summaryrefslogtreecommitdiffstats
diff options
context:
space:
mode:
Diffstat
-rw-r--r--llvm/test/CodeGen/X86/avx512dqvl-intrinsics-upgrade.ll1251
-rw-r--r--llvm/test/CodeGen/X86/avx512dqvl-intrinsics.ll1248
2 files changed, 1251 insertions, 1248 deletions
diff --git a/llvm/test/CodeGen/X86/avx512dqvl-intrinsics-upgrade.ll b/llvm/test/CodeGen/X86/avx512dqvl-intrinsics-upgrade.ll
new file mode 100644
index 00000000000..a6217d19dab
--- /dev/null
+++ b/llvm/test/CodeGen/X86/avx512dqvl-intrinsics-upgrade.ll
@@ -0,0 +1,1251 @@
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+; RUN: llc < %s -mtriple=x86_64-apple-darwin -mcpu=knl -mattr=+avx512dq -mattr=+avx512vl --show-mc-encoding| FileCheck %s
+
+define <4 x float> @test_mask_andnot_ps_rr_128(<4 x float> %a, <4 x float> %b) {
+; CHECK-LABEL: test_mask_andnot_ps_rr_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vandnps %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7c,0x08,0x55,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
+ %res = call <4 x float> @llvm.x86.avx512.mask.andn.ps.128(<4 x float> %a, <4 x float> %b, <4 x float> zeroinitializer, i8 -1)
+ ret <4 x float> %res
+}
+
+define <4 x float> @test_mask_andnot_ps_rrk_128(<4 x float> %a, <4 x float> %b, <4 x float> %passThru, i8 %mask) {
+; CHECK-LABEL: test_mask_andnot_ps_rrk_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovb %edi, %k1 ## encoding: [0xc5,0xf9,0x92,0xcf]
+; CHECK-NEXT: vandnps %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf1,0x7c,0x09,0x55,0xd1]
+; CHECK-NEXT: vmovaps %xmm2, %xmm0 ## encoding: [0x62,0xf1,0x7c,0x08,0x28,0xc2]
+; CHECK-NEXT: retq ## encoding: [0xc3]
+ %res = call <4 x float> @llvm.x86.avx512.mask.andn.ps.128(<4 x float> %a, <4 x float> %b, <4 x float> %passThru, i8 %mask)
+ ret <4 x float> %res
+}
+
+define <4 x float> @test_mask_andnot_ps_rrkz_128(<4 x float> %a, <4 x float> %b, i8 %mask) {
+; CHECK-LABEL: test_mask_andnot_ps_rrkz_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovb %edi, %k1 ## encoding: [0xc5,0xf9,0x92,0xcf]
+; CHECK-NEXT: vandnps %xmm1, %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0x89,0x55,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
+ %res = call <4 x float> @llvm.x86.avx512.mask.andn.ps.128(<4 x float> %a, <4 x float> %b, <4 x float> zeroinitializer, i8 %mask)
+ ret <4 x float> %res
+}
+
+define <4 x float> @test_mask_andnot_ps_rm_128(<4 x float> %a, <4 x float>* %ptr_b) {
+; CHECK-LABEL: test_mask_andnot_ps_rm_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vandnps (%rdi), %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7c,0x08,0x55,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
+ %b = load <4 x float>, <4 x float>* %ptr_b
+ %res = call <4 x float> @llvm.x86.avx512.mask.andn.ps.128(<4 x float> %a, <4 x float> %b, <4 x float> zeroinitializer, i8 -1)
+ ret <4 x float> %res
+}
+
+define <4 x float> @test_mask_andnot_ps_rmk_128(<4 x float> %a, <4 x float>* %ptr_b, <4 x float> %passThru, i8 %mask) {
+; CHECK-LABEL: test_mask_andnot_ps_rmk_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovb %esi, %k1 ## encoding: [0xc5,0xf9,0x92,0xce]
+; CHECK-NEXT: vandnps (%rdi), %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf1,0x7c,0x09,0x55,0x0f]
+; CHECK-NEXT: vmovaps %xmm1, %xmm0 ## encoding: [0x62,0xf1,0x7c,0x08,0x28,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
+ %b = load <4 x float>, <4 x float>* %ptr_b
+ %res = call <4 x float> @llvm.x86.avx512.mask.andn.ps.128(<4 x float> %a, <4 x float> %b, <4 x float> %passThru, i8 %mask)
+ ret <4 x float> %res
+}
+
+define <4 x float> @test_mask_andnot_ps_rmkz_128(<4 x float> %a, <4 x float>* %ptr_b, i8 %mask) {
+; CHECK-LABEL: test_mask_andnot_ps_rmkz_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovb %esi, %k1 ## encoding: [0xc5,0xf9,0x92,0xce]
+; CHECK-NEXT: vandnps (%rdi), %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0x89,0x55,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
+ %b = load <4 x float>, <4 x float>* %ptr_b
+ %res = call <4 x float> @llvm.x86.avx512.mask.andn.ps.128(<4 x float> %a, <4 x float> %b, <4 x float> zeroinitializer, i8 %mask)
+ ret <4 x float> %res
+}
+
+define <4 x float> @test_mask_andnot_ps_rmb_128(<4 x float> %a, float* %ptr_b) {
+; CHECK-LABEL: test_mask_andnot_ps_rmb_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vandnps (%rdi){1to4}, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7c,0x18,0x55,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
+ %q = load float, float* %ptr_b
+ %vecinit.i = insertelement <4 x float> undef, float %q, i32 0
+ %b = shufflevector <4 x float> %vecinit.i, <4 x float> undef, <4 x i32> zeroinitializer
+ %res = call <4 x float> @llvm.x86.avx512.mask.andn.ps.128(<4 x float> %a, <4 x float> %b, <4 x float> zeroinitializer, i8 -1)
+ ret <4 x float> %res
+}
+
+define <4 x float> @test_mask_andnot_ps_rmbk_128(<4 x float> %a, float* %ptr_b, <4 x float> %passThru, i8 %mask) {
+; CHECK-LABEL: test_mask_andnot_ps_rmbk_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovb %esi, %k1 ## encoding: [0xc5,0xf9,0x92,0xce]
+; CHECK-NEXT: vandnps (%rdi){1to4}, %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf1,0x7c,0x19,0x55,0x0f]
+; CHECK-NEXT: vmovaps %xmm1, %xmm0 ## encoding: [0x62,0xf1,0x7c,0x08,0x28,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
+ %q = load float, float* %ptr_b
+ %vecinit.i = insertelement <4 x float> undef, float %q, i32 0
+ %b = shufflevector <4 x float> %vecinit.i, <4 x float> undef, <4 x i32> zeroinitializer
+ %res = call <4 x float> @llvm.x86.avx512.mask.andn.ps.128(<4 x float> %a, <4 x float> %b, <4 x float> %passThru, i8 %mask)
+ ret <4 x float> %res
+}
+
+define <4 x float> @test_mask_andnot_ps_rmbkz_128(<4 x float> %a, float* %ptr_b, i8 %mask) {
+; CHECK-LABEL: test_mask_andnot_ps_rmbkz_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovb %esi, %k1 ## encoding: [0xc5,0xf9,0x92,0xce]
+; CHECK-NEXT: vandnps (%rdi){1to4}, %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0x99,0x55,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
+ %q = load float, float* %ptr_b
+ %vecinit.i = insertelement <4 x float> undef, float %q, i32 0
+ %b = shufflevector <4 x float> %vecinit.i, <4 x float> undef, <4 x i32> zeroinitializer
+ %res = call <4 x float> @llvm.x86.avx512.mask.andn.ps.128(<4 x float> %a, <4 x float> %b, <4 x float> zeroinitializer, i8 %mask)
+ ret <4 x float> %res
+}
+
+declare <4 x float> @llvm.x86.avx512.mask.andn.ps.128(<4 x float>, <4 x float>, <4 x float>, i8)
+
+define <8 x float> @test_mask_andnot_ps_rr_256(<8 x float> %a, <8 x float> %b) {
+; CHECK-LABEL: test_mask_andnot_ps_rr_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vandnps %ymm1, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7c,0x28,0x55,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
+ %res = call <8 x float> @llvm.x86.avx512.mask.andn.ps.256(<8 x float> %a, <8 x float> %b, <8 x float> zeroinitializer, i8 -1)
+ ret <8 x float> %res
+}
+
+define <8 x float> @test_mask_andnot_ps_rrk_256(<8 x float> %a, <8 x float> %b, <8 x float> %passThru, i8 %mask) {
+; CHECK-LABEL: test_mask_andnot_ps_rrk_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovb %edi, %k1 ## encoding: [0xc5,0xf9,0x92,0xcf]
+; CHECK-NEXT: vandnps %ymm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf1,0x7c,0x29,0x55,0xd1]
+; CHECK-NEXT: vmovaps %ymm2, %ymm0 ## encoding: [0x62,0xf1,0x7c,0x28,0x28,0xc2]
+; CHECK-NEXT: retq ## encoding: [0xc3]
+ %res = call <8 x float> @llvm.x86.avx512.mask.andn.ps.256(<8 x float> %a, <8 x float> %b, <8 x float> %passThru, i8 %mask)
+ ret <8 x float> %res
+}
+
+define <8 x float> @test_mask_andnot_ps_rrkz_256(<8 x float> %a, <8 x float> %b, i8 %mask) {
+; CHECK-LABEL: test_mask_andnot_ps_rrkz_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovb %edi, %k1 ## encoding: [0xc5,0xf9,0x92,0xcf]
+; CHECK-NEXT: vandnps %ymm1, %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0xa9,0x55,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
+ %res = call <8 x float> @llvm.x86.avx512.mask.andn.ps.256(<8 x float> %a, <8 x float> %b, <8 x float> zeroinitializer, i8 %mask)
+ ret <8 x float> %res
+}
+
+define <8 x float> @test_mask_andnot_ps_rm_256(<8 x float> %a, <8 x float>* %ptr_b) {
+; CHECK-LABEL: test_mask_andnot_ps_rm_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vandnps (%rdi), %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7c,0x28,0x55,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
+ %b = load <8 x float>, <8 x float>* %ptr_b
+ %res = call <8 x float> @llvm.x86.avx512.mask.andn.ps.256(<8 x float> %a, <8 x float> %b, <8 x float> zeroinitializer, i8 -1)
+ ret <8 x float> %res
+}
+
+define <8 x float> @test_mask_andnot_ps_rmk_256(<8 x float> %a, <8 x float>* %ptr_b, <8 x float> %passThru, i8 %mask) {
+; CHECK-LABEL: test_mask_andnot_ps_rmk_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovb %esi, %k1 ## encoding: [0xc5,0xf9,0x92,0xce]
+; CHECK-NEXT: vandnps (%rdi), %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf1,0x7c,0x29,0x55,0x0f]
+; CHECK-NEXT: vmovaps %ymm1, %ymm0 ## encoding: [0x62,0xf1,0x7c,0x28,0x28,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
+ %b = load <8 x float>, <8 x float>* %ptr_b
+ %res = call <8 x float> @llvm.x86.avx512.mask.andn.ps.256(<8 x float> %a, <8 x float> %b, <8 x float> %passThru, i8 %mask)
+ ret <8 x float> %res
+}
+
+define <8 x float> @test_mask_andnot_ps_rmkz_256(<8 x float> %a, <8 x float>* %ptr_b, i8 %mask) {
+; CHECK-LABEL: test_mask_andnot_ps_rmkz_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovb %esi, %k1 ## encoding: [0xc5,0xf9,0x92,0xce]
+; CHECK-NEXT: vandnps (%rdi), %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0xa9,0x55,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
+ %b = load <8 x float>, <8 x float>* %ptr_b
+ %res = call <8 x float> @llvm.x86.avx512.mask.andn.ps.256(<8 x float> %a, <8 x float> %b, <8 x float> zeroinitializer, i8 %mask)
+ ret <8 x float> %res
+}
+
+define <8 x float> @test_mask_andnot_ps_rmb_256(<8 x float> %a, float* %ptr_b) {
+; CHECK-LABEL: test_mask_andnot_ps_rmb_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vandnps (%rdi){1to8}, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7c,0x38,0x55,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
+ %q = load float, float* %ptr_b
+ %vecinit.i = insertelement <8 x float> undef, float %q, i32 0
+ %b = shufflevector <8 x float> %vecinit.i, <8 x float> undef, <8 x i32> zeroinitializer
+ %res = call <8 x float> @llvm.x86.avx512.mask.andn.ps.256(<8 x float> %a, <8 x float> %b, <8 x float> zeroinitializer, i8 -1)
+ ret <8 x float> %res
+}
+
+define <8 x float> @test_mask_andnot_ps_rmbk_256(<8 x float> %a, float* %ptr_b, <8 x float> %passThru, i8 %mask) {
+; CHECK-LABEL: test_mask_andnot_ps_rmbk_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovb %esi, %k1 ## encoding: [0xc5,0xf9,0x92,0xce]
+; CHECK-NEXT: vandnps (%rdi){1to8}, %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf1,0x7c,0x39,0x55,0x0f]
+; CHECK-NEXT: vmovaps %ymm1, %ymm0 ## encoding: [0x62,0xf1,0x7c,0x28,0x28,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
+ %q = load float, float* %ptr_b
+ %vecinit.i = insertelement <8 x float> undef, float %q, i32 0
+ %b = shufflevector <8 x float> %vecinit.i, <8 x float> undef, <8 x i32> zeroinitializer
+ %res = call <8 x float> @llvm.x86.avx512.mask.andn.ps.256(<8 x float> %a, <8 x float> %b, <8 x float> %passThru, i8 %mask)
+ ret <8 x float> %res
+}
+
+define <8 x float> @test_mask_andnot_ps_rmbkz_256(<8 x float> %a, float* %ptr_b, i8 %mask) {
+; CHECK-LABEL: test_mask_andnot_ps_rmbkz_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovb %esi, %k1 ## encoding: [0xc5,0xf9,0x92,0xce]
+; CHECK-NEXT: vandnps (%rdi){1to8}, %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0xb9,0x55,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
+ %q = load float, float* %ptr_b
+ %vecinit.i = insertelement <8 x float> undef, float %q, i32 0
+ %b = shufflevector <8 x float> %vecinit.i, <8 x float> undef, <8 x i32> zeroinitializer
+ %res = call <8 x float> @llvm.x86.avx512.mask.andn.ps.256(<8 x float> %a, <8 x float> %b, <8 x float> zeroinitializer, i8 %mask)
+ ret <8 x float> %res
+}
+
+declare <8 x float> @llvm.x86.avx512.mask.andn.ps.256(<8 x float>, <8 x float>, <8 x float>, i8)
+
+define <16 x float> @test_mask_andnot_ps_rr_512(<16 x float> %a, <16 x float> %b) {
+; CHECK-LABEL: test_mask_andnot_ps_rr_512:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vandnps %zmm1, %zmm0, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x55,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
+ %res = call <16 x float> @llvm.x86.avx512.mask.andn.ps.512(<16 x float> %a, <16 x float> %b, <16 x float> zeroinitializer, i16 -1)
+ ret <16 x float> %res
+}
+
+define <16 x float> @test_mask_andnot_ps_rrk_512(<16 x float> %a, <16 x float> %b, <16 x float> %passThru, i16 %mask) {
+; CHECK-LABEL: test_mask_andnot_ps_rrk_512:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vandnps %zmm1, %zmm0, %zmm2 {%k1} ## encoding: [0x62,0xf1,0x7c,0x49,0x55,0xd1]
+; CHECK-NEXT: vmovaps %zmm2, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc2]
+; CHECK-NEXT: retq ## encoding: [0xc3]
+ %res = call <16 x float> @llvm.x86.avx512.mask.andn.ps.512(<16 x float> %a, <16 x float> %b, <16 x float> %passThru, i16 %mask)
+ ret <16 x float> %res
+}
+
+define <16 x float> @test_mask_andnot_ps_rrkz_512(<16 x float> %a, <16 x float> %b, i16 %mask) {
+; CHECK-LABEL: test_mask_andnot_ps_rrkz_512:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vandnps %zmm1, %zmm0, %zmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0xc9,0x55,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
+ %res = call <16 x float> @llvm.x86.avx512.mask.andn.ps.512(<16 x float> %a, <16 x float> %b, <16 x float> zeroinitializer, i16 %mask)
+ ret <16 x float> %res
+}
+
+define <16 x float> @test_mask_andnot_ps_rm_512(<16 x float> %a, <16 x float>* %ptr_b) {
+; CHECK-LABEL: test_mask_andnot_ps_rm_512:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vandnps (%rdi), %zmm0, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x55,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
+ %b = load <16 x float>, <16 x float>* %ptr_b
+ %res = call <16 x float> @llvm.x86.avx512.mask.andn.ps.512(<16 x float> %a, <16 x float> %b, <16 x float> zeroinitializer, i16 -1)
+ ret <16 x float> %res
+}
+
+define <16 x float> @test_mask_andnot_ps_rmk_512(<16 x float> %a, <16 x float>* %ptr_b, <16 x float> %passThru, i16 %mask) {
+; CHECK-LABEL: test_mask_andnot_ps_rmk_512:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
+; CHECK-NEXT: vandnps (%rdi), %zmm0, %zmm1 {%k1} ## encoding: [0x62,0xf1,0x7c,0x49,0x55,0x0f]
+; CHECK-NEXT: vmovaps %zmm1, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
+ %b = load <16 x float>, <16 x float>* %ptr_b
+ %res = call <16 x float> @llvm.x86.avx512.mask.andn.ps.512(<16 x float> %a, <16 x float> %b, <16 x float> %passThru, i16 %mask)
+ ret <16 x float> %res
+}
+
+define <16 x float> @test_mask_andnot_ps_rmkz_512(<16 x float> %a, <16 x float>* %ptr_b, i16 %mask) {
+; CHECK-LABEL: test_mask_andnot_ps_rmkz_512:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
+; CHECK-NEXT: vandnps (%rdi), %zmm0, %zmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0xc9,0x55,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
+ %b = load <16 x float>, <16 x float>* %ptr_b
+ %res = call <16 x float> @llvm.x86.avx512.mask.andn.ps.512(<16 x float> %a, <16 x float> %b, <16 x float> zeroinitializer, i16 %mask)
+ ret <16 x float> %res
+}
+
+define <16 x float> @test_mask_andnot_ps_rmb_512(<16 x float> %a, float* %ptr_b) {
+; CHECK-LABEL: test_mask_andnot_ps_rmb_512:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vandnps (%rdi){1to16}, %zmm0, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x58,0x55,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
+ %q = load float, float* %ptr_b
+ %vecinit.i = insertelement <16 x float> undef, float %q, i32 0
+ %b = shufflevector <16 x float> %vecinit.i, <16 x float> undef, <16 x i32> zeroinitializer
+ %res = call <16 x float> @llvm.x86.avx512.mask.andn.ps.512(<16 x float> %a, <16 x float> %b, <16 x float> zeroinitializer, i16 -1)
+ ret <16 x float> %res
+}
+
+define <16 x float> @test_mask_andnot_ps_rmbk_512(<16 x float> %a, float* %ptr_b, <16 x float> %passThru, i16 %mask) {
+; CHECK-LABEL: test_mask_andnot_ps_rmbk_512:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
+; CHECK-NEXT: vandnps (%rdi){1to16}, %zmm0, %zmm1 {%k1} ## encoding: [0x62,0xf1,0x7c,0x59,0x55,0x0f]
+; CHECK-NEXT: vmovaps %zmm1, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
+ %q = load float, float* %ptr_b
+ %vecinit.i = insertelement <16 x float> undef, float %q, i32 0
+ %b = shufflevector <16 x float> %vecinit.i, <16 x float> undef, <16 x i32> zeroinitializer
+ %res = call <16 x float> @llvm.x86.avx512.mask.andn.ps.512(<16 x float> %a, <16 x float> %b, <16 x float> %passThru, i16 %mask)
