[AVX512] add PSLLD and PSLLQ Intrinsic

[oota-llvm.git] / test / CodeGen / X86 / avx512vl-intrinsics.ll

diff --git a/test/CodeGen/X86/avx512vl-intrinsics.ll b/test/CodeGen/X86/avx512vl-intrinsics.ll
index 7812148de1c8981a325e57711ca3aa6a0460c7c3..8ab34bd8c436e73d714817ae12356efc1481ff10 100644 (file)
--- a/test/CodeGen/X86/avx512vl-intrinsics.ll
+++ b/test/CodeGen/X86/avx512vl-intrinsics.ll
@@ -1867,7 +1867,7 @@ define <4 x i32> @test_mask_xor_epi32_rmbk_128(<4 x i32> %a, i32* %ptr_b, <4 x i
 
 define <4 x i32> @test_mask_xor_epi32_rmbkz_128(<4 x i32> %a, i32* %ptr_b, i8 %mask) {
   ;CHECK-LABEL: test_mask_xor_epi32_rmbkz_128
-  ;CHECK: vpxord  (%rdi){1to4}, %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0x99,0xef,0x07]  
+  ;CHECK: vpxord  (%rdi){1to4}, %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0x99,0xef,0x07]
   %q = load i32, i32* %ptr_b
   %vecinit.i = insertelement <4 x i32> undef, i32 %q, i32 0
   %b = shufflevector <4 x i32> %vecinit.i, <4 x i32> undef, <4 x i32> zeroinitializer
@@ -2299,7 +2299,7 @@ define <8 x float> @test_mm512_maskz_add_ps_256(<8 x float> %a0, <8 x float> %a1
 
 define <8 x float> @test_mm512_mask_add_ps_256(<8 x float> %a0, <8 x float> %a1, <8 x float> %src, i8 %mask) {
   ;CHECK-LABEL: test_mm512_mask_add_ps_256
-  ;CHECK: vaddps %ymm1, %ymm0, %ymm2 {%k1} 
+  ;CHECK: vaddps %ymm1, %ymm0, %ymm2 {%k1}
   %res = call <8 x float> @llvm.x86.avx512.mask.add.ps.256(<8 x float> %a0, <8 x float> %a1, <8 x float> %src, i8 %mask)
   ret <8 x float> %res
 }
@@ -2321,7 +2321,7 @@ define <4 x float> @test_mm512_maskz_add_ps_128(<4 x float> %a0, <4 x float> %a1
 
 define <4 x float> @test_mm512_mask_add_ps_128(<4 x float> %a0, <4 x float> %a1, <4 x float> %src, i8 %mask) {
   ;CHECK-LABEL: test_mm512_mask_add_ps_128
-  ;CHECK: vaddps %xmm1, %xmm0, %xmm2 {%k1} 
+  ;CHECK: vaddps %xmm1, %xmm0, %xmm2 {%k1}
   %res = call <4 x float> @llvm.x86.avx512.mask.add.ps.128(<4 x float> %a0, <4 x float> %a1, <4 x float> %src, i8 %mask)
   ret <4 x float> %res
 }
@@ -2343,7 +2343,7 @@ define <8 x float> @test_mm512_maskz_sub_ps_256(<8 x float> %a0, <8 x float> %a1
 
 define <8 x float> @test_mm512_mask_sub_ps_256(<8 x float> %a0, <8 x float> %a1, <8 x float> %src, i8 %mask) {
   ;CHECK-LABEL: test_mm512_mask_sub_ps_256
-  ;CHECK: vsubps %ymm1, %ymm0, %ymm2 {%k1} 
+  ;CHECK: vsubps %ymm1, %ymm0, %ymm2 {%k1}
   %res = call <8 x float> @llvm.x86.avx512.mask.sub.ps.256(<8 x float> %a0, <8 x float> %a1, <8 x float> %src, i8 %mask)
   ret <8 x float> %res
 }
@@ -2365,7 +2365,7 @@ define <4 x float> @test_mm512_maskz_sub_ps_128(<4 x float> %a0, <4 x float> %a1
 
 define <4 x float> @test_mm512_mask_sub_ps_128(<4 x float> %a0, <4 x float> %a1, <4 x float> %src, i8 %mask) {
   ;CHECK-LABEL: test_mm512_mask_sub_ps_128
-  ;CHECK: vsubps %xmm1, %xmm0, %xmm2 {%k1} 
+  ;CHECK: vsubps %xmm1, %xmm0, %xmm2 {%k1}
   %res = call <4 x float> @llvm.x86.avx512.mask.sub.ps.128(<4 x float> %a0, <4 x float> %a1, <4 x float> %src, i8 %mask)
   ret <4 x float> %res
 }
@@ -2387,7 +2387,7 @@ define <8 x float> @test_mm512_maskz_mul_ps_256(<8 x float> %a0, <8 x float> %a1
 
 define <8 x float> @test_mm512_mask_mul_ps_256(<8 x float> %a0, <8 x float> %a1, <8 x float> %src, i8 %mask) {
   ;CHECK-LABEL: test_mm512_mask_mul_ps_256
-  ;CHECK: vmulps %ymm1, %ymm0, %ymm2 {%k1} 
+  ;CHECK: vmulps %ymm1, %ymm0, %ymm2 {%k1}
   %res = call <8 x float> @llvm.x86.avx512.mask.mul.ps.256(<8 x float> %a0, <8 x float> %a1, <8 x float> %src, i8 %mask)
   ret <8 x float> %res
 }
@@ -2409,7 +2409,7 @@ define <4 x float> @test_mm512_maskz_mul_ps_128(<4 x float> %a0, <4 x float> %a1
 
 define <4 x float> @test_mm512_mask_mul_ps_128(<4 x float> %a0, <4 x float> %a1, <4 x float> %src, i8 %mask) {
   ;CHECK-LABEL: test_mm512_mask_mul_ps_128
-  ;CHECK: vmulps %xmm1, %xmm0, %xmm2 {%k1} 
+  ;CHECK: vmulps %xmm1, %xmm0, %xmm2 {%k1}
   %res = call <4 x float> @llvm.x86.avx512.mask.mul.ps.128(<4 x float> %a0, <4 x float> %a1, <4 x float> %src, i8 %mask)
   ret <4 x float> %res
 }
@@ -2431,7 +2431,7 @@ define <8 x float> @test_mm512_maskz_div_ps_256(<8 x float> %a0, <8 x float> %a1
 
 define <8 x float> @test_mm512_mask_div_ps_256(<8 x float> %a0, <8 x float> %a1, <8 x float> %src, i8 %mask) {
   ;CHECK-LABEL: test_mm512_mask_div_ps_256
-  ;CHECK: vdivps %ymm1, %ymm0, %ymm2 {%k1} 
+  ;CHECK: vdivps %ymm1, %ymm0, %ymm2 {%k1}
   %res = call <8 x float> @llvm.x86.avx512.mask.div.ps.256(<8 x float> %a0, <8 x float> %a1, <8 x float> %src, i8 %mask)
   ret <8 x float> %res
 }
@@ -2453,7 +2453,7 @@ define <4 x float> @test_mm512_maskz_div_ps_128(<4 x float> %a0, <4 x float> %a1
 
 define <4 x float> @test_mm512_mask_div_ps_128(<4 x float> %a0, <4 x float> %a1, <4 x float> %src, i8 %mask) {
   ;CHECK-LABEL: test_mm512_mask_div_ps_128
-  ;CHECK: vdivps %xmm1, %xmm0, %xmm2 {%k1} 
+  ;CHECK: vdivps %xmm1, %xmm0, %xmm2 {%k1}
   %res = call <4 x float> @llvm.x86.avx512.mask.div.ps.128(<4 x float> %a0, <4 x float> %a1, <4 x float> %src, i8 %mask)
   ret <4 x float> %res
 }
@@ -2475,7 +2475,7 @@ define <8 x float> @test_mm512_maskz_max_ps_256(<8 x float> %a0, <8 x float> %a1
 
 define <8 x float> @test_mm512_mask_max_ps_256(<8 x float> %a0, <8 x float> %a1, <8 x float> %src, i8 %mask) {
   ;CHECK-LABEL: test_mm512_mask_max_ps_256
-  ;CHECK: vmaxps %ymm1, %ymm0, %ymm2 {%k1} 
+  ;CHECK: vmaxps %ymm1, %ymm0, %ymm2 {%k1}
   %res = call <8 x float> @llvm.x86.avx512.mask.max.ps.256(<8 x float> %a0, <8 x float> %a1, <8 x float> %src, i8 %mask)
   ret <8 x float> %res
 }
@@ -2497,7 +2497,7 @@ define <4 x float> @test_mm512_maskz_max_ps_128(<4 x float> %a0, <4 x float> %a1
 
 define <4 x float> @test_mm512_mask_max_ps_128(<4 x float> %a0, <4 x float> %a1, <4 x float> %src, i8 %mask) {
   ;CHECK-LABEL: test_mm512_mask_max_ps_128
-  ;CHECK: vmaxps %xmm1, %xmm0, %xmm2 {%k1} 
+  ;CHECK: vmaxps %xmm1, %xmm0, %xmm2 {%k1}
   %res = call <4 x float> @llvm.x86.avx512.mask.max.ps.128(<4 x float> %a0, <4 x float> %a1, <4 x float> %src, i8 %mask)
   ret <4 x float> %res
 }
@@ -2519,7 +2519,7 @@ define <8 x float> @test_mm512_maskz_min_ps_256(<8 x float> %a0, <8 x float> %a1
 
 define <8 x float> @test_mm512_mask_min_ps_256(<8 x float> %a0, <8 x float> %a1, <8 x float> %src, i8 %mask) {
   ;CHECK-LABEL: test_mm512_mask_min_ps_256
-  ;CHECK: vminps %ymm1, %ymm0, %ymm2 {%k1} 
+  ;CHECK: vminps %ymm1, %ymm0, %ymm2 {%k1}
   %res = call <8 x float> @llvm.x86.avx512.mask.min.ps.256(<8 x float> %a0, <8 x float> %a1, <8 x float> %src, i8 %mask)
   ret <8 x float> %res
 }
@@ -2541,7 +2541,7 @@ define <4 x float> @test_mm512_maskz_min_ps_128(<4 x float> %a0, <4 x float> %a1
 
 define <4 x float> @test_mm512_mask_min_ps_128(<4 x float> %a0, <4 x float> %a1, <4 x float> %src, i8 %mask) {
   ;CHECK-LABEL: test_mm512_mask_min_ps_128
-  ;CHECK: vminps %xmm1, %xmm0, %xmm2 {%k1} 
+  ;CHECK: vminps %xmm1, %xmm0, %xmm2 {%k1}
   %res = call <4 x float> @llvm.x86.avx512.mask.min.ps.128(<4 x float> %a0, <4 x float> %a1, <4 x float> %src, i8 %mask)
   ret <4 x float> %res
 }
@@ -2591,9 +2591,9 @@ declare <8 x float> @llvm.x86.avx512.mask.getexp.ps.256(<8 x float>, <8 x float>
 declare <4 x i32> @llvm.x86.avx512.mask.pmaxs.d.128(<4 x i32>, <4 x i32>, <4 x i32>, i8)
 
