512_
; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT: vmovaps %xmm2, %xmm3 ## encoding: [0x62,0xf1,0x7c,0x08,0x28,0xda]
; CHECK-NEXT: vfmsub231pd %xmm1, %xmm0, %xmm3 {%k1} ## encoding: [0x62,0xf2,0xfd,0x09,0xba,0xd9]
-; CHECK-NEXT: vfmsub213pd %xmm2, %xmm1, %xmm0 ## encoding: [0x62,0xf2,0xf5,0x08,0xaa,0xc2]
-; CHECK-NEXT: vaddpd %xmm0, %xmm3, %xmm0 ## encoding: [0x62,0xf1,0xe5,0x08,0x58,0xc0]
+; CHECK-NEXT: vfmsub213pd %xmm2, %xmm0, %xmm1 ## encoding: [0x62,0xf2,0xfd,0x08,0xaa,0xca]
+; CHECK-NEXT: vaddpd %xmm1, %xmm3, %xmm0 ## encoding: [0x62,0xf1,0xe5,0x08,0x58,0xc1]
; CHECK-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)
%res1 = call <2 x double> @llvm.x86.avx512.mask3.vfmsub.pd.128(<2 x double> %x0, <2 x double> %x1, <2 x double> %x2, i8 -1)
@@ -1156,8 +1156,8 @@ define <4 x double>@test_int_x86_avx512_
; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT: vmovaps %ymm2, %ymm3 ## encoding: [0x62,0xf1,0x7c,0x28,0x28,0xda]
; CHECK-NEXT: vfmsub231pd %ymm1, %ymm0, %ymm3 {%k1} ## encoding: [0x62,0xf2,0xfd,0x29,0xba,0xd9]
-; CHECK-NEXT: vfmsub213pd %ymm2, %ymm1, %ymm0 ## encoding: [0x62,0xf2,0xf5,0x28,0xaa,0xc2]
-; CHECK-NEXT: vaddpd %ymm0, %ymm3, %ymm0 ## encoding: [0x62,0xf1,0xe5,0x28,0x58,0xc0]
+; CHECK-NEXT: vfmsub213pd %ymm2, %ymm0, %ymm1 ## encoding: [0x62,0xf2,0xfd,0x28,0xaa,0xca]
+; CHECK-NEXT: vaddpd %ymm1, %ymm3, %ymm0 ## encoding: [0x62,0xf1,0xe5,0x28,0x58,0xc1]
; CHECK-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)
%res1 = call <4 x double> @llvm.x86.avx512.mask3.vfmsub.pd.256(<4 x double> %x0, <4 x double> %x1, <4 x double> %x2, i8 -1)
@@ -1173,8 +1173,8 @@ define <4 x float>@test_int_x86_avx512_m
; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT: vmovaps %xmm2, %xmm3 ## encoding: [0x62,0xf1,0x7c,0x08,0x28,0xda]
; CHECK-NEXT: vfmsub231ps %xmm1, %xmm0, %xmm3 {%k1} ## encoding: [0x62,0xf2,0x7d,0x09,0xba,0xd9]
-; CHECK-NEXT: vfmsub213ps %xmm2, %xmm1, %xmm0 ## encoding: [0x62,0xf2,0x75,0x08,0xaa,0xc2]
-; CHECK-NEXT: vaddps %xmm0, %xmm3, %xmm0 ## encoding: [0x62,0xf1,0x64,0x08,0x58,0xc0]
+; CHECK-NEXT: vfmsub213ps %xmm2, %xmm0, %xmm1 ## encoding: [0x62,0xf2,0x7d,0x08,0xaa,0xca]
+; CHECK-NEXT: vaddps %xmm1, %xmm3, %xmm0 ## encoding: [0x62,0xf1,0x64,0x08,0x58,0xc1]
; CHECK-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)
%res1 = call <4 x float> @llvm.x86.avx512.mask3.vfmsub.ps.128(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2, i8 -1)
@@ -1190,8 +1190,8 @@ define <8 x float>@test_int_x86_avx512_m
; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT: vmovaps %ymm2, %ymm3 ## encoding: [0x62,0xf1,0x7c,0x28,0x28,0xda]
; CHECK-NEXT: vfmsub231ps %ymm1, %ymm0, %ymm3 {%k1} ## encoding: [0x62,0xf2,0x7d,0x29,0xba,0xd9]
-; CHECK-NEXT: vfmsub213ps %ymm2, %ymm1, %ymm0 ## encoding: [0x62,0xf2,0x75,0x28,0xaa,0xc2]
-; CHECK-NEXT: vaddps %ymm0, %ymm3, %ymm0 ## encoding: [0x62,0xf1,0x64,0x28,0x58,0xc0]
+; CHECK-NEXT: vfmsub213ps %ymm2, %ymm0, %ymm1 ## encoding: [0x62,0xf2,0x7d,0x28,0xaa,0xca]
+; CHECK-NEXT: vaddps %ymm1, %ymm3, %ymm0 ## encoding: [0x62,0xf1,0x64,0x28,0x58,0xc1]
; CHECK-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)
%res1 = call <8 x float> @llvm.x86.avx512.mask3.vfmsub.ps.256(<8 x float> %x0, <8 x float> %x1, <8 x float> %x2, i8 -1)
@@ -1205,7 +1205,7 @@ define <8 x float> @test_mask_vfnmadd256
; CHECK-LABEL: test_mask_vfnmadd256_ps:
; CHECK: ## BB#0:
; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfnmadd213ps %ymm2, %ymm1, %ymm0 {%k1} ## encoding: [0x62,0xf2,0x75,0x29,0xac,0xc2]
+; CHECK-NEXT: vfnmadd132ps %ymm1, %ymm2, %ymm0 {%k1} ## encoding: [0x62,0xf2,0x6d,0x29,0x9c,0xc1]
; CHECK-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
@@ -1217,7 +1217,7 @@ define <4 x float> @test_mask_vfnmadd128
; CHECK-LABEL: test_mask_vfnmadd128_ps:
; CHECK: ## BB#0:
; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfnmadd213ps %xmm2, %xmm1, %xmm0 {%k1} ## encoding: [0x62,0xf2,0x75,0x09,0xac,0xc2]
+; CHECK-NEXT: vfnmadd132ps %xmm1, %xmm2, %xmm0 {%k1} ## encoding: [0x62,0xf2,0x6d,0x09,0x9c,0xc1]
; CHECK-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
@@ -1229,7 +1229,7 @@ define <4 x double> @test_mask_vfnmadd25
; CHECK-LABEL: test_mask_vfnmadd256_pd:
; CHECK: ## BB#0:
; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfnmadd213pd %ymm2, %ymm1, %ymm0 {%k1} ## encoding: [0x62,0xf2,0xf5,0x29,0xac,0xc2]
+; CHECK-NEXT: vfnmadd132pd %ymm1, %ymm2, %ymm0 {%k1} ## encoding: [0x62,0xf2,0xed,0x29,0x9c,0xc1]
; CHECK-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
@@ -1241,7 +1241,7 @@ define <2 x double> @test_mask_vfnmadd12
; CHECK-LABEL: test_mask_vfnmadd128_pd:
; CHECK: ## BB#0:
; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfnmadd213pd %xmm2, %xmm1, %xmm0 {%k1} ## encoding: [0x62,0xf2,0xf5,0x09,0xac,0xc2]
+; CHECK-NEXT: vfnmadd132pd %xmm1, %xmm2, %xmm0 {%k1} ## encoding: [0x62,0xf2,0xed,0x09,0x9c,0xc1]
; CHECK-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
@@ -1253,7 +1253,7 @@ define <8 x float> @test_mask_vfnmsub256
; CHECK-LABEL: test_mask_vfnmsub256_ps:
; CHECK: ## BB#0:
; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfnmsub213ps %ymm2, %ymm1, %ymm0 {%k1} ## encoding: [0x62,0xf2,0x75,0x29,0xae,0xc2]