+ ret <16 x float> %res
+}
+
+define <16 x float> @test_mask_andnot_ps_rmbkz_512(<16 x float> %a, float* %ptr_b, i16 %mask) {
+; CHECK-LABEL: test_mask_andnot_ps_rmbkz_512:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
+; CHECK-NEXT: vandnps (%rdi){1to16}, %zmm0, %zmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0xd9,0x55,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
+ %q = load float, float* %ptr_b
+ %vecinit.i = insertelement <16 x float> undef, float %q, i32 0
+ %b = shufflevector <16 x float> %vecinit.i, <16 x float> undef, <16 x i32> zeroinitializer
+ %res = call <16 x float> @llvm.x86.avx512.mask.andn.ps.512(<16 x float> %a, <16 x float> %b, <16 x float> zeroinitializer, i16 %mask)
+ ret <16 x float> %res
+}
+
+declare <16 x float> @llvm.x86.avx512.mask.andn.ps.512(<16 x float>, <16 x float>, <16 x float>, i16)
+
+define <4 x float> @test_mask_and_ps_rr_128(<4 x float> %a, <4 x float> %b) {
+; CHECK-LABEL: test_mask_and_ps_rr_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vandps %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7c,0x08,0x54,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
+ %res = call <4 x float> @llvm.x86.avx512.mask.and.ps.128(<4 x float> %a, <4 x float> %b, <4 x float> zeroinitializer, i8 -1)
+ ret <4 x float> %res
+}
+
+define <4 x float> @test_mask_and_ps_rrk_128(<4 x float> %a, <4 x float> %b, <4 x float> %passThru, i8 %mask) {
+; CHECK-LABEL: test_mask_and_ps_rrk_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovb %edi, %k1 ## encoding: [0xc5,0xf9,0x92,0xcf]
+; CHECK-NEXT: vandps %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf1,0x7c,0x09,0x54,0xd1]
+; CHECK-NEXT: vmovaps %xmm2, %xmm0 ## encoding: [0x62,0xf1,0x7c,0x08,0x28,0xc2]
+; CHECK-NEXT: retq ## encoding: [0xc3]
+ %res = call <4 x float> @llvm.x86.avx512.mask.and.ps.128(<4 x float> %a, <4 x float> %b, <4 x float> %passThru, i8 %mask)
+ ret <4 x float> %res
+}
+
+define <4 x float> @test_mask_and_ps_rrkz_128(<4 x float> %a, <4 x float> %b, i8 %mask) {
+; CHECK-LABEL: test_mask_and_ps_rrkz_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovb %edi, %k1 ## encoding: [0xc5,0xf9,0x92,0xcf]
+; CHECK-NEXT: vandps %xmm1, %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0x89,0x54,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
+ %res = call <4 x float> @llvm.x86.avx512.mask.and.ps.128(<4 x float> %a, <4 x float> %b, <4 x float> zeroinitializer, i8 %mask)
+ ret <4 x float> %res
+}
+
+define <4 x float> @test_mask_and_ps_rm_128(<4 x float> %a, <4 x float>* %ptr_b) {
+; CHECK-LABEL: test_mask_and_ps_rm_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vandps (%rdi), %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7c,0x08,0x54,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
+ %b = load <4 x float>, <4 x float>* %ptr_b
+ %res = call <4 x float> @llvm.x86.avx512.mask.and.ps.128(<4 x float> %a, <4 x float> %b, <4 x float> zeroinitializer, i8 -1)
+ ret <4 x float> %res
+}
+
+define <4 x float> @test_mask_and_ps_rmk_128(<4 x float> %a, <4 x float>* %ptr_b, <4 x float> %passThru, i8 %mask) {
+; CHECK-LABEL: test_mask_and_ps_rmk_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovb %esi, %k1 ## encoding: [0xc5,0xf9,0x92,0xce]
+; CHECK-NEXT: vandps (%rdi), %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf1,0x7c,0x09,0x54,0x0f]
+; CHECK-NEXT: vmovaps %xmm1, %xmm0 ## encoding: [0x62,0xf1,0x7c,0x08,0x28,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
+ %b = load <4 x float>, <4 x float>* %ptr_b
+ %res = call <4 x float> @llvm.x86.avx512.mask.and.ps.128(<4 x float> %a, <4 x float> %b, <4 x float> %passThru, i8 %mask)
+ ret <4 x float> %res
+}
+
+define <4 x float> @test_mask_and_ps_rmkz_128(<4 x float> %a, <4 x float>* %ptr_b, i8 %mask) {
+; CHECK-LABEL: test_mask_and_ps_rmkz_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovb %esi, %k1 ## encoding: [0xc5,0xf9,0x92,0xce]
+; CHECK-NEXT: vandps (%rdi), %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0x89,0x54,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
+ %b = load <4 x float>, <4 x float>* %ptr_b
+ %res = call <4 x float> @llvm.x86.avx512.mask.and.ps.128(<4 x float> %a, <4 x float> %b, <4 x float> zeroinitializer, i8 %mask)
+ ret <4 x float> %res
+}
+
+define <4 x float> @test_mask_and_ps_rmb_128(<4 x float> %a, float* %ptr_b) {
+; CHECK-LABEL: test_mask_and_ps_rmb_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vandps (%rdi){1to4}, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7c,0x18,0x54,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
+ %q = load float, float* %ptr_b
+ %vecinit.i = insertelement <4 x float> undef, float %q, i32 0
+ %b = shufflevector <4 x float> %vecinit.i, <4 x float> undef, <4 x i32> zeroinitializer
+ %res = call <4 x float> @llvm.x86.avx512.mask.and.ps.128(<4 x float> %a, <4 x float> %b, <4 x float> zeroinitializer, i8 -1)
+ ret <4 x float> %res
+}
+
+define <4 x float> @test_mask_and_ps_rmbk_128(<4 x float> %a, float* %ptr_b, <4 x float> %passThru, i8 %mask) {
+; CHECK-LABEL: test_mask_and_ps_rmbk_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovb %esi, %k1 ## encoding: [0xc5,0xf9,0x92,0xce]
+; CHECK-NEXT: vandps (%rdi){1to4}, %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf1,0x7c,0x19,0x54,0x0f]
+; CHECK-NEXT: vmovaps %xmm1, %xmm0 ## encoding: [0x62,0xf1,0x7c,0x08,0x28,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
+ %q = load float, float* %ptr_b
+ %vecinit.i = insertelement <4 x float> undef, float %q, i32 0
+ %b = shufflevector <4 x float> %vecinit.i, <4 x float> undef, <4 x i32> zeroinitializer
+ %res = call <4 x float> @llvm.x86.avx512.mask.and.ps.128(<4 x float> %a, <4 x float> %b, <4 x float> %passThru, i8 %mask)
+ ret <4 x float> %res
+}
+
+define <4 x float> @test_mask_and_ps_rmbkz_128(<4 x float> %a, float* %ptr_b, i8 %mask) {
+; CHECK-LABEL: test_mask_and_ps_rmbkz_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovb %esi, %k1 ## encoding: [0xc5,0xf9,0x92,0xce]
+; CHECK-NEXT: vandps (%rdi){1to4}, %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0x99,0x54,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
+ %q = load float, float* %ptr_b
+ %vecinit.i = insertelement <4 x float> undef, float %q, i32 0
+ %b = shufflevector <4 x float> %vecinit.i, <4 x float> undef, <4 x i32> zeroinitializer
+ %res = call <4 x float> @llvm.x86.avx512.mask.and.ps.128(<4 x float> %a, <4 x float> %b, <4 x float> zeroinitializer, i8 %mask)
+ ret <4 x float> %res
+}
+
+declare <4 x float> @llvm.x86.avx512.mask.and.ps.128(<4 x float>, <4 x float>, <4 x float>, i8)
+
+define <8 x float> @test_mask_and_ps_rr_256(<8 x float> %a, <8 x float> %b) {
+; CHECK-LABEL: test_mask_and_ps_rr_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vandps %ymm1, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7c,0x28,0x54,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
+ %res = call <8 x float> @llvm.x86.avx512.mask.and.ps.256(<8 x float> %a, <8 x float> %b, <8 x float> zeroinitializer, i8 -1)
+ ret <8 x float> %res
+}
+
+define <8 x float> @test_mask_and_ps_rrk_256(<8 x float> %a, <8 x float> %b, <8 x float> %passThru, i8 %mask) {
+; CHECK-LABEL: test_mask_and_ps_rrk_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovb %edi, %k1 ## encoding: [0xc5,0xf9,0x92,0xcf]
+; CHECK-NEXT: vandps %ymm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf1,0x7c,0x29,0x54,0xd1]
+; CHECK-NEXT: vmovaps %ymm2, %ymm0 ## encoding: [0x62,0xf1,0x7c,0x28,0x28,0xc2]
+; CHECK-NEXT: retq ## encoding: [0xc3]
+ %res = call <8 x float> @llvm.x86.avx512.mask.and.ps.256(<8 x float> %a, <8 x float> %b, <8 x float> %passThru, i8 %mask)
+ ret <8 x float> %res
+}
+
+define <8 x float> @test_mask_and_ps_rrkz_256(<8 x float> %a, <8 x float> %b, i8 %mask) {
+; CHECK-LABEL: test_mask_and_ps_rrkz_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovb %edi, %k1 ## encoding: [0xc5,0xf9,0x92,0xcf]
+; CHECK-NEXT: vandps %ymm1, %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0xa9,0x54,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
+ %res = call <8 x float> @llvm.x86.avx512.mask.and.ps.256(<8 x float> %a, <8 x float> %b, <8 x float> zeroinitializer, i8 %mask)
+ ret <8 x float> %res
+}
+
+define <8 x float> @test_mask_and_ps_rm_256(<8 x float> %a, <8 x float>* %ptr_b) {
+; CHECK-LABEL: test_mask_and_ps_rm_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vandps (%rdi), %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7c,0x28,0x54,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
+ %b = load <8 x float>, <8 x float>* %ptr_b
+ %res = call <8 x float> @llvm.x86.avx512.mask.and.ps.256(<8 x float> %a, <8 x float> %b, <8 x float> zeroinitializer, i8 -1)
+ ret <8 x float> %res
+}
+
+define <8 x float> @test_mask_and_ps_rmk_256(<8 x float> %a, <8 x float>* %ptr_b, <8 x float> %passThru, i8 %mask) {
+; CHECK-LABEL: test_mask_and_ps_rmk_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovb %esi, %k1 ## encoding: [0xc5,0xf9,0x92,0xce]
+; CHECK-NEXT: vandps (%rdi), %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf1,0x7c,0x29,0x54,0x0f]
+; CHECK-NEXT: vmovaps %ymm1, %ymm0 ## encoding: [0x62,0xf1,0x7c,0x28,0x28,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
+ %b = load <8 x float>, <8 x float>* %ptr_b
+ %res = call <8 x float> @llvm.x86.avx512.mask.and.ps.256(<8 x float> %a, <8 x float> %b, <8 x float> %passThru, i8 %mask)
+ ret <8 x float> %res
+}
+
+define <8 x float> @test_mask_and_ps_rmkz_256(<8 x float> %a, <8 x float>* %ptr_b, i8 %mask) {
+; CHECK-LABEL: test_mask_and_ps_rmkz_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovb %esi, %k1 ## encoding: [0xc5,0xf9,0x92,0xce]
+; CHECK-NEXT: vandps (%rdi), %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0xa9,0x54,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
+ %b = load <8 x float>, <8 x float>* %ptr_b
+ %res = call <8 x float> @llvm.x86.avx512.mask.and.ps.256(<8 x float> %a, <8 x float> %b, <8 x float> zeroinitializer, i8 %mask)
+ ret <8 x float> %res
+}
+
+define <8 x float> @test_mask_and_ps_rmb_256(<8 x float> %a, float* %ptr_b) {
+; CHECK-LABEL: test_mask_and_ps_rmb_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vandps (%rdi){1to8}, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7c,0x38,0x54,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
+ %q = load float, float* %ptr_b
+ %vecinit.i = insertelement <8 x float> undef, float %q, i32 0
+ %b = shufflevector <8 x float> %vecinit.i, <8 x float> undef, <8 x i32> zeroinitializer
+ %res = call <8 x float> @llvm.x86.avx512.mask.and.ps.256(<8 x float> %a, <8 x float> %b, <8 x float> zeroinitializer, i8 -1)
+ ret <8 x float> %res
+}
+
+define <8 x float> @test_mask_and_ps_rmbk_256(<8 x float> %a, float* %ptr_b, <8 x float> %passThru, i8 %mask) {
+; CHECK-LABEL: test_mask_and_ps_rmbk_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovb %esi, %k1 ## encoding: [0xc5,0xf9,0x92,0xce]
+; CHECK-NEXT: vandps (%rdi){1to8}, %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf1,0x7c,0x39,0x54,0x0f]
+; CHECK-NEXT: vmovaps %ymm1, %ymm0 ## encoding: [0x62,0xf1,0x7c,0x28,0x28,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
+ %q = load float, float* %ptr_b
+ %vecinit.i = insertelement <8 x float> undef, float %q, i32 0
+ %b = shufflevector <8 x float> %vecinit.i, <8 x float> undef, <8 x i32> zeroinitializer
+ %res = call <8 x float> @llvm.x86.avx512.mask.and.ps.256(<8 x float> %a, <8 x float> %b, <8 x float> %passThru, i8 %mask)
+ ret <8 x float> %res
+}
+
+define <8 x float> @test_mask_and_ps_rmbkz_256(<8 x float> %a, float* %ptr_b, i8 %mask) {
+; CHECK-LABEL: test_mask_and_ps_rmbkz_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovb %esi, %k1 ## encoding: [0xc5,0xf9,0x92,0xce]
+; CHECK-NEXT: vandps (%rdi){1to8}, %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0xb9,0x54,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
+ %q = load float, float* %ptr_b
+ %vecinit.i = insertelement <8 x float> undef, float %q, i32 0
+ %b = shufflevector <8 x float> %vecinit.i, <8 x float> undef, <8 x i32> zeroinitializer
+ %res = call <8 x float> @llvm.x86.avx512.mask.and.ps.256(<8 x float> %a, <8 x float> %b, <8 x float> zeroinitializer, i8 %mask)
+ ret <8 x float> %res
+}
+
+declare <8 x float> @llvm.x86.avx512.mask.and.ps.256(<8 x float>, <8 x float>, <8 x float>, i8)
+
+define <16 x float> @test_mask_and_ps_rr_512(<16 x float> %a, <16 x float> %b) {
+; CHECK-LABEL: test_mask_and_ps_rr_512:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vandps %zmm1, %zmm0, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x54,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
+ %res = call <16 x float> @llvm.x86.avx512.mask.and.ps.512(<16 x float> %a, <16 x float> %b, <16 x float> zeroinitializer, i16 -1)
+ ret <16 x float> %res
+}
+
+define <16 x float> @test_mask_and_ps_rrk_512(<16 x float> %a, <16 x float> %b, <16 x float> %passThru, i16 %mask) {
+; CHECK-LABEL: test_mask_and_ps_rrk_512:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vandps %zmm1, %zmm0, %zmm2 {%k1} ## encoding: [0x62,0xf1,0x7c,0x49,0x54,0xd1]
+; CHECK-NEXT: vmovaps %zmm2, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc2]
+; CHECK-NEXT: retq ## encoding: [0xc3]
+ %res = call <16 x float> @llvm.x86.avx512.mask.and.ps.512(<16 x float> %a, <16 x float> %b, <16 x float> %passThru, i16 %mask)
+ ret <16 x float> %res
+}
+
+define <16 x float> @test_mask_and_ps_rrkz_512(<16 x float> %a, <16 x float> %b, i16 %mask) {
+; CHECK-LABEL: test_mask_and_ps_rrkz_512:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vandps %zmm1, %zmm0, %zmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0xc9,0x54,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
+ %res = call <16 x float> @llvm.x86.avx512.mask.and.ps.512(<16 x float> %a, <16 x float> %b, <16 x float> zeroinitializer, i16 %mask)
+ ret <16 x float> %res
+}
+
+define <16 x float> @test_mask_and_ps_rm_512(<16 x float> %a, <16 x float>* %ptr_b) {
+; CHECK-LABEL: test_mask_and_ps_rm_512:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vandps (%rdi), %zmm0, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x54,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
+ %b = load <16 x float>, <16 x float>* %ptr_b
+ %res = call <16 x float> @llvm.x86.avx512.mask.and.ps.512(<16 x float> %a, <16 x float> %b, <16 x float> zeroinitializer, i16 -1)
+ ret <16 x float> %res
+}
+
+define <16 x float> @test_mask_and_ps_rmk_512(<16 x float> %a, <16 x float>* %ptr_b, <16 x float> %passThru, i16 %mask) {
+; CHECK-LABEL: test_mask_and_ps_rmk_512:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
+; CHECK-NEXT: vandps (%rdi), %zmm0, %zmm1 {%k1} ## encoding: [0x62,0xf1,0x7c,0x49,0x54,0x0f]
+; CHECK-NEXT: vmovaps %zmm1, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
+ %b = load <16 x float>, <16 x float>* %ptr_b
+ %res = call <16 x float> @llvm.x86.avx512.mask.and.ps.512(<16 x float> %a, <16 x float> %b, <16 x float> %passThru, i16 %mask)
+ ret <16 x float> %res
+}
+
+define <16 x float> @test_mask_and_ps_rmkz_512(<16 x float> %a, <16 x float>* %ptr_b, i16 %mask) {
+; CHECK-LABEL: test_mask_and_ps_rmkz_512:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
+; CHECK-NEXT: vandps (%rdi), %zmm0, %zmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0xc9,0x54,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
+ %b = load <16 x float>, <16 x float>* %ptr_b
+ %res = call <16 x float> @llvm.x86.avx512.mask.and.ps.512(<16 x float> %a, <16 x float> %b, <16 x float> zeroinitializer, i16 %mask)
+ ret <16 x float> %res
+}
+
+define <16 x float> @test_mask_and_ps_rmb_512(<16 x float> %a, float* %ptr_b) {
+; CHECK-LABEL: test_mask_and_ps_rmb_512:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vandps (%rdi){1to16}, %zmm0, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x58,0x54,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
+ %q = load float, float* %ptr_b
+ %vecinit.i = insertelement <16 x float> undef, float %q, i32 0
+ %b = shufflevector <16 x float> %vecinit.i, <16 x float> undef, <16 x i32> zeroinitializer
+ %res = call <16 x float> @llvm.x86.avx512.mask.and.ps.512(<16 x float> %a, <16 x float> %b, <16 x float> zeroinitializer, i16 -1)
+ ret <16 x float> %res
+}
+
+define <16 x float> @test_mask_and_ps_rmbk_512(<16 x float> %a, float* %ptr_b, <16 x float> %passThru, i16 %mask) {
+; CHECK-LABEL: test_mask_and_ps_rmbk_512:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
+; CHECK-NEXT: vandps (%rdi){1to16}, %zmm0, %zmm1 {%k1} ## encoding: [0x62,0xf1,0x7c,0x59,0x54,0x0f]
+; CHECK-NEXT: vmovaps %zmm1, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
+ %q = load float, float* %ptr_b
+ %vecinit.i = insertelement <16 x float> undef, float %q, i32 0
+ %b = shufflevector <16 x float> %vecinit.i, <16 x float> undef, <16 x i32> zeroinitializer
+ %res = call <16 x float> @llvm.x86.avx512.mask.and.ps.512(<16 x float> %a, <16 x float> %b, <16 x float> %passThru, i16 %mask)
+ ret <16 x float> %res
+}
+