 ; CHECK-LABEL: @test_int_x86_avx512_mask_pmaxs_d_128
-; CHECK-NOT: call 
+; CHECK-NOT: call
 ; CHECK: vpmaxsd %xmm
-; CHECK: {%k1} 
+; CHECK: {%k1}
 define <4 x i32>@test_int_x86_avx512_mask_pmaxs_d_128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x2, i8 %mask) {
   %res = call <4 x i32> @llvm.x86.avx512.mask.pmaxs.d.128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x2 ,i8 %mask)
   %res1 = call <4 x i32> @llvm.x86.avx512.mask.pmaxs.d.128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> zeroinitializer, i8 %mask)
@@ -2604,9 +2604,9 @@ define <4 x i32>@test_int_x86_avx512_mask_pmaxs_d_128(<4 x i32> %x0, <4 x i32> %
 declare <8 x i32> @llvm.x86.avx512.mask.pmaxs.d.256(<8 x i32>, <8 x i32>, <8 x i32>, i8)
 
 ; CHECK-LABEL: @test_int_x86_avx512_mask_pmaxs_d_256
-; CHECK-NOT: call 
+; CHECK-NOT: call
 ; CHECK: vpmaxsd %ymm
-; CHECK: {%k1} 
+; CHECK: {%k1}
 define <8 x i32>@test_int_x86_avx512_mask_pmaxs_d_256(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x2, i8 %x3) {
   %res = call <8 x i32> @llvm.x86.avx512.mask.pmaxs.d.256(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x2, i8 %x3)
   %res1 = call <8 x i32> @llvm.x86.avx512.mask.pmaxs.d.256(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x2, i8 -1)
@@ -2617,9 +2617,9 @@ define <8 x i32>@test_int_x86_avx512_mask_pmaxs_d_256(<8 x i32> %x0, <8 x i32> %
 declare <2 x i64> @llvm.x86.avx512.mask.pmaxs.q.128(<2 x i64>, <2 x i64>, <2 x i64>, i8)
 
 ; CHECK-LABEL: @test_int_x86_avx512_mask_pmaxs_q_128
-; CHECK-NOT: call 
+; CHECK-NOT: call
 ; CHECK: vpmaxsq %xmm
-; CHECK: {%k1} 
+; CHECK: {%k1}
 define <2 x i64>@test_int_x86_avx512_mask_pmaxs_q_128(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> %x2, i8 %x3) {
   %res = call <2 x i64> @llvm.x86.avx512.mask.pmaxs.q.128(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> %x2, i8 %x3)
   %res1 = call <2 x i64> @llvm.x86.avx512.mask.pmaxs.q.128(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> %x2, i8 -1)
@@ -2630,9 +2630,9 @@ define <2 x i64>@test_int_x86_avx512_mask_pmaxs_q_128(<2 x i64> %x0, <2 x i64> %
 declare <4 x i64> @llvm.x86.avx512.mask.pmaxs.q.256(<4 x i64>, <4 x i64>, <4 x i64>, i8)
 
 ; CHECK-LABEL: @test_int_x86_avx512_mask_pmaxs_q_256
-; CHECK-NOT: call 
+; CHECK-NOT: call
 ; CHECK: vpmaxsq %ymm
-; CHECK: {%k1} 
+; CHECK: {%k1}
 define <4 x i64>@test_int_x86_avx512_mask_pmaxs_q_256(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> %x2, i8 %mask) {
   %res = call <4 x i64> @llvm.x86.avx512.mask.pmaxs.q.256(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> %x2, i8 %mask)
   %res1 = call <4 x i64> @llvm.x86.avx512.mask.pmaxs.q.256(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> zeroinitializer, i8 %mask)
@@ -2643,9 +2643,9 @@ define <4 x i64>@test_int_x86_avx512_mask_pmaxs_q_256(<4 x i64> %x0, <4 x i64> %
 declare <4 x i32> @llvm.x86.avx512.mask.pmaxu.d.128(<4 x i32>, <4 x i32>, <4 x i32>, i8)
 
 ; CHECK-LABEL: @test_int_x86_avx512_mask_pmaxu_d_128
-; CHECK-NOT: call 
+; CHECK-NOT: call
 ; CHECK: vpmaxud %xmm
-; CHECK: {%k1} 
+; CHECK: {%k1}
 define <4 x i32>@test_int_x86_avx512_mask_pmaxu_d_128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x2,i8 %mask) {
   %res = call <4 x i32> @llvm.x86.avx512.mask.pmaxu.d.128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x2, i8 %mask)
   %res1 = call <4 x i32> @llvm.x86.avx512.mask.pmaxu.d.128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> zeroinitializer, i8 %mask)
@@ -2656,9 +2656,9 @@ define <4 x i32>@test_int_x86_avx512_mask_pmaxu_d_128(<4 x i32> %x0, <4 x i32> %
 declare <8 x i32> @llvm.x86.avx512.mask.pmaxu.d.256(<8 x i32>, <8 x i32>, <8 x i32>, i8)
 
 ; CHECK-LABEL: @test_int_x86_avx512_mask_pmaxu_d_256
-; CHECK-NOT: call 
+; CHECK-NOT: call
 ; CHECK: vpmaxud %ymm
-; CHECK: {%k1} 
+; CHECK: {%k1}
 define <8 x i32>@test_int_x86_avx512_mask_pmaxu_d_256(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x2, i8 %x3) {
   %res = call <8 x i32> @llvm.x86.avx512.mask.pmaxu.d.256(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x2, i8 %x3)
   %res1 = call <8 x i32> @llvm.x86.avx512.mask.pmaxu.d.256(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x2, i8 -1)
@@ -2669,9 +2669,9 @@ define <8 x i32>@test_int_x86_avx512_mask_pmaxu_d_256(<8 x i32> %x0, <8 x i32> %
 declare <2 x i64> @llvm.x86.avx512.mask.pmaxu.q.128(<2 x i64>, <2 x i64>, <2 x i64>, i8)
 
 ; CHECK-LABEL: @test_int_x86_avx512_mask_pmaxu_q_128
-; CHECK-NOT: call 
+; CHECK-NOT: call
 ; CHECK: vpmaxuq %xmm
-; CHECK: {%k1} 
+; CHECK: {%k1}
 define <2 x i64>@test_int_x86_avx512_mask_pmaxu_q_128(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> %x2, i8 %x3) {
   %res = call <2 x i64> @llvm.x86.avx512.mask.pmaxu.q.128(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> %x2, i8 %x3)
   %res1 = call <2 x i64> @llvm.x86.avx512.mask.pmaxu.q.128(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> %x2, i8 -1)
@@ -2682,9 +2682,9 @@ define <2 x i64>@test_int_x86_avx512_mask_pmaxu_q_128(<2 x i64> %x0, <2 x i64> %
 declare <4 x i64> @llvm.x86.avx512.mask.pmaxu.q.256(<4 x i64>, <4 x i64>, <4 x i64>, i8)
 
 ; CHECK-LABEL: @test_int_x86_avx512_mask_pmaxu_q_256
-; CHECK-NOT: call 
+; CHECK-NOT: call
 ; CHECK: vpmaxuq %ymm
-; CHECK: {%k1} 
+; CHECK: {%k1}
 define <4 x i64>@test_int_x86_avx512_mask_pmaxu_q_256(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> %x2, i8 %mask) {
   %res = call <4 x i64> @llvm.x86.avx512.mask.pmaxu.q.256(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> %x2, i8 %mask)
   %res1 = call <4 x i64> @llvm.x86.avx512.mask.pmaxu.q.256(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> zeroinitializer, i8 %mask)
@@ -2695,9 +2695,9 @@ define <4 x i64>@test_int_x86_avx512_mask_pmaxu_q_256(<4 x i64> %x0, <4 x i64> %
 declare <4 x i32> @llvm.x86.avx512.mask.pmins.d.128(<4 x i32>, <4 x i32>, <4 x i32>, i8)
 
 ; CHECK-LABEL: @test_int_x86_avx512_mask_pmins_d_128
-; CHECK-NOT: call 
+; CHECK-NOT: call
 ; CHECK: vpminsd %xmm
-; CHECK: {%k1} 
+; CHECK: {%k1}
 define <4 x i32>@test_int_x86_avx512_mask_pmins_d_128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x2, i8 %mask) {
   %res = call <4 x i32> @llvm.x86.avx512.mask.pmins.d.128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x2, i8 %mask)
   %res1 = call <4 x i32> @llvm.x86.avx512.mask.pmins.d.128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> zeroinitializer, i8 %mask)
@@ -2708,9 +2708,9 @@ define <4 x i32>@test_int_x86_avx512_mask_pmins_d_128(<4 x i32> %x0, <4 x i32> %
 declare <8 x i32> @llvm.x86.avx512.mask.pmins.d.256(<8 x i32>, <8 x i32>, <8 x i32>, i8)
 
 ; CHECK-LABEL: @test_int_x86_avx512_mask_pmins_d_256
-; CHECK-NOT: call 
+; CHECK-NOT: call
 ; CHECK: vpminsd %ymm
-; CHECK: {%k1} 
+; CHECK: {%k1}
 define <8 x i32>@test_int_x86_avx512_mask_pmins_d_256(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x2, i8 %x3) {
   %res = call <8 x i32> @llvm.x86.avx512.mask.pmins.d.256(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x2, i8 %x3)
   %res1 = call <8 x i32> @llvm.x86.avx512.mask.pmins.d.256(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x2, i8 -1)
@@ -2721,9 +2721,9 @@ define <8 x i32>@test_int_x86_avx512_mask_pmins_d_256(<8 x i32> %x0, <8 x i32> %
 declare <2 x i64> @llvm.x86.avx512.mask.pmins.q.128(<2 x i64>, <2 x i64>, <2 x i64>, i8)
 