+; CHECK-NEXT: vfnmsub132ps %ymm1, %ymm2, %ymm0 {%k1} ## encoding: [0x62,0xf2,0x6d,0x29,0x9e,0xc1]
; CHECK-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
@@ -1265,7 +1265,7 @@ define <4 x float> @test_mask_vfnmsub128
; CHECK-LABEL: test_mask_vfnmsub128_ps:
; CHECK: ## BB#0:
; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfnmsub213ps %xmm2, %xmm1, %xmm0 {%k1} ## encoding: [0x62,0xf2,0x75,0x09,0xae,0xc2]
+; CHECK-NEXT: vfnmsub132ps %xmm1, %xmm2, %xmm0 {%k1} ## encoding: [0x62,0xf2,0x6d,0x09,0x9e,0xc1]
; CHECK-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
@@ -1277,7 +1277,7 @@ define <4 x double> @test_mask_vfnmsub25
; CHECK-LABEL: test_mask_vfnmsub256_pd:
; CHECK: ## BB#0:
; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfnmsub213pd %ymm2, %ymm1, %ymm0 {%k1} ## encoding: [0x62,0xf2,0xf5,0x29,0xae,0xc2]
+; CHECK-NEXT: vfnmsub132pd %ymm1, %ymm2, %ymm0 {%k1} ## encoding: [0x62,0xf2,0xed,0x29,0x9e,0xc1]
; CHECK-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
@@ -1289,7 +1289,7 @@ define <2 x double> @test_mask_vfnmsub12
; CHECK-LABEL: test_mask_vfnmsub128_pd:
; CHECK: ## BB#0:
; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfnmsub213pd %xmm2, %xmm1, %xmm0 {%k1} ## encoding: [0x62,0xf2,0xf5,0x09,0xae,0xc2]
+; CHECK-NEXT: vfnmsub132pd %xmm1, %xmm2, %xmm0 {%k1} ## encoding: [0x62,0xf2,0xed,0x09,0x9e,0xc1]
; CHECK-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
@@ -1301,9 +1301,9 @@ define <2 x double>@test_int_x86_avx512_
; CHECK: ## BB#0:
; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT: vmovaps %xmm0, %xmm3 ## encoding: [0x62,0xf1,0x7c,0x08,0x28,0xd8]
-; CHECK-NEXT: vfnmsub213pd %xmm2, %xmm1, %xmm3 {%k1} ## encoding: [0x62,0xf2,0xf5,0x09,0xae,0xda]
-; CHECK-NEXT: vfnmsub213pd %xmm2, %xmm1, %xmm0 ## encoding: [0x62,0xf2,0xf5,0x08,0xae,0xc2]
-; CHECK-NEXT: vaddpd %xmm0, %xmm3, %xmm0 ## encoding: [0x62,0xf1,0xe5,0x08,0x58,0xc0]
+; CHECK-NEXT: vfnmsub132pd %xmm1, %xmm2, %xmm3 {%k1} ## encoding: [0x62,0xf2,0xed,0x09,0x9e,0xd9]
+; CHECK-NEXT: vfnmsub213pd %xmm2, %xmm0, %xmm1 ## encoding: [0x62,0xf2,0xfd,0x08,0xae,0xca]
+; CHECK-NEXT: vaddpd %xmm1, %xmm3, %xmm0 ## encoding: [0x62,0xf1,0xe5,0x08,0x58,0xc1]
; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <2 x double> @llvm.x86.avx512.mask.vfnmsub.pd.128(<2 x double> %x0, <2 x double> %x1, <2 x double> %x2, i8 %x3)
%res1 = call <2 x double> @llvm.x86.avx512.mask.vfnmsub.pd.128(<2 x double> %x0, <2 x double> %x1, <2 x double> %x2, i8 -1)
@@ -1319,8 +1319,8 @@ define <2 x double>@test_int_x86_avx512_
; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT: vmovaps %xmm2, %xmm3 ## encoding: [0x62,0xf1,0x7c,0x08,0x28,0xda]
; CHECK-NEXT: vfnmsub231pd %xmm1, %xmm0, %xmm3 {%k1} ## encoding: [0x62,0xf2,0xfd,0x09,0xbe,0xd9]
-; CHECK-NEXT: vfnmsub213pd %xmm2, %xmm1, %xmm0 ## encoding: [0x62,0xf2,0xf5,0x08,0xae,0xc2]
-; CHECK-NEXT: vaddpd %xmm0, %xmm3, %xmm0 ## encoding: [0x62,0xf1,0xe5,0x08,0x58,0xc0]
+; CHECK-NEXT: vfnmsub213pd %xmm2, %xmm0, %xmm1 ## encoding: [0x62,0xf2,0xfd,0x08,0xae,0xca]
+; CHECK-NEXT: vaddpd %xmm1, %xmm3, %xmm0 ## encoding: [0x62,0xf1,0xe5,0x08,0x58,0xc1]
; CHECK-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)
%res1 = call <2 x double> @llvm.x86.avx512.mask3.vfnmsub.pd.128(<2 x double> %x0, <2 x double> %x1, <2 x double> %x2, i8 -1)
@@ -1333,9 +1333,9 @@ define <4 x double>@test_int_x86_avx512_
; CHECK: ## BB#0:
; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT: vmovaps %ymm0, %ymm3 ## encoding: [0x62,0xf1,0x7c,0x28,0x28,0xd8]
-; CHECK-NEXT: vfnmsub213pd %ymm2, %ymm1, %ymm3 {%k1} ## encoding: [0x62,0xf2,0xf5,0x29,0xae,0xda]
-; CHECK-NEXT: vfnmsub213pd %ymm2, %ymm1, %ymm0 ## encoding: [0x62,0xf2,0xf5,0x28,0xae,0xc2]
-; CHECK-NEXT: vaddpd %ymm0, %ymm3, %ymm0 ## encoding: [0x62,0xf1,0xe5,0x28,0x58,0xc0]
+; CHECK-NEXT: vfnmsub132pd %ymm1, %ymm2, %ymm3 {%k1} ## encoding: [0x62,0xf2,0xed,0x29,0x9e,0xd9]
+; CHECK-NEXT: vfnmsub213pd %ymm2, %ymm0, %ymm1 ## encoding: [0x62,0xf2,0xfd,0x28,0xae,0xca]
+; CHECK-NEXT: vaddpd %ymm1, %ymm3, %ymm0 ## encoding: [0x62,0xf1,0xe5,0x28,0x58,0xc1]
; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <4 x double> @llvm.x86.avx512.mask.vfnmsub.pd.256(<4 x double> %x0, <4 x double> %x1, <4 x double> %x2, i8 %x3)
%res1 = call <4 x double> @llvm.x86.avx512.mask.vfnmsub.pd.256(<4 x double> %x0, <4 x double> %x1, <4 x double> %x2, i8 -1)
@@ -1351,8 +1351,8 @@ define <4 x double>@test_int_x86_avx512_
; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT: vmovaps %ymm2, %ymm3 ## encoding: [0x62,0xf1,0x7c,0x28,0x28,0xda]
; CHECK-NEXT: vfnmsub231pd %ymm1, %ymm0, %ymm3 {%k1} ## encoding: [0x62,0xf2,0xfd,0x29,0xbe,0xd9]
-; CHECK-NEXT: vfnmsub213pd %ymm2, %ymm1, %ymm0 ## encoding: [0x62,0xf2,0xf5,0x28,0xae,0xc2]
-; CHECK-NEXT: vaddpd %ymm0, %ymm3, %ymm0 ## encoding: [0x62,0xf1,0xe5,0x28,0x58,0xc0]
+; CHECK-NEXT: vfnmsub213pd %ymm2, %ymm0, %ymm1 ## encoding: [0x62,0xf2,0xfd,0x28,0xae,0xca]
+; CHECK-NEXT: vaddpd %ymm1, %ymm3, %ymm0 ## encoding: [0x62,0xf1,0xe5,0x28,0x58,0xc1]
; CHECK-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)
%res1 = call <4 x double> @llvm.x86.avx512.mask3.vfnmsub.pd.256(<4 x double> %x0, <4 x double> %x1, <4 x double> %x2, i8 -1)