+define <16 x float> @test_mask_and_ps_rmbkz_512(<16 x float> %a, float* %ptr_b, i16 %mask) {
+; CHECK-LABEL: test_mask_and_ps_rmbkz_512:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
+; CHECK-NEXT: vandps (%rdi){1to16}, %zmm0, %zmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0xd9,0x54,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
+ %q = load float, float* %ptr_b
+ %vecinit.i = insertelement <16 x float> undef, float %q, i32 0
+ %b = shufflevector <16 x float> %vecinit.i, <16 x float> undef, <16 x i32> zeroinitializer
+ %res = call <16 x float> @llvm.x86.avx512.mask.and.ps.512(<16 x float> %a, <16 x float> %b, <16 x float> zeroinitializer, i16 %mask)
+ ret <16 x float> %res
+}
+
+declare <16 x float> @llvm.x86.avx512.mask.and.ps.512(<16 x float>, <16 x float>, <16 x float>, i16)
+
+define <4 x float> @test_mask_or_ps_rr_128(<4 x float> %a, <4 x float> %b) {
+; CHECK-LABEL: test_mask_or_ps_rr_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vorps %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7c,0x08,0x56,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
+ %res = call <4 x float> @llvm.x86.avx512.mask.or.ps.128(<4 x float> %a, <4 x float> %b, <4 x float> zeroinitializer, i8 -1)
+ ret <4 x float> %res
+}
+
+define <4 x float> @test_mask_or_ps_rrk_128(<4 x float> %a, <4 x float> %b, <4 x float> %passThru, i8 %mask) {
+; CHECK-LABEL: test_mask_or_ps_rrk_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovb %edi, %k1 ## encoding: [0xc5,0xf9,0x92,0xcf]
+; CHECK-NEXT: vorps %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf1,0x7c,0x09,0x56,0xd1]
+; CHECK-NEXT: vmovaps %xmm2, %xmm0 ## encoding: [0x62,0xf1,0x7c,0x08,0x28,0xc2]
+; CHECK-NEXT: retq ## encoding: [0xc3]
+ %res = call <4 x float> @llvm.x86.avx512.mask.or.ps.128(<4 x float> %a, <4 x float> %b, <4 x float> %passThru, i8 %mask)
+ ret <4 x float> %res
+}
+
+define <4 x float> @test_mask_or_ps_rrkz_128(<4 x float> %a, <4 x float> %b, i8 %mask) {
+; CHECK-LABEL: test_mask_or_ps_rrkz_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovb %edi, %k1 ## encoding: [0xc5,0xf9,0x92,0xcf]
+; CHECK-NEXT: vorps %xmm1, %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0x89,0x56,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
+ %res = call <4 x float> @llvm.x86.avx512.mask.or.ps.128(<4 x float> %a, <4 x float> %b, <4 x float> zeroinitializer, i8 %mask)
+ ret <4 x float> %res
+}
+
+define <4 x float> @test_mask_or_ps_rm_128(<4 x float> %a, <4 x float>* %ptr_b) {
+; CHECK-LABEL: test_mask_or_ps_rm_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vorps (%rdi), %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7c,0x08,0x56,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
+ %b = load <4 x float>, <4 x float>* %ptr_b
+ %res = call <4 x float> @llvm.x86.avx512.mask.or.ps.128(<4 x float> %a, <4 x float> %b, <4 x float> zeroinitializer, i8 -1)
+ ret <4 x float> %res
+}
+
+define <4 x float> @test_mask_or_ps_rmk_128(<4 x float> %a, <4 x float>* %ptr_b, <4 x float> %passThru, i8 %mask) {
+; CHECK-LABEL: test_mask_or_ps_rmk_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovb %esi, %k1 ## encoding: [0xc5,0xf9,0x92,0xce]
+; CHECK-NEXT: vorps (%rdi), %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf1,0x7c,0x09,0x56,0x0f]
+; CHECK-NEXT: vmovaps %xmm1, %xmm0 ## encoding: [0x62,0xf1,0x7c,0x08,0x28,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
+ %b = load <4 x float>, <4 x float>* %ptr_b
+ %res = call <4 x float> @llvm.x86.avx512.mask.or.ps.128(<4 x float> %a, <4 x float> %b, <4 x float> %passThru, i8 %mask)
+ ret <4 x float> %res
+}
+
+define <4 x float> @test_mask_or_ps_rmkz_128(<4 x float> %a, <4 x float>* %ptr_b, i8 %mask) {
+; CHECK-LABEL: test_mask_or_ps_rmkz_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovb %esi, %k1 ## encoding: [0xc5,0xf9,0x92,0xce]
+; CHECK-NEXT: vorps (%rdi), %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0x89,0x56,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
+ %b = load <4 x float>, <4 x float>* %ptr_b
+ %res = call <4 x float> @llvm.x86.avx512.mask.or.ps.128(<4 x float> %a, <4 x float> %b, <4 x float> zeroinitializer, i8 %mask)
+ ret <4 x float> %res
+}
+
+define <4 x float> @test_mask_or_ps_rmb_128(<4 x float> %a, float* %ptr_b) {
+; CHECK-LABEL: test_mask_or_ps_rmb_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vorps (%rdi){1to4}, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7c,0x18,0x56,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
+ %q = load float, float* %ptr_b
+ %vecinit.i = insertelement <4 x float> undef, float %q, i32 0
+ %b = shufflevector <4 x float> %vecinit.i, <4 x float> undef, <4 x i32> zeroinitializer
+ %res = call <4 x float> @llvm.x86.avx512.mask.or.ps.128(<4 x float> %a, <4 x float> %b, <4 x float> zeroinitializer, i8 -1)
+ ret <4 x float> %res
+}
+
+define <4 x float> @test_mask_or_ps_rmbk_128(<4 x float> %a, float* %ptr_b, <4 x float> %passThru, i8 %mask) {
+; CHECK-LABEL: test_mask_or_ps_rmbk_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovb %esi, %k1 ## encoding: [0xc5,0xf9,0x92,0xce]
+; CHECK-NEXT: vorps (%rdi){1to4}, %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf1,0x7c,0x19,0x56,0x0f]
+; CHECK-NEXT: vmovaps %xmm1, %xmm0 ## encoding: [0x62,0xf1,0x7c,0x08,0x28,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
+ %q = load float, float* %ptr_b
+ %vecinit.i = insertelement <4 x float> undef, float %q, i32 0
+ %b = shufflevector <4 x float> %vecinit.i, <4 x float> undef, <4 x i32> zeroinitializer
+ %res = call <4 x float> @llvm.x86.avx512.mask.or.ps.128(<4 x float> %a, <4 x float> %b, <4 x float> %passThru, i8 %mask)
+ ret <4 x float> %res
+}
+
+define <4 x float> @test_mask_or_ps_rmbkz_128(<4 x float> %a, float* %ptr_b, i8 %mask) {
+; CHECK-LABEL: test_mask_or_ps_rmbkz_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovb %esi, %k1 ## encoding: [0xc5,0xf9,0x92,0xce]
+; CHECK-NEXT: vorps (%rdi){1to4}, %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0x99,0x56,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
+ %q = load float, float* %ptr_b
+ %vecinit.i = insertelement <4 x float> undef, float %q, i32 0
+ %b = shufflevector <4 x float> %vecinit.i, <4 x float> undef, <4 x i32> zeroinitializer
+ %res = call <4 x float> @llvm.x86.avx512.mask.or.ps.128(<4 x float> %a, <4 x float> %b, <4 x float> zeroinitializer, i8 %mask)
+ ret <4 x float> %res
+}
+
+declare <4 x float> @llvm.x86.avx512.mask.or.ps.128(<4 x float>, <4 x float>, <4 x float>, i8)
+
+define <8 x float> @test_mask_or_ps_rr_256(<8 x float> %a, <8 x float> %b) {
+; CHECK-LABEL: test_mask_or_ps_rr_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vorps %ymm1, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7c,0x28,0x56,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
+ %res = call <8 x float> @llvm.x86.avx512.mask.or.ps.256(<8 x float> %a, <8 x float> %b, <8 x float> zeroinitializer, i8 -1)
+ ret <8 x float> %res
+}
+
+define <8 x float> @test_mask_or_ps_rrk_256(<8 x float> %a, <8 x float> %b, <8 x float> %passThru, i8 %mask) {
+; CHECK-LABEL: test_mask_or_ps_rrk_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovb %edi, %k1 ## encoding: [0xc5,0xf9,0x92,0xcf]
+; CHECK-NEXT: vorps %ymm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf1,0x7c,0x29,0x56,0xd1]
+; CHECK-NEXT: vmovaps %ymm2, %ymm0 ## encoding: [0x62,0xf1,0x7c,0x28,0x28,0xc2]
+; CHECK-NEXT: retq ## encoding: [0xc3]
+ %res = call <8 x float> @llvm.x86.avx512.mask.or.ps.256(<8 x float> %a, <8 x float> %b, <8 x float> %passThru, i8 %mask)
+ ret <8 x float> %res
+}
+
+define <8 x float> @test_mask_or_ps_rrkz_256(<8 x float> %a, <8 x float> %b, i8 %mask) {
+; CHECK-LABEL: test_mask_or_ps_rrkz_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovb %edi, %k1 ## encoding: [0xc5,0xf9,0x92,0xcf]
+; CHECK-NEXT: vorps %ymm1, %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0xa9,0x56,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
+ %res = call <8 x float> @llvm.x86.avx512.mask.or.ps.256(<8 x float> %a, <8 x float> %b, <8 x float> zeroinitializer, i8 %mask)
+ ret <8 x float> %res
+}
+
+define <8 x float> @test_mask_or_ps_rm_256(<8 x float> %a, <8 x float>* %ptr_b) {
+; CHECK-LABEL: test_mask_or_ps_rm_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vorps (%rdi), %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7c,0x28,0x56,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
+ %b = load <8 x float>, <8 x float>* %ptr_b
+ %res = call <8 x float> @llvm.x86.avx512.mask.or.ps.256(<8 x float> %a, <8 x float> %b, <8 x float> zeroinitializer, i8 -1)
+ ret <8 x float> %res
+}
+
+define <8 x float> @test_mask_or_ps_rmk_256(<8 x float> %a, <8 x float>* %ptr_b, <8 x float> %passThru, i8 %mask) {
+; CHECK-LABEL: test_mask_or_ps_rmk_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovb %esi, %k1 ## encoding: [0xc5,0xf9,0x92,0xce]
+; CHECK-NEXT: vorps (%rdi), %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf1,0x7c,0x29,0x56,0x0f]
+; CHECK-NEXT: vmovaps %ymm1, %ymm0 ## encoding: [0x62,0xf1,0x7c,0x28,0x28,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
+ %b = load <8 x float>, <8 x float>* %ptr_b
+ %res = call <8 x float> @llvm.x86.avx512.mask.or.ps.256(<8 x float> %a, <8 x float> %b, <8 x float> %passThru, i8 %mask)
+ ret <8 x float> %res
+}
+
+define <8 x float> @test_mask_or_ps_rmkz_256(<8 x float> %a, <8 x float>* %ptr_b, i8 %mask) {
+; CHECK-LABEL: test_mask_or_ps_rmkz_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovb %esi, %k1 ## encoding: [0xc5,0xf9,0x92,0xce]
+; CHECK-NEXT: vorps (%rdi), %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0xa9,0x56,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
+ %b = load <8 x float>, <8 x float>* %ptr_b
+ %res = call <8 x float> @llvm.x86.avx512.mask.or.ps.256(<8 x float> %a, <8 x float> %b, <8 x float> zeroinitializer, i8 %mask)
+ ret <8 x float> %res
+}
+
+define <8 x float> @test_mask_or_ps_rmb_256(<8 x float> %a, float* %ptr_b) {
+; CHECK-LABEL: test_mask_or_ps_rmb_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vorps (%rdi){1to8}, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7c,0x38,0x56,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
+ %q = load float, float* %ptr_b
+ %vecinit.i = insertelement <8 x float> undef, float %q, i32 0
+ %b = shufflevector <8 x float> %vecinit.i, <8 x float> undef, <8 x i32> zeroinitializer
+ %res = call <8 x float> @llvm.x86.avx512.mask.or.ps.256(<8 x float> %a, <8 x float> %b, <8 x float> zeroinitializer, i8 -1)
+ ret <8 x float> %res
+}
+
+define <8 x float> @test_mask_or_ps_rmbk_256(<8 x float> %a, float* %ptr_b, <8 x float> %passThru, i8 %mask) {
+; CHECK-LABEL: test_mask_or_ps_rmbk_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovb %esi, %k1 ## encoding: [0xc5,0xf9,0x92,0xce]
+; CHECK-NEXT: vorps (%rdi){1to8}, %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf1,0x7c,0x39,0x56,0x0f]
+; CHECK-NEXT: vmovaps %ymm1, %ymm0 ## encoding: [0x62,0xf1,0x7c,0x28,0x28,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
+ %q = load float, float* %ptr_b
+ %vecinit.i = insertelement <8 x float> undef, float %q, i32 0
+ %b = shufflevector <8 x float> %vecinit.i, <8 x float> undef, <8 x i32> zeroinitializer
+ %res = call <8 x float> @llvm.x86.avx512.mask.or.ps.256(<8 x float> %a, <8 x float> %b, <8 x float> %passThru, i8 %mask)
+ ret <8 x float> %res
+}
+
+define <8 x float> @test_mask_or_ps_rmbkz_256(<8 x float> %a, float* %ptr_b, i8 %mask) {
+; CHECK-LABEL: test_mask_or_ps_rmbkz_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovb %esi, %k1 ## encoding: [0xc5,0xf9,0x92,0xce]
+; CHECK-NEXT: vorps (%rdi){1to8}, %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0xb9,0x56,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
+ %q = load float, float* %ptr_b
+ %vecinit.i = insertelement <8 x float> undef, float %q, i32 0
+ %b = shufflevector <8 x float> %vecinit.i, <8 x float> undef, <8 x i32> zeroinitializer
+ %res = call <8 x float> @llvm.x86.avx512.mask.or.ps.256(<8 x float> %a, <8 x float> %b, <8 x float> zeroinitializer, i8 %mask)
+ ret <8 x float> %res
+}
+
+declare <8 x float> @llvm.x86.avx512.mask.or.ps.256(<8 x float>, <8 x float>, <8 x float>, i8)
+
+define <16 x float> @test_mask_or_ps_rr_512(<16 x float> %a, <16 x float> %b) {
+; CHECK-LABEL: test_mask_or_ps_rr_512:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vorps %zmm1, %zmm0, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x56,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
+ %res = call <16 x float> @llvm.x86.avx512.mask.or.ps.512(<16 x float> %a, <16 x float> %b, <16 x float> zeroinitializer, i16 -1)
+ ret <16 x float> %res
+}
+
+define <16 x float> @test_mask_or_ps_rrk_512(<16 x float> %a, <16 x float> %b, <16 x float> %passThru, i16 %mask) {
+; CHECK-LABEL: test_mask_or_ps_rrk_512:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vorps %zmm1, %zmm0, %zmm2 {%k1} ## encoding: [0x62,0xf1,0x7c,0x49,0x56,0xd1]
+; CHECK-NEXT: vmovaps %zmm2, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc2]
+; CHECK-NEXT: retq ## encoding: [0xc3]
+ %res = call <16 x float> @llvm.x86.avx512.mask.or.ps.512(<16 x float> %a, <16 x float> %b, <16 x float> %passThru, i16 %mask)
+ ret <16 x float> %res
+}
+
+define <16 x float> @test_mask_or_ps_rrkz_512(<16 x float> %a, <16 x float> %b, i16 %mask) {
+; CHECK-LABEL: test_mask_or_ps_rrkz_512:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vorps %zmm1, %zmm0, %zmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0xc9,0x56,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
+ %res = call <16 x float> @llvm.x86.avx512.mask.or.ps.512(<16 x float> %a, <16 x float> %b, <16 x float> zeroinitializer, i16 %mask)
+ ret <16 x float> %res
+}
+
+define <16 x float> @test_mask_or_ps_rm_512(<16 x float> %a, <16 x float>* %ptr_b) {
+; CHECK-LABEL: test_mask_or_ps_rm_512:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vorps (%rdi), %zmm0, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x56,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
+ %b = load <16 x float>, <16 x float>* %ptr_b
+ %res = call <16 x float> @llvm.x86.avx512.mask.or.ps.512(<16 x float> %a, <16 x float> %b, <16 x float> zeroinitializer, i16 -1)
+ ret <16 x float> %res
+}
+
+define <16 x float> @test_mask_or_ps_rmk_512(<16 x float> %a, <16 x float>* %ptr_b, <16 x float> %passThru, i16 %mask) {
+; CHECK-LABEL: test_mask_or_ps_rmk_512:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
+; CHECK-NEXT: vorps (%rdi), %zmm0, %zmm1 {%k1} ## encoding: [0x62,0xf1,0x7c,0x49,0x56,0x0f]
+; CHECK-NEXT: vmovaps %zmm1, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
+ %b = load <16 x float>, <16 x float>* %ptr_b
+ %res = call <16 x float> @llvm.x86.avx512.mask.or.ps.512(<16 x float> %a, <16 x float> %b, <16 x float> %passThru, i16 %mask)
+ ret <16 x float> %res
+}
+
+define <16 x float> @test_mask_or_ps_rmkz_512(<16 x float> %a, <16 x float>* %ptr_b, i16 %mask) {
+; CHECK-LABEL: test_mask_or_ps_rmkz_512:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
+; CHECK-NEXT: vorps (%rdi), %zmm0, %zmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0xc9,0x56,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
+ %b = load <16 x float>, <16 x float>* %ptr_b
+ %res = call <16 x float> @llvm.x86.avx512.mask.or.ps.512(<16 x float> %a, <16 x float> %b, <16 x float> zeroinitializer, i16 %mask)
+ ret <16 x float> %res
+}
+
+define <16 x float> @test_mask_or_ps_rmb_512(<16 x float> %a, float* %ptr_b) {
+; CHECK-LABEL: test_mask_or_ps_rmb_512:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vorps (%rdi){1to16}, %zmm0, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x58,0x56,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
+ %q = load float, float* %ptr_b
+ %vecinit.i = insertelement <16 x float> undef, float %q, i32 0
+ %b = shufflevector <16 x float> %vecinit.i, <16 x float> undef, <16 x i32> zeroinitializer
+ %res = call <16 x float> @llvm.x86.avx512.mask.or.ps.512(<16 x float> %a, <16 x float> %b, <16 x float> zeroinitializer, i16 -1)
+ ret <16 x float> %res
+}
+
+define <16 x float> @test_mask_or_ps_rmbk_512(<16 x float> %a, float* %ptr_b, <16 x float> %passThru, i16 %mask) {
+; CHECK-LABEL: test_mask_or_ps_rmbk_512:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
+; CHECK-NEXT: vorps (%rdi){1to16}, %zmm0, %zmm1 {%k1} ## encoding: [0x62,0xf1,0x7c,0x59,0x56,0x0f]
+; CHECK-NEXT: vmovaps %zmm1, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
+ %q = load float, float* %ptr_b
+ %vecinit.i = insertelement <16 x float> undef, float %q, i32 0
+ %b = shufflevector <16 x float> %vecinit.i, <16 x float> undef, <16 x i32> zeroinitializer
+ %res = call <16 x float> @llvm.x86.avx512.mask.or.ps.512(<16 x float> %a, <16 x float> %b, <16 x float> %passThru, i16 %mask)
+ ret <16 x float> %res
+}
+
+define <16 x float> @test_mask_or_ps_rmbkz_512(<16 x float> %a, float* %ptr_b, i16 %mask) {
+; CHECK-LABEL: test_mask_or_ps_rmbkz_512:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