 ; CHECK-LABEL: @test_int_x86_avx512_mask_pmins_q_128
-; CHECK-NOT: call 
+; CHECK-NOT: call
 ; CHECK: vpminsq %xmm
-; CHECK: {%k1} 
+; CHECK: {%k1}
 define <2 x i64>@test_int_x86_avx512_mask_pmins_q_128(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> %x2, i8 %x3) {
   %res = call <2 x i64> @llvm.x86.avx512.mask.pmins.q.128(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> %x2, i8 %x3)
   %res1 = call <2 x i64> @llvm.x86.avx512.mask.pmins.q.128(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> %x2, i8 -1)
@@ -2734,9 +2734,9 @@ define <2 x i64>@test_int_x86_avx512_mask_pmins_q_128(<2 x i64> %x0, <2 x i64> %
 declare <4 x i64> @llvm.x86.avx512.mask.pmins.q.256(<4 x i64>, <4 x i64>, <4 x i64>, i8)
 
 ; CHECK-LABEL: @test_int_x86_avx512_mask_pmins_q_256
-; CHECK-NOT: call 
+; CHECK-NOT: call
 ; CHECK: vpminsq %ymm
-; CHECK: {%k1} 
+; CHECK: {%k1}
 define <4 x i64>@test_int_x86_avx512_mask_pmins_q_256(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> %x2, i8 %mask) {
   %res = call <4 x i64> @llvm.x86.avx512.mask.pmins.q.256(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> %x2, i8 %mask)
   %res1 = call <4 x i64> @llvm.x86.avx512.mask.pmins.q.256(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> zeroinitializer, i8 %mask)
@@ -2747,9 +2747,9 @@ define <4 x i64>@test_int_x86_avx512_mask_pmins_q_256(<4 x i64> %x0, <4 x i64> %
 declare <4 x i32> @llvm.x86.avx512.mask.pminu.d.128(<4 x i32>, <4 x i32>, <4 x i32>, i8)
 
 ; CHECK-LABEL: @test_int_x86_avx512_mask_pminu_d_128
-; CHECK-NOT: call 
+; CHECK-NOT: call
 ; CHECK: vpminud %xmm
-; CHECK: {%k1} 
+; CHECK: {%k1}
 define <4 x i32>@test_int_x86_avx512_mask_pminu_d_128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x2, i8 %mask) {
   %res = call <4 x i32> @llvm.x86.avx512.mask.pminu.d.128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x2, i8 %mask)
   %res1 = call <4 x i32> @llvm.x86.avx512.mask.pminu.d.128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> zeroinitializer, i8 %mask)
@@ -2760,9 +2760,9 @@ define <4 x i32>@test_int_x86_avx512_mask_pminu_d_128(<4 x i32> %x0, <4 x i32> %
 declare <8 x i32> @llvm.x86.avx512.mask.pminu.d.256(<8 x i32>, <8 x i32>, <8 x i32>, i8)
 
 ; CHECK-LABEL: @test_int_x86_avx512_mask_pminu_d_256
-; CHECK-NOT: call 
+; CHECK-NOT: call
 ; CHECK: vpminud %ymm
-; CHECK: {%k1} 
+; CHECK: {%k1}
 define <8 x i32>@test_int_x86_avx512_mask_pminu_d_256(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x2, i8 %x3) {
   %res = call <8 x i32> @llvm.x86.avx512.mask.pminu.d.256(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x2, i8 %x3)
   %res1 = call <8 x i32> @llvm.x86.avx512.mask.pminu.d.256(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x2, i8 -1)
@@ -2773,9 +2773,9 @@ define <8 x i32>@test_int_x86_avx512_mask_pminu_d_256(<8 x i32> %x0, <8 x i32> %
 declare <2 x i64> @llvm.x86.avx512.mask.pminu.q.128(<2 x i64>, <2 x i64>, <2 x i64>, i8)
 
 ; CHECK-LABEL: @test_int_x86_avx512_mask_pminu_q_128
-; CHECK-NOT: call 
+; CHECK-NOT: call
 ; CHECK: vpminuq %xmm
-; CHECK: {%k1} 
+; CHECK: {%k1}
 define <2 x i64>@test_int_x86_avx512_mask_pminu_q_128(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> %x2, i8 %x3) {
   %res = call <2 x i64> @llvm.x86.avx512.mask.pminu.q.128(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> %x2, i8 %x3)
   %res1 = call <2 x i64> @llvm.x86.avx512.mask.pminu.q.128(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> %x2, i8 -1)
@@ -2786,9 +2786,9 @@ define <2 x i64>@test_int_x86_avx512_mask_pminu_q_128(<2 x i64> %x0, <2 x i64> %
 declare <4 x i64> @llvm.x86.avx512.mask.pminu.q.256(<4 x i64>, <4 x i64>, <4 x i64>, i8)
 
 ; CHECK-LABEL: @test_int_x86_avx512_mask_pminu_q_256
-; CHECK-NOT: call 
+; CHECK-NOT: call
 ; CHECK: vpminuq %ymm
-; CHECK: {%k1} 
+; CHECK: {%k1}
 define <4 x i64>@test_int_x86_avx512_mask_pminu_q_256(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> %x2, i8 %mask) {
   %res = call <4 x i64> @llvm.x86.avx512.mask.pminu.q.256(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> %x2, i8 %mask)
   %res1 = call <4 x i64> @llvm.x86.avx512.mask.pminu.q.256(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> zeroinitializer, i8 %mask)
@@ -2799,8 +2799,8 @@ define <4 x i64>@test_int_x86_avx512_mask_pminu_q_256(<4 x i64> %x0, <4 x i64> %
 declare <4 x i32> @llvm.x86.avx512.mask.vpermt2var.d.128(<4 x i32>, <4 x i32>, <4 x i32>, i8)
 
 ; CHECK-LABEL: @test_int_x86_avx512_mask_vpermt2var_d_128
-; CHECK-NOT: call 
-; CHECK: kmov 
+; CHECK-NOT: call
+; CHECK: kmov
 ; CHECK: vpermt2d %xmm{{.*}}{%k1}
 ; CHECK-NOT: {z}
 define <4 x i32>@test_int_x86_avx512_mask_vpermt2var_d_128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x2, i8 %x3) {
@@ -2813,8 +2813,8 @@ define <4 x i32>@test_int_x86_avx512_mask_vpermt2var_d_128(<4 x i32> %x0, <4 x i
 declare <4 x i32> @llvm.x86.avx512.maskz.vpermt2var.d.128(<4 x i32>, <4 x i32>, <4 x i32>, i8)
 
 ; CHECK-LABEL: @test_int_x86_avx512_maskz_vpermt2var_d_128
-; CHECK-NOT: call 
-; CHECK: kmov 
+; CHECK-NOT: call
+; CHECK: kmov
 ; CHECK: vpermt2d %xmm{{.*}}{%k1} {z}
 define <4 x i32>@test_int_x86_avx512_maskz_vpermt2var_d_128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x2, i8 %x3) {
   %res = call <4 x i32> @llvm.x86.avx512.maskz.vpermt2var.d.128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x2, i8 %x3)
@@ -2826,8 +2826,8 @@ define <4 x i32>@test_int_x86_avx512_maskz_vpermt2var_d_128(<4 x i32> %x0, <4 x
 declare <8 x i32> @llvm.x86.avx512.mask.vpermt2var.d.256(<8 x i32>, <8 x i32>, <8 x i32>, i8)
 
 ; CHECK-LABEL: @test_int_x86_avx512_mask_vpermt2var_d_256
-; CHECK-NOT: call 
-; CHECK: kmov 
+; CHECK-NOT: call
+; CHECK: kmov
 ; CHECK: vpermt2d %ymm{{.*}}{%k1}
 ; CHECK-NOT: {z}
 define <8 x i32>@test_int_x86_avx512_mask_vpermt2var_d_256(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x2, i8 %x3) {
@@ -2840,8 +2840,8 @@ define <8 x i32>@test_int_x86_avx512_mask_vpermt2var_d_256(<8 x i32> %x0, <8 x i
 declare <8 x i32> @llvm.x86.avx512.maskz.vpermt2var.d.256(<8 x i32>, <8 x i32>, <8 x i32>, i8)
 
 ; CHECK-LABEL: @test_int_x86_avx512_maskz_vpermt2var_d_256
-; CHECK-NOT: call 
-; CHECK: kmov 
+; CHECK-NOT: call
+; CHECK: kmov
 ; CHECK: vpermt2d {{.*}}{%k1} {z}
 define <8 x i32>@test_int_x86_avx512_maskz_vpermt2var_d_256(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x2, i8 %x3) {
   %res = call <8 x i32> @llvm.x86.avx512.maskz.vpermt2var.d.256(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x2, i8 %x3)
@@ -2853,9 +2853,9 @@ define <8 x i32>@test_int_x86_avx512_maskz_vpermt2var_d_256(<8 x i32> %x0, <8 x
 declare <2 x double> @llvm.x86.avx512.mask.vpermi2var.pd.128(<2 x double>, <2 x i64>, <2 x double>, i8)
 
 ; CHECK-LABEL: @test_int_x86_avx512_mask_vpermi2var_pd_128
-; CHECK-NOT: call 
-; CHECK: kmov 
-; CHECK: vpermi2pd %xmm{{.*}}{%k1} 
+; CHECK-NOT: call
+; CHECK: kmov
+; CHECK: vpermi2pd %xmm{{.*}}{%k1}
 define <2 x double>@test_int_x86_avx512_mask_vpermi2var_pd_128(<2 x double> %x0, <2 x i64> %x1, <2 x double> %x2, i8 %x3) {
   %res = call <2 x double> @llvm.x86.avx512.mask.vpermi2var.pd.128(<2 x double> %x0, <2 x i64> %x1, <2 x double> %x2, i8 %x3)
   %res1 = call <2 x double> @llvm.x86.avx512.mask.vpermi2var.pd.128(<2 x double> %x0, <2 x i64> %x1, <2 x double> %x2, i8 -1)
@@ -2866,9 +2866,9 @@ define <2 x double>@test_int_x86_avx512_mask_vpermi2var_pd_128(<2 x double> %x0,
 declare <4 x double> @llvm.x86.avx512.mask.vpermi2var.pd.256(<4 x double>, <4 x i64>, <4 x double>, i8)
 