@@ -1365,9 +1365,9 @@ define <4 x float>@test_int_x86_avx512_m
; CHECK: ## BB#0:
; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT: vmovaps %xmm0, %xmm3 ## encoding: [0x62,0xf1,0x7c,0x08,0x28,0xd8]
-; CHECK-NEXT: vfnmsub213ps %xmm2, %xmm1, %xmm3 {%k1} ## encoding: [0x62,0xf2,0x75,0x09,0xae,0xda]
-; CHECK-NEXT: vfnmsub213ps %xmm2, %xmm1, %xmm0 ## encoding: [0x62,0xf2,0x75,0x08,0xae,0xc2]
-; CHECK-NEXT: vaddps %xmm0, %xmm3, %xmm0 ## encoding: [0x62,0xf1,0x64,0x08,0x58,0xc0]
+; CHECK-NEXT: vfnmsub132ps %xmm1, %xmm2, %xmm3 {%k1} ## encoding: [0x62,0xf2,0x6d,0x09,0x9e,0xd9]
+; CHECK-NEXT: vfnmsub213ps %xmm2, %xmm0, %xmm1 ## encoding: [0x62,0xf2,0x7d,0x08,0xae,0xca]
+; CHECK-NEXT: vaddps %xmm1, %xmm3, %xmm0 ## encoding: [0x62,0xf1,0x64,0x08,0x58,0xc1]
; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <4 x float> @llvm.x86.avx512.mask.vfnmsub.ps.128(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2, i8 %x3)
%res1 = call <4 x float> @llvm.x86.avx512.mask.vfnmsub.ps.128(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2, i8 -1)
@@ -1383,8 +1383,8 @@ define <4 x float>@test_int_x86_avx512_m
; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT: vmovaps %xmm2, %xmm3 ## encoding: [0x62,0xf1,0x7c,0x08,0x28,0xda]
; CHECK-NEXT: vfnmsub231ps %xmm1, %xmm0, %xmm3 {%k1} ## encoding: [0x62,0xf2,0x7d,0x09,0xbe,0xd9]
-; CHECK-NEXT: vfnmsub213ps %xmm2, %xmm1, %xmm0 ## encoding: [0x62,0xf2,0x75,0x08,0xae,0xc2]
-; CHECK-NEXT: vaddps %xmm0, %xmm3, %xmm0 ## encoding: [0x62,0xf1,0x64,0x08,0x58,0xc0]
+; CHECK-NEXT: vfnmsub213ps %xmm2, %xmm0, %xmm1 ## encoding: [0x62,0xf2,0x7d,0x08,0xae,0xca]
+; CHECK-NEXT: vaddps %xmm1, %xmm3, %xmm0 ## encoding: [0x62,0xf1,0x64,0x08,0x58,0xc1]
; CHECK-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)
%res1 = call <4 x float> @llvm.x86.avx512.mask3.vfnmsub.ps.128(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2, i8 -1)
@@ -1397,9 +1397,9 @@ define <8 x float>@test_int_x86_avx512_m
; CHECK: ## BB#0:
; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT: vmovaps %ymm0, %ymm3 ## encoding: [0x62,0xf1,0x7c,0x28,0x28,0xd8]
-; CHECK-NEXT: vfnmsub213ps %ymm2, %ymm1, %ymm3 {%k1} ## encoding: [0x62,0xf2,0x75,0x29,0xae,0xda]
-; CHECK-NEXT: vfnmsub213ps %ymm2, %ymm1, %ymm0 ## encoding: [0x62,0xf2,0x75,0x28,0xae,0xc2]
-; CHECK-NEXT: vaddps %ymm0, %ymm3, %ymm0 ## encoding: [0x62,0xf1,0x64,0x28,0x58,0xc0]
+; CHECK-NEXT: vfnmsub132ps %ymm1, %ymm2, %ymm3 {%k1} ## encoding: [0x62,0xf2,0x6d,0x29,0x9e,0xd9]
+; CHECK-NEXT: vfnmsub213ps %ymm2, %ymm0, %ymm1 ## encoding: [0x62,0xf2,0x7d,0x28,0xae,0xca]
+; CHECK-NEXT: vaddps %ymm1, %ymm3, %ymm0 ## encoding: [0x62,0xf1,0x64,0x28,0x58,0xc1]
; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <8 x float> @llvm.x86.avx512.mask.vfnmsub.ps.256(<8 x float> %x0, <8 x float> %x1, <8 x float> %x2, i8 %x3)
%res1 = call <8 x float> @llvm.x86.avx512.mask.vfnmsub.ps.256(<8 x float> %x0, <8 x float> %x1, <8 x float> %x2, i8 -1)
@@ -1415,8 +1415,8 @@ define <8 x float>@test_int_x86_avx512_m
; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT: vmovaps %ymm2, %ymm3 ## encoding: [0x62,0xf1,0x7c,0x28,0x28,0xda]
; CHECK-NEXT: vfnmsub231ps %ymm1, %ymm0, %ymm3 {%k1} ## encoding: [0x62,0xf2,0x7d,0x29,0xbe,0xd9]
-; CHECK-NEXT: vfnmsub213ps %ymm2, %ymm1, %ymm0 ## encoding: [0x62,0xf2,0x75,0x28,0xae,0xc2]
-; CHECK-NEXT: vaddps %ymm0, %ymm3, %ymm0 ## encoding: [0x62,0xf1,0x64,0x28,0x58,0xc0]
+; CHECK-NEXT: vfnmsub213ps %ymm2, %ymm0, %ymm1 ## encoding: [0x62,0xf2,0x7d,0x28,0xae,0xca]
+; CHECK-NEXT: vaddps %ymm1, %ymm3, %ymm0 ## encoding: [0x62,0xf1,0x64,0x28,0x58,0xc1]
; CHECK-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)
%res1 = call <8 x float> @llvm.x86.avx512.mask3.vfnmsub.ps.256(<8 x float> %x0, <8 x float> %x1, <8 x float> %x2, i8 -1)
@@ -1429,9 +1429,9 @@ define <2 x double>@test_int_x86_avx512_
; CHECK: ## BB#0:
; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT: vmovaps %xmm0, %xmm3 ## encoding: [0x62,0xf1,0x7c,0x08,0x28,0xd8]
-; CHECK-NEXT: vfnmadd213pd %xmm2, %xmm1, %xmm3 {%k1} ## encoding: [0x62,0xf2,0xf5,0x09,0xac,0xda]
-; CHECK-NEXT: vfnmadd213pd %xmm2, %xmm1, %xmm0 ## encoding: [0x62,0xf2,0xf5,0x08,0xac,0xc2]
-; CHECK-NEXT: vaddpd %xmm0, %xmm3, %xmm0 ## encoding: [0x62,0xf1,0xe5,0x08,0x58,0xc0]
+; CHECK-NEXT: vfnmadd132pd %xmm1, %xmm2, %xmm3 {%k1} ## encoding: [0x62,0xf2,0xed,0x09,0x9c,0xd9]
+; CHECK-NEXT: vfnmadd213pd %xmm2, %xmm0, %xmm1 ## encoding: [0x62,0xf2,0xfd,0x08,0xac,0xca]
+; CHECK-NEXT: vaddpd %xmm1, %xmm3, %xmm0 ## encoding: [0x62,0xf1,0xe5,0x08,0x58,0xc1]
; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <2 x double> @llvm.x86.avx512.mask.vfnmadd.pd.128(<2 x double> %x0, <2 x double> %x1, <2 x double> %x2, i8 %x3)
%res1 = call <2 x double> @llvm.x86.avx512.mask.vfnmadd.pd.128(<2 x double> %x0, <2 x double> %x1, <2 x double> %x2, i8 -1)
@@ -1444,9 +1444,9 @@ define <4 x double>@test_int_x86_avx512_
; CHECK: ## BB#0:
; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT: vmovaps %ymm0, %ymm3 ## encoding: [0x62,0xf1,0x7c,0x28,0x28,0xd8]
-; CHECK-NEXT: vfnmadd213pd %ymm2, %ymm1, %ymm3 {%k1} ## encoding: [0x62,0xf2,0xf5,0x29,0xac,0xda]