+; CHECK-NEXT: vorps (%rdi){1to16}, %zmm0, %zmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0xd9,0x56,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
+ %q = load float, float* %ptr_b
+ %vecinit.i = insertelement <16 x float> undef, float %q, i32 0
+ %b = shufflevector <16 x float> %vecinit.i, <16 x float> undef, <16 x i32> zeroinitializer
+ %res = call <16 x float> @llvm.x86.avx512.mask.or.ps.512(<16 x float> %a, <16 x float> %b, <16 x float> zeroinitializer, i16 %mask)
+ ret <16 x float> %res
+}
+
+declare <16 x float> @llvm.x86.avx512.mask.or.ps.512(<16 x float>, <16 x float>, <16 x float>, i16)
+
+define <4 x float> @test_mask_xor_ps_rr_128(<4 x float> %a, <4 x float> %b) {
+; CHECK-LABEL: test_mask_xor_ps_rr_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vxorps %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7c,0x08,0x57,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
+ %res = call <4 x float> @llvm.x86.avx512.mask.xor.ps.128(<4 x float> %a, <4 x float> %b, <4 x float> zeroinitializer, i8 -1)
+ ret <4 x float> %res
+}
+
+define <4 x float> @test_mask_xor_ps_rrk_128(<4 x float> %a, <4 x float> %b, <4 x float> %passThru, i8 %mask) {
+; CHECK-LABEL: test_mask_xor_ps_rrk_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovb %edi, %k1 ## encoding: [0xc5,0xf9,0x92,0xcf]
+; CHECK-NEXT: vxorps %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf1,0x7c,0x09,0x57,0xd1]
+; CHECK-NEXT: vmovaps %xmm2, %xmm0 ## encoding: [0x62,0xf1,0x7c,0x08,0x28,0xc2]
+; CHECK-NEXT: retq ## encoding: [0xc3]
+ %res = call <4 x float> @llvm.x86.avx512.mask.xor.ps.128(<4 x float> %a, <4 x float> %b, <4 x float> %passThru, i8 %mask)
+ ret <4 x float> %res
+}
+
+define <4 x float> @test_mask_xor_ps_rrkz_128(<4 x float> %a, <4 x float> %b, i8 %mask) {
+; CHECK-LABEL: test_mask_xor_ps_rrkz_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovb %edi, %k1 ## encoding: [0xc5,0xf9,0x92,0xcf]
+; CHECK-NEXT: vxorps %xmm1, %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0x89,0x57,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
+ %res = call <4 x float> @llvm.x86.avx512.mask.xor.ps.128(<4 x float> %a, <4 x float> %b, <4 x float> zeroinitializer, i8 %mask)
+ ret <4 x float> %res
+}
+
+define <4 x float> @test_mask_xor_ps_rm_128(<4 x float> %a, <4 x float>* %ptr_b) {
+; CHECK-LABEL: test_mask_xor_ps_rm_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vxorps (%rdi), %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7c,0x08,0x57,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
+ %b = load <4 x float>, <4 x float>* %ptr_b
+ %res = call <4 x float> @llvm.x86.avx512.mask.xor.ps.128(<4 x float> %a, <4 x float> %b, <4 x float> zeroinitializer, i8 -1)
+ ret <4 x float> %res
+}
+
+define <4 x float> @test_mask_xor_ps_rmk_128(<4 x float> %a, <4 x float>* %ptr_b, <4 x float> %passThru, i8 %mask) {
+; CHECK-LABEL: test_mask_xor_ps_rmk_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovb %esi, %k1 ## encoding: [0xc5,0xf9,0x92,0xce]
+; CHECK-NEXT: vxorps (%rdi), %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf1,0x7c,0x09,0x57,0x0f]
+; CHECK-NEXT: vmovaps %xmm1, %xmm0 ## encoding: [0x62,0xf1,0x7c,0x08,0x28,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
+ %b = load <4 x float>, <4 x float>* %ptr_b
+ %res = call <4 x float> @llvm.x86.avx512.mask.xor.ps.128(<4 x float> %a, <4 x float> %b, <4 x float> %passThru, i8 %mask)
+ ret <4 x float> %res
+}
+
+define <4 x float> @test_mask_xor_ps_rmkz_128(<4 x float> %a, <4 x float>* %ptr_b, i8 %mask) {
+; CHECK-LABEL: test_mask_xor_ps_rmkz_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovb %esi, %k1 ## encoding: [0xc5,0xf9,0x92,0xce]
+; CHECK-NEXT: vxorps (%rdi), %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0x89,0x57,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
+ %b = load <4 x float>, <4 x float>* %ptr_b
+ %res = call <4 x float> @llvm.x86.avx512.mask.xor.ps.128(<4 x float> %a, <4 x float> %b, <4 x float> zeroinitializer, i8 %mask)
+ ret <4 x float> %res
+}
+
+define <4 x float> @test_mask_xor_ps_rmb_128(<4 x float> %a, float* %ptr_b) {
+; CHECK-LABEL: test_mask_xor_ps_rmb_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vxorps (%rdi){1to4}, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7c,0x18,0x57,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
+ %q = load float, float* %ptr_b
+ %vecinit.i = insertelement <4 x float> undef, float %q, i32 0
+ %b = shufflevector <4 x float> %vecinit.i, <4 x float> undef, <4 x i32> zeroinitializer
+ %res = call <4 x float> @llvm.x86.avx512.mask.xor.ps.128(<4 x float> %a, <4 x float> %b, <4 x float> zeroinitializer, i8 -1)
+ ret <4 x float> %res
+}
+
+define <4 x float> @test_mask_xor_ps_rmbk_128(<4 x float> %a, float* %ptr_b, <4 x float> %passThru, i8 %mask) {
+; CHECK-LABEL: test_mask_xor_ps_rmbk_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovb %esi, %k1 ## encoding: [0xc5,0xf9,0x92,0xce]
+; CHECK-NEXT: vxorps (%rdi){1to4}, %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf1,0x7c,0x19,0x57,0x0f]
+; CHECK-NEXT: vmovaps %xmm1, %xmm0 ## encoding: [0x62,0xf1,0x7c,0x08,0x28,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
+ %q = load float, float* %ptr_b
+ %vecinit.i = insertelement <4 x float> undef, float %q, i32 0
+ %b = shufflevector <4 x float> %vecinit.i, <4 x float> undef, <4 x i32> zeroinitializer
+ %res = call <4 x float> @llvm.x86.avx512.mask.xor.ps.128(<4 x float> %a, <4 x float> %b, <4 x float> %passThru, i8 %mask)
+ ret <4 x float> %res
+}
+
+define <4 x float> @test_mask_xor_ps_rmbkz_128(<4 x float> %a, float* %ptr_b, i8 %mask) {
+; CHECK-LABEL: test_mask_xor_ps_rmbkz_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovb %esi, %k1 ## encoding: [0xc5,0xf9,0x92,0xce]
+; CHECK-NEXT: vxorps (%rdi){1to4}, %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0x99,0x57,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
+ %q = load float, float* %ptr_b
+ %vecinit.i = insertelement <4 x float> undef, float %q, i32 0
+ %b = shufflevector <4 x float> %vecinit.i, <4 x float> undef, <4 x i32> zeroinitializer
+ %res = call <4 x float> @llvm.x86.avx512.mask.xor.ps.128(<4 x float> %a, <4 x float> %b, <4 x float> zeroinitializer, i8 %mask)
+ ret <4 x float> %res
+}
+
+declare <4 x float> @llvm.x86.avx512.mask.xor.ps.128(<4 x float>, <4 x float>, <4 x float>, i8)
+
+define <8 x float> @test_mask_xor_ps_rr_256(<8 x float> %a, <8 x float> %b) {
+; CHECK-LABEL: test_mask_xor_ps_rr_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vxorps %ymm1, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7c,0x28,0x57,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
+ %res = call <8 x float> @llvm.x86.avx512.mask.xor.ps.256(<8 x float> %a, <8 x float> %b, <8 x float> zeroinitializer, i8 -1)
+ ret <8 x float> %res
+}
+
+define <8 x float> @test_mask_xor_ps_rrk_256(<8 x float> %a, <8 x float> %b, <8 x float> %passThru, i8 %mask) {
+; CHECK-LABEL: test_mask_xor_ps_rrk_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovb %edi, %k1 ## encoding: [0xc5,0xf9,0x92,0xcf]
+; CHECK-NEXT: vxorps %ymm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf1,0x7c,0x29,0x57,0xd1]
+; CHECK-NEXT: vmovaps %ymm2, %ymm0 ## encoding: [0x62,0xf1,0x7c,0x28,0x28,0xc2]
+; CHECK-NEXT: retq ## encoding: [0xc3]
+ %res = call <8 x float> @llvm.x86.avx512.mask.xor.ps.256(<8 x float> %a, <8 x float> %b, <8 x float> %passThru, i8 %mask)
+ ret <8 x float> %res
+}
+
+define <8 x float> @test_mask_xor_ps_rrkz_256(<8 x float> %a, <8 x float> %b, i8 %mask) {
+; CHECK-LABEL: test_mask_xor_ps_rrkz_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovb %edi, %k1 ## encoding: [0xc5,0xf9,0x92,0xcf]
+; CHECK-NEXT: vxorps %ymm1, %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0xa9,0x57,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
+ %res = call <8 x float> @llvm.x86.avx512.mask.xor.ps.256(<8 x float> %a, <8 x float> %b, <8 x float> zeroinitializer, i8 %mask)
+ ret <8 x float> %res
+}
+
+define <8 x float> @test_mask_xor_ps_rm_256(<8 x float> %a, <8 x float>* %ptr_b) {
+; CHECK-LABEL: test_mask_xor_ps_rm_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vxorps (%rdi), %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7c,0x28,0x57,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
+ %b = load <8 x float>, <8 x float>* %ptr_b
+ %res = call <8 x float> @llvm.x86.avx512.mask.xor.ps.256(<8 x float> %a, <8 x float> %b, <8 x float> zeroinitializer, i8 -1)
+ ret <8 x float> %res
+}
+
+define <8 x float> @test_mask_xor_ps_rmk_256(<8 x float> %a, <8 x float>* %ptr_b, <8 x float> %passThru, i8 %mask) {
+; CHECK-LABEL: test_mask_xor_ps_rmk_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovb %esi, %k1 ## encoding: [0xc5,0xf9,0x92,0xce]
+; CHECK-NEXT: vxorps (%rdi), %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf1,0x7c,0x29,0x57,0x0f]
+; CHECK-NEXT: vmovaps %ymm1, %ymm0 ## encoding: [0x62,0xf1,0x7c,0x28,0x28,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
+ %b = load <8 x float>, <8 x float>* %ptr_b
+ %res = call <8 x float> @llvm.x86.avx512.mask.xor.ps.256(<8 x float> %a, <8 x float> %b, <8 x float> %passThru, i8 %mask)
+ ret <8 x float> %res
+}
+
+define <8 x float> @test_mask_xor_ps_rmkz_256(<8 x float> %a, <8 x float>* %ptr_b, i8 %mask) {
+; CHECK-LABEL: test_mask_xor_ps_rmkz_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovb %esi, %k1 ## encoding: [0xc5,0xf9,0x92,0xce]
+; CHECK-NEXT: vxorps (%rdi), %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0xa9,0x57,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
+ %b = load <8 x float>, <8 x float>* %ptr_b
+ %res = call <8 x float> @llvm.x86.avx512.mask.xor.ps.256(<8 x float> %a, <8 x float> %b, <8 x float> zeroinitializer, i8 %mask)
+ ret <8 x float> %res
+}
+
+define <8 x float> @test_mask_xor_ps_rmb_256(<8 x float> %a, float* %ptr_b) {
+; CHECK-LABEL: test_mask_xor_ps_rmb_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vxorps (%rdi){1to8}, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7c,0x38,0x57,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
+ %q = load float, float* %ptr_b
+ %vecinit.i = insertelement <8 x float> undef, float %q, i32 0
+ %b = shufflevector <8 x float> %vecinit.i, <8 x float> undef, <8 x i32> zeroinitializer
+ %res = call <8 x float> @llvm.x86.avx512.mask.xor.ps.256(<8 x float> %a, <8 x float> %b, <8 x float> zeroinitializer, i8 -1)
+ ret <8 x float> %res
+}
+
+define <8 x float> @test_mask_xor_ps_rmbk_256(<8 x float> %a, float* %ptr_b, <8 x float> %passThru, i8 %mask) {
+; CHECK-LABEL: test_mask_xor_ps_rmbk_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovb %esi, %k1 ## encoding: [0xc5,0xf9,0x92,0xce]
+; CHECK-NEXT: vxorps (%rdi){1to8}, %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf1,0x7c,0x39,0x57,0x0f]
+; CHECK-NEXT: vmovaps %ymm1, %ymm0 ## encoding: [0x62,0xf1,0x7c,0x28,0x28,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
+ %q = load float, float* %ptr_b
+ %vecinit.i = insertelement <8 x float> undef, float %q, i32 0
+ %b = shufflevector <8 x float> %vecinit.i, <8 x float> undef, <8 x i32> zeroinitializer
+ %res = call <8 x float> @llvm.x86.avx512.mask.xor.ps.256(<8 x float> %a, <8 x float> %b, <8 x float> %passThru, i8 %mask)
+ ret <8 x float> %res
+}
+
+define <8 x float> @test_mask_xor_ps_rmbkz_256(<8 x float> %a, float* %ptr_b, i8 %mask) {
+; CHECK-LABEL: test_mask_xor_ps_rmbkz_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovb %esi, %k1 ## encoding: [0xc5,0xf9,0x92,0xce]
+; CHECK-NEXT: vxorps (%rdi){1to8}, %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0xb9,0x57,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
+ %q = load float, float* %ptr_b
+ %vecinit.i = insertelement <8 x float> undef, float %q, i32 0
+ %b = shufflevector <8 x float> %vecinit.i, <8 x float> undef, <8 x i32> zeroinitializer
+ %res = call <8 x float> @llvm.x86.avx512.mask.xor.ps.256(<8 x float> %a, <8 x float> %b, <8 x float> zeroinitializer, i8 %mask)
+ ret <8 x float> %res
+}
+
+declare <8 x float> @llvm.x86.avx512.mask.xor.ps.256(<8 x float>, <8 x float>, <8 x float>, i8)
+
+define <16 x float> @test_mask_xor_ps_rr_512(<16 x float> %a, <16 x float> %b) {
+; CHECK-LABEL: test_mask_xor_ps_rr_512:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vxorps %zmm1, %zmm0, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x57,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
+ %res = call <16 x float> @llvm.x86.avx512.mask.xor.ps.512(<16 x float> %a, <16 x float> %b, <16 x float> zeroinitializer, i16 -1)
+ ret <16 x float> %res
+}
+
+define <16 x float> @test_mask_xor_ps_rrk_512(<16 x float> %a, <16 x float> %b, <16 x float> %passThru, i16 %mask) {
+; CHECK-LABEL: test_mask_xor_ps_rrk_512:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vxorps %zmm1, %zmm0, %zmm2 {%k1} ## encoding: [0x62,0xf1,0x7c,0x49,0x57,0xd1]
+; CHECK-NEXT: vmovaps %zmm2, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc2]
+; CHECK-NEXT: retq ## encoding: [0xc3]
+ %res = call <16 x float> @llvm.x86.avx512.mask.xor.ps.512(<16 x float> %a, <16 x float> %b, <16 x float> %passThru, i16 %mask)
+ ret <16 x float> %res
+}
+
+define <16 x float> @test_mask_xor_ps_rrkz_512(<16 x float> %a, <16 x float> %b, i16 %mask) {
+; CHECK-LABEL: test_mask_xor_ps_rrkz_512:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vxorps %zmm1, %zmm0, %zmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0xc9,0x57,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
+ %res = call <16 x float> @llvm.x86.avx512.mask.xor.ps.512(<16 x float> %a, <16 x float> %b, <16 x float> zeroinitializer, i16 %mask)
+ ret <16 x float> %res
+}
+
+define <16 x float> @test_mask_xor_ps_rm_512(<16 x float> %a, <16 x float>* %ptr_b) {
+; CHECK-LABEL: test_mask_xor_ps_rm_512:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vxorps (%rdi), %zmm0, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x57,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
+ %b = load <16 x float>, <16 x float>* %ptr_b
+ %res = call <16 x float> @llvm.x86.avx512.mask.xor.ps.512(<16 x float> %a, <16 x float> %b, <16 x float> zeroinitializer, i16 -1)
+ ret <16 x float> %res
+}
+
+define <16 x float> @test_mask_xor_ps_rmk_512(<16 x float> %a, <16 x float>* %ptr_b, <16 x float> %passThru, i16 %mask) {
+; CHECK-LABEL: test_mask_xor_ps_rmk_512:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
+; CHECK-NEXT: vxorps (%rdi), %zmm0, %zmm1 {%k1} ## encoding: [0x62,0xf1,0x7c,0x49,0x57,0x0f]
+; CHECK-NEXT: vmovaps %zmm1, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
+ %b = load <16 x float>, <16 x float>* %ptr_b
+ %res = call <16 x float> @llvm.x86.avx512.mask.xor.ps.512(<16 x float> %a, <16 x float> %b, <16 x float> %passThru, i16 %mask)
+ ret <16 x float> %res
+}
+
+define <16 x float> @test_mask_xor_ps_rmkz_512(<16 x float> %a, <16 x float>* %ptr_b, i16 %mask) {
+; CHECK-LABEL: test_mask_xor_ps_rmkz_512:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
+; CHECK-NEXT: vxorps (%rdi), %zmm0, %zmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0xc9,0x57,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
+ %b = load <16 x float>, <16 x float>* %ptr_b
+ %res = call <16 x float> @llvm.x86.avx512.mask.xor.ps.512(<16 x float> %a, <16 x float> %b, <16 x float> zeroinitializer, i16 %mask)
+ ret <16 x float> %res
+}
+
+define <16 x float> @test_mask_xor_ps_rmb_512(<16 x float> %a, float* %ptr_b) {
+; CHECK-LABEL: test_mask_xor_ps_rmb_512:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vxorps (%rdi){1to16}, %zmm0, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x58,0x57,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
+ %q = load float, float* %ptr_b
+ %vecinit.i = insertelement <16 x float> undef, float %q, i32 0
+ %b = shufflevector <16 x float> %vecinit.i, <16 x float> undef, <16 x i32> zeroinitializer
+ %res = call <16 x float> @llvm.x86.avx512.mask.xor.ps.512(<16 x float> %a, <16 x float> %b, <16 x float> zeroinitializer, i16 -1)
+ ret <16 x float> %res
+}
+
+define <16 x float> @test_mask_xor_ps_rmbk_512(<16 x float> %a, float* %ptr_b, <16 x float> %passThru, i16 %mask) {
+; CHECK-LABEL: test_mask_xor_ps_rmbk_512:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
+; CHECK-NEXT: vxorps (%rdi){1to16}, %zmm0, %zmm1 {%k1} ## encoding: [0x62,0xf1,0x7c,0x59,0x57,0x0f]
+; CHECK-NEXT: vmovaps %zmm1, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
+ %q = load float, float* %ptr_b
+ %vecinit.i = insertelement <16 x float> undef, float %q, i32 0
+ %b = shufflevector <16 x float> %vecinit.i, <16 x float> undef, <16 x i32> zeroinitializer
+ %res = call <16 x float> @llvm.x86.avx512.mask.xor.ps.512(<16 x float> %a, <16 x float> %b, <16 x float> %passThru, i16 %mask)
+ ret <16 x float> %res
+}
+
+define <16 x float> @test_mask_xor_ps_rmbkz_512(<16 x float> %a, float* %ptr_b, i16 %mask) {
+; CHECK-LABEL: test_mask_xor_ps_rmbkz_512:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
+; CHECK-NEXT: vxorps (%rdi){1to16}, %zmm0, %zmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0xd9,0x57,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