 ; CHECK-LABEL: @test_int_x86_avx512_mask_vpermi2var_pd_256
-; CHECK-NOT: call 
-; CHECK: kmov 
-; CHECK: vpermi2pd %ymm{{.*}}{%k1} 
+; CHECK-NOT: call
+; CHECK: kmov
+; CHECK: vpermi2pd %ymm{{.*}}{%k1}
 define <4 x double>@test_int_x86_avx512_mask_vpermi2var_pd_256(<4 x double> %x0, <4 x i64> %x1, <4 x double> %x2, i8 %x3) {
   %res = call <4 x double> @llvm.x86.avx512.mask.vpermi2var.pd.256(<4 x double> %x0, <4 x i64> %x1, <4 x double> %x2, i8 %x3)
   %res1 = call <4 x double> @llvm.x86.avx512.mask.vpermi2var.pd.256(<4 x double> %x0, <4 x i64> %x1, <4 x double> %x2, i8 -1)
@@ -2879,9 +2879,9 @@ define <4 x double>@test_int_x86_avx512_mask_vpermi2var_pd_256(<4 x double> %x0,
 declare <4 x float> @llvm.x86.avx512.mask.vpermi2var.ps.128(<4 x float>, <4 x i32>, <4 x float>, i8)
 
 ; CHECK-LABEL: @test_int_x86_avx512_mask_vpermi2var_ps_128
-; CHECK-NOT: call 
-; CHECK: kmov 
-; CHECK: vpermi2ps %xmm{{.*}}{%k1} 
+; CHECK-NOT: call
+; CHECK: kmov
+; CHECK: vpermi2ps %xmm{{.*}}{%k1}
 define <4 x float>@test_int_x86_avx512_mask_vpermi2var_ps_128(<4 x float> %x0, <4 x i32> %x1, <4 x float> %x2, i8 %x3) {
   %res = call <4 x float> @llvm.x86.avx512.mask.vpermi2var.ps.128(<4 x float> %x0, <4 x i32> %x1, <4 x float> %x2, i8 %x3)
   %res1 = call <4 x float> @llvm.x86.avx512.mask.vpermi2var.ps.128(<4 x float> %x0, <4 x i32> %x1, <4 x float> %x2, i8 -1)
@@ -2892,9 +2892,9 @@ define <4 x float>@test_int_x86_avx512_mask_vpermi2var_ps_128(<4 x float> %x0, <
 declare <8 x float> @llvm.x86.avx512.mask.vpermi2var.ps.256(<8 x float>, <8 x i32>, <8 x float>, i8)
 
 ; CHECK-LABEL: @test_int_x86_avx512_mask_vpermi2var_ps_256
-; CHECK-NOT: call 
-; CHECK: kmov 
-; CHECK: vpermi2ps %ymm{{.*}}{%k1} 
+; CHECK-NOT: call
+; CHECK: kmov
+; CHECK: vpermi2ps %ymm{{.*}}{%k1}
 define <8 x float>@test_int_x86_avx512_mask_vpermi2var_ps_256(<8 x float> %x0, <8 x i32> %x1, <8 x float> %x2, i8 %x3) {
   %res = call <8 x float> @llvm.x86.avx512.mask.vpermi2var.ps.256(<8 x float> %x0, <8 x i32> %x1, <8 x float> %x2, i8 %x3)
   %res1 = call <8 x float> @llvm.x86.avx512.mask.vpermi2var.ps.256(<8 x float> %x0, <8 x i32> %x1, <8 x float> %x2, i8 -1)
@@ -2905,9 +2905,9 @@ define <8 x float>@test_int_x86_avx512_mask_vpermi2var_ps_256(<8 x float> %x0, <
 declare <2 x i64> @llvm.x86.avx512.mask.pabs.q.128(<2 x i64>, <2 x i64>, i8)
 
 ; CHECK-LABEL: @test_int_x86_avx512_mask_pabs_q_128
-; CHECK-NOT: call 
-; CHECK: kmov 
-; CHECK: vpabsq{{.*}}{%k1} 
+; CHECK-NOT: call
+; CHECK: kmov
+; CHECK: vpabsq{{.*}}{%k1}
 define <2 x i64>@test_int_x86_avx512_mask_pabs_q_128(<2 x i64> %x0, <2 x i64> %x1, i8 %x2) {
   %res = call <2 x i64> @llvm.x86.avx512.mask.pabs.q.128(<2 x i64> %x0, <2 x i64> %x1, i8 %x2)
   %res1 = call <2 x i64> @llvm.x86.avx512.mask.pabs.q.128(<2 x i64> %x0, <2 x i64> %x1, i8 -1)
@@ -2918,9 +2918,9 @@ define <2 x i64>@test_int_x86_avx512_mask_pabs_q_128(<2 x i64> %x0, <2 x i64> %x
 declare <4 x i64> @llvm.x86.avx512.mask.pabs.q.256(<4 x i64>, <4 x i64>, i8)
 
 ; CHECK-LABEL: @test_int_x86_avx512_mask_pabs_q_256
-; CHECK-NOT: call 
-; CHECK: kmov 
-; CHECK: vpabsq{{.*}}{%k1} 
+; CHECK-NOT: call
+; CHECK: kmov
+; CHECK: vpabsq{{.*}}{%k1}
 define <4 x i64>@test_int_x86_avx512_mask_pabs_q_256(<4 x i64> %x0, <4 x i64> %x1, i8 %x2) {
   %res = call <4 x i64> @llvm.x86.avx512.mask.pabs.q.256(<4 x i64> %x0, <4 x i64> %x1, i8 %x2)
   %res1 = call <4 x i64> @llvm.x86.avx512.mask.pabs.q.256(<4 x i64> %x0, <4 x i64> %x1, i8 -1)
@@ -2931,9 +2931,9 @@ define <4 x i64>@test_int_x86_avx512_mask_pabs_q_256(<4 x i64> %x0, <4 x i64> %x
 declare <4 x i32> @llvm.x86.avx512.mask.pabs.d.128(<4 x i32>, <4 x i32>, i8)
 
 ; CHECK-LABEL: @test_int_x86_avx512_mask_pabs_d_128
-; CHECK-NOT: call 
-; CHECK: kmov 
-; CHECK: vpabsd{{.*}}{%k1} 
+; CHECK-NOT: call
+; CHECK: kmov
+; CHECK: vpabsd{{.*}}{%k1}
 define <4 x i32>@test_int_x86_avx512_mask_pabs_d_128(<4 x i32> %x0, <4 x i32> %x1, i8 %x2) {
   %res = call <4 x i32> @llvm.x86.avx512.mask.pabs.d.128(<4 x i32> %x0, <4 x i32> %x1, i8 %x2)
   %res1 = call <4 x i32> @llvm.x86.avx512.mask.pabs.d.128(<4 x i32> %x0, <4 x i32> %x1, i8 -1)
@@ -2944,9 +2944,9 @@ define <4 x i32>@test_int_x86_avx512_mask_pabs_d_128(<4 x i32> %x0, <4 x i32> %x
 declare <8 x i32> @llvm.x86.avx512.mask.pabs.d.256(<8 x i32>, <8 x i32>, i8)
 
 ; CHECK-LABEL: @test_int_x86_avx512_mask_pabs_d_256
-; CHECK-NOT: call 
-; CHECK: kmov 
-; CHECK: vpabsd{{.*}}{%k1} 
+; CHECK-NOT: call
+; CHECK: kmov
+; CHECK: vpabsd{{.*}}{%k1}
 define <8 x i32>@test_int_x86_avx512_mask_pabs_d_256(<8 x i32> %x0, <8 x i32> %x1, i8 %x2) {
   %res = call <8 x i32> @llvm.x86.avx512.mask.pabs.d.256(<8 x i32> %x0, <8 x i32> %x1, i8 %x2)
   %res1 = call <8 x i32> @llvm.x86.avx512.mask.pabs.d.256(<8 x i32> %x0, <8 x i32> %x1, i8 -1)
@@ -2958,9 +2958,9 @@ define <8 x i32>@test_int_x86_avx512_mask_pabs_d_256(<8 x i32> %x0, <8 x i32> %x
 declare <2 x double> @llvm.x86.avx512.mask.scalef.pd.128(<2 x double>, <2 x double>, <2 x double>, i8)
 
 ; CHECK-LABEL: @test_int_x86_avx512_mask_scalef_pd_128
-; CHECK-NOT: call 
-; CHECK: kmov 
-; CHECK: vscalefpd{{.*}}{%k1} 
+; CHECK-NOT: call
+; CHECK: kmov
+; CHECK: vscalefpd{{.*}}{%k1}
 define <2 x double>@test_int_x86_avx512_mask_scalef_pd_128(<2 x double> %x0, <2 x double> %x1, <2 x double> %x2, i8 %x3) {
   %res = call <2 x double> @llvm.x86.avx512.mask.scalef.pd.128(<2 x double> %x0, <2 x double> %x1, <2 x double> %x2, i8 %x3)
   %res1 = call <2 x double> @llvm.x86.avx512.mask.scalef.pd.128(<2 x double> %x0, <2 x double> %x1, <2 x double> %x2, i8 -1)
@@ -2971,9 +2971,9 @@ define <2 x double>@test_int_x86_avx512_mask_scalef_pd_128(<2 x double> %x0, <2
 declare <4 x double> @llvm.x86.avx512.mask.scalef.pd.256(<4 x double>, <4 x double>, <4 x double>, i8)
 