-; CHECK-NEXT: vfnmadd213pd %ymm2, %ymm1, %ymm0 ## encoding: [0x62,0xf2,0xf5,0x28,0xac,0xc2]
-; CHECK-NEXT: vaddpd %ymm0, %ymm3, %ymm0 ## encoding: [0x62,0xf1,0xe5,0x28,0x58,0xc0]
+; CHECK-NEXT: vfnmadd132pd %ymm1, %ymm2, %ymm3 {%k1} ## encoding: [0x62,0xf2,0xed,0x29,0x9c,0xd9]
+; CHECK-NEXT: vfnmadd213pd %ymm2, %ymm0, %ymm1 ## encoding: [0x62,0xf2,0xfd,0x28,0xac,0xca]
+; CHECK-NEXT: vaddpd %ymm1, %ymm3, %ymm0 ## encoding: [0x62,0xf1,0xe5,0x28,0x58,0xc1]
; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <4 x double> @llvm.x86.avx512.mask.vfnmadd.pd.256(<4 x double> %x0, <4 x double> %x1, <4 x double> %x2, i8 %x3)
%res1 = call <4 x double> @llvm.x86.avx512.mask.vfnmadd.pd.256(<4 x double> %x0, <4 x double> %x1, <4 x double> %x2, i8 -1)
@@ -1459,9 +1459,9 @@ define <4 x float>@test_int_x86_avx512_m
; CHECK: ## BB#0:
; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT: vmovaps %xmm0, %xmm3 ## encoding: [0x62,0xf1,0x7c,0x08,0x28,0xd8]
-; CHECK-NEXT: vfnmadd213ps %xmm2, %xmm1, %xmm3 {%k1} ## encoding: [0x62,0xf2,0x75,0x09,0xac,0xda]
-; CHECK-NEXT: vfnmadd213ps %xmm2, %xmm1, %xmm0 ## encoding: [0x62,0xf2,0x75,0x08,0xac,0xc2]
-; CHECK-NEXT: vaddps %xmm0, %xmm3, %xmm0 ## encoding: [0x62,0xf1,0x64,0x08,0x58,0xc0]
+; CHECK-NEXT: vfnmadd132ps %xmm1, %xmm2, %xmm3 {%k1} ## encoding: [0x62,0xf2,0x6d,0x09,0x9c,0xd9]
+; CHECK-NEXT: vfnmadd213ps %xmm2, %xmm0, %xmm1 ## encoding: [0x62,0xf2,0x7d,0x08,0xac,0xca]
+; CHECK-NEXT: vaddps %xmm1, %xmm3, %xmm0 ## encoding: [0x62,0xf1,0x64,0x08,0x58,0xc1]
; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <4 x float> @llvm.x86.avx512.mask.vfnmadd.ps.128(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2, i8 %x3)
%res1 = call <4 x float> @llvm.x86.avx512.mask.vfnmadd.ps.128(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2, i8 -1)
@@ -1474,9 +1474,9 @@ define <8 x float>@test_int_x86_avx512_m
; CHECK: ## BB#0:
; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT: vmovaps %ymm0, %ymm3 ## encoding: [0x62,0xf1,0x7c,0x28,0x28,0xd8]
-; CHECK-NEXT: vfnmadd213ps %ymm2, %ymm1, %ymm3 {%k1} ## encoding: [0x62,0xf2,0x75,0x29,0xac,0xda]
-; CHECK-NEXT: vfnmadd213ps %ymm2, %ymm1, %ymm0 ## encoding: [0x62,0xf2,0x75,0x28,0xac,0xc2]
-; CHECK-NEXT: vaddps %ymm0, %ymm3, %ymm0 ## encoding: [0x62,0xf1,0x64,0x28,0x58,0xc0]
+; CHECK-NEXT: vfnmadd132ps %ymm1, %ymm2, %ymm3 {%k1} ## encoding: [0x62,0xf2,0x6d,0x29,0x9c,0xd9]
+; CHECK-NEXT: vfnmadd213ps %ymm2, %ymm0, %ymm1 ## encoding: [0x62,0xf2,0x7d,0x28,0xac,0xca]
+; CHECK-NEXT: vaddps %ymm1, %ymm3, %ymm0 ## encoding: [0x62,0xf1,0x64,0x28,0x58,0xc1]
; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <8 x float> @llvm.x86.avx512.mask.vfnmadd.ps.256(<8 x float> %x0, <8 x float> %x1, <8 x float> %x2, i8 %x3)
%res1 = call <8 x float> @llvm.x86.avx512.mask.vfnmadd.ps.256(<8 x float> %x0, <8 x float> %x1, <8 x float> %x2, i8 -1)
@@ -1490,7 +1490,7 @@ define <8 x float> @test_mask_fmaddsub25
; CHECK-LABEL: test_mask_fmaddsub256_ps:
; CHECK: ## BB#0:
; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfmaddsub213ps %ymm2, %ymm1, %ymm0 {%k1} ## encoding: [0x62,0xf2,0x75,0x29,0xa6,0xc2]
+; CHECK-NEXT: vfmaddsub132ps %ymm1, %ymm2, %ymm0 {%k1} ## encoding: [0x62,0xf2,0x6d,0x29,0x96,0xc1]
; CHECK-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
@@ -1502,7 +1502,7 @@ define <4 x float> @test_mask_fmaddsub12
; CHECK-LABEL: test_mask_fmaddsub128_ps:
; CHECK: ## BB#0:
; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfmaddsub213ps %xmm2, %xmm1, %xmm0 {%k1} ## encoding: [0x62,0xf2,0x75,0x09,0xa6,0xc2]
+; CHECK-NEXT: vfmaddsub132ps %xmm1, %xmm2, %xmm0 {%k1} ## encoding: [0x62,0xf2,0x6d,0x09,0x96,0xc1]
; CHECK-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
@@ -1514,7 +1514,7 @@ define <4 x double> @test_mask_vfmaddsub
; CHECK-LABEL: test_mask_vfmaddsub256_pd:
; CHECK: ## BB#0:
; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfmaddsub213pd %ymm2, %ymm1, %ymm0 {%k1} ## encoding: [0x62,0xf2,0xf5,0x29,0xa6,0xc2]
+; CHECK-NEXT: vfmaddsub132pd %ymm1, %ymm2, %ymm0 {%k1} ## encoding: [0x62,0xf2,0xed,0x29,0x96,0xc1]
; CHECK-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
@@ -1526,7 +1526,7 @@ define <2 x double> @test_mask_vfmaddsub
; CHECK-LABEL: test_mask_vfmaddsub128_pd:
; CHECK: ## BB#0:
; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfmaddsub213pd %xmm2, %xmm1, %xmm0 {%k1} ## encoding: [0x62,0xf2,0xf5,0x09,0xa6,0xc2]
+; CHECK-NEXT: vfmaddsub132pd %xmm1, %xmm2, %xmm0 {%k1} ## encoding: [0x62,0xf2,0xed,0x09,0x96,0xc1]
; CHECK-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
@@ -1537,9 +1537,9 @@ define <2 x double>@test_int_x86_avx512_
; CHECK: ## BB#0:
; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT: vmovaps %xmm0, %xmm3 ## encoding: [0x62,0xf1,0x7c,0x08,0x28,0xd8]
-; CHECK-NEXT: vfmaddsub213pd %xmm2, %xmm1, %xmm3 {%k1} ## encoding: [0x62,0xf2,0xf5,0x09,0xa6,0xda]
-; CHECK-NEXT: vfmaddsub213pd %xmm2, %xmm1, %xmm0 ## encoding: [0x62,0xf2,0xf5,0x08,0xa6,0xc2]
-; CHECK-NEXT: vaddpd %xmm0, %xmm3, %xmm0 ## encoding: [0x62,0xf1,0xe5,0x08,0x58,0xc0]
+; CHECK-NEXT: vfmaddsub132pd %xmm1, %xmm2, %xmm3 {%k1} ## encoding: [0x62,0xf2,0xed,0x09,0x96,0xd9]
+; CHECK-NEXT: vfmaddsub213pd %xmm2, %xmm0, %xmm1 ## encoding: [0x62,0xf2,0xfd,0x08,0xa6,0xca]