+ %q = load float, float* %ptr_b
+ %vecinit.i = insertelement <16 x float> undef, float %q, i32 0
+ %b = shufflevector <16 x float> %vecinit.i, <16 x float> undef, <16 x i32> zeroinitializer
+ %res = call <16 x float> @llvm.x86.avx512.mask.xor.ps.512(<16 x float> %a, <16 x float> %b, <16 x float> zeroinitializer, i16 %mask)
+ ret <16 x float> %res
+}
+
+declare <16 x float> @llvm.x86.avx512.mask.xor.ps.512(<16 x float>, <16 x float>, <16 x float>, i16)
+
diff --git a/llvm/test/CodeGen/X86/avx512dqvl-intrinsics.ll b/llvm/test/CodeGen/X86/avx512dqvl-intrinsics.ll
index cc9d445d056..f8cac8a4c8b 100644
--- a/llvm/test/CodeGen/X86/avx512dqvl-intrinsics.ll
+++ b/llvm/test/CodeGen/X86/avx512dqvl-intrinsics.ll
@@ -312,1254 +312,6 @@ define <2 x i64> @test_mask_mullo_epi64_rmbkz_128(<2 x i64> %a, i64* %ptr_b, i8
declare <2 x i64> @llvm.x86.avx512.mask.pmull.q.128(<2 x i64>, <2 x i64>, <2 x i64>, i8)
-define <4 x float> @test_mask_andnot_ps_rr_128(<4 x float> %a, <4 x float> %b) {
-; CHECK-LABEL: test_mask_andnot_ps_rr_128:
-; CHECK: ## BB#0:
-; CHECK-NEXT: vandnps %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7c,0x08,0x55,0xc1]
-; CHECK-NEXT: retq ## encoding: [0xc3]
- %res = call <4 x float> @llvm.x86.avx512.mask.andn.ps.128(<4 x float> %a, <4 x float> %b, <4 x float> zeroinitializer, i8 -1)
- ret <4 x float> %res
-}
-
-define <4 x float> @test_mask_andnot_ps_rrk_128(<4 x float> %a, <4 x float> %b, <4 x float> %passThru, i8 %mask) {
-; CHECK-LABEL: test_mask_andnot_ps_rrk_128:
-; CHECK: ## BB#0:
-; CHECK-NEXT: kmovb %edi, %k1 ## encoding: [0xc5,0xf9,0x92,0xcf]
-; CHECK-NEXT: vandnps %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf1,0x7c,0x09,0x55,0xd1]
-; CHECK-NEXT: vmovaps %xmm2, %xmm0 ## encoding: [0x62,0xf1,0x7c,0x08,0x28,0xc2]
-; CHECK-NEXT: retq ## encoding: [0xc3]
- %res = call <4 x float> @llvm.x86.avx512.mask.andn.ps.128(<4 x float> %a, <4 x float> %b, <4 x float> %passThru, i8 %mask)
- ret <4 x float> %res
-}
-
-define <4 x float> @test_mask_andnot_ps_rrkz_128(<4 x float> %a, <4 x float> %b, i8 %mask) {
-; CHECK-LABEL: test_mask_andnot_ps_rrkz_128:
-; CHECK: ## BB#0:
-; CHECK-NEXT: kmovb %edi, %k1 ## encoding: [0xc5,0xf9,0x92,0xcf]
-; CHECK-NEXT: vandnps %xmm1, %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0x89,0x55,0xc1]
-; CHECK-NEXT: retq ## encoding: [0xc3]
- %res = call <4 x float> @llvm.x86.avx512.mask.andn.ps.128(<4 x float> %a, <4 x float> %b, <4 x float> zeroinitializer, i8 %mask)
- ret <4 x float> %res
-}
-
-define <4 x float> @test_mask_andnot_ps_rm_128(<4 x float> %a, <4 x float>* %ptr_b) {
-; CHECK-LABEL: test_mask_andnot_ps_rm_128:
-; CHECK: ## BB#0:
-; CHECK-NEXT: vandnps (%rdi), %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7c,0x08,0x55,0x07]
-; CHECK-NEXT: retq ## encoding: [0xc3]
- %b = load <4 x float>, <4 x float>* %ptr_b
- %res = call <4 x float> @llvm.x86.avx512.mask.andn.ps.128(<4 x float> %a, <4 x float> %b, <4 x float> zeroinitializer, i8 -1)
- ret <4 x float> %res
-}
-
-define <4 x float> @test_mask_andnot_ps_rmk_128(<4 x float> %a, <4 x float>* %ptr_b, <4 x float> %passThru, i8 %mask) {
-; CHECK-LABEL: test_mask_andnot_ps_rmk_128:
-; CHECK: ## BB#0:
-; CHECK-NEXT: kmovb %esi, %k1 ## encoding: [0xc5,0xf9,0x92,0xce]
-; CHECK-NEXT: vandnps (%rdi), %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf1,0x7c,0x09,0x55,0x0f]
-; CHECK-NEXT: vmovaps %xmm1, %xmm0 ## encoding: [0x62,0xf1,0x7c,0x08,0x28,0xc1]
-; CHECK-NEXT: retq ## encoding: [0xc3]
- %b = load <4 x float>, <4 x float>* %ptr_b
- %res = call <4 x float> @llvm.x86.avx512.mask.andn.ps.128(<4 x float> %a, <4 x float> %b, <4 x float> %passThru, i8 %mask)
- ret <4 x float> %res
-}
-
-define <4 x float> @test_mask_andnot_ps_rmkz_128(<4 x float> %a, <4 x float>* %ptr_b, i8 %mask) {
-; CHECK-LABEL: test_mask_andnot_ps_rmkz_128:
-; CHECK: ## BB#0:
-; CHECK-NEXT: kmovb %esi, %k1 ## encoding: [0xc5,0xf9,0x92,0xce]
-; CHECK-NEXT: vandnps (%rdi), %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0x89,0x55,0x07]
-; CHECK-NEXT: retq ## encoding: [0xc3]
- %b = load <4 x float>, <4 x float>* %ptr_b
- %res = call <4 x float> @llvm.x86.avx512.mask.andn.ps.128(<4 x float> %a, <4 x float> %b, <4 x float> zeroinitializer, i8 %mask)
- ret <4 x float> %res
-}
-
-define <4 x float> @test_mask_andnot_ps_rmb_128(<4 x float> %a, float* %ptr_b) {
-; CHECK-LABEL: test_mask_andnot_ps_rmb_128:
-; CHECK: ## BB#0:
-; CHECK-NEXT: vandnps (%rdi){1to4}, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7c,0x18,0x55,0x07]
-; CHECK-NEXT: retq ## encoding: [0xc3]
- %q = load float, float* %ptr_b
- %vecinit.i = insertelement <4 x float> undef, float %q, i32 0
- %b = shufflevector <4 x float> %vecinit.i, <4 x float> undef, <4 x i32> zeroinitializer
- %res = call <4 x float> @llvm.x86.avx512.mask.andn.ps.128(<4 x float> %a, <4 x float> %b, <4 x float> zeroinitializer, i8 -1)
- ret <4 x float> %res
-}
-
-define <4 x float> @test_mask_andnot_ps_rmbk_128(<4 x float> %a, float* %ptr_b, <4 x float> %passThru, i8 %mask) {
-; CHECK-LABEL: test_mask_andnot_ps_rmbk_128:
-; CHECK: ## BB#0:
-; CHECK-NEXT: kmovb %esi, %k1 ## encoding: [0xc5,0xf9,0x92,0xce]
-; CHECK-NEXT: vandnps (%rdi){1to4}, %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf1,0x7c,0x19,0x55,0x0f]
-; CHECK-NEXT: vmovaps %xmm1, %xmm0 ## encoding: [0x62,0xf1,0x7c,0x08,0x28,0xc1]
-; CHECK-NEXT: retq ## encoding: [0xc3]
- %q = load float, float* %ptr_b
- %vecinit.i = insertelement <4 x float> undef, float %q, i32 0
- %b = shufflevector <4 x float> %vecinit.i, <4 x float> undef, <4 x i32> zeroinitializer
- %res = call <4 x float> @llvm.x86.avx512.mask.andn.ps.128(<4 x float> %a, <4 x float> %b, <4 x float> %passThru, i8 %mask)
- ret <4 x float> %res
-}
-
-define <4 x float> @test_mask_andnot_ps_rmbkz_128(<4 x float> %a, float* %ptr_b, i8 %mask) {
-; CHECK-LABEL: test_mask_andnot_ps_rmbkz_128:
-; CHECK: ## BB#0:
-; CHECK-NEXT: kmovb %esi, %k1 ## encoding: [0xc5,0xf9,0x92,0xce]
-; CHECK-NEXT: vandnps (%rdi){1to4}, %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0x99,0x55,0x07]
-; CHECK-NEXT: retq ## encoding: [0xc3]
- %q = load float, float* %ptr_b
- %vecinit.i = insertelement <4 x float> undef, float %q, i32 0
- %b = shufflevector <4 x float> %vecinit.i, <4 x float> undef, <4 x i32> zeroinitializer
- %res = call <4 x float> @llvm.x86.avx512.mask.andn.ps.128(<4 x float> %a, <4 x float> %b, <4 x float> zeroinitializer, i8 %mask)
- ret <4 x float> %res
-}
-
-declare <4 x float> @llvm.x86.avx512.mask.andn.ps.128(<4 x float>, <4 x float>, <4 x float>, i8)
-
-define <8 x float> @test_mask_andnot_ps_rr_256(<8 x float> %a, <8 x float> %b) {
-; CHECK-LABEL: test_mask_andnot_ps_rr_256:
-; CHECK: ## BB#0:
-; CHECK-NEXT: vandnps %ymm1, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7c,0x28,0x55,0xc1]
-; CHECK-NEXT: retq ## encoding: [0xc3]
- %res = call <8 x float> @llvm.x86.avx512.mask.andn.ps.256(<8 x float> %a, <8 x float> %b, <8 x float> zeroinitializer, i8 -1)
- ret <8 x float> %res
-}
-
-define <8 x float> @test_mask_andnot_ps_rrk_256(<8 x float> %a, <8 x float> %b, <8 x float> %passThru, i8 %mask) {
-; CHECK-LABEL: test_mask_andnot_ps_rrk_256:
-; CHECK: ## BB#0:
-; CHECK-NEXT: kmovb %edi, %k1 ## encoding: [0xc5,0xf9,0x92,0xcf]
-; CHECK-NEXT: vandnps %ymm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf1,0x7c,0x29,0x55,0xd1]
-; CHECK-NEXT: vmovaps %ymm2, %ymm0 ## encoding: [0x62,0xf1,0x7c,0x28,0x28,0xc2]
-; CHECK-NEXT: retq ## encoding: [0xc3]
- %res = call <8 x float> @llvm.x86.avx512.mask.andn.ps.256(<8 x float> %a, <8 x float> %b, <8 x float> %passThru, i8 %mask)
- ret <8 x float> %res
-}
-
-define <8 x float> @test_mask_andnot_ps_rrkz_256(<8 x float> %a, <8 x float> %b, i8 %mask) {
-; CHECK-LABEL: test_mask_andnot_ps_rrkz_256:
-; CHECK: ## BB#0:
-; CHECK-NEXT: kmovb %edi, %k1 ## encoding: [0xc5,0xf9,0x92,0xcf]
-; CHECK-NEXT: vandnps %ymm1, %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0xa9,0x55,0xc1]
-; CHECK-NEXT: retq ## encoding: [0xc3]
- %res = call <8 x float> @llvm.x86.avx512.mask.andn.ps.256(<8 x float> %a, <8 x float> %b, <8 x float> zeroinitializer, i8 %mask)
- ret <8 x float> %res
-}
-
-define <8 x float> @test_mask_andnot_ps_rm_256(<8 x float> %a, <8 x float>* %ptr_b) {
-; CHECK-LABEL: test_mask_andnot_ps_rm_256:
-; CHECK: ## BB#0:
-; CHECK-NEXT: vandnps (%rdi), %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7c,0x28,0x55,0x07]
-; CHECK-NEXT: retq ## encoding: [0xc3]
- %b = load <8 x float>, <8 x float>* %ptr_b
- %res = call <8 x float> @llvm.x86.avx512.mask.andn.ps.256(<8 x float> %a, <8 x float> %b, <8 x float> zeroinitializer, i8 -1)
- ret <8 x float> %res
-}
-
-define <8 x float> @test_mask_andnot_ps_rmk_256(<8 x float> %a, <8 x float>* %ptr_b, <8 x float> %passThru, i8 %mask) {
-; CHECK-LABEL: test_mask_andnot_ps_rmk_256:
-; CHECK: ## BB#0:
-; CHECK-NEXT: kmovb %esi, %k1 ## encoding: [0xc5,0xf9,0x92,0xce]
-; CHECK-NEXT: vandnps (%rdi), %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf1,0x7c,0x29,0x55,0x0f]
-; CHECK-NEXT: vmovaps %ymm1, %ymm0 ## encoding: [0x62,0xf1,0x7c,0x28,0x28,0xc1]
-; CHECK-NEXT: retq ## encoding: [0xc3]
- %b = load <8 x float>, <8 x float>* %ptr_b
- %res = call <8 x float> @llvm.x86.avx512.mask.andn.ps.256(<8 x float> %a, <8 x float> %b, <8 x float> %passThru, i8 %mask)
- ret <8 x float> %res
-}
-
-define <8 x float> @test_mask_andnot_ps_rmkz_256(<8 x float> %a, <8 x float>* %ptr_b, i8 %mask) {
-; CHECK-LABEL: test_mask_andnot_ps_rmkz_256:
-; CHECK: ## BB#0:
-; CHECK-NEXT: kmovb %esi, %k1 ## encoding: [0xc5,0xf9,0x92,0xce]
-; CHECK-NEXT: vandnps (%rdi), %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0xa9,0x55,0x07]
-; CHECK-NEXT: retq ## encoding: [0xc3]
- %b = load <8 x float>, <8 x float>* %ptr_b
- %res = call <8 x float> @llvm.x86.avx512.mask.andn.ps.256(<8 x float> %a, <8 x float> %b, <8 x float> zeroinitializer, i8 %mask)
- ret <8 x float> %res
-}
-
-define <8 x float> @test_mask_andnot_ps_rmb_256(<8 x float> %a, float* %ptr_b) {
-; CHECK-LABEL: test_mask_andnot_ps_rmb_256:
-; CHECK: ## BB#0:
-; CHECK-NEXT: vandnps (%rdi){1to8}, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7c,0x38,0x55,0x07]
-; CHECK-NEXT: retq ## encoding: [0xc3]
- %q = load float, float* %ptr_b
- %vecinit.i = insertelement <8 x float> undef, float %q, i32 0
- %b = shufflevector <8 x float> %vecinit.i, <8 x float> undef, <8 x i32> zeroinitializer
- %res = call <8 x float> @llvm.x86.avx512.mask.andn.ps.256(<8 x float> %a, <8 x float> %b, <8 x float> zeroinitializer, i8 -1)
- ret <8 x float> %res
-}
-
-define <8 x float> @test_mask_andnot_ps_rmbk_256(<8 x float> %a, float* %ptr_b, <8 x float> %passThru, i8 %mask) {
-; CHECK-LABEL: test_mask_andnot_ps_rmbk_256:
-; CHECK: ## BB#0:
-; CHECK-NEXT: kmovb %esi, %k1 ## encoding: [0xc5,0xf9,0x92,0xce]
-; CHECK-NEXT: vandnps (%rdi){1to8}, %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf1,0x7c,0x39,0x55,0x0f]
-; CHECK-NEXT: vmovaps %ymm1, %ymm0 ## encoding: [0x62,0xf1,0x7c,0x28,0x28,0xc1]
-; CHECK-NEXT: retq ## encoding: [0xc3]
- %q = load float, float* %ptr_b
- %vecinit.i = insertelement <8 x float> undef, float %q, i32 0
- %b = shufflevector <8 x float> %vecinit.i, <8 x float> undef, <8 x i32> zeroinitializer
- %res = call <8 x float> @llvm.x86.avx512.mask.andn.ps.256(<8 x float> %a, <8 x float> %b, <8 x float> %passThru, i8 %mask)
- ret <8 x float> %res
-}
-
-define <8 x float> @test_mask_andnot_ps_rmbkz_256(<8 x float> %a, float* %ptr_b, i8 %mask) {
-; CHECK-LABEL: test_mask_andnot_ps_rmbkz_256:
-; CHECK: ## BB#0:
-; CHECK-NEXT: kmovb %esi, %k1 ## encoding: [0xc5,0xf9,0x92,0xce]
-; CHECK-NEXT: vandnps (%rdi){1to8}, %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0xb9,0x55,0x07]
-; CHECK-NEXT: retq ## encoding: [0xc3]
- %q = load float, float* %ptr_b
- %vecinit.i = insertelement <8 x float> undef, float %q, i32 0
- %b = shufflevector <8 x float> %vecinit.i, <8 x float> undef, <8 x i32> zeroinitializer
- %res = call <8 x float> @llvm.x86.avx512.mask.andn.ps.256(<8 x float> %a, <8 x float> %b, <8 x float> zeroinitializer, i8 %mask)
- ret <8 x float> %res
-}
-
-declare <8 x float> @llvm.x86.avx512.mask.andn.ps.256(<8 x float>, <8 x float>, <8 x float>, i8)
-
-define <16 x float> @test_mask_andnot_ps_rr_512(<16 x float> %a, <16 x float> %b) {
-; CHECK-LABEL: test_mask_andnot_ps_rr_512:
-; CHECK: ## BB#0:
-; CHECK-NEXT: vandnps %zmm1, %zmm0, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x55,0xc1]
-; CHECK-NEXT: retq ## encoding: [0xc3]
- %res = call <16 x float> @llvm.x86.avx512.mask.andn.ps.512(<16 x float> %a, <16 x float> %b, <16 x float> zeroinitializer, i16 -1)
- ret <16 x float> %res
-}
-
-define <16 x float> @test_mask_andnot_ps_rrk_512(<16 x float> %a, <16 x float> %b, <16 x float> %passThru, i16 %mask) {
-; CHECK-LABEL: test_mask_andnot_ps_rrk_512:
-; CHECK: ## BB#0:
-; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vandnps %zmm1, %zmm0, %zmm2 {%k1} ## encoding: [0x62,0xf1,0x7c,0x49,0x55,0xd1]
-; CHECK-NEXT: vmovaps %zmm2, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc2]
-; CHECK-NEXT: retq ## encoding: [0xc3]
- %res = call <16 x float> @llvm.x86.avx512.mask.andn.ps.512(<16 x float> %a, <16 x float> %b, <16 x float> %passThru, i16 %mask)
- ret <16 x float> %res
-}
-
-define <16 x float> @test_mask_andnot_ps_rrkz_512(<16 x float> %a, <16 x float> %b, i16 %mask) {
-; CHECK-LABEL: test_mask_andnot_ps_rrkz_512:
-; CHECK: ## BB#0:
-; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vandnps %zmm1, %zmm0, %zmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0xc9,0x55,0xc1]
-; CHECK-NEXT: retq ## encoding: [0xc3]
- %res = call <16 x float> @llvm.x86.avx512.mask.andn.ps.512(<16 x float> %a, <16 x float> %b, <16 x float> zeroinitializer, i16 %mask)
- ret <16 x float> %res
-}
-
-define <16 x float> @test_mask_andnot_ps_rm_512(<16 x float> %a, <16 x float>* %ptr_b) {
-; CHECK-LABEL: test_mask_andnot_ps_rm_512:
-; CHECK: ## BB#0:
-; CHECK-NEXT: vandnps (%rdi), %zmm0, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x55,0x07]
-; CHECK-NEXT: retq ## encoding: [0xc3]
- %b = load <16 x float>, <16 x float>* %ptr_b
- %res = call <16 x float> @llvm.x86.avx512.mask.andn.ps.512(<16 x float> %a, <16 x float> %b, <16 x float> zeroinitializer, i16 -1)
- ret <16 x float> %res
-}
-
-define <16 x float> @test_mask_andnot_ps_rmk_512(<16 x float> %a, <16 x float>* %ptr_b, <16 x float> %passThru, i16 %mask) {
-; CHECK-LABEL: test_mask_andnot_ps_rmk_512:
-; CHECK: ## BB#0:
-; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
-; CHECK-NEXT: vandnps (%rdi), %zmm0, %zmm1 {%k1} ## encoding: [0x62,0xf1,0x7c,0x49,0x55,0x0f]
-; CHECK-NEXT: vmovaps %zmm1, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc1]
-; CHECK-NEXT: retq ## encoding: [0xc3]
- %b = load <16 x float>, <16 x float>* %ptr_b
- %res = call <16 x float> @llvm.x86.avx512.mask.andn.ps.512(<16 x float> %a, <16 x float> %b, <16 x float> %passThru, i16 %mask)
- ret <16 x float> %res
-}
-
-define <16 x float> @test_mask_andnot_ps_rmkz_512(<16 x float> %a, <16 x float>* %ptr_b, i16 %mask) {
-; CHECK-LABEL: test_mask_andnot_ps_rmkz_512:
-; CHECK: ## BB#0:
-; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
-; CHECK-NEXT: vandnps (%rdi), %zmm0, %zmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0xc9,0x55,0x07]
-; CHECK-NEXT: retq ## encoding: [0xc3]
- %b = load <16 x float>, <16 x float>* %ptr_b
- %res = call <16 x float> @llvm.x86.avx512.mask.andn.ps.512(<16 x float> %a, <16 x float> %b, <16 x float> zeroinitializer, i16 %mask)
- ret <16 x float> %res
-}
-
-define <16 x float> @test_mask_andnot_ps_rmb_512(<16 x float> %a, float* %ptr_b) {
-; CHECK-LABEL: test_mask_andnot_ps_rmb_512:
-; CHECK: ## BB#0:
-; CHECK-NEXT: vandnps (%rdi){1to16}, %zmm0, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x58,0x55,0x07]
-; CHECK-NEXT: retq ## encoding: [0xc3]
- %q = load float, float* %ptr_b
- %vecinit.i = insertelement <16 x float> undef, float %q, i32 0
- %b = shufflevector <16 x float> %vecinit.i, <16 x float> undef, <16 x i32> zeroinitializer
- %res = call <16 x float> @llvm.x86.avx512.mask.andn.ps.512(<16 x float> %a, <16 x float> %b, <16 x float> zeroinitializer, i16 -1)
- ret <16 x float> %res
-}
-
-define <16 x float> @test_mask_andnot_ps_rmbk_512(<16 x float> %a, float* %ptr_b, <16 x float> %passThru, i16 %mask) {
-; CHECK-LABEL: test_mask_andnot_ps_rmbk_512:
-; CHECK: ## BB#0:
-; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
-; CHECK-NEXT: vandnps (%rdi){1to16}, %zmm0, %zmm1 {%k1} ## encoding: [0x62,0xf1,0x7c,0x59,0x55,0x0f]
-; CHECK-NEXT: vmovaps %zmm1, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc1]
-; CHECK-NEXT: retq ## encoding: [0xc3]
- %q = load float, float* %ptr_b
- %vecinit.i = insertelement <16 x float> undef, float %q, i32 0
- %b = shufflevector <16 x float> %vecinit.i, <16 x float> undef, <16 x i32> zeroinitializer
- %res = call <16 x float> @llvm.x86.avx512.mask.andn.ps.512(<16 x float> %a, <16 x float> %b, <16 x float> %passThru, i16 %mask)