 ; CHECK-LABEL: @test_int_x86_avx512_mask_scalef_pd_256
-; CHECK-NOT: call 
-; CHECK: kmov 
-; CHECK: vscalefpd{{.*}}{%k1} 
+; CHECK-NOT: call
+; CHECK: kmov
+; CHECK: vscalefpd{{.*}}{%k1}
 define <4 x double>@test_int_x86_avx512_mask_scalef_pd_256(<4 x double> %x0, <4 x double> %x1, <4 x double> %x2, i8 %x3) {
   %res = call <4 x double> @llvm.x86.avx512.mask.scalef.pd.256(<4 x double> %x0, <4 x double> %x1, <4 x double> %x2, i8 %x3)
   %res1 = call <4 x double> @llvm.x86.avx512.mask.scalef.pd.256(<4 x double> %x0, <4 x double> %x1, <4 x double> %x2, i8 -1)
@@ -2983,9 +2983,9 @@ define <4 x double>@test_int_x86_avx512_mask_scalef_pd_256(<4 x double> %x0, <4
 
 declare <4 x float> @llvm.x86.avx512.mask.scalef.ps.128(<4 x float>, <4 x float>, <4 x float>, i8)
 ; CHECK-LABEL: @test_int_x86_avx512_mask_scalef_ps_128
-; CHECK-NOT: call 
-; CHECK: kmov 
-; CHECK: vscalefps{{.*}}{%k1} 
+; CHECK-NOT: call
+; CHECK: kmov
+; CHECK: vscalefps{{.*}}{%k1}
 define <4 x float>@test_int_x86_avx512_mask_scalef_ps_128(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2, i8 %x3) {
   %res = call <4 x float> @llvm.x86.avx512.mask.scalef.ps.128(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2, i8 %x3)
   %res1 = call <4 x float> @llvm.x86.avx512.mask.scalef.ps.128(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2, i8 -1)
@@ -2995,9 +2995,9 @@ define <4 x float>@test_int_x86_avx512_mask_scalef_ps_128(<4 x float> %x0, <4 x
 
 declare <8 x float> @llvm.x86.avx512.mask.scalef.ps.256(<8 x float>, <8 x float>, <8 x float>, i8)
 ; CHECK-LABEL: @test_int_x86_avx512_mask_scalef_ps_256
-; CHECK-NOT: call 
-; CHECK: kmov 
-; CHECK: vscalefps{{.*}}{%k1} 
+; CHECK-NOT: call
+; CHECK: kmov
+; CHECK: vscalefps{{.*}}{%k1}
 define <8 x float>@test_int_x86_avx512_mask_scalef_ps_256(<8 x float> %x0, <8 x float> %x1, <8 x float> %x2, i8 %x3) {
   %res = call <8 x float> @llvm.x86.avx512.mask.scalef.ps.256(<8 x float> %x0, <8 x float> %x1, <8 x float> %x2, i8 %x3)
   %res1 = call <8 x float> @llvm.x86.avx512.mask.scalef.ps.256(<8 x float> %x0, <8 x float> %x1, <8 x float> %x2, i8 -1)
@@ -3005,6 +3005,232 @@ define <8 x float>@test_int_x86_avx512_mask_scalef_ps_256(<8 x float> %x0, <8 x
   ret <8 x float> %res2
 }
 