+; CHECK-NEXT: vaddpd %xmm1, %xmm3, %xmm0 ## encoding: [0x62,0xf1,0xe5,0x08,0x58,0xc1]
; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <2 x double> @llvm.x86.avx512.mask.vfmaddsub.pd.128(<2 x double> %x0, <2 x double> %x1, <2 x double> %x2, i8 %x3)
%res1 = call <2 x double> @llvm.x86.avx512.mask.vfmaddsub.pd.128(<2 x double> %x0, <2 x double> %x1, <2 x double> %x2, i8 -1)
@@ -1555,8 +1555,8 @@ define <2 x double>@test_int_x86_avx512_
; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT: vmovaps %xmm2, %xmm3 ## encoding: [0x62,0xf1,0x7c,0x08,0x28,0xda]
; CHECK-NEXT: vfmaddsub231pd %xmm1, %xmm0, %xmm3 {%k1} ## encoding: [0x62,0xf2,0xfd,0x09,0xb6,0xd9]
-; CHECK-NEXT: vfmaddsub213pd %xmm2, %xmm1, %xmm0 ## encoding: [0x62,0xf2,0xf5,0x08,0xa6,0xc2]
-; CHECK-NEXT: vaddpd %xmm0, %xmm3, %xmm0 ## encoding: [0x62,0xf1,0xe5,0x08,0x58,0xc0]
+; CHECK-NEXT: vfmaddsub213pd %xmm2, %xmm0, %xmm1 ## encoding: [0x62,0xf2,0xfd,0x08,0xa6,0xca]
+; CHECK-NEXT: vaddpd %xmm1, %xmm3, %xmm0 ## encoding: [0x62,0xf1,0xe5,0x08,0x58,0xc1]
; CHECK-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)
%res1 = call <2 x double> @llvm.x86.avx512.mask3.vfmaddsub.pd.128(<2 x double> %x0, <2 x double> %x1, <2 x double> %x2, i8 -1)
@@ -1570,10 +1570,10 @@ define <2 x double>@test_int_x86_avx512_
; CHECK-LABEL: test_int_x86_avx512_maskz_vfmaddsub_pd_128:
; CHECK: ## BB#0:
; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vmovaps %xmm0, %xmm3 ## encoding: [0x62,0xf1,0x7c,0x08,0x28,0xd8]
-; CHECK-NEXT: vfmaddsub213pd %xmm2, %xmm1, %xmm3 {%k1} {z} ## encoding: [0x62,0xf2,0xf5,0x89,0xa6,0xda]
-; CHECK-NEXT: vfmaddsub213pd %xmm2, %xmm1, %xmm0 ## encoding: [0x62,0xf2,0xf5,0x08,0xa6,0xc2]
-; CHECK-NEXT: vaddpd %xmm0, %xmm3, %xmm0 ## encoding: [0x62,0xf1,0xe5,0x08,0x58,0xc0]
+; CHECK-NEXT: vmovaps %xmm1, %xmm3 ## encoding: [0x62,0xf1,0x7c,0x08,0x28,0xd9]
+; CHECK-NEXT: vfmaddsub213pd %xmm2, %xmm0, %xmm3 {%k1} {z} ## encoding: [0x62,0xf2,0xfd,0x89,0xa6,0xda]
+; CHECK-NEXT: vfmaddsub213pd %xmm2, %xmm0, %xmm1 ## encoding: [0x62,0xf2,0xfd,0x08,0xa6,0xca]
+; CHECK-NEXT: vaddpd %xmm1, %xmm3, %xmm0 ## encoding: [0x62,0xf1,0xe5,0x08,0x58,0xc1]
; CHECK-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)
%res1 = call <2 x double> @llvm.x86.avx512.maskz.vfmaddsub.pd.128(<2 x double> %x0, <2 x double> %x1, <2 x double> %x2, i8 -1)
@@ -1586,9 +1586,9 @@ define <4 x double>@test_int_x86_avx512_
; CHECK: ## BB#0:
; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT: vmovaps %ymm0, %ymm3 ## encoding: [0x62,0xf1,0x7c,0x28,0x28,0xd8]
-; CHECK-NEXT: vfmaddsub213pd %ymm2, %ymm1, %ymm3 {%k1} ## encoding: [0x62,0xf2,0xf5,0x29,0xa6,0xda]
-; CHECK-NEXT: vfmaddsub213pd %ymm2, %ymm1, %ymm0 ## encoding: [0x62,0xf2,0xf5,0x28,0xa6,0xc2]
-; CHECK-NEXT: vaddpd %ymm0, %ymm3, %ymm0 ## encoding: [0x62,0xf1,0xe5,0x28,0x58,0xc0]
+; CHECK-NEXT: vfmaddsub132pd %ymm1, %ymm2, %ymm3 {%k1} ## encoding: [0x62,0xf2,0xed,0x29,0x96,0xd9]
+; CHECK-NEXT: vfmaddsub213pd %ymm2, %ymm0, %ymm1 ## encoding: [0x62,0xf2,0xfd,0x28,0xa6,0xca]
+; CHECK-NEXT: vaddpd %ymm1, %ymm3, %ymm0 ## encoding: [0x62,0xf1,0xe5,0x28,0x58,0xc1]
; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <4 x double> @llvm.x86.avx512.mask.vfmaddsub.pd.256(<4 x double> %x0, <4 x double> %x1, <4 x double> %x2, i8 %x3)
%res1 = call <4 x double> @llvm.x86.avx512.mask.vfmaddsub.pd.256(<4 x double> %x0, <4 x double> %x1, <4 x double> %x2, i8 -1)
@@ -1604,8 +1604,8 @@ define <4 x double>@test_int_x86_avx512_
; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT: vmovaps %ymm2, %ymm3 ## encoding: [0x62,0xf1,0x7c,0x28,0x28,0xda]
; CHECK-NEXT: vfmaddsub231pd %ymm1, %ymm0, %ymm3 {%k1} ## encoding: [0x62,0xf2,0xfd,0x29,0xb6,0xd9]
-; CHECK-NEXT: vfmaddsub213pd %ymm2, %ymm1, %ymm0 ## encoding: [0x62,0xf2,0xf5,0x28,0xa6,0xc2]
-; CHECK-NEXT: vaddpd %ymm0, %ymm3, %ymm0 ## encoding: [0x62,0xf1,0xe5,0x28,0x58,0xc0]
+; CHECK-NEXT: vfmaddsub213pd %ymm2, %ymm0, %ymm1 ## encoding: [0x62,0xf2,0xfd,0x28,0xa6,0xca]
+; CHECK-NEXT: vaddpd %ymm1, %ymm3, %ymm0 ## encoding: [0x62,0xf1,0xe5,0x28,0x58,0xc1]
; CHECK-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)
%res1 = call <4 x double> @llvm.x86.avx512.mask3.vfmaddsub.pd.256(<4 x double> %x0, <4 x double> %x1, <4 x double> %x2, i8 -1)
@@ -1619,10 +1619,10 @@ define <4 x double>@test_int_x86_avx512_
; CHECK-LABEL: test_int_x86_avx512_maskz_vfmaddsub_pd_256:
; CHECK: ## BB#0:
; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vmovaps %ymm0, %ymm3 ## encoding: [0x62,0xf1,0x7c,0x28,0x28,0xd8]
-; CHECK-NEXT: vfmaddsub213pd %ymm2, %ymm1, %ymm3 {%k1} {z} ## encoding: [0x62,0xf2,0xf5,0xa9,0xa6,0xda]
-; CHECK-NEXT: vfmaddsub213pd %ymm2, %ymm1, %ymm0 ## encoding: [0x62,0xf2,0xf5,0x28,0xa6,0xc2]
-; CHECK-NEXT: vaddpd %ymm0, %ymm3, %ymm0 ## encoding: [0x62,0xf1,0xe5,0x28,0x58,0xc0]
+; CHECK-NEXT: vmovaps %ymm1, %ymm3 ## encoding: [0x62,0xf1,0x7c,0x28,0x28,0xd9]
+; CHECK-NEXT: vfmaddsub213pd %ymm2, %ymm0, %ymm3 {%k1} {z} ## encoding: [0x62,0xf2,0xfd,0xa9,0xa6,0xda]
+; CHECK-NEXT: vfmaddsub213pd %ymm2, %ymm0, %ymm1 ## encoding: [0x62,0xf2,0xfd,0x28,0xa6,0xca]
+; CHECK-NEXT: vaddpd %ymm1, %ymm3, %ymm0 ## encoding: [0x62,0xf1,0xe5,0x28,0x58,0xc1]