- ret <16 x float> %res
-}
-
-define <16 x float> @test_mask_andnot_ps_rmbkz_512(<16 x float> %a, float* %ptr_b, i16 %mask) {
-; CHECK-LABEL: test_mask_andnot_ps_rmbkz_512:
-; CHECK: ## BB#0:
-; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
-; CHECK-NEXT: vandnps (%rdi){1to16}, %zmm0, %zmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0xd9,0x55,0x07]
-; CHECK-NEXT: retq ## encoding: [0xc3]
- %q = load float, float* %ptr_b
- %vecinit.i = insertelement <16 x float> undef, float %q, i32 0
- %b = shufflevector <16 x float> %vecinit.i, <16 x float> undef, <16 x i32> zeroinitializer
- %res = call <16 x float> @llvm.x86.avx512.mask.andn.ps.512(<16 x float> %a, <16 x float> %b, <16 x float> zeroinitializer, i16 %mask)
- ret <16 x float> %res
-}
-
-declare <16 x float> @llvm.x86.avx512.mask.andn.ps.512(<16 x float>, <16 x float>, <16 x float>, i16)
-
-define <4 x float> @test_mask_and_ps_rr_128(<4 x float> %a, <4 x float> %b) {
-; CHECK-LABEL: test_mask_and_ps_rr_128:
-; CHECK: ## BB#0:
-; CHECK-NEXT: vandps %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7c,0x08,0x54,0xc1]
-; CHECK-NEXT: retq ## encoding: [0xc3]
- %res = call <4 x float> @llvm.x86.avx512.mask.and.ps.128(<4 x float> %a, <4 x float> %b, <4 x float> zeroinitializer, i8 -1)
- ret <4 x float> %res
-}
-
-define <4 x float> @test_mask_and_ps_rrk_128(<4 x float> %a, <4 x float> %b, <4 x float> %passThru, i8 %mask) {
-; CHECK-LABEL: test_mask_and_ps_rrk_128:
-; CHECK: ## BB#0:
-; CHECK-NEXT: kmovb %edi, %k1 ## encoding: [0xc5,0xf9,0x92,0xcf]
-; CHECK-NEXT: vandps %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf1,0x7c,0x09,0x54,0xd1]
-; CHECK-NEXT: vmovaps %xmm2, %xmm0 ## encoding: [0x62,0xf1,0x7c,0x08,0x28,0xc2]
-; CHECK-NEXT: retq ## encoding: [0xc3]
- %res = call <4 x float> @llvm.x86.avx512.mask.and.ps.128(<4 x float> %a, <4 x float> %b, <4 x float> %passThru, i8 %mask)
- ret <4 x float> %res
-}
-
-define <4 x float> @test_mask_and_ps_rrkz_128(<4 x float> %a, <4 x float> %b, i8 %mask) {
-; CHECK-LABEL: test_mask_and_ps_rrkz_128:
-; CHECK: ## BB#0:
-; CHECK-NEXT: kmovb %edi, %k1 ## encoding: [0xc5,0xf9,0x92,0xcf]
-; CHECK-NEXT: vandps %xmm1, %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0x89,0x54,0xc1]
-; CHECK-NEXT: retq ## encoding: [0xc3]
- %res = call <4 x float> @llvm.x86.avx512.mask.and.ps.128(<4 x float> %a, <4 x float> %b, <4 x float> zeroinitializer, i8 %mask)
- ret <4 x float> %res
-}
-
-define <4 x float> @test_mask_and_ps_rm_128(<4 x float> %a, <4 x float>* %ptr_b) {
-; CHECK-LABEL: test_mask_and_ps_rm_128:
-; CHECK: ## BB#0:
-; CHECK-NEXT: vandps (%rdi), %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7c,0x08,0x54,0x07]
-; CHECK-NEXT: retq ## encoding: [0xc3]
- %b = load <4 x float>, <4 x float>* %ptr_b
- %res = call <4 x float> @llvm.x86.avx512.mask.and.ps.128(<4 x float> %a, <4 x float> %b, <4 x float> zeroinitializer, i8 -1)
- ret <4 x float> %res
-}
-
-define <4 x float> @test_mask_and_ps_rmk_128(<4 x float> %a, <4 x float>* %ptr_b, <4 x float> %passThru, i8 %mask) {
-; CHECK-LABEL: test_mask_and_ps_rmk_128:
-; CHECK: ## BB#0:
-; CHECK-NEXT: kmovb %esi, %k1 ## encoding: [0xc5,0xf9,0x92,0xce]
-; CHECK-NEXT: vandps (%rdi), %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf1,0x7c,0x09,0x54,0x0f]
-; CHECK-NEXT: vmovaps %xmm1, %xmm0 ## encoding: [0x62,0xf1,0x7c,0x08,0x28,0xc1]
-; CHECK-NEXT: retq ## encoding: [0xc3]
- %b = load <4 x float>, <4 x float>* %ptr_b
- %res = call <4 x float> @llvm.x86.avx512.mask.and.ps.128(<4 x float> %a, <4 x float> %b, <4 x float> %passThru, i8 %mask)
- ret <4 x float> %res
-}
-
-define <4 x float> @test_mask_and_ps_rmkz_128(<4 x float> %a, <4 x float>* %ptr_b, i8 %mask) {
-; CHECK-LABEL: test_mask_and_ps_rmkz_128:
-; CHECK: ## BB#0:
-; CHECK-NEXT: kmovb %esi, %k1 ## encoding: [0xc5,0xf9,0x92,0xce]
-; CHECK-NEXT: vandps (%rdi), %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0x89,0x54,0x07]
-; CHECK-NEXT: retq ## encoding: [0xc3]
- %b = load <4 x float>, <4 x float>* %ptr_b
- %res = call <4 x float> @llvm.x86.avx512.mask.and.ps.128(<4 x float> %a, <4 x float> %b, <4 x float> zeroinitializer, i8 %mask)
- ret <4 x float> %res
-}
-
-define <4 x float> @test_mask_and_ps_rmb_128(<4 x float> %a, float* %ptr_b) {
-; CHECK-LABEL: test_mask_and_ps_rmb_128:
-; CHECK: ## BB#0:
-; CHECK-NEXT: vandps (%rdi){1to4}, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7c,0x18,0x54,0x07]
-; CHECK-NEXT: retq ## encoding: [0xc3]
- %q = load float, float* %ptr_b
- %vecinit.i = insertelement <4 x float> undef, float %q, i32 0
- %b = shufflevector <4 x float> %vecinit.i, <4 x float> undef, <4 x i32> zeroinitializer
- %res = call <4 x float> @llvm.x86.avx512.mask.and.ps.128(<4 x float> %a, <4 x float> %b, <4 x float> zeroinitializer, i8 -1)
- ret <4 x float> %res
-}
-
-define <4 x float> @test_mask_and_ps_rmbk_128(<4 x float> %a, float* %ptr_b, <4 x float> %passThru, i8 %mask) {
-; CHECK-LABEL: test_mask_and_ps_rmbk_128:
-; CHECK: ## BB#0:
-; CHECK-NEXT: kmovb %esi, %k1 ## encoding: [0xc5,0xf9,0x92,0xce]
-; CHECK-NEXT: vandps (%rdi){1to4}, %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf1,0x7c,0x19,0x54,0x0f]
-; CHECK-NEXT: vmovaps %xmm1, %xmm0 ## encoding: [0x62,0xf1,0x7c,0x08,0x28,0xc1]
-; CHECK-NEXT: retq ## encoding: [0xc3]
- %q = load float, float* %ptr_b
- %vecinit.i = insertelement <4 x float> undef, float %q, i32 0
- %b = shufflevector <4 x float> %vecinit.i, <4 x float> undef, <4 x i32> zeroinitializer
- %res = call <4 x float> @llvm.x86.avx512.mask.and.ps.128(<4 x float> %a, <4 x float> %b, <4 x float> %passThru, i8 %mask)
- ret <4 x float> %res
-}
-
-define <4 x float> @test_mask_and_ps_rmbkz_128(<4 x float> %a, float* %ptr_b, i8 %mask) {
-; CHECK-LABEL: test_mask_and_ps_rmbkz_128:
-; CHECK: ## BB#0:
-; CHECK-NEXT: kmovb %esi, %k1 ## encoding: [0xc5,0xf9,0x92,0xce]
-; CHECK-NEXT: vandps (%rdi){1to4}, %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0x99,0x54,0x07]
-; CHECK-NEXT: retq ## encoding: [0xc3]
- %q = load float, float* %ptr_b
- %vecinit.i = insertelement <4 x float> undef, float %q, i32 0
- %b = shufflevector <4 x float> %vecinit.i, <4 x float> undef, <4 x i32> zeroinitializer
- %res = call <4 x float> @llvm.x86.avx512.mask.and.ps.128(<4 x float> %a, <4 x float> %b, <4 x float> zeroinitializer, i8 %mask)
- ret <4 x float> %res
-}
-
-declare <4 x float> @llvm.x86.avx512.mask.and.ps.128(<4 x float>, <4 x float>, <4 x float>, i8)
-
-define <8 x float> @test_mask_and_ps_rr_256(<8 x float> %a, <8 x float> %b) {
-; CHECK-LABEL: test_mask_and_ps_rr_256:
-; CHECK: ## BB#0:
-; CHECK-NEXT: vandps %ymm1, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7c,0x28,0x54,0xc1]
-; CHECK-NEXT: retq ## encoding: [0xc3]
- %res = call <8 x float> @llvm.x86.avx512.mask.and.ps.256(<8 x float> %a, <8 x float> %b, <8 x float> zeroinitializer, i8 -1)
- ret <8 x float> %res
-}
-
-define <8 x float> @test_mask_and_ps_rrk_256(<8 x float> %a, <8 x float> %b, <8 x float> %passThru, i8 %mask) {
-; CHECK-LABEL: test_mask_and_ps_rrk_256:
-; CHECK: ## BB#0:
-; CHECK-NEXT: kmovb %edi, %k1 ## encoding: [0xc5,0xf9,0x92,0xcf]
-; CHECK-NEXT: vandps %ymm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf1,0x7c,0x29,0x54,0xd1]
-; CHECK-NEXT: vmovaps %ymm2, %ymm0 ## encoding: [0x62,0xf1,0x7c,0x28,0x28,0xc2]
-; CHECK-NEXT: retq ## encoding: [0xc3]
- %res = call <8 x float> @llvm.x86.avx512.mask.and.ps.256(<8 x float> %a, <8 x float> %b, <8 x float> %passThru, i8 %mask)
- ret <8 x float> %res
-}
-
-define <8 x float> @test_mask_and_ps_rrkz_256(<8 x float> %a, <8 x float> %b, i8 %mask) {
-; CHECK-LABEL: test_mask_and_ps_rrkz_256:
-; CHECK: ## BB#0:
-; CHECK-NEXT: kmovb %edi, %k1 ## encoding: [0xc5,0xf9,0x92,0xcf]
-; CHECK-NEXT: vandps %ymm1, %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0xa9,0x54,0xc1]
-; CHECK-NEXT: retq ## encoding: [0xc3]
- %res = call <8 x float> @llvm.x86.avx512.mask.and.ps.256(<8 x float> %a, <8 x float> %b, <8 x float> zeroinitializer, i8 %mask)
- ret <8 x float> %res
-}
-
-define <8 x float> @test_mask_and_ps_rm_256(<8 x float> %a, <8 x float>* %ptr_b) {
-; CHECK-LABEL: test_mask_and_ps_rm_256:
-; CHECK: ## BB#0:
-; CHECK-NEXT: vandps (%rdi), %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7c,0x28,0x54,0x07]
-; CHECK-NEXT: retq ## encoding: [0xc3]
- %b = load <8 x float>, <8 x float>* %ptr_b
- %res = call <8 x float> @llvm.x86.avx512.mask.and.ps.256(<8 x float> %a, <8 x float> %b, <8 x float> zeroinitializer, i8 -1)
- ret <8 x float> %res
-}
-
-define <8 x float> @test_mask_and_ps_rmk_256(<8 x float> %a, <8 x float>* %ptr_b, <8 x float> %passThru, i8 %mask) {
-; CHECK-LABEL: test_mask_and_ps_rmk_256:
-; CHECK: ## BB#0:
-; CHECK-NEXT: kmovb %esi, %k1 ## encoding: [0xc5,0xf9,0x92,0xce]
-; CHECK-NEXT: vandps (%rdi), %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf1,0x7c,0x29,0x54,0x0f]
-; CHECK-NEXT: vmovaps %ymm1, %ymm0 ## encoding: [0x62,0xf1,0x7c,0x28,0x28,0xc1]
-; CHECK-NEXT: retq ## encoding: [0xc3]
- %b = load <8 x float>, <8 x float>* %ptr_b
- %res = call <8 x float> @llvm.x86.avx512.mask.and.ps.256(<8 x float> %a, <8 x float> %b, <8 x float> %passThru, i8 %mask)
- ret <8 x float> %res
-}
-
-define <8 x float> @test_mask_and_ps_rmkz_256(<8 x float> %a, <8 x float>* %ptr_b, i8 %mask) {
-; CHECK-LABEL: test_mask_and_ps_rmkz_256:
-; CHECK: ## BB#0:
-; CHECK-NEXT: kmovb %esi, %k1 ## encoding: [0xc5,0xf9,0x92,0xce]
-; CHECK-NEXT: vandps (%rdi), %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0xa9,0x54,0x07]
-; CHECK-NEXT: retq ## encoding: [0xc3]
- %b = load <8 x float>, <8 x float>* %ptr_b
- %res = call <8 x float> @llvm.x86.avx512.mask.and.ps.256(<8 x float> %a, <8 x float> %b, <8 x float> zeroinitializer, i8 %mask)
- ret <8 x float> %res
-}
-
-define <8 x float> @test_mask_and_ps_rmb_256(<8 x float> %a, float* %ptr_b) {
-; CHECK-LABEL: test_mask_and_ps_rmb_256:
-; CHECK: ## BB#0:
-; CHECK-NEXT: vandps (%rdi){1to8}, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7c,0x38,0x54,0x07]
-; CHECK-NEXT: retq ## encoding: [0xc3]
- %q = load float, float* %ptr_b
- %vecinit.i = insertelement <8 x float> undef, float %q, i32 0
- %b = shufflevector <8 x float> %vecinit.i, <8 x float> undef, <8 x i32> zeroinitializer
- %res = call <8 x float> @llvm.x86.avx512.mask.and.ps.256(<8 x float> %a, <8 x float> %b, <8 x float> zeroinitializer, i8 -1)
- ret <8 x float> %res
-}
-
-define <8 x float> @test_mask_and_ps_rmbk_256(<8 x float> %a, float* %ptr_b, <8 x float> %passThru, i8 %mask) {
-; CHECK-LABEL: test_mask_and_ps_rmbk_256:
-; CHECK: ## BB#0:
-; CHECK-NEXT: kmovb %esi, %k1 ## encoding: [0xc5,0xf9,0x92,0xce]
-; CHECK-NEXT: vandps (%rdi){1to8}, %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf1,0x7c,0x39,0x54,0x0f]
-; CHECK-NEXT: vmovaps %ymm1, %ymm0 ## encoding: [0x62,0xf1,0x7c,0x28,0x28,0xc1]
-; CHECK-NEXT: retq ## encoding: [0xc3]
- %q = load float, float* %ptr_b
- %vecinit.i = insertelement <8 x float> undef, float %q, i32 0
- %b = shufflevector <8 x float> %vecinit.i, <8 x float> undef, <8 x i32> zeroinitializer
- %res = call <8 x float> @llvm.x86.avx512.mask.and.ps.256(<8 x float> %a, <8 x float> %b, <8 x float> %passThru, i8 %mask)
- ret <8 x float> %res
-}
-
-define <8 x float> @test_mask_and_ps_rmbkz_256(<8 x float> %a, float* %ptr_b, i8 %mask) {
-; CHECK-LABEL: test_mask_and_ps_rmbkz_256:
-; CHECK: ## BB#0:
-; CHECK-NEXT: kmovb %esi, %k1 ## encoding: [0xc5,0xf9,0x92,0xce]
-; CHECK-NEXT: vandps (%rdi){1to8}, %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0xb9,0x54,0x07]
-; CHECK-NEXT: retq ## encoding: [0xc3]
- %q = load float, float* %ptr_b
- %vecinit.i = insertelement <8 x float> undef, float %q, i32 0
- %b = shufflevector <8 x float> %vecinit.i, <8 x float> undef, <8 x i32> zeroinitializer
- %res = call <8 x float> @llvm.x86.avx512.mask.and.ps.256(<8 x float> %a, <8 x float> %b, <8 x float> zeroinitializer, i8 %mask)
- ret <8 x float> %res
-}
-
-declare <8 x float> @llvm.x86.avx512.mask.and.ps.256(<8 x float>, <8 x float>, <8 x float>, i8)
-
-define <16 x float> @test_mask_and_ps_rr_512(<16 x float> %a, <16 x float> %b) {
-; CHECK-LABEL: test_mask_and_ps_rr_512:
-; CHECK: ## BB#0:
-; CHECK-NEXT: vandps %zmm1, %zmm0, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x54,0xc1]
-; CHECK-NEXT: retq ## encoding: [0xc3]
- %res = call <16 x float> @llvm.x86.avx512.mask.and.ps.512(<16 x float> %a, <16 x float> %b, <16 x float> zeroinitializer, i16 -1)
- ret <16 x float> %res
-}
-
-define <16 x float> @test_mask_and_ps_rrk_512(<16 x float> %a, <16 x float> %b, <16 x float> %passThru, i16 %mask) {
-; CHECK-LABEL: test_mask_and_ps_rrk_512:
-; CHECK: ## BB#0:
-; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vandps %zmm1, %zmm0, %zmm2 {%k1} ## encoding: [0x62,0xf1,0x7c,0x49,0x54,0xd1]
-; CHECK-NEXT: vmovaps %zmm2, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc2]
-; CHECK-NEXT: retq ## encoding: [0xc3]
- %res = call <16 x float> @llvm.x86.avx512.mask.and.ps.512(<16 x float> %a, <16 x float> %b, <16 x float> %passThru, i16 %mask)
- ret <16 x float> %res
-}
-
-define <16 x float> @test_mask_and_ps_rrkz_512(<16 x float> %a, <16 x float> %b, i16 %mask) {
-; CHECK-LABEL: test_mask_and_ps_rrkz_512:
-; CHECK: ## BB#0:
-; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vandps %zmm1, %zmm0, %zmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0xc9,0x54,0xc1]
-; CHECK-NEXT: retq ## encoding: [0xc3]
- %res = call <16 x float> @llvm.x86.avx512.mask.and.ps.512(<16 x float> %a, <16 x float> %b, <16 x float> zeroinitializer, i16 %mask)
- ret <16 x float> %res
-}
-
-define <16 x float> @test_mask_and_ps_rm_512(<16 x float> %a, <16 x float>* %ptr_b) {
-; CHECK-LABEL: test_mask_and_ps_rm_512:
-; CHECK: ## BB#0:
-; CHECK-NEXT: vandps (%rdi), %zmm0, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x54,0x07]
-; CHECK-NEXT: retq ## encoding: [0xc3]
- %b = load <16 x float>, <16 x float>* %ptr_b
- %res = call <16 x float> @llvm.x86.avx512.mask.and.ps.512(<16 x float> %a, <16 x float> %b, <16 x float> zeroinitializer, i16 -1)
- ret <16 x float> %res
-}
-
-define <16 x float> @test_mask_and_ps_rmk_512(<16 x float> %a, <16 x float>* %ptr_b, <16 x float> %passThru, i16 %mask) {
-; CHECK-LABEL: test_mask_and_ps_rmk_512:
-; CHECK: ## BB#0:
-; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
-; CHECK-NEXT: vandps (%rdi), %zmm0, %zmm1 {%k1} ## encoding: [0x62,0xf1,0x7c,0x49,0x54,0x0f]
-; CHECK-NEXT: vmovaps %zmm1, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc1]
-; CHECK-NEXT: retq ## encoding: [0xc3]
- %b = load <16 x float>, <16 x float>* %ptr_b
- %res = call <16 x float> @llvm.x86.avx512.mask.and.ps.512(<16 x float> %a, <16 x float> %b, <16 x float> %passThru, i16 %mask)
- ret <16 x float> %res
-}
-
-define <16 x float> @test_mask_and_ps_rmkz_512(<16 x float> %a, <16 x float>* %ptr_b, i16 %mask) {
-; CHECK-LABEL: test_mask_and_ps_rmkz_512:
-; CHECK: ## BB#0:
-; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
-; CHECK-NEXT: vandps (%rdi), %zmm0, %zmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0xc9,0x54,0x07]
-; CHECK-NEXT: retq ## encoding: [0xc3]
- %b = load <16 x float>, <16 x float>* %ptr_b
- %res = call <16 x float> @llvm.x86.avx512.mask.and.ps.512(<16 x float> %a, <16 x float> %b, <16 x float> zeroinitializer, i16 %mask)
- ret <16 x float> %res
-}
-
-define <16 x float> @test_mask_and_ps_rmb_512(<16 x float> %a, float* %ptr_b) {
-; CHECK-LABEL: test_mask_and_ps_rmb_512:
-; CHECK: ## BB#0:
-; CHECK-NEXT: vandps (%rdi){1to16}, %zmm0, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x58,0x54,0x07]
-; CHECK-NEXT: retq ## encoding: [0xc3]
- %q = load float, float* %ptr_b
- %vecinit.i = insertelement <16 x float> undef, float %q, i32 0
- %b = shufflevector <16 x float> %vecinit.i, <16 x float> undef, <16 x i32> zeroinitializer
- %res = call <16 x float> @llvm.x86.avx512.mask.and.ps.512(<16 x float> %a, <16 x float> %b, <16 x float> zeroinitializer, i16 -1)
- ret <16 x float> %res
-}
-
-define <16 x float> @test_mask_and_ps_rmbk_512(<16 x float> %a, float* %ptr_b, <16 x float> %passThru, i16 %mask) {
-; CHECK-LABEL: test_mask_and_ps_rmbk_512:
-; CHECK: ## BB#0:
-; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
-; CHECK-NEXT: vandps (%rdi){1to16}, %zmm0, %zmm1 {%k1} ## encoding: [0x62,0xf1,0x7c,0x59,0x54,0x0f]
-; CHECK-NEXT: vmovaps %zmm1, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc1]
-; CHECK-NEXT: retq ## encoding: [0xc3]
- %q = load float, float* %ptr_b
- %vecinit.i = insertelement <16 x float> undef, float %q, i32 0
- %b = shufflevector <16 x float> %vecinit.i, <16 x float> undef, <16 x i32> zeroinitializer
- %res = call <16 x float> @llvm.x86.avx512.mask.and.ps.512(<16 x float> %a, <16 x float> %b, <16 x float> %passThru, i16 %mask)
- ret <16 x float> %res
-}
-
-define <16 x float> @test_mask_and_ps_rmbkz_512(<16 x float> %a, float* %ptr_b, i16 %mask) {
-; CHECK-LABEL: test_mask_and_ps_rmbkz_512:
-; CHECK: ## BB#0:
-; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
-; CHECK-NEXT: vandps (%rdi){1to16}, %zmm0, %zmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0xd9,0x54,0x07]