+declare <2 x double> @llvm.x86.avx512.mask.unpckh.pd.128(<2 x double>, <2 x double>, <2 x double>, i8)
+
+define <2 x double>@test_int_x86_avx512_mask_unpckh_pd_128(<2 x double> %x0, <2 x double> %x1, <2 x double> %x2, i8 %x3) {
+; CHECK-LABEL: test_int_x86_avx512_mask_unpckh_pd_128:
+; CHECK:         vunpckhpd %xmm1, %xmm0, %xmm2 {%k1}
+; CHECK-NEXT:    ## xmm2 = xmm2[1],k1[1]
+; CHECK-NEXT:    vunpckhpd %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0xfd,0x08,0x15,0xc1]
+; CHECK-NEXT:    ## xmm0 = xmm0[1],xmm1[1]
+  %res = call <2 x double> @llvm.x86.avx512.mask.unpckh.pd.128(<2 x double> %x0, <2 x double> %x1, <2 x double> %x2, i8 %x3)
+  %res1 = call <2 x double> @llvm.x86.avx512.mask.unpckh.pd.128(<2 x double> %x0, <2 x double> %x1, <2 x double> %x2, i8 -1)
+  %res2 = fadd <2 x double> %res, %res1
+  ret <2 x double> %res2
+}
+
+declare <4 x double> @llvm.x86.avx512.mask.unpckh.pd.256(<4 x double>, <4 x double>, <4 x double>, i8)
+
+define <4 x double>@test_int_x86_avx512_mask_unpckh_pd_256(<4 x double> %x0, <4 x double> %x1, <4 x double> %x2, i8 %x3) {
+; CHECK-LABEL: test_int_x86_avx512_mask_unpckh_pd_256:
+; CHECK:         vunpckhpd %ymm1, %ymm0, %ymm2 {%k1}
+; CHECK-NEXT:    ## ymm2 = ymm2[1],k1[1],ymm2[3],k1[3]
+; CHECK-NEXT:    vunpckhpd %ymm1, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0xfd,0x28,0x15,0xc1]
+; CHECK-NEXT:    ## ymm0 = ymm0[1],ymm1[1],ymm0[3],ymm1[3]
+  %res = call <4 x double> @llvm.x86.avx512.mask.unpckh.pd.256(<4 x double> %x0, <4 x double> %x1, <4 x double> %x2, i8 %x3)
+  %res1 = call <4 x double> @llvm.x86.avx512.mask.unpckh.pd.256(<4 x double> %x0, <4 x double> %x1, <4 x double> %x2, i8 -1)
+  %res2 = fadd <4 x double> %res, %res1
+  ret <4 x double> %res2
+}
+
+declare <4 x float> @llvm.x86.avx512.mask.unpckh.ps.128(<4 x float>, <4 x float>, <4 x float>, i8)
+
+define <4 x float>@test_int_x86_avx512_mask_unpckh_ps_128(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2, i8 %x3) {
+; CHECK-LABEL: test_int_x86_avx512_mask_unpckh_ps_128:
+; CHECK:         vunpckhps %xmm1, %xmm0, %xmm2 {%k1}
+; CHECK-NEXT:    ## xmm2 = xmm2[2],k1[2],xmm2[3],k1[3]
+; CHECK-NEXT:    vunpckhps %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7c,0x08,0x15,0xc1]
+; CHECK-NEXT:    ## xmm0 = xmm0[2],xmm1[2],xmm0[3],xmm1[3]
+  %res = call <4 x float> @llvm.x86.avx512.mask.unpckh.ps.128(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2, i8 %x3)
+  %res1 = call <4 x float> @llvm.x86.avx512.mask.unpckh.ps.128(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2, i8 -1)
+  %res2 = fadd <4 x float> %res, %res1
+  ret <4 x float> %res2
+}
+
+declare <8 x float> @llvm.x86.avx512.mask.unpckh.ps.256(<8 x float>, <8 x float>, <8 x float>, i8)
+
+define <8 x float>@test_int_x86_avx512_mask_unpckh_ps_256(<8 x float> %x0, <8 x float> %x1, <8 x float> %x2, i8 %x3) {
+; CHECK-LABEL: test_int_x86_avx512_mask_unpckh_ps_256:
+; CHECK:       ## BB#0:
+; CHECK:         vunpckhps %ymm1, %ymm0, %ymm2 {%k1}
+; CHECK-NEXT:    ## ymm2 = ymm2[2],k1[2],ymm2[3],k1[3],ymm2[6],k1[6],ymm2[7],k1[7]
+; CHECK-NEXT:    vunpckhps %ymm1, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7c,0x28,0x15,0xc1]
+; CHECK-NEXT:    ## ymm0 = ymm0[2],ymm1[2],ymm0[3],ymm1[3],ymm0[6],ymm1[6],ymm0[7],ymm1[7]
+  %res = call <8 x float> @llvm.x86.avx512.mask.unpckh.ps.256(<8 x float> %x0, <8 x float> %x1, <8 x float> %x2, i8 %x3)
+  %res1 = call <8 x float> @llvm.x86.avx512.mask.unpckh.ps.256(<8 x float> %x0, <8 x float> %x1, <8 x float> %x2, i8 -1)
+  %res2 = fadd <8 x float> %res, %res1
+  ret <8 x float> %res2
+}
+
+declare <2 x double> @llvm.x86.avx512.mask.unpckl.pd.128(<2 x double>, <2 x double>, <2 x double>, i8)
+
+define <2 x double>@test_int_x86_avx512_mask_unpckl_pd_128(<2 x double> %x0, <2 x double> %x1, <2 x double> %x2, i8 %x3) {
+; CHECK-LABEL: test_int_x86_avx512_mask_unpckl_pd_128:
+; CHECK:         vunpcklpd %xmm1, %xmm0, %xmm2 {%k1}
+; CHECK-NEXT:    ## xmm2 = xmm2[0],k1[0]
+; CHECK-NEXT:    vunpcklpd %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0xfd,0x08,0x14,0xc1]
+; CHECK-NEXT:    ## xmm0 = xmm0[0],xmm1[0]
+  %res = call <2 x double> @llvm.x86.avx512.mask.unpckl.pd.128(<2 x double> %x0, <2 x double> %x1, <2 x double> %x2, i8 %x3)
+  %res1 = call <2 x double> @llvm.x86.avx512.mask.unpckl.pd.128(<2 x double> %x0, <2 x double> %x1, <2 x double> %x2, i8 -1)
+  %res2 = fadd <2 x double> %res, %res1
+  ret <2 x double> %res2
+}
+
+declare <4 x double> @llvm.x86.avx512.mask.unpckl.pd.256(<4 x double>, <4 x double>, <4 x double>, i8)
+
+define <4 x double>@test_int_x86_avx512_mask_unpckl_pd_256(<4 x double> %x0, <4 x double> %x1, <4 x double> %x2, i8 %x3) {
+; CHECK-LABEL: test_int_x86_avx512_mask_unpckl_pd_256:
+; CHECK:         vunpcklpd %ymm1, %ymm0, %ymm2 {%k1}
+; CHECK-NEXT:    ## ymm2 = ymm2[0],k1[0],ymm2[2],k1[2]
+; CHECK-NEXT:    vunpcklpd %ymm1, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0xfd,0x28,0x14,0xc1]
+; CHECK-NEXT:    ## ymm0 = ymm0[0],ymm1[0],ymm0[2],ymm1[2]
+  %res = call <4 x double> @llvm.x86.avx512.mask.unpckl.pd.256(<4 x double> %x0, <4 x double> %x1, <4 x double> %x2, i8 %x3)
+  %res1 = call <4 x double> @llvm.x86.avx512.mask.unpckl.pd.256(<4 x double> %x0, <4 x double> %x1, <4 x double> %x2, i8 -1)
+  %res2 = fadd <4 x double> %res, %res1
+  ret <4 x double> %res2
+}
+
+declare <4 x float> @llvm.x86.avx512.mask.unpckl.ps.128(<4 x float>, <4 x float>, <4 x float>, i8)
+
+define <4 x float>@test_int_x86_avx512_mask_unpckl_ps_128(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2, i8 %x3) {
+; CHECK-LABEL: test_int_x86_avx512_mask_unpckl_ps_128:
+; CHECK:         vunpcklps %xmm1, %xmm0, %xmm2 {%k1}
+; CHECK-NEXT:    ## xmm2 = xmm2[0],k1[0],xmm2[1],k1[1]
+; CHECK-NEXT:    vunpcklps %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7c,0x08,0x14,0xc1]
+; CHECK-NEXT:    ## xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1]
+  %res = call <4 x float> @llvm.x86.avx512.mask.unpckl.ps.128(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2, i8 %x3)
+  %res1 = call <4 x float> @llvm.x86.avx512.mask.unpckl.ps.128(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2, i8 -1)
+  %res2 = fadd <4 x float> %res, %res1
+  ret <4 x float> %res2
+}
+
+declare <8 x float> @llvm.x86.avx512.mask.unpckl.ps.256(<8 x float>, <8 x float>, <8 x float>, i8)
+
+define <8 x float>@test_int_x86_avx512_mask_unpckl_ps_256(<8 x float> %x0, <8 x float> %x1, <8 x float> %x2, i8 %x3) {
+; CHECK-LABEL: test_int_x86_avx512_mask_unpckl_ps_256:
+; CHECK:         vunpcklps %ymm1, %ymm0, %ymm2 {%k1}
+; CHECK-NEXT:    ## ymm2 = ymm2[0],k1[0],ymm2[1],k1[1],ymm2[4],k1[4],ymm2[5],k1[5]
+; CHECK-NEXT:    vunpcklps %ymm1, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7c,0x28,0x14,0xc1]
+; CHECK-NEXT:    ## ymm0 = ymm0[0],ymm1[0],ymm0[1],ymm1[1],ymm0[4],ymm1[4],ymm0[5],ymm1[5]
+  %res = call <8 x float> @llvm.x86.avx512.mask.unpckl.ps.256(<8 x float> %x0, <8 x float> %x1, <8 x float> %x2, i8 %x3)
+  %res1 = call <8 x float> @llvm.x86.avx512.mask.unpckl.ps.256(<8 x float> %x0, <8 x float> %x1, <8 x float> %x2, i8 -1)
+  %res2 = fadd <8 x float> %res, %res1
+  ret <8 x float> %res2
+}
+
+declare <4 x i32> @llvm.x86.avx512.mask.punpckhd.q.128(<4 x i32>, <4 x i32>, <4 x i32>, i8)
+
+define <4 x i32>@test_int_x86_avx512_mask_punpckhd_q_128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x2, i8 %x3) {
+; CHECK-LABEL: test_int_x86_avx512_mask_punpckhd_q_128:
+; CHECK:         vpunpckhdq %xmm1, %xmm0, %xmm2 {%k1}
+; CHECK-NEXT:    ## xmm2 = xmm2[2],k1[2],xmm2[3],k1[3]
+; CHECK-NEXT:    vpunpckhdq %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0x6a,0xc1]
+; CHECK-NEXT:    ## xmm0 = xmm0[2],xmm1[2],xmm0[3],xmm1[3]
+  %res = call <4 x i32> @llvm.x86.avx512.mask.punpckhd.q.128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x2, i8 %x3)
+  %res1 = call <4 x i32> @llvm.x86.avx512.mask.punpckhd.q.128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x2, i8 -1)
+  %res2 = add <4 x i32> %res, %res1
+  ret <4 x i32> %res2
+}
+
+declare <4 x i32> @llvm.x86.avx512.mask.punpckld.q.128(<4 x i32>, <4 x i32>, <4 x i32>, i8)
+
+define <4 x i32>@test_int_x86_avx512_mask_punpckld_q_128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x2, i8 %x3) {
+; CHECK-LABEL: test_int_x86_avx512_mask_punpckld_q_128:
+; CHECK:         vpunpckldq %xmm1, %xmm0, %xmm2 {%k1}
+; CHECK-NEXT:    ## xmm2 = xmm2[0],k1[0],xmm2[1],k1[1]
+; CHECK-NEXT:    vpunpckldq %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0x62,0xc1]
+; CHECK-NEXT:    ## xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1]
+  %res = call <4 x i32> @llvm.x86.avx512.mask.punpckld.q.128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x2, i8 %x3)
+  %res1 = call <4 x i32> @llvm.x86.avx512.mask.punpckld.q.128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x2, i8 -1)
+  %res2 = add <4 x i32> %res, %res1
+  ret <4 x i32> %res2
+}
+
+declare <8 x i32> @llvm.x86.avx512.mask.punpckhd.q.256(<8 x i32>, <8 x i32>, <8 x i32>, i8)
+
+define <8 x i32>@test_int_x86_avx512_mask_punpckhd_q_256(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x2, i8 %x3) {
+; CHECK-LABEL: test_int_x86_avx512_mask_punpckhd_q_256:
+; CHECK:       ## BB#0:
+; CHECK:         vpunpckhdq %ymm1, %ymm0, %ymm2 {%k1}
+; CHECK-NEXT:    ## ymm2 = ymm2[2],k1[2],ymm2[3],k1[3],ymm2[6],k1[6],ymm2[7],k1[7]
+; CHECK-NEXT:    vpunpckhdq %ymm1, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7d,0x28,0x6a,0xc1]
+; CHECK-NEXT:    ## ymm0 = ymm0[2],ymm1[2],ymm0[3],ymm1[3],ymm0[6],ymm1[6],ymm0[7],ymm1[7]
+  %res = call <8 x i32> @llvm.x86.avx512.mask.punpckhd.q.256(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x2, i8 %x3)
+  %res1 = call <8 x i32> @llvm.x86.avx512.mask.punpckhd.q.256(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x2, i8 -1)
+  %res2 = add <8 x i32> %res, %res1
+  ret <8 x i32> %res2
+}
+
+declare <8 x i32> @llvm.x86.avx512.mask.punpckld.q.256(<8 x i32>, <8 x i32>, <8 x i32>, i8)
+
+define <8 x i32>@test_int_x86_avx512_mask_punpckld_q_256(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x2, i8 %x3) {
+; CHECK-LABEL: test_int_x86_avx512_mask_punpckld_q_256:
+; CHECK:         vpunpckldq %ymm1, %ymm0, %ymm2 {%k1}
+; CHECK-NEXT:    ## ymm2 = ymm2[0],k1[0],ymm2[1],k1[1],ymm2[4],k1[4],ymm2[5],k1[5]
+; CHECK-NEXT:    vpunpckldq %ymm1, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7d,0x28,0x62,0xc1]
+; CHECK-NEXT:    ## ymm0 = ymm0[0],ymm1[0],ymm0[1],ymm1[1],ymm0[4],ymm1[4],ymm0[5],ymm1[5]
+  %res = call <8 x i32> @llvm.x86.avx512.mask.punpckld.q.256(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x2, i8 %x3)
+  %res1 = call <8 x i32> @llvm.x86.avx512.mask.punpckld.q.256(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x2, i8 -1)
+  %res2 = add <8 x i32> %res, %res1
+  ret <8 x i32> %res2
+}
+
+declare <2 x i64> @llvm.x86.avx512.mask.punpckhqd.q.128(<2 x i64>, <2 x i64>, <2 x i64>, i8)
+
+define <2 x i64>@test_int_x86_avx512_mask_punpckhqd_q_128(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> %x2, i8 %x3) {
+; CHECK-LABEL: test_int_x86_avx512_mask_punpckhqd_q_128:
+; CHECK:         vpunpckhqdq %xmm1, %xmm0, %xmm2 {%k1}
+; CHECK-NEXT:    ## xmm2 = xmm2[1],k1[1]
+; CHECK-NEXT:    vpunpckhqdq %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0xfd,0x08,0x6d,0xc1]
+; CHECK-NEXT:    ## xmm0 = xmm0[1],xmm1[1]
+  %res = call <2 x i64> @llvm.x86.avx512.mask.punpckhqd.q.128(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> %x2, i8 %x3)
+  %res1 = call <2 x i64> @llvm.x86.avx512.mask.punpckhqd.q.128(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> %x2, i8 -1)
+  %res2 = add <2 x i64> %res, %res1
+  ret <2 x i64> %res2
+}
+
+declare <2 x i64> @llvm.x86.avx512.mask.punpcklqd.q.128(<2 x i64>, <2 x i64>, <2 x i64>, i8)
+
+define <2 x i64>@test_int_x86_avx512_mask_punpcklqd_q_128(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> %x2, i8 %x3) {
+; CHECK-LABEL: test_int_x86_avx512_mask_punpcklqd_q_128:
+; CHECK:         vpunpcklqdq %xmm1, %xmm0, %xmm2 {%k1}
+; CHECK-NEXT:    ## xmm2 = xmm2[0],k1[0]
+; CHECK-NEXT:    vpunpcklqdq %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0xfd,0x08,0x6c,0xc1]
+; CHECK-NEXT:    ## xmm0 = xmm0[0],xmm1[0]
+  %res = call <2 x i64> @llvm.x86.avx512.mask.punpcklqd.q.128(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> %x2, i8 %x3)
+  %res1 = call <2 x i64> @llvm.x86.avx512.mask.punpcklqd.q.128(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> %x2, i8 -1)
+  %res2 = add <2 x i64> %res, %res1
+  ret <2 x i64> %res2
+}
+
+declare <4 x i64> @llvm.x86.avx512.mask.punpcklqd.q.256(<4 x i64>, <4 x i64>, <4 x i64>, i8)
+
+define <4 x i64>@test_int_x86_avx512_mask_punpcklqd_q_256(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> %x2, i8 %x3) {
+; CHECK-LABEL: test_int_x86_avx512_mask_punpcklqd_q_256:
+; CHECK:         vpunpcklqdq %ymm1, %ymm0, %ymm2 {%k1}
+; CHECK-NEXT:    ## ymm2 = ymm2[0],k1[0],ymm2[2],k1[2]
+; CHECK-NEXT:    vpunpcklqdq %ymm1, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0xfd,0x28,0x6c,0xc1]
+; CHECK-NEXT:    ## ymm0 = ymm0[0],ymm1[0],ymm0[2],ymm1[2]
+  %res = call <4 x i64> @llvm.x86.avx512.mask.punpcklqd.q.256(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> %x2, i8 %x3)
+  %res1 = call <4 x i64> @llvm.x86.avx512.mask.punpcklqd.q.256(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> %x2, i8 -1)
+  %res2 = add <4 x i64> %res, %res1
+  ret <4 x i64> %res2
+}
+
+declare <4 x i64> @llvm.x86.avx512.mask.punpckhqd.q.256(<4 x i64>, <4 x i64>, <4 x i64>, i8)
+
+define <4 x i64>@test_int_x86_avx512_mask_punpckhqd_q_256(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> %x2, i8 %x3) {
+; CHECK-LABEL: test_int_x86_avx512_mask_punpckhqd_q_256:
+; CHECK:         vpunpckhqdq %ymm1, %ymm0, %ymm2 {%k1}
+; CHECK-NEXT:    ## ymm2 = ymm2[1],k1[1],ymm2[3],k1[3]
+; CHECK-NEXT:    vpunpckhqdq %ymm1, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0xfd,0x28,0x6d,0xc1]
+; CHECK-NEXT:    ## ymm0 = ymm0[1],ymm1[1],ymm0[3],ymm1[3]
+  %res = call <4 x i64> @llvm.x86.avx512.mask.punpckhqd.q.256(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> %x2, i8 %x3)
+  %res1 = call <4 x i64> @llvm.x86.avx512.mask.punpckhqd.q.256(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> %x2, i8 -1)
+  %res2 = add <4 x i64> %res, %res1
+  ret <4 x i64> %res2
+}
+
 declare <16 x i8> @llvm.x86.avx512.mask.pmov.qb.128(<2 x i64>, <16 x i8>, i8)
 
 define <16 x i8>@test_int_x86_avx512_mask_pmov_qb_128(<2 x i64> %x0, <16 x i8> %x1, i8 %x2) {
@@ -4263,9 +4489,9 @@ define <8 x float>@test_int_x86_avx512_mask_cvt_udq2ps_256(<8 x i32> %x0, <8 x f
 
 declare <2 x double> @llvm.x86.avx512.mask.rndscale.pd.128(<2 x double>, i32, <2 x double>, i8)
 ; CHECK-LABEL: @test_int_x86_avx512_mask_rndscale_pd_128
-; CHECK-NOT: call 
-; CHECK: kmov 
-; CHECK: vrndscalepd {{.*}}{%k1} 
+; CHECK-NOT: call
+; CHECK: kmov
+; CHECK: vrndscalepd {{.*}}{%k1}
 ; CHECK: vrndscalepd
 define <2 x double>@test_int_x86_avx512_mask_rndscale_pd_128(<2 x double> %x0, <2 x double> %x2, i8 %x3) {
   %res = call <2 x double> @llvm.x86.avx512.mask.rndscale.pd.128(<2 x double> %x0, i32 4, <2 x double> %x2, i8 %x3)
@@ -4276,9 +4502,9 @@ define <2 x double>@test_int_x86_avx512_mask_rndscale_pd_128(<2 x double> %x0, <
 
 declare <4 x double> @llvm.x86.avx512.mask.rndscale.pd.256(<4 x double>, i32, <4 x double>, i8)
 ; CHECK-LABEL: @test_int_x86_avx512_mask_rndscale_pd_256
-; CHECK-NOT: call 
-; CHECK: kmov 
-; CHECK: vrndscalepd {{.*}}{%k1} 
+; CHECK-NOT: call
+; CHECK: kmov
+; CHECK: vrndscalepd {{.*}}{%k1}
 ; CHECK: vrndscalepd
 define <4 x double>@test_int_x86_avx512_mask_rndscale_pd_256(<4 x double> %x0, <4 x double> %x2, i8 %x3) {
   %res = call <4 x double> @llvm.x86.avx512.mask.rndscale.pd.256(<4 x double> %x0, i32 4, <4 x double> %x2, i8 %x3)
@@ -4289,9 +4515,9 @@ define <4 x double>@test_int_x86_avx512_mask_rndscale_pd_256(<4 x double> %x0, <
 
 declare <4 x float> @llvm.x86.avx512.mask.rndscale.ps.128(<4 x float>, i32, <4 x float>, i8)
 ; CHECK-LABEL: @test_int_x86_avx512_mask_rndscale_ps_128
-; CHECK-NOT: call 
-; CHECK: kmov 
-; CHECK: vrndscaleps {{.*}}{%k1} 
+; CHECK-NOT: call
+; CHECK: kmov
+; CHECK: vrndscaleps {{.*}}{%k1}
 ; CHECK: vrndscaleps
 define <4 x float>@test_int_x86_avx512_mask_rndscale_ps_128(<4 x float> %x0, <4 x float> %x2, i8 %x3) {
   %res = call <4 x float> @llvm.x86.avx512.mask.rndscale.ps.128(<4 x float> %x0, i32 88, <4 x float> %x2, i8 %x3)
@@ -4303,9 +4529,9 @@ define <4 x float>@test_int_x86_avx512_mask_rndscale_ps_128(<4 x float> %x0, <4
 declare <8 x float> @llvm.x86.avx512.mask.rndscale.ps.256(<8 x float>, i32, <8 x float>, i8)
 
 ; CHECK-LABEL: @test_int_x86_avx512_mask_rndscale_ps_256
-; CHECK-NOT: call 
-; CHECK: kmov 
-; CHECK: vrndscaleps {{.*}}{%k1} 
+; CHECK-NOT: call
+; CHECK: kmov
+; CHECK: vrndscaleps {{.*}}{%k1}
 ; CHECK: vrndscaleps
 define <8 x float>@test_int_x86_avx512_mask_rndscale_ps_256(<8 x float> %x0, <8 x float> %x2, i8 %x3) {
   %res = call <8 x float> @llvm.x86.avx512.mask.rndscale.ps.256(<8 x float> %x0, i32 5, <8 x float> %x2, i8 %x3)
@@ -4313,3 +4539,1851 @@ define <8 x float>@test_int_x86_avx512_mask_rndscale_ps_256(<8 x float> %x0, <8
   %res2 = fadd <8 x float> %res, %res1
   ret <8 x float> %res2
 }
+
+declare <8 x float> @llvm.x86.avx512.mask.shuf.f32x4.256(<8 x float>, <8 x float>, i32, <8 x float>, i8)
+
+define <8 x float>@test_int_x86_avx512_mask_shuf_f32x4_256(<8 x float> %x0, <8 x float> %x1, <8 x float> %x3, i8 %x4) {
+; CHECK-LABEL: test_int_x86_avx512_mask_shuf_f32x4_256:
+; CHECK:       ## BB#0:
+; CHECK-NEXT:    movzbl %dil, %eax
+; CHECK-NEXT:    kmovw %eax, %k1
+; CHECK-NEXT:    vshuff32x4 $22, %ymm1, %ymm0, %ymm2 {%k1}
+; CHECK-NEXT:    ## ymm2 = ymm0[0,1,2,3],ymm1[4,5,6,7]
+; CHECK-NEXT:    vshuff32x4 $22, %ymm1, %ymm0, %ymm3 {%k1} {z}
+; CHECK-NEXT:    ## ymm3 = ymm0[0,1,2,3],ymm1[4,5,6,7]
+; CHECK-NEXT:    vshuff32x4 $22, %ymm1, %ymm0, %ymm0
+; CHECK-NEXT:    ## ymm0 = ymm0[0,1,2,3],ymm1[4,5,6,7]
+; CHECK-NEXT:    vaddps %ymm0, %ymm2, %ymm0
+; CHECK-NEXT:    vaddps %ymm0, %ymm3, %ymm0
+; CHECK-NEXT:    retq
+  %res = call <8 x float> @llvm.x86.avx512.mask.shuf.f32x4.256(<8 x float> %x0, <8 x float> %x1, i32 22, <8 x float> %x3, i8 %x4)
+  %res1 = call <8 x float> @llvm.x86.avx512.mask.shuf.f32x4.256(<8 x float> %x0, <8 x float> %x1, i32 22, <8 x float> %x3, i8 -1)
+  %res2 = call <8 x float> @llvm.x86.avx512.mask.shuf.f32x4.256(<8 x float> %x0, <8 x float> %x1, i32 22, <8 x float> zeroinitializer, i8 %x4)
+  %res3 = fadd <8 x float> %res, %res1
+  %res4 = fadd <8 x float> %res2, %res3
+  ret <8 x float> %res4
+}
+
+declare <4 x double> @llvm.x86.avx512.mask.shuf.f64x2.256(<4 x double>, <4 x double>, i32, <4 x double>, i8)
+
+define <4 x double>@test_int_x86_avx512_mask_shuf_f64x2_256(<4 x double> %x0, <4 x double> %x1, <4 x double> %x3, i8 %x4) {
+; CHECK-LABEL: test_int_x86_avx512_mask_shuf_f64x2_256:
+; CHECK:       ## BB#0:
+; CHECK-NEXT:    movzbl %dil, %eax
+; CHECK-NEXT:    kmovw %eax, %k1
+; CHECK-NEXT:    vshuff64x2 $22, %ymm1, %ymm0, %ymm2 {%k1}
+; CHECK-NEXT:    ## ymm2 = ymm0[0,1],ymm1[2,3]
+; CHECK-NEXT:    vshuff64x2 $22, %ymm1, %ymm0, %ymm3 {%k1} {z}
+; CHECK-NEXT:    ## ymm3 = ymm0[0,1],ymm1[2,3]
+; CHECK-NEXT:    vshuff64x2 $22, %ymm1, %ymm0, %ymm0
+; CHECK-NEXT:    ## ymm0 = ymm0[0,1],ymm1[2,3]
+; CHECK-NEXT:    vaddpd %ymm0, %ymm2, %ymm0
+; CHECK-NEXT:    vaddpd %ymm0, %ymm3, %ymm0
+; CHECK-NEXT:    retq
+  %res = call <4 x double> @llvm.x86.avx512.mask.shuf.f64x2.256(<4 x double> %x0, <4 x double> %x1, i32 22, <4 x double> %x3, i8 %x4)
+  %res1 = call <4 x double> @llvm.x86.avx512.mask.shuf.f64x2.256(<4 x double> %x0, <4 x double> %x1, i32 22, <4 x double> %x3, i8 -1)
+  %res2 = call <4 x double> @llvm.x86.avx512.mask.shuf.f64x2.256(<4 x double> %x0, <4 x double> %x1, i32 22, <4 x double> zeroinitializer, i8 %x4)
+  %res3 = fadd <4 x double> %res, %res1
+  %res4 = fadd <4 x double> %res2, %res3
+  ret <4 x double> %res4
+}
+
+declare <8 x i32> @llvm.x86.avx512.mask.shuf.i32x4.256(<8 x i32>, <8 x i32>, i32, <8 x i32>, i8)
+
+define <8 x i32>@test_int_x86_avx512_mask_shuf_i32x4_256(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x3, i8 %x4) {
+; CHECK-LABEL: test_int_x86_avx512_mask_shuf_i32x4_256:
+; CHECK:       ## BB#0:
+; CHECK-NEXT:    movzbl %dil, %eax
+; CHECK-NEXT:    kmovw %eax, %k1
+; CHECK-NEXT:    vshufi32x4 $22, %ymm1, %ymm0, %ymm2 {%k1}
+; CHECK-NEXT:    ## ymm2 = ymm0[0,1,2,3],ymm1[4,5,6,7]
+; CHECK-NEXT:    vshufi32x4 $22, %ymm1, %ymm0, %ymm0
+; CHECK-NEXT:    ## ymm0 = ymm0[0,1,2,3],ymm1[4,5,6,7]
+; CHECK-NEXT:    vpaddd %ymm0, %ymm2, %ymm0
+; CHECK-NEXT:    retq
+  %res = call <8 x i32> @llvm.x86.avx512.mask.shuf.i32x4.256(<8 x i32> %x0, <8 x i32> %x1, i32 22, <8 x i32> %x3, i8 %x4)
+  %res1 = call <8 x i32> @llvm.x86.avx512.mask.shuf.i32x4.256(<8 x i32> %x0, <8 x i32> %x1, i32 22, <8 x i32> %x3, i8 -1)
+  %res2 = add <8 x i32> %res, %res1
+  ret <8 x i32> %res2
+}
+
+declare <4 x i64> @llvm.x86.avx512.mask.shuf.i64x2.256(<4 x i64>, <4 x i64>, i32, <4 x i64>, i8)
+
+define <4 x i64>@test_int_x86_avx512_mask_shuf_i64x2_256(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> %x3, i8 %x4) {
+; CHECK-LABEL: test_int_x86_avx512_mask_shuf_i64x2_256:
+; CHECK:       ## BB#0:
+; CHECK-NEXT:    movzbl %dil, %eax
+; CHECK-NEXT:    kmovw %eax, %k1
+; CHECK-NEXT:    vshufi64x2 $22, %ymm1, %ymm0, %ymm2 {%k1}
+; CHECK-NEXT:    ## ymm2 = ymm0[0,1],ymm1[2,3]
+; CHECK-NEXT:    vshufi64x2 $22, %ymm1, %ymm0, %ymm0
+; CHECK-NEXT:    ## ymm0 = ymm0[0,1],ymm1[2,3]
+; CHECK-NEXT:    vpaddq %ymm0, %ymm2, %ymm0
+; CHECK-NEXT:    retq
+  %res = call <4 x i64> @llvm.x86.avx512.mask.shuf.i64x2.256(<4 x i64> %x0, <4 x i64> %x1, i32 22, <4 x i64> %x3, i8 %x4)
+  %res1 = call <4 x i64> @llvm.x86.avx512.mask.shuf.i64x2.256(<4 x i64> %x0, <4 x i64> %x1, i32 22, <4 x i64> %x3, i8 -1)
+  %res2 = add <4 x i64> %res, %res1
+  ret <4 x i64> %res2
+}
+
+declare <4 x float> @llvm.x86.avx512.mask.vextractf32x4.256(<8 x float>, i32, <4 x float>, i8)
+
+define <4 x float>@test_int_x86_avx512_mask_vextractf32x4_256(<8 x float> %x0, <4 x float> %x2, i8 %x3) {
+; CHECK-LABEL: test_int_x86_avx512_mask_vextractf32x4_256:
+; CHECK:       ## BB#0:
+; CHECK-NEXT:    kmovw %edi, %k1
+; CHECK-NEXT:    vextractf32x4 $1, %ymm0, %xmm1 {%k1}
+; CHECK-NEXT:    vextractf32x4 $1, %ymm0, %xmm2 {%k1} {z}
+; CHECK-NEXT:    vextractf32x4 $1, %ymm0, %xmm0
+; CHECK-NEXT:    vaddps %xmm2, %xmm1, %xmm1
+; CHECK-NEXT:    vaddps %xmm1, %xmm0, %xmm0
+; CHECK-NEXT:    retq
+  %res  = call <4 x float> @llvm.x86.avx512.mask.vextractf32x4.256(<8 x float> %x0, i32 1, <4 x float> %x2, i8 %x3)
+  %res1 = call <4 x float> @llvm.x86.avx512.mask.vextractf32x4.256(<8 x float> %x0, i32 1, <4 x float> zeroinitializer, i8 %x3)
+  %res2 = call <4 x float> @llvm.x86.avx512.mask.vextractf32x4.256(<8 x float> %x0, i32 1, <4 x float> zeroinitializer, i8 -1)
+  %res3 = fadd <4 x float> %res, %res1
+  %res4 = fadd <4 x float> %res2, %res3
+  ret <4 x float> %res4
+}
+
+declare <2 x double> @llvm.x86.avx512.mask.getmant.pd.128(<2 x double>, i32, <2 x double>, i8)
+
+define <2 x double>@test_int_x86_avx512_mask_getmant_pd_128(<2 x double> %x0, <2 x double> %x2, i8 %x3) {
+; CHECK-LABEL: test_int_x86_avx512_mask_getmant_pd_128:
+; CHECK:       ## BB#0:
+; CHECK-NEXT:    movzbl %dil, %eax
+; CHECK-NEXT:    kmovw %eax, %k1
+; CHECK-NEXT:    vgetmantpd $11, %xmm0, %xmm1 {%k1}
+; CHECK-NEXT:    vgetmantpd $11, %xmm0, %xmm2 {%k1} {z}
+; CHECK-NEXT:    vgetmantpd $11, %xmm0, %xmm0
+; CHECK-NEXT:    vaddpd %xmm0, %xmm1, %xmm0
+; CHECK-NEXT:    vaddpd %xmm0, %xmm2, %xmm0
+; CHECK-NEXT:    retq
+  %res = call <2 x double> @llvm.x86.avx512.mask.getmant.pd.128(<2 x double> %x0, i32 11, <2 x double> %x2, i8 %x3)
+  %res2 = call <2 x double> @llvm.x86.avx512.mask.getmant.pd.128(<2 x double> %x0, i32 11, <2 x double> zeroinitializer, i8 %x3)
+  %res1 = call <2 x double> @llvm.x86.avx512.mask.getmant.pd.128(<2 x double> %x0, i32 11, <2 x double> %x2, i8 -1)
+  %res3 = fadd <2 x double> %res, %res1
+  %res4 = fadd <2 x double> %res2, %res3
+  ret <2 x double> %res4
+}
+
+declare <4 x double> @llvm.x86.avx512.mask.getmant.pd.256(<4 x double>, i32, <4 x double>, i8)
+
+define <4 x double>@test_int_x86_avx512_mask_getmant_pd_256(<4 x double> %x0, <4 x double> %x2, i8 %x3) {
+; CHECK-LABEL: test_int_x86_avx512_mask_getmant_pd_256:
+; CHECK:       ## BB#0:
+; CHECK-NEXT:    movzbl %dil, %eax
+; CHECK-NEXT:    kmovw %eax, %k1
+; CHECK-NEXT:    vgetmantpd $11, %ymm0, %ymm1 {%k1}
+; CHECK-NEXT:    vgetmantpd $11, %ymm0, %ymm0
+; CHECK-NEXT:    vaddpd %ymm0, %ymm1, %ymm0
+; CHECK-NEXT:    retq
+  %res = call <4 x double> @llvm.x86.avx512.mask.getmant.pd.256(<4 x double> %x0, i32 11, <4 x double> %x2, i8 %x3)
+  %res1 = call <4 x double> @llvm.x86.avx512.mask.getmant.pd.256(<4 x double> %x0, i32 11, <4 x double> %x2, i8 -1)
+  %res2 = fadd <4 x double> %res, %res1
+  ret <4 x double> %res2
+}
+
+declare <4 x float> @llvm.x86.avx512.mask.getmant.ps.128(<4 x float>, i32, <4 x float>, i8)
+
+define <4 x float>@test_int_x86_avx512_mask_getmant_ps_128(<4 x float> %x0, <4 x float> %x2, i8 %x3) {
+; CHECK-LABEL: test_int_x86_avx512_mask_getmant_ps_128:
+; CHECK:       ## BB#0:
+; CHECK-NEXT:    movzbl %dil, %eax
+; CHECK-NEXT:    kmovw %eax, %k1
+; CHECK-NEXT:    vgetmantps $11, %xmm0, %xmm1 {%k1}
+; CHECK-NEXT:    vgetmantps $11, %xmm0, %xmm0
+; CHECK-NEXT:    vaddps %xmm0, %xmm1, %xmm0
+; CHECK-NEXT:    retq
+  %res = call <4 x float> @llvm.x86.avx512.mask.getmant.ps.128(<4 x float> %x0, i32 11, <4 x float> %x2, i8 %x3)
+  %res1 = call <4 x float> @llvm.x86.avx512.mask.getmant.ps.128(<4 x float> %x0, i32 11, <4 x float> %x2, i8 -1)
+  %res2 = fadd <4 x float> %res, %res1
+  ret <4 x float> %res2
+}
+
+declare <8 x float> @llvm.x86.avx512.mask.getmant.ps.256(<8 x float>, i32, <8 x float>, i8)
+
+define <8 x float>@test_int_x86_avx512_mask_getmant_ps_256(<8 x float> %x0, <8 x float> %x2, i8 %x3) {
+; CHECK-LABEL: test_int_x86_avx512_mask_getmant_ps_256:
+; CHECK:       ## BB#0:
+; CHECK-NEXT:    movzbl %dil, %eax
+; CHECK-NEXT:    kmovw %eax, %k1
+; CHECK-NEXT:    vgetmantps $11, %ymm0, %ymm1 {%k1}
+; CHECK-NEXT:    vgetmantps $11, %ymm0, %ymm0
+; CHECK-NEXT:    vaddps %ymm0, %ymm1, %ymm0
+; CHECK-NEXT:    retq
+  %res = call <8 x float> @llvm.x86.avx512.mask.getmant.ps.256(<8 x float> %x0, i32 11, <8 x float> %x2, i8 %x3)
+  %res1 = call <8 x float> @llvm.x86.avx512.mask.getmant.ps.256(<8 x float> %x0, i32 11, <8 x float> %x2, i8 -1)
+  %res2 = fadd <8 x float> %res, %res1
+  ret <8 x float> %res2
+}
+
+declare <2 x double> @llvm.x86.avx512.mask.shuf.pd.128(<2 x double>, <2 x double>, i32, <2 x double>, i8)
+
+define <2 x double>@test_int_x86_avx512_mask_shuf_pd_128(<2 x double> %x0, <2 x double> %x1, <2 x double> %x3, i8 %x4) {