; CHECK-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)
%res1 = call <4 x double> @llvm.x86.avx512.maskz.vfmaddsub.pd.256(<4 x double> %x0, <4 x double> %x1, <4 x double> %x2, i8 -1)
@@ -1635,9 +1635,9 @@ define <4 x float>@test_int_x86_avx512_m
; CHECK: ## BB#0:
; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT: vmovaps %xmm0, %xmm3 ## encoding: [0x62,0xf1,0x7c,0x08,0x28,0xd8]
-; CHECK-NEXT: vfmaddsub213ps %xmm2, %xmm1, %xmm3 {%k1} ## encoding: [0x62,0xf2,0x75,0x09,0xa6,0xda]
-; CHECK-NEXT: vfmaddsub213ps %xmm2, %xmm1, %xmm0 ## encoding: [0x62,0xf2,0x75,0x08,0xa6,0xc2]
-; CHECK-NEXT: vaddps %xmm0, %xmm3, %xmm0 ## encoding: [0x62,0xf1,0x64,0x08,0x58,0xc0]
+; CHECK-NEXT: vfmaddsub132ps %xmm1, %xmm2, %xmm3 {%k1} ## encoding: [0x62,0xf2,0x6d,0x09,0x96,0xd9]
+; CHECK-NEXT: vfmaddsub213ps %xmm2, %xmm0, %xmm1 ## encoding: [0x62,0xf2,0x7d,0x08,0xa6,0xca]
+; CHECK-NEXT: vaddps %xmm1, %xmm3, %xmm0 ## encoding: [0x62,0xf1,0x64,0x08,0x58,0xc1]
; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <4 x float> @llvm.x86.avx512.mask.vfmaddsub.ps.128(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2, i8 %x3)
%res1 = call <4 x float> @llvm.x86.avx512.mask.vfmaddsub.ps.128(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2, i8 -1)
@@ -1653,8 +1653,8 @@ define <4 x float>@test_int_x86_avx512_m
; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT: vmovaps %xmm2, %xmm3 ## encoding: [0x62,0xf1,0x7c,0x08,0x28,0xda]
; CHECK-NEXT: vfmaddsub231ps %xmm1, %xmm0, %xmm3 {%k1} ## encoding: [0x62,0xf2,0x7d,0x09,0xb6,0xd9]
-; CHECK-NEXT: vfmaddsub213ps %xmm2, %xmm1, %xmm0 ## encoding: [0x62,0xf2,0x75,0x08,0xa6,0xc2]
-; CHECK-NEXT: vaddps %xmm0, %xmm3, %xmm0 ## encoding: [0x62,0xf1,0x64,0x08,0x58,0xc0]
+; CHECK-NEXT: vfmaddsub213ps %xmm2, %xmm0, %xmm1 ## encoding: [0x62,0xf2,0x7d,0x08,0xa6,0xca]
+; CHECK-NEXT: vaddps %xmm1, %xmm3, %xmm0 ## encoding: [0x62,0xf1,0x64,0x08,0x58,0xc1]
; CHECK-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)
%res1 = call <4 x float> @llvm.x86.avx512.mask3.vfmaddsub.ps.128(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2, i8 -1)
@@ -1668,10 +1668,10 @@ define <4 x float>@test_int_x86_avx512_m
; CHECK-LABEL: test_int_x86_avx512_maskz_vfmaddsub_ps_128:
; CHECK: ## BB#0:
; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vmovaps %xmm0, %xmm3 ## encoding: [0x62,0xf1,0x7c,0x08,0x28,0xd8]
-; CHECK-NEXT: vfmaddsub213ps %xmm2, %xmm1, %xmm3 {%k1} {z} ## encoding: [0x62,0xf2,0x75,0x89,0xa6,0xda]
-; CHECK-NEXT: vfmaddsub213ps %xmm2, %xmm1, %xmm0 ## encoding: [0x62,0xf2,0x75,0x08,0xa6,0xc2]
-; CHECK-NEXT: vaddps %xmm0, %xmm3, %xmm0 ## encoding: [0x62,0xf1,0x64,0x08,0x58,0xc0]
+; CHECK-NEXT: vmovaps %xmm1, %xmm3 ## encoding: [0x62,0xf1,0x7c,0x08,0x28,0xd9]
+; CHECK-NEXT: vfmaddsub213ps %xmm2, %xmm0, %xmm3 {%k1} {z} ## encoding: [0x62,0xf2,0x7d,0x89,0xa6,0xda]
+; CHECK-NEXT: vfmaddsub213ps %xmm2, %xmm0, %xmm1 ## encoding: [0x62,0xf2,0x7d,0x08,0xa6,0xca]
+; CHECK-NEXT: vaddps %xmm1, %xmm3, %xmm0 ## encoding: [0x62,0xf1,0x64,0x08,0x58,0xc1]
; CHECK-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)
%res1 = call <4 x float> @llvm.x86.avx512.maskz.vfmaddsub.ps.128(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2, i8 -1)
@@ -1684,9 +1684,9 @@ define <8 x float>@test_int_x86_avx512_m
; CHECK: ## BB#0:
; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT: vmovaps %ymm0, %ymm3 ## encoding: [0x62,0xf1,0x7c,0x28,0x28,0xd8]
-; CHECK-NEXT: vfmaddsub213ps %ymm2, %ymm1, %ymm3 {%k1} ## encoding: [0x62,0xf2,0x75,0x29,0xa6,0xda]
-; CHECK-NEXT: vfmaddsub213ps %ymm2, %ymm1, %ymm0 ## encoding: [0x62,0xf2,0x75,0x28,0xa6,0xc2]
-; CHECK-NEXT: vaddps %ymm0, %ymm3, %ymm0 ## encoding: [0x62,0xf1,0x64,0x28,0x58,0xc0]
+; CHECK-NEXT: vfmaddsub132ps %ymm1, %ymm2, %ymm3 {%k1} ## encoding: [0x62,0xf2,0x6d,0x29,0x96,0xd9]
+; CHECK-NEXT: vfmaddsub213ps %ymm2, %ymm0, %ymm1 ## encoding: [0x62,0xf2,0x7d,0x28,0xa6,0xca]
+; CHECK-NEXT: vaddps %ymm1, %ymm3, %ymm0 ## encoding: [0x62,0xf1,0x64,0x28,0x58,0xc1]
; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <8 x float> @llvm.x86.avx512.mask.vfmaddsub.ps.256(<8 x float> %x0, <8 x float> %x1, <8 x float> %x2, i8 %x3)
%res1 = call <8 x float> @llvm.x86.avx512.mask.vfmaddsub.ps.256(<8 x float> %x0, <8 x float> %x1, <8 x float> %x2, i8 -1)
@@ -1702,8 +1702,8 @@ define <8 x float>@test_int_x86_avx512_m
; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT: vmovaps %ymm2, %ymm3 ## encoding: [0x62,0xf1,0x7c,0x28,0x28,0xda]
; CHECK-NEXT: vfmaddsub231ps %ymm1, %ymm0, %ymm3 {%k1} ## encoding: [0x62,0xf2,0x7d,0x29,0xb6,0xd9]
-; CHECK-NEXT: vfmaddsub213ps %ymm2, %ymm1, %ymm0 ## encoding: [0x62,0xf2,0x75,0x28,0xa6,0xc2]
-; CHECK-NEXT: vaddps %ymm0, %ymm3, %ymm0 ## encoding: [0x62,0xf1,0x64,0x28,0x58,0xc0]
+; CHECK-NEXT: vfmaddsub213ps %ymm2, %ymm0, %ymm1 ## encoding: [0x62,0xf2,0x7d,0x28,0xa6,0xca]
+; CHECK-NEXT: vaddps %ymm1, %ymm3, %ymm0 ## encoding: [0x62,0xf1,0x64,0x28,0x58,0xc1]
; CHECK-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)
%res1 = call <8 x float> @llvm.x86.avx512.mask3.vfmaddsub.ps.256(<8 x float> %x0, <8 x float> %x1, <8 x float> %x2, i8 -1)
@@ -1717,10 +1717,10 @@ define <8 x float>@test_int_x86_avx512_m
; CHECK-LABEL: test_int_x86_avx512_maskz_vfmaddsub_ps_256:
; CHECK: ## BB#0:
; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vmovaps %ymm0, %ymm3 ## encoding: [0x62,0xf1,0x7c,0x28,0x28,0xd8]
-; CHECK-NEXT: vfmaddsub213ps %ymm2, %ymm1, %ymm3 {%k1} {z} ## encoding: [0x62,0xf2,0x75,0xa9,0xa6,0xda]
-; CHECK-NEXT: vfmaddsub213ps %ymm2, %ymm1, %ymm0 ## encoding: [0x62,0xf2,0x75,0x28,0xa6,0xc2]
-; CHECK-NEXT: vaddps %ymm0, %ymm3, %ymm0 ## encoding: [0x62,0xf1,0x64,0x28,0x58,0xc0]
+; CHECK-NEXT: vmovaps %ymm1, %ymm3 ## encoding: [0x62,0xf1,0x7c,0x28,0x28,0xd9]
+; CHECK-NEXT: vfmaddsub213ps %ymm2, %ymm0, %ymm3 {%k1} {z} ## encoding: [0x62,0xf2,0x7d,0xa9,0xa6,0xda]
+; CHECK-NEXT: vfmaddsub213ps %ymm2, %ymm0, %ymm1 ## encoding: [0x62,0xf2,0x7d,0x28,0xa6,0xca]
+; CHECK-NEXT: vaddps %ymm1, %ymm3, %ymm0 ## encoding: [0x62,0xf1,0x64,0x28,0x58,0xc1]
; CHECK-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)
%res1 = call <8 x float> @llvm.x86.avx512.maskz.vfmaddsub.ps.256(<8 x float> %x0, <8 x float> %x1, <8 x float> %x2, i8 -1)
@@ -1736,8 +1736,8 @@ define <2 x double>@test_int_x86_avx512_
; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT: vmovaps %xmm2, %xmm3 ## encoding: [0x62,0xf1,0x7c,0x08,0x28,0xda]
; CHECK-NEXT: vfmsubadd231pd %xmm1, %xmm0, %xmm3 {%k1} ## encoding: [0x62,0xf2,0xfd,0x09,0xb7,0xd9]
-; CHECK-NEXT: vfmsubadd213pd %xmm2, %xmm1, %xmm0 ## encoding: [0x62,0xf2,0xf5,0x08,0xa7,0xc2]
-; CHECK-NEXT: vaddpd %xmm0, %xmm3, %xmm0 ## encoding: [0x62,0xf1,0xe5,0x08,0x58,0xc0]
+; CHECK-NEXT: vfmsubadd213pd %xmm2, %xmm0, %xmm1 ## encoding: [0x62,0xf2,0xfd,0x08,0xa7,0xca]
+; CHECK-NEXT: vaddpd %xmm1, %xmm3, %xmm0 ## encoding: [0x62,0xf1,0xe5,0x08,0x58,0xc1]
; CHECK-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)
%res1 = call <2 x double> @llvm.x86.avx512.mask3.vfmsubadd.pd.128(<2 x double> %x0, <2 x double> %x1, <2 x double> %x2, i8 -1)
@@ -1753,8 +1753,8 @@ define <4 x double>@test_int_x86_avx512_
; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT: vmovaps %ymm2, %ymm3 ## encoding: [0x62,0xf1,0x7c,0x28,0x28,0xda]
; CHECK-NEXT: vfmsubadd231pd %ymm1, %ymm0, %ymm3 {%k1} ## encoding: [0x62,0xf2,0xfd,0x29,0xb7,0xd9]
-; CHECK-NEXT: vfmsubadd213pd %ymm2, %ymm1, %ymm0 ## encoding: [0x62,0xf2,0xf5,0x28,0xa7,0xc2]
-; CHECK-NEXT: vaddpd %ymm0, %ymm3, %ymm0 ## encoding: [0x62,0xf1,0xe5,0x28,0x58,0xc0]
+; CHECK-NEXT: vfmsubadd213pd %ymm2, %ymm0, %ymm1 ## encoding: [0x62,0xf2,0xfd,0x28,0xa7,0xca]
+; CHECK-NEXT: vaddpd %ymm1, %ymm3, %ymm0 ## encoding: [0x62,0xf1,0xe5,0x28,0x58,0xc1]
; CHECK-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)
%res1 = call <4 x double> @llvm.x86.avx512.mask3.vfmsubadd.pd.256(<4 x double> %x0, <4 x double> %x1, <4 x double> %x2, i8 -1)
@@ -1770,8 +1770,8 @@ define <4 x float>@test_int_x86_avx512_m
; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT: vmovaps %xmm2, %xmm3 ## encoding: [0x62,0xf1,0x7c,0x08,0x28,0xda]
; CHECK-NEXT: vfmsubadd231ps %xmm1, %xmm0, %xmm3 {%k1} ## encoding: [0x62,0xf2,0x7d,0x09,0xb7,0xd9]
-; CHECK-NEXT: vfmsubadd213ps %xmm2, %xmm1, %xmm0 ## encoding: [0x62,0xf2,0x75,0x08,0xa7,0xc2]
-; CHECK-NEXT: vaddps %xmm0, %xmm3, %xmm0 ## encoding: [0x62,0xf1,0x64,0x08,0x58,0xc0]
+; CHECK-NEXT: vfmsubadd213ps %xmm2, %xmm0, %xmm1 ## encoding: [0x62,0xf2,0x7d,0x08,0xa7,0xca]
+; CHECK-NEXT: vaddps %xmm1, %xmm3, %xmm0 ## encoding: [0x62,0xf1,0x64,0x08,0x58,0xc1]
; CHECK-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)
%res1 = call <4 x float> @llvm.x86.avx512.mask3.vfmsubadd.ps.128(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2, i8 -1)
@@ -1787,8 +1787,8 @@ define <8 x float>@test_int_x86_avx512_m
; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT: vmovaps %ymm2, %ymm3 ## encoding: [0x62,0xf1,0x7c,0x28,0x28,0xda]
; CHECK-NEXT: vfmsubadd231ps %ymm1, %ymm0, %ymm3 {%k1} ## encoding: [0x62,0xf2,0x7d,0x29,0xb7,0xd9]
-; CHECK-NEXT: vfmsubadd213ps %ymm2, %ymm1, %ymm0 ## encoding: [0x62,0xf2,0x75,0x28,0xa7,0xc2]
-; CHECK-NEXT: vaddps %ymm0, %ymm3, %ymm0 ## encoding: [0x62,0xf1,0x64,0x28,0x58,0xc0]
+; CHECK-NEXT: vfmsubadd213ps %ymm2, %ymm0, %ymm1 ## encoding: [0x62,0xf2,0x7d,0x28,0xa7,0xca]
+; CHECK-NEXT: vaddps %ymm1, %ymm3, %ymm0 ## encoding: [0x62,0xf1,0x64,0x28,0x58,0xc1]
; CHECK-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)
%res1 = call <8 x float> @llvm.x86.avx512.mask3.vfmsubadd.ps.256(<8 x float> %x0, <8 x float> %x1, <8 x float> %x2, i8 -1)
@@ -1801,7 +1801,7 @@ define <4 x float> @test_mask_vfmadd128_
; CHECK-LABEL: test_mask_vfmadd128_ps_r:
; CHECK: ## BB#0:
; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfmadd213ps %xmm2, %xmm1, %xmm0 {%k1} ## encoding: [0x62,0xf2,0x75,0x09,0xa8,0xc2]
+; CHECK-NEXT: vfmadd132ps %xmm1, %xmm2, %xmm0 {%k1} ## encoding: [0x62,0xf2,0x6d,0x09,0x98,0xc1]
; CHECK-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
@@ -1820,7 +1820,8 @@ define <4 x float> @test_mask_vfmadd128_
; CHECK-LABEL: test_mask_vfmadd128_ps_rmk:
; CHECK: ## BB#0:
; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
-; CHECK-NEXT: vfmadd213ps (%rdi), %xmm1, %xmm0 {%k1} ## encoding: [0x62,0xf2,0x75,0x09,0xa8,0x07]
+; CHECK-NEXT: vmovaps (%rdi), %xmm2 ## encoding: [0x62,0xf1,0x7c,0x08,0x28,0x17]