-; CHECK-NEXT: retq ## encoding: [0xc3]
- %q = load float, float* %ptr_b
- %vecinit.i = insertelement <16 x float> undef, float %q, i32 0
- %b = shufflevector <16 x float> %vecinit.i, <16 x float> undef, <16 x i32> zeroinitializer
- %res = call <16 x float> @llvm.x86.avx512.mask.and.ps.512(<16 x float> %a, <16 x float> %b, <16 x float> zeroinitializer, i16 %mask)
- ret <16 x float> %res
-}
-
-declare <16 x float> @llvm.x86.avx512.mask.and.ps.512(<16 x float>, <16 x float>, <16 x float>, i16)
-
-define <4 x float> @test_mask_or_ps_rr_128(<4 x float> %a, <4 x float> %b) {
-; CHECK-LABEL: test_mask_or_ps_rr_128:
-; CHECK: ## BB#0:
-; CHECK-NEXT: vorps %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7c,0x08,0x56,0xc1]
-; CHECK-NEXT: retq ## encoding: [0xc3]
- %res = call <4 x float> @llvm.x86.avx512.mask.or.ps.128(<4 x float> %a, <4 x float> %b, <4 x float> zeroinitializer, i8 -1)
- ret <4 x float> %res
-}
-
-define <4 x float> @test_mask_or_ps_rrk_128(<4 x float> %a, <4 x float> %b, <4 x float> %passThru, i8 %mask) {
-; CHECK-LABEL: test_mask_or_ps_rrk_128:
-; CHECK: ## BB#0:
-; CHECK-NEXT: kmovb %edi, %k1 ## encoding: [0xc5,0xf9,0x92,0xcf]
-; CHECK-NEXT: vorps %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf1,0x7c,0x09,0x56,0xd1]
-; CHECK-NEXT: vmovaps %xmm2, %xmm0 ## encoding: [0x62,0xf1,0x7c,0x08,0x28,0xc2]
-; CHECK-NEXT: retq ## encoding: [0xc3]
- %res = call <4 x float> @llvm.x86.avx512.mask.or.ps.128(<4 x float> %a, <4 x float> %b, <4 x float> %passThru, i8 %mask)
- ret <4 x float> %res
-}
-
-define <4 x float> @test_mask_or_ps_rrkz_128(<4 x float> %a, <4 x float> %b, i8 %mask) {
-; CHECK-LABEL: test_mask_or_ps_rrkz_128:
-; CHECK: ## BB#0:
-; CHECK-NEXT: kmovb %edi, %k1 ## encoding: [0xc5,0xf9,0x92,0xcf]
-; CHECK-NEXT: vorps %xmm1, %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0x89,0x56,0xc1]
-; CHECK-NEXT: retq ## encoding: [0xc3]
- %res = call <4 x float> @llvm.x86.avx512.mask.or.ps.128(<4 x float> %a, <4 x float> %b, <4 x float> zeroinitializer, i8 %mask)
- ret <4 x float> %res
-}
-
-define <4 x float> @test_mask_or_ps_rm_128(<4 x float> %a, <4 x float>* %ptr_b) {
-; CHECK-LABEL: test_mask_or_ps_rm_128:
-; CHECK: ## BB#0:
-; CHECK-NEXT: vorps (%rdi), %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7c,0x08,0x56,0x07]
-; CHECK-NEXT: retq ## encoding: [0xc3]
- %b = load <4 x float>, <4 x float>* %ptr_b
- %res = call <4 x float> @llvm.x86.avx512.mask.or.ps.128(<4 x float> %a, <4 x float> %b, <4 x float> zeroinitializer, i8 -1)
- ret <4 x float> %res
-}
-
-define <4 x float> @test_mask_or_ps_rmk_128(<4 x float> %a, <4 x float>* %ptr_b, <4 x float> %passThru, i8 %mask) {
-; CHECK-LABEL: test_mask_or_ps_rmk_128:
-; CHECK: ## BB#0:
-; CHECK-NEXT: kmovb %esi, %k1 ## encoding: [0xc5,0xf9,0x92,0xce]
-; CHECK-NEXT: vorps (%rdi), %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf1,0x7c,0x09,0x56,0x0f]
-; CHECK-NEXT: vmovaps %xmm1, %xmm0 ## encoding: [0x62,0xf1,0x7c,0x08,0x28,0xc1]
-; CHECK-NEXT: retq ## encoding: [0xc3]
- %b = load <4 x float>, <4 x float>* %ptr_b
- %res = call <4 x float> @llvm.x86.avx512.mask.or.ps.128(<4 x float> %a, <4 x float> %b, <4 x float> %passThru, i8 %mask)
- ret <4 x float> %res
-}
-
-define <4 x float> @test_mask_or_ps_rmkz_128(<4 x float> %a, <4 x float>* %ptr_b, i8 %mask) {
-; CHECK-LABEL: test_mask_or_ps_rmkz_128:
-; CHECK: ## BB#0:
-; CHECK-NEXT: kmovb %esi, %k1 ## encoding: [0xc5,0xf9,0x92,0xce]
-; CHECK-NEXT: vorps (%rdi), %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0x89,0x56,0x07]
-; CHECK-NEXT: retq ## encoding: [0xc3]
- %b = load <4 x float>, <4 x float>* %ptr_b
- %res = call <4 x float> @llvm.x86.avx512.mask.or.ps.128(<4 x float> %a, <4 x float> %b, <4 x float> zeroinitializer, i8 %mask)
- ret <4 x float> %res
-}
-
-define <4 x float> @test_mask_or_ps_rmb_128(<4 x float> %a, float* %ptr_b) {
-; CHECK-LABEL: test_mask_or_ps_rmb_128:
-; CHECK: ## BB#0:
-; CHECK-NEXT: vorps (%rdi){1to4}, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7c,0x18,0x56,0x07]
-; CHECK-NEXT: retq ## encoding: [0xc3]
- %q = load float, float* %ptr_b
- %vecinit.i = insertelement <4 x float> undef, float %q, i32 0
- %b = shufflevector <4 x float> %vecinit.i, <4 x float> undef, <4 x i32> zeroinitializer
- %res = call <4 x float> @llvm.x86.avx512.mask.or.ps.128(<4 x float> %a, <4 x float> %b, <4 x float> zeroinitializer, i8 -1)
- ret <4 x float> %res
-}
-
-define <4 x float> @test_mask_or_ps_rmbk_128(<4 x float> %a, float* %ptr_b, <4 x float> %passThru, i8 %mask) {
-; CHECK-LABEL: test_mask_or_ps_rmbk_128:
-; CHECK: ## BB#0:
-; CHECK-NEXT: kmovb %esi, %k1 ## encoding: [0xc5,0xf9,0x92,0xce]
-; CHECK-NEXT: vorps (%rdi){1to4}, %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf1,0x7c,0x19,0x56,0x0f]
-; CHECK-NEXT: vmovaps %xmm1, %xmm0 ## encoding: [0x62,0xf1,0x7c,0x08,0x28,0xc1]
-; CHECK-NEXT: retq ## encoding: [0xc3]
- %q = load float, float* %ptr_b
- %vecinit.i = insertelement <4 x float> undef, float %q, i32 0
- %b = shufflevector <4 x float> %vecinit.i, <4 x float> undef, <4 x i32> zeroinitializer
- %res = call <4 x float> @llvm.x86.avx512.mask.or.ps.128(<4 x float> %a, <4 x float> %b, <4 x float> %passThru, i8 %mask)
- ret <4 x float> %res
-}
-
-define <4 x float> @test_mask_or_ps_rmbkz_128(<4 x float> %a, float* %ptr_b, i8 %mask) {
-; CHECK-LABEL: test_mask_or_ps_rmbkz_128:
-; CHECK: ## BB#0:
-; CHECK-NEXT: kmovb %esi, %k1 ## encoding: [0xc5,0xf9,0x92,0xce]
-; CHECK-NEXT: vorps (%rdi){1to4}, %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0x99,0x56,0x07]
-; CHECK-NEXT: retq ## encoding: [0xc3]
- %q = load float, float* %ptr_b
- %vecinit.i = insertelement <4 x float> undef, float %q, i32 0
- %b = shufflevector <4 x float> %vecinit.i, <4 x float> undef, <4 x i32> zeroinitializer
- %res = call <4 x float> @llvm.x86.avx512.mask.or.ps.128(<4 x float> %a, <4 x float> %b, <4 x float> zeroinitializer, i8 %mask)
- ret <4 x float> %res
-}
-
-declare <4 x float> @llvm.x86.avx512.mask.or.ps.128(<4 x float>, <4 x float>, <4 x float>, i8)
-
-define <8 x float> @test_mask_or_ps_rr_256(<8 x float> %a, <8 x float> %b) {
-; CHECK-LABEL: test_mask_or_ps_rr_256:
-; CHECK: ## BB#0:
-; CHECK-NEXT: vorps %ymm1, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7c,0x28,0x56,0xc1]
-; CHECK-NEXT: retq ## encoding: [0xc3]
- %res = call <8 x float> @llvm.x86.avx512.mask.or.ps.256(<8 x float> %a, <8 x float> %b, <8 x float> zeroinitializer, i8 -1)
- ret <8 x float> %res
-}
-
-define <8 x float> @test_mask_or_ps_rrk_256(<8 x float> %a, <8 x float> %b, <8 x float> %passThru, i8 %mask) {
-; CHECK-LABEL: test_mask_or_ps_rrk_256:
-; CHECK: ## BB#0:
-; CHECK-NEXT: kmovb %edi, %k1 ## encoding: [0xc5,0xf9,0x92,0xcf]
-; CHECK-NEXT: vorps %ymm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf1,0x7c,0x29,0x56,0xd1]
-; CHECK-NEXT: vmovaps %ymm2, %ymm0 ## encoding: [0x62,0xf1,0x7c,0x28,0x28,0xc2]
-; CHECK-NEXT: retq ## encoding: [0xc3]
- %res = call <8 x float> @llvm.x86.avx512.mask.or.ps.256(<8 x float> %a, <8 x float> %b, <8 x float> %passThru, i8 %mask)
- ret <8 x float> %res
-}
-
-define <8 x float> @test_mask_or_ps_rrkz_256(<8 x float> %a, <8 x float> %b, i8 %mask) {
-; CHECK-LABEL: test_mask_or_ps_rrkz_256:
-; CHECK: ## BB#0:
-; CHECK-NEXT: kmovb %edi, %k1 ## encoding: [0xc5,0xf9,0x92,0xcf]
-; CHECK-NEXT: vorps %ymm1, %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0xa9,0x56,0xc1]
-; CHECK-NEXT: retq ## encoding: [0xc3]
- %res = call <8 x float> @llvm.x86.avx512.mask.or.ps.256(<8 x float> %a, <8 x float> %b, <8 x float> zeroinitializer, i8 %mask)
- ret <8 x float> %res
-}
-
-define <8 x float> @test_mask_or_ps_rm_256(<8 x float> %a, <8 x float>* %ptr_b) {
-; CHECK-LABEL: test_mask_or_ps_rm_256:
-; CHECK: ## BB#0:
-; CHECK-NEXT: vorps (%rdi), %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7c,0x28,0x56,0x07]
-; CHECK-NEXT: retq ## encoding: [0xc3]
- %b = load <8 x float>, <8 x float>* %ptr_b
- %res = call <8 x float> @llvm.x86.avx512.mask.or.ps.256(<8 x float> %a, <8 x float> %b, <8 x float> zeroinitializer, i8 -1)
- ret <8 x float> %res
-}
-
-define <8 x float> @test_mask_or_ps_rmk_256(<8 x float> %a, <8 x float>* %ptr_b, <8 x float> %passThru, i8 %mask) {
-; CHECK-LABEL: test_mask_or_ps_rmk_256:
-; CHECK: ## BB#0:
-; CHECK-NEXT: kmovb %esi, %k1 ## encoding: [0xc5,0xf9,0x92,0xce]
-; CHECK-NEXT: vorps (%rdi), %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf1,0x7c,0x29,0x56,0x0f]
-; CHECK-NEXT: vmovaps %ymm1, %ymm0 ## encoding: [0x62,0xf1,0x7c,0x28,0x28,0xc1]
-; CHECK-NEXT: retq ## encoding: [0xc3]
- %b = load <8 x float>, <8 x float>* %ptr_b
- %res = call <8 x float> @llvm.x86.avx512.mask.or.ps.256(<8 x float> %a, <8 x float> %b, <8 x float> %passThru, i8 %mask)
- ret <8 x float> %res
-}
-
-define <8 x float> @test_mask_or_ps_rmkz_256(<8 x float> %a, <8 x float>* %ptr_b, i8 %mask) {
-; CHECK-LABEL: test_mask_or_ps_rmkz_256:
-; CHECK: ## BB#0:
-; CHECK-NEXT: kmovb %esi, %k1 ## encoding: [0xc5,0xf9,0x92,0xce]
-; CHECK-NEXT: vorps (%rdi), %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0xa9,0x56,0x07]
-; CHECK-NEXT: retq ## encoding: [0xc3]
- %b = load <8 x float>, <8 x float>* %ptr_b
- %res = call <8 x float> @llvm.x86.avx512.mask.or.ps.256(<8 x float> %a, <8 x float> %b, <8 x float> zeroinitializer, i8 %mask)
- ret <8 x float> %res
-}
-
-define <8 x float> @test_mask_or_ps_rmb_256(<8 x float> %a, float* %ptr_b) {
-; CHECK-LABEL: test_mask_or_ps_rmb_256:
-; CHECK: ## BB#0:
-; CHECK-NEXT: vorps (%rdi){1to8}, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7c,0x38,0x56,0x07]
-; CHECK-NEXT: retq ## encoding: [0xc3]
- %q = load float, float* %ptr_b
- %vecinit.i = insertelement <8 x float> undef, float %q, i32 0
- %b = shufflevector <8 x float> %vecinit.i, <8 x float> undef, <8 x i32> zeroinitializer
- %res = call <8 x float> @llvm.x86.avx512.mask.or.ps.256(<8 x float> %a, <8 x float> %b, <8 x float> zeroinitializer, i8 -1)
- ret <8 x float> %res
-}
-
-define <8 x float> @test_mask_or_ps_rmbk_256(<8 x float> %a, float* %ptr_b, <8 x float> %passThru, i8 %mask) {
-; CHECK-LABEL: test_mask_or_ps_rmbk_256:
-; CHECK: ## BB#0:
-; CHECK-NEXT: kmovb %esi, %k1 ## encoding: [0xc5,0xf9,0x92,0xce]
-; CHECK-NEXT: vorps (%rdi){1to8}, %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf1,0x7c,0x39,0x56,0x0f]
-; CHECK-NEXT: vmovaps %ymm1, %ymm0 ## encoding: [0x62,0xf1,0x7c,0x28,0x28,0xc1]
-; CHECK-NEXT: retq ## encoding: [0xc3]
- %q = load float, float* %ptr_b
- %vecinit.i = insertelement <8 x float> undef, float %q, i32 0
- %b = shufflevector <8 x float> %vecinit.i, <8 x float> undef, <8 x i32> zeroinitializer
- %res = call <8 x float> @llvm.x86.avx512.mask.or.ps.256(<8 x float> %a, <8 x float> %b, <8 x float> %passThru, i8 %mask)
- ret <8 x float> %res
-}
-
-define <8 x float> @test_mask_or_ps_rmbkz_256(<8 x float> %a, float* %ptr_b, i8 %mask) {
-; CHECK-LABEL: test_mask_or_ps_rmbkz_256:
-; CHECK: ## BB#0:
-; CHECK-NEXT: kmovb %esi, %k1 ## encoding: [0xc5,0xf9,0x92,0xce]
-; CHECK-NEXT: vorps (%rdi){1to8}, %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0xb9,0x56,0x07]
-; CHECK-NEXT: retq ## encoding: [0xc3]
- %q = load float, float* %ptr_b
- %vecinit.i = insertelement <8 x float> undef, float %q, i32 0
- %b = shufflevector <8 x float> %vecinit.i, <8 x float> undef, <8 x i32> zeroinitializer
- %res = call <8 x float> @llvm.x86.avx512.mask.or.ps.256(<8 x float> %a, <8 x float> %b, <8 x float> zeroinitializer, i8 %mask)
- ret <8 x float> %res
-}
-
-declare <8 x float> @llvm.x86.avx512.mask.or.ps.256(<8 x float>, <8 x float>, <8 x float>, i8)
-
-define <16 x float> @test_mask_or_ps_rr_512(<16 x float> %a, <16 x float> %b) {
-; CHECK-LABEL: test_mask_or_ps_rr_512:
-; CHECK: ## BB#0:
-; CHECK-NEXT: vorps %zmm1, %zmm0, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x56,0xc1]
-; CHECK-NEXT: retq ## encoding: [0xc3]
- %res = call <16 x float> @llvm.x86.avx512.mask.or.ps.512(<16 x float> %a, <16 x float> %b, <16 x float> zeroinitializer, i16 -1)
- ret <16 x float> %res
-}
-
-define <16 x float> @test_mask_or_ps_rrk_512(<16 x float> %a, <16 x float> %b, <16 x float> %passThru, i16 %mask) {
-; CHECK-LABEL: test_mask_or_ps_rrk_512:
-; CHECK: ## BB#0:
-; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vorps %zmm1, %zmm0, %zmm2 {%k1} ## encoding: [0x62,0xf1,0x7c,0x49,0x56,0xd1]
-; CHECK-NEXT: vmovaps %zmm2, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc2]
-; CHECK-NEXT: retq ## encoding: [0xc3]
- %res = call <16 x float> @llvm.x86.avx512.mask.or.ps.512(<16 x float> %a, <16 x float> %b, <16 x float> %passThru, i16 %mask)
- ret <16 x float> %res
-}
-
-define <16 x float> @test_mask_or_ps_rrkz_512(<16 x float> %a, <16 x float> %b, i16 %mask) {
-; CHECK-LABEL: test_mask_or_ps_rrkz_512:
-; CHECK: ## BB#0:
-; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vorps %zmm1, %zmm0, %zmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0xc9,0x56,0xc1]
-; CHECK-NEXT: retq ## encoding: [0xc3]
- %res = call <16 x float> @llvm.x86.avx512.mask.or.ps.512(<16 x float> %a, <16 x float> %b, <16 x float> zeroinitializer, i16 %mask)
- ret <16 x float> %res
-}
-
-define <16 x float> @test_mask_or_ps_rm_512(<16 x float> %a, <16 x float>* %ptr_b) {
-; CHECK-LABEL: test_mask_or_ps_rm_512:
-; CHECK: ## BB#0:
-; CHECK-NEXT: vorps (%rdi), %zmm0, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x56,0x07]
-; CHECK-NEXT: retq ## encoding: [0xc3]
- %b = load <16 x float>, <16 x float>* %ptr_b
- %res = call <16 x float> @llvm.x86.avx512.mask.or.ps.512(<16 x float> %a, <16 x float> %b, <16 x float> zeroinitializer, i16 -1)
- ret <16 x float> %res
-}
-
-define <16 x float> @test_mask_or_ps_rmk_512(<16 x float> %a, <16 x float>* %ptr_b, <16 x float> %passThru, i16 %mask) {
-; CHECK-LABEL: test_mask_or_ps_rmk_512:
-; CHECK: ## BB#0:
-; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
-; CHECK-NEXT: vorps (%rdi), %zmm0, %zmm1 {%k1} ## encoding: [0x62,0xf1,0x7c,0x49,0x56,0x0f]
-; CHECK-NEXT: vmovaps %zmm1, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc1]
-; CHECK-NEXT: retq ## encoding: [0xc3]
- %b = load <16 x float>, <16 x float>* %ptr_b
- %res = call <16 x float> @llvm.x86.avx512.mask.or.ps.512(<16 x float> %a, <16 x float> %b, <16 x float> %passThru, i16 %mask)
- ret <16 x float> %res
-}
-
-define <16 x float> @test_mask_or_ps_rmkz_512(<16 x float> %a, <16 x float>* %ptr_b, i16 %mask) {
-; CHECK-LABEL: test_mask_or_ps_rmkz_512:
-; CHECK: ## BB#0:
-; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
-; CHECK-NEXT: vorps (%rdi), %zmm0, %zmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0xc9,0x56,0x07]
-; CHECK-NEXT: retq ## encoding: [0xc3]
- %b = load <16 x float>, <16 x float>* %ptr_b
- %res = call <16 x float> @llvm.x86.avx512.mask.or.ps.512(<16 x float> %a, <16 x float> %b, <16 x float> zeroinitializer, i16 %mask)
- ret <16 x float> %res
-}
-
-define <16 x float> @test_mask_or_ps_rmb_512(<16 x float> %a, float* %ptr_b) {
-; CHECK-LABEL: test_mask_or_ps_rmb_512:
-; CHECK: ## BB#0:
-; CHECK-NEXT: vorps (%rdi){1to16}, %zmm0, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x58,0x56,0x07]
-; CHECK-NEXT: retq ## encoding: [0xc3]
- %q = load float, float* %ptr_b
- %vecinit.i = insertelement <16 x float> undef, float %q, i32 0
- %b = shufflevector <16 x float> %vecinit.i, <16 x float> undef, <16 x i32> zeroinitializer
- %res = call <16 x float> @llvm.x86.avx512.mask.or.ps.512(<16 x float> %a, <16 x float> %b, <16 x float> zeroinitializer, i16 -1)
- ret <16 x float> %res
-}
-
-define <16 x float> @test_mask_or_ps_rmbk_512(<16 x float> %a, float* %ptr_b, <16 x float> %passThru, i16 %mask) {
-; CHECK-LABEL: test_mask_or_ps_rmbk_512:
-; CHECK: ## BB#0:
-; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
-; CHECK-NEXT: vorps (%rdi){1to16}, %zmm0, %zmm1 {%k1} ## encoding: [0x62,0xf1,0x7c,0x59,0x56,0x0f]
-; CHECK-NEXT: vmovaps %zmm1, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc1]
-; CHECK-NEXT: retq ## encoding: [0xc3]
- %q = load float, float* %ptr_b
- %vecinit.i = insertelement <16 x float> undef, float %q, i32 0
- %b = shufflevector <16 x float> %vecinit.i, <16 x float> undef, <16 x i32> zeroinitializer
- %res = call <16 x float> @llvm.x86.avx512.mask.or.ps.512(<16 x float> %a, <16 x float> %b, <16 x float> %passThru, i16 %mask)
- ret <16 x float> %res
-}
-
-define <16 x float> @test_mask_or_ps_rmbkz_512(<16 x float> %a, float* %ptr_b, i16 %mask) {
-; CHECK-LABEL: test_mask_or_ps_rmbkz_512:
-; CHECK: ## BB#0:
-; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
-; CHECK-NEXT: vorps (%rdi){1to16}, %zmm0, %zmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0xd9,0x56,0x07]
-; CHECK-NEXT: retq ## encoding: [0xc3]
- %q = load float, float* %ptr_b
- %vecinit.i = insertelement <16 x float> undef, float %q, i32 0
- %b = shufflevector <16 x float> %vecinit.i, <16 x float> undef, <16 x i32> zeroinitializer
- %res = call <16 x float> @llvm.x86.avx512.mask.or.ps.512(<16 x float> %a, <16 x float> %b, <16 x float> zeroinitializer, i16 %mask)
- ret <16 x float> %res
-}
-
-declare <16 x float> @llvm.x86.avx512.mask.or.ps.512(<16 x float>, <16 x float>, <16 x float>, i16)
-
-define <4 x float> @test_mask_xor_ps_rr_128(<4 x float> %a, <4 x float> %b) {
-; CHECK-LABEL: test_mask_xor_ps_rr_128:
-; CHECK: ## BB#0:
-; CHECK-NEXT: vxorps %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7c,0x08,0x57,0xc1]
-; CHECK-NEXT: retq ## encoding: [0xc3]
- %res = call <4 x float> @llvm.x86.avx512.mask.xor.ps.128(<4 x float> %a, <4 x float> %b, <4 x float> zeroinitializer, i8 -1)
- ret <4 x float> %res
-}
-
-define <4 x float> @test_mask_xor_ps_rrk_128(<4 x float> %a, <4 x float> %b, <4 x float> %passThru, i8 %mask) {
-; CHECK-LABEL: test_mask_xor_ps_rrk_128:
-; CHECK: ## BB#0:
-; CHECK-NEXT: kmovb %edi, %k1 ## encoding: [0xc5,0xf9,0x92,0xcf]
-; CHECK-NEXT: vxorps %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf1,0x7c,0x09,0x57,0xd1]
-; CHECK-NEXT: vmovaps %xmm2, %xmm0 ## encoding: [0x62,0xf1,0x7c,0x08,0x28,0xc2]
-; CHECK-NEXT: retq ## encoding: [0xc3]
- %res = call <4 x float> @llvm.x86.avx512.mask.xor.ps.128(<4 x float> %a, <4 x float> %b, <4 x float> %passThru, i8 %mask)
- ret <4 x float> %res
-}
-
-define <4 x float> @test_mask_xor_ps_rrkz_128(<4 x float> %a, <4 x float> %b, i8 %mask) {
-; CHECK-LABEL: test_mask_xor_ps_rrkz_128:
-; CHECK: ## BB#0:
-; CHECK-NEXT: kmovb %edi, %k1 ## encoding: [0xc5,0xf9,0x92,0xcf]
-; CHECK-NEXT: vxorps %xmm1, %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0x89,0x57,0xc1]
-; CHECK-NEXT: retq ## encoding: [0xc3]
- %res = call <4 x float> @llvm.x86.avx512.mask.xor.ps.128(<4 x float> %a, <4 x float> %b, <4 x float> zeroinitializer, i8 %mask)
- ret <4 x float> %res
-}
-
-define <4 x float> @test_mask_xor_ps_rm_128(<4 x float> %a, <4 x float>* %ptr_b) {
-; CHECK-LABEL: test_mask_xor_ps_rm_128:
-; CHECK: ## BB#0:
-; CHECK-NEXT: vxorps (%rdi), %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7c,0x08,0x57,0x07]
-; CHECK-NEXT: retq ## encoding: [0xc3]
- %b = load <4 x float>, <4 x float>* %ptr_b
- %res = call <4 x float> @llvm.x86.avx512.mask.xor.ps.128(<4 x float> %a, <4 x float> %b, <4 x float> zeroinitializer, i8 -1)
- ret <4 x float> %res
-}
-
-define <4 x float> @test_mask_xor_ps_rmk_128(<4 x float> %a, <4 x float>* %ptr_b, <4 x float> %passThru, i8 %mask) {
-; CHECK-LABEL: test_mask_xor_ps_rmk_128:
-; CHECK: ## BB#0:
-; CHECK-NEXT: kmovb %esi, %k1 ## encoding: [0xc5,0xf9,0x92,0xce]
-; CHECK-NEXT: vxorps (%rdi), %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf1,0x7c,0x09,0x57,0x0f]
-; CHECK-NEXT: vmovaps %xmm1, %xmm0 ## encoding: [0x62,0xf1,0x7c,0x08,0x28,0xc1]
-; CHECK-NEXT: retq ## encoding: [0xc3]
- %b = load <4 x float>, <4 x float>* %ptr_b
- %res = call <4 x float> @llvm.x86.avx512.mask.xor.ps.128(<4 x float> %a, <4 x float> %b, <4 x float> %passThru, i8 %mask)
- ret <4 x float> %res
-}
-
-define <4 x float> @test_mask_xor_ps_rmkz_128(<4 x float> %a, <4 x float>* %ptr_b, i8 %mask) {
-; CHECK-LABEL: test_mask_xor_ps_rmkz_128:
-; CHECK: ## BB#0:
-; CHECK-NEXT: kmovb %esi, %k1 ## encoding: [0xc5,0xf9,0x92,0xce]
-; CHECK-NEXT: vxorps (%rdi), %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0x89,0x57,0x07]
-; CHECK-NEXT: retq ## encoding: [0xc3]
- %b = load <4 x float>, <4 x float>* %ptr_b
- %res = call <4 x float> @llvm.x86.avx512.mask.xor.ps.128(<4 x float> %a, <4 x float> %b, <4 x float> zeroinitializer, i8 %mask)
- ret <4 x float> %res
-}
-
-define <4 x float> @test_mask_xor_ps_rmb_128(<4 x float> %a, float* %ptr_b) {
-; CHECK-LABEL: test_mask_xor_ps_rmb_128:
-; CHECK: ## BB#0:
-; CHECK-NEXT: vxorps (%rdi){1to4}, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7c,0x18,0x57,0x07]
-; CHECK-NEXT: retq ## encoding: [0xc3]
- %q = load float, float* %ptr_b
- %vecinit.i = insertelement <4 x float> undef, float %q, i32 0
- %b = shufflevector <4 x float> %vecinit.i, <4 x float> undef, <4 x i32> zeroinitializer
- %res = call <4 x float> @llvm.x86.avx512.mask.xor.ps.128(<4 x float> %a, <4 x float> %b, <4 x float> zeroinitializer, i8 -1)
- ret <4 x float> %res
-}
-
-define <4 x float> @test_mask_xor_ps_rmbk_128(<4 x float> %a, float* %ptr_b, <4 x float> %passThru, i8 %mask) {
-; CHECK-LABEL: test_mask_xor_ps_rmbk_128:
-; CHECK: ## BB#0:
-; CHECK-NEXT: kmovb %esi, %k1 ## encoding: [0xc5,0xf9,0x92,0xce]
-; CHECK-NEXT: vxorps (%rdi){1to4}, %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf1,0x7c,0x19,0x57,0x0f]
-; CHECK-NEXT: vmovaps %xmm1, %xmm0 ## encoding: [0x62,0xf1,0x7c,0x08,0x28,0xc1]
-; CHECK-NEXT: retq ## encoding: [0xc3]
- %q = load float, float* %ptr_b
- %vecinit.i = insertelement <4 x float> undef, float %q, i32 0
- %b = shufflevector <4 x float> %vecinit.i, <4 x float> undef, <4 x i32> zeroinitializer
- %res = call <4 x float> @llvm.x86.avx512.mask.xor.ps.128(<4 x float> %a, <4 x float> %b, <4 x float> %passThru, i8 %mask)
- ret <4 x float> %res
-}
-
-define <4 x float> @test_mask_xor_ps_rmbkz_128(<4 x float> %a, float* %ptr_b, i8 %mask) {
-; CHECK-LABEL: test_mask_xor_ps_rmbkz_128:
-; CHECK: ## BB#0:
-; CHECK-NEXT: kmovb %esi, %k1 ## encoding: [0xc5,0xf9,0x92,0xce]
-; CHECK-NEXT: vxorps (%rdi){1to4}, %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0x99,0x57,0x07]
-; CHECK-NEXT: retq ## encoding: [0xc3]
- %q = load float, float* %ptr_b
- %vecinit.i = insertelement <4 x float> undef, float %q, i32 0
- %b = shufflevector <4 x float> %vecinit.i, <4 x float> undef, <4 x i32> zeroinitializer
- %res = call <4 x float> @llvm.x86.avx512.mask.xor.ps.128(<4 x float> %a, <4 x float> %b, <4 x float> zeroinitializer, i8 %mask)
- ret <4 x float> %res
-}
-
-declare <4 x float> @llvm.x86.avx512.mask.xor.ps.128(<4 x float>, <4 x float>, <4 x float>, i8)
-
-define <8 x float> @test_mask_xor_ps_rr_256(<8 x float> %a, <8 x float> %b) {
-; CHECK-LABEL: test_mask_xor_ps_rr_256:
-; CHECK: ## BB#0:
-; CHECK-NEXT: vxorps %ymm1, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7c,0x28,0x57,0xc1]
-; CHECK-NEXT: retq ## encoding: [0xc3]
- %res = call <8 x float> @llvm.x86.avx512.mask.xor.ps.256(<8 x float> %a, <8 x float> %b, <8 x float> zeroinitializer, i8 -1)
- ret <8 x float> %res
-}
-
-define <8 x float> @test_mask_xor_ps_rrk_256(<8 x float> %a, <8 x float> %b, <8 x float> %passThru, i8 %mask) {
-; CHECK-LABEL: test_mask_xor_ps_rrk_256:
-; CHECK: ## BB#0:
-; CHECK-NEXT: kmovb %edi, %k1 ## encoding: [0xc5,0xf9,0x92,0xcf]
-; CHECK-NEXT: vxorps %ymm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf1,0x7c,0x29,0x57,0xd1]
-; CHECK-NEXT: vmovaps %ymm2, %ymm0 ## encoding: [0x62,0xf1,0x7c,0x28,0x28,0xc2]
-; CHECK-NEXT: retq ## encoding: [0xc3]
- %res = call <8 x float> @llvm.x86.avx512.mask.xor.ps.256(<8 x float> %a, <8 x float> %b, <8 x float> %passThru, i8 %mask)
- ret <8 x float> %res
-}
-
-define <8 x float> @test_mask_xor_ps_rrkz_256(<8 x float> %a, <8 x float> %b, i8 %mask) {
-; CHECK-LABEL: test_mask_xor_ps_rrkz_256:
-; CHECK: ## BB#0:
-; CHECK-NEXT: kmovb %edi, %k1 ## encoding: [0xc5,0xf9,0x92,0xcf]
-; CHECK-NEXT: vxorps %ymm1, %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0xa9,0x57,0xc1]
-; CHECK-NEXT: retq ## encoding: [0xc3]
- %res = call <8 x float> @llvm.x86.avx512.mask.xor.ps.256(<8 x float> %a, <8 x float> %b, <8 x float> zeroinitializer, i8 %mask)
- ret <8 x float> %res
-}
-
-define <8 x float> @test_mask_xor_ps_rm_256(<8 x float> %a, <8 x float>* %ptr_b) {
-; CHECK-LABEL: test_mask_xor_ps_rm_256:
-; CHECK: ## BB#0:
-; CHECK-NEXT: vxorps (%rdi), %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7c,0x28,0x57,0x07]
-; CHECK-NEXT: retq ## encoding: [0xc3]
- %b = load <8 x float>, <8 x float>* %ptr_b
- %res = call <8 x float> @llvm.x86.avx512.mask.xor.ps.256(<8 x float> %a, <8 x float> %b, <8 x float> zeroinitializer, i8 -1)
- ret <8 x float> %res
-}
-
-define <8 x float> @test_mask_xor_ps_rmk_256(<8 x float> %a, <8 x float>* %ptr_b, <8 x float> %passThru, i8 %mask) {
-; CHECK-LABEL: test_mask_xor_ps_rmk_256:
-; CHECK: ## BB#0:
-; CHECK-NEXT: kmovb %esi, %k1 ## encoding: [0xc5,0xf9,0x92,0xce]
-; CHECK-NEXT: vxorps (%rdi), %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf1,0x7c,0x29,0x57,0x0f]
-; CHECK-NEXT: vmovaps %ymm1, %ymm0 ## encoding: [0x62,0xf1,0x7c,0x28,0x28,0xc1]
-; CHECK-NEXT: retq ## encoding: [0xc3]
- %b = load <8 x float>, <8 x float>* %ptr_b
- %res = call <8 x float> @llvm.x86.avx512.mask.xor.ps.256(<8 x float> %a, <8 x float> %b, <8 x float> %passThru, i8 %mask)
- ret <8 x float> %res
-}
-
-define <8 x float> @test_mask_xor_ps_rmkz_256(<8 x float> %a, <8 x float>* %ptr_b, i8 %mask) {
-; CHECK-LABEL: test_mask_xor_ps_rmkz_256:
-; CHECK: ## BB#0:
-; CHECK-NEXT: kmovb %esi, %k1 ## encoding: [0xc5,0xf9,0x92,0xce]
-; CHECK-NEXT: vxorps (%rdi), %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0xa9,0x57,0x07]
-; CHECK-NEXT: retq ## encoding: [0xc3]
- %b = load <8 x float>, <8 x float>* %ptr_b
- %res = call <8 x float> @llvm.x86.avx512.mask.xor.ps.256(<8 x float> %a, <8 x float> %b, <8 x float> zeroinitializer, i8 %mask)
- ret <8 x float> %res
-}
-
-define <8 x float> @test_mask_xor_ps_rmb_256(<8 x float> %a, float* %ptr_b) {
-; CHECK-LABEL: test_mask_xor_ps_rmb_256:
-; CHECK: ## BB#0:
-; CHECK-NEXT: vxorps (%rdi){1to8}, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7c,0x38,0x57,0x07]
-; CHECK-NEXT: retq ## encoding: [0xc3]
- %q = load float, float* %ptr_b
- %vecinit.i = insertelement <8 x float> undef, float %q, i32 0
- %b = shufflevector <8 x float> %vecinit.i, <8 x float> undef, <8 x i32> zeroinitializer
- %res = call <8 x float> @llvm.x86.avx512.mask.xor.ps.256(<8 x float> %a, <8 x float> %b, <8 x float> zeroinitializer, i8 -1)
- ret <8 x float> %res
-}
-
-define <8 x float> @test_mask_xor_ps_rmbk_256(<8 x float> %a, float* %ptr_b, <8 x float> %passThru, i8 %mask) {
-; CHECK-LABEL: test_mask_xor_ps_rmbk_256:
-; CHECK: ## BB#0:
-; CHECK-NEXT: kmovb %esi, %k1 ## encoding: [0xc5,0xf9,0x92,0xce]
-; CHECK-NEXT: vxorps (%rdi){1to8}, %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf1,0x7c,0x39,0x57,0x0f]
-; CHECK-NEXT: vmovaps %ymm1, %ymm0 ## encoding: [0x62,0xf1,0x7c,0x28,0x28,0xc1]
-; CHECK-NEXT: retq ## encoding: [0xc3]
- %q = load float, float* %ptr_b
- %vecinit.i = insertelement <8 x float> undef, float %q, i32 0
- %b = shufflevector <8 x float> %vecinit.i, <8 x float> undef, <8 x i32> zeroinitializer
- %res = call <8 x float> @llvm.x86.avx512.mask.xor.ps.256(<8 x float> %a, <8 x float> %b, <8 x float> %passThru, i8 %mask)
- ret <8 x float> %res
-}
-
-define <8 x float> @test_mask_xor_ps_rmbkz_256(<8 x float> %a, float* %ptr_b, i8 %mask) {
-; CHECK-LABEL: test_mask_xor_ps_rmbkz_256:
-; CHECK: ## BB#0:
-; CHECK-NEXT: kmovb %esi, %k1 ## encoding: [0xc5,0xf9,0x92,0xce]
-; CHECK-NEXT: vxorps (%rdi){1to8}, %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0xb9,0x57,0x07]
-; CHECK-NEXT: retq ## encoding: [0xc3]
- %q = load float, float* %ptr_b
- %vecinit.i = insertelement <8 x float> undef, float %q, i32 0
- %b = shufflevector <8 x float> %vecinit.i, <8 x float> undef, <8 x i32> zeroinitializer
- %res = call <8 x float> @llvm.x86.avx512.mask.xor.ps.256(<8 x float> %a, <8 x float> %b, <8 x float> zeroinitializer, i8 %mask)
- ret <8 x float> %res
-}
-
-declare <8 x float> @llvm.x86.avx512.mask.xor.ps.256(<8 x float>, <8 x float>, <8 x float>, i8)
-
-define <16 x float> @test_mask_xor_ps_rr_512(<16 x float> %a, <16 x float> %b) {
-; CHECK-LABEL: test_mask_xor_ps_rr_512:
-; CHECK: ## BB#0:
-; CHECK-NEXT: vxorps %zmm1, %zmm0, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x57,0xc1]
-; CHECK-NEXT: retq ## encoding: [0xc3]
- %res = call <16 x float> @llvm.x86.avx512.mask.xor.ps.512(<16 x float> %a, <16 x float> %b, <16 x float> zeroinitializer, i16 -1)
- ret <16 x float> %res
-}
-
-define <16 x float> @test_mask_xor_ps_rrk_512(<16 x float> %a, <16 x float> %b, <16 x float> %passThru, i16 %mask) {
-; CHECK-LABEL: test_mask_xor_ps_rrk_512:
-; CHECK: ## BB#0:
-; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vxorps %zmm1, %zmm0, %zmm2 {%k1} ## encoding: [0x62,0xf1,0x7c,0x49,0x57,0xd1]
-; CHECK-NEXT: vmovaps %zmm2, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc2]
-; CHECK-NEXT: retq ## encoding: [0xc3]
- %res = call <16 x float> @llvm.x86.avx512.mask.xor.ps.512(<16 x float> %a, <16 x float> %b, <16 x float> %passThru, i16 %mask)
- ret <16 x float> %res
-}
-
-define <16 x float> @test_mask_xor_ps_rrkz_512(<16 x float> %a, <16 x float> %b, i16 %mask) {
-; CHECK-LABEL: test_mask_xor_ps_rrkz_512:
-; CHECK: ## BB#0:
-; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vxorps %zmm1, %zmm0, %zmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0xc9,0x57,0xc1]
-; CHECK-NEXT: retq ## encoding: [0xc3]
- %res = call <16 x float> @llvm.x86.avx512.mask.xor.ps.512(<16 x float> %a, <16 x float> %b, <16 x float> zeroinitializer, i16 %mask)
- ret <16 x float> %res
-}
-
-define <16 x float> @test_mask_xor_ps_rm_512(<16 x float> %a, <16 x float>* %ptr_b) {
-; CHECK-LABEL: test_mask_xor_ps_rm_512:
-; CHECK: ## BB#0:
-; CHECK-NEXT: vxorps (%rdi), %zmm0, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x57,0x07]
-; CHECK-NEXT: retq ## encoding: [0xc3]
- %b = load <16 x float>, <16 x float>* %ptr_b
- %res = call <16 x float> @llvm.x86.avx512.mask.xor.ps.512(<16 x float> %a, <16 x float> %b, <16 x float> zeroinitializer, i16 -1)
- ret <16 x float> %res
-}
-
-define <16 x float> @test_mask_xor_ps_rmk_512(<16 x float> %a, <16 x float>* %ptr_b, <16 x float> %passThru, i16 %mask) {
-; CHECK-LABEL: test_mask_xor_ps_rmk_512:
-; CHECK: ## BB#0:
-; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
-; CHECK-NEXT: vxorps (%rdi), %zmm0, %zmm1 {%k1} ## encoding: [0x62,0xf1,0x7c,0x49,0x57,0x0f]
-; CHECK-NEXT: vmovaps %zmm1, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc1]
-; CHECK-NEXT: retq ## encoding: [0xc3]
- %b = load <16 x float>, <16 x float>* %ptr_b
- %res = call <16 x float> @llvm.x86.avx512.mask.xor.ps.512(<16 x float> %a, <16 x float> %b, <16 x float> %passThru, i16 %mask)
- ret <16 x float> %res
-}
-
-define <16 x float> @test_mask_xor_ps_rmkz_512(<16 x float> %a, <16 x float>* %ptr_b, i16 %mask) {
-; CHECK-LABEL: test_mask_xor_ps_rmkz_512:
-; CHECK: ## BB#0:
-; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
-; CHECK-NEXT: vxorps (%rdi), %zmm0, %zmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0xc9,0x57,0x07]
-; CHECK-NEXT: retq ## encoding: [0xc3]
- %b = load <16 x float>, <16 x float>* %ptr_b
- %res = call <16 x float> @llvm.x86.avx512.mask.xor.ps.512(<16 x float> %a, <16 x float> %b, <16 x float> zeroinitializer, i16 %mask)
- ret <16 x float> %res
-}
-
-define <16 x float> @test_mask_xor_ps_rmb_512(<16 x float> %a, float* %ptr_b) {
-; CHECK-LABEL: test_mask_xor_ps_rmb_512:
-; CHECK: ## BB#0:
-; CHECK-NEXT: vxorps (%rdi){1to16}, %zmm0, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x58,0x57,0x07]
-; CHECK-NEXT: retq ## encoding: [0xc3]
- %q = load float, float* %ptr_b
- %vecinit.i = insertelement <16 x float> undef, float %q, i32 0
- %b = shufflevector <16 x float> %vecinit.i, <16 x float> undef, <16 x i32> zeroinitializer
- %res = call <16 x float> @llvm.x86.avx512.mask.xor.ps.512(<16 x float> %a, <16 x float> %b, <16 x float> zeroinitializer, i16 -1)
- ret <16 x float> %res
-}
-
-define <16 x float> @test_mask_xor_ps_rmbk_512(<16 x float> %a, float* %ptr_b, <16 x float> %passThru, i16 %mask) {
-; CHECK-LABEL: test_mask_xor_ps_rmbk_512:
-; CHECK: ## BB#0:
-; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
-; CHECK-NEXT: vxorps (%rdi){1to16}, %zmm0, %zmm1 {%k1} ## encoding: [0x62,0xf1,0x7c,0x59,0x57,0x0f]
-; CHECK-NEXT: vmovaps %zmm1, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc1]
-; CHECK-NEXT: retq ## encoding: [0xc3]
- %q = load float, float* %ptr_b
- %vecinit.i = insertelement <16 x float> undef, float %q, i32 0
- %b = shufflevector <16 x float> %vecinit.i, <16 x float> undef, <16 x i32> zeroinitializer
- %res = call <16 x float> @llvm.x86.avx512.mask.xor.ps.512(<16 x float> %a, <16 x float> %b, <16 x float> %passThru, i16 %mask)
- ret <16 x float> %res
-}
-
-define <16 x float> @test_mask_xor_ps_rmbkz_512(<16 x float> %a, float* %ptr_b, i16 %mask) {
-; CHECK-LABEL: test_mask_xor_ps_rmbkz_512:
-; CHECK: ## BB#0:
-; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
-; CHECK-NEXT: vxorps (%rdi){1to16}, %zmm0, %zmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0xd9,0x57,0x07]
-; CHECK-NEXT: retq ## encoding: [0xc3]
- %q = load float, float* %ptr_b
- %vecinit.i = insertelement <16 x float> undef, float %q, i32 0
- %b = shufflevector <16 x float> %vecinit.i, <16 x float> undef, <16 x i32> zeroinitializer
- %res = call <16 x float> @llvm.x86.avx512.mask.xor.ps.512(<16 x float> %a, <16 x float> %b, <16 x float> zeroinitializer, i16 %mask)
- ret <16 x float> %res
-}
-
-declare <16 x float> @llvm.x86.avx512.mask.xor.ps.512(<16 x float>, <16 x float>, <16 x float>, i16)
-
declare <2 x i64> @llvm.x86.avx512.mask.cvtpd2qq.128(<2 x double>, <2 x i64>, i8)
define <2 x i64>@test_int_x86_avx512_mask_cvt_pd2qq_128(<2 x double> %x0, <2 x i64> %x1, i8 %x2) {
OpenPOWER on IntegriCloud