+; CHECK-NEXT: vfmadd132ps %xmm1, %xmm2, %xmm0 {%k1} ## encoding: [0x62,0xf2,0x6d,0x09,0x98,0xc1]
; CHECK-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
@@ -1831,7 +1832,8 @@ define <4 x float> @test_mask_vfmadd128_
; CHECK-LABEL: test_mask_vfmadd128_ps_rmka:
; CHECK: ## BB#0:
; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
-; CHECK-NEXT: vfmadd213ps (%rdi), %xmm1, %xmm0 {%k1} ## encoding: [0x62,0xf2,0x75,0x09,0xa8,0x07]
+; CHECK-NEXT: vmovups (%rdi), %xmm2 ## encoding: [0x62,0xf1,0x7c,0x08,0x10,0x17]
+; CHECK-NEXT: vfmadd132ps %xmm1, %xmm2, %xmm0 {%k1} ## encoding: [0x62,0xf2,0x6d,0x09,0x98,0xc1]
; CHECK-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
@@ -1862,7 +1864,8 @@ define <4 x float> @test_mask_vfmadd128_
; CHECK-LABEL: test_mask_vfmadd128_ps_rmb:
; CHECK: ## BB#0:
; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
-; CHECK-NEXT: vfmadd213ps (%rdi){1to4}, %xmm1, %xmm0 {%k1} ## encoding: [0x62,0xf2,0x75,0x19,0xa8,0x07]
+; CHECK-NEXT: vbroadcastss (%rdi), %xmm2 ## encoding: [0x62,0xf2,0x7d,0x08,0x18,0x17]
+; CHECK-NEXT: vfmadd132ps %xmm1, %xmm2, %xmm0 {%k1} ## encoding: [0x62,0xf2,0x6d,0x09,0x98,0xc1]
; CHECK-NEXT: retq ## encoding: [0xc3]
%q = load float, float* %ptr_a2
%vecinit.i = insertelement <4 x float> undef, float %q, i32 0
@@ -1877,7 +1880,8 @@ define <4 x float> @test_mask_vfmadd128_
; CHECK-LABEL: test_mask_vfmadd128_ps_rmba:
; CHECK: ## BB#0:
; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
-; CHECK-NEXT: vfmadd213ps (%rdi){1to4}, %xmm1, %xmm0 {%k1} ## encoding: [0x62,0xf2,0x75,0x19,0xa8,0x07]
+; CHECK-NEXT: vbroadcastss (%rdi), %xmm2 ## encoding: [0x62,0xf2,0x7d,0x08,0x18,0x17]
+; CHECK-NEXT: vfmadd132ps %xmm1, %xmm2, %xmm0 {%k1} ## encoding: [0x62,0xf2,0x6d,0x09,0x98,0xc1]
; CHECK-NEXT: retq ## encoding: [0xc3]
%q = load float, float* %ptr_a2, align 4
%vecinit.i = insertelement <4 x float> undef, float %q, i32 0
@@ -1891,7 +1895,8 @@ define <4 x float> @test_mask_vfmadd128_
define <4 x float> @test_mask_vfmadd128_ps_rmbz(<4 x float> %a0, <4 x float> %a1, float* %ptr_a2) {
; CHECK-LABEL: test_mask_vfmadd128_ps_rmbz:
; CHECK: ## BB#0:
-; CHECK-NEXT: vfmadd213ps (%rdi){1to4}, %xmm1, %xmm0 ## encoding: [0x62,0xf2,0x75,0x18,0xa8,0x07]
+; CHECK-NEXT: vfmadd213ps (%rdi){1to4}, %xmm0, %xmm1 ## encoding: [0x62,0xf2,0x7d,0x18,0xa8,0x0f]
+; CHECK-NEXT: vmovaps %xmm1, %xmm0 ## encoding: [0x62,0xf1,0x7c,0x08,0x28,0xc1]
; CHECK-NEXT: retq ## encoding: [0xc3]
%q = load float, float* %ptr_a2
%vecinit.i = insertelement <4 x float> undef, float %q, i32 0
@@ -1905,7 +1910,8 @@ define <4 x float> @test_mask_vfmadd128_
define <4 x float> @test_mask_vfmadd128_ps_rmbza(<4 x float> %a0, <4 x float> %a1, float* %ptr_a2) {
; CHECK-LABEL: test_mask_vfmadd128_ps_rmbza:
; CHECK: ## BB#0:
-; CHECK-NEXT: vfmadd213ps (%rdi){1to4}, %xmm1, %xmm0 ## encoding: [0x62,0xf2,0x75,0x18,0xa8,0x07]
+; CHECK-NEXT: vfmadd213ps (%rdi){1to4}, %xmm0, %xmm1 ## encoding: [0x62,0xf2,0x7d,0x18,0xa8,0x0f]
+; CHECK-NEXT: vmovaps %xmm1, %xmm0 ## encoding: [0x62,0xf1,0x7c,0x08,0x28,0xc1]
; CHECK-NEXT: retq ## encoding: [0xc3]
%q = load float, float* %ptr_a2, align 4
%vecinit.i = insertelement <4 x float> undef, float %q, i32 0
@@ -1920,7 +1926,7 @@ define <2 x double> @test_mask_vfmadd128
; CHECK-LABEL: test_mask_vfmadd128_pd_r:
; CHECK: ## BB#0:
; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfmadd213pd %xmm2, %xmm1, %xmm0 {%k1} ## encoding: [0x62,0xf2,0xf5,0x09,0xa8,0xc2]
+; CHECK-NEXT: vfmadd132pd %xmm1, %xmm2, %xmm0 {%k1} ## encoding: [0x62,0xf2,0xed,0x09,0x98,0xc1]
; CHECK-NEXT: retq ## encoding: [0xc3]
%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
ret <2 x double> %res
@@ -1939,7 +1945,8 @@ define <2 x double> @test_mask_vfmadd128
; CHECK-LABEL: test_mask_vfmadd128_pd_rmk:
; CHECK: ## BB#0:
; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
-; CHECK-NEXT: vfmadd213pd (%rdi), %xmm1, %xmm0 {%k1} ## encoding: [0x62,0xf2,0xf5,0x09,0xa8,0x07]
+; CHECK-NEXT: vmovapd (%rdi), %xmm2 ## encoding: [0x62,0xf1,0xfd,0x08,0x28,0x17]
+; CHECK-NEXT: vfmadd132pd %xmm1, %xmm2, %xmm0 {%k1} ## encoding: [0x62,0xf2,0xed,0x09,0x98,0xc1]
; CHECK-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
@@ -1960,7 +1967,7 @@ define <4 x double> @test_mask_vfmadd256
; CHECK-LABEL: test_mask_vfmadd256_pd_r:
; CHECK: ## BB#0:
; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfmadd213pd %ymm2, %ymm1, %ymm0 {%k1} ## encoding: [0x62,0xf2,0xf5,0x29,0xa8,0xc2]
+; CHECK-NEXT: vfmadd132pd %ymm1, %ymm2, %ymm0 {%k1} ## encoding: [0x62,0xf2,0xed,0x29,0x98,0xc1]
; CHECK-NEXT: retq ## encoding: [0xc3]
%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
ret <4 x double> %res
@@ -1979,7 +1986,8 @@ define <4 x double> @test_mask_vfmadd256
; CHECK-LABEL: test_mask_vfmadd256_pd_rmk:
; CHECK: ## BB#0:
; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
-; CHECK-NEXT: vfmadd213pd (%rdi), %ymm1, %ymm0 {%k1} ## encoding: [0x62,0xf2,0xf5,0x29,0xa8,0x07]
+; CHECK-NEXT: vmovapd (%rdi), %ymm2 ## encoding: [0x62,0xf1,0xfd,0x28,0x28,0x17]
+; CHECK-NEXT: vfmadd132pd %ymm1, %ymm2, %ymm0 {%k1} ## encoding: [0x62,0xf2,0xed,0x29,0x98,0xc1]
; CHECK-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
More information about the llvm-commits
mailing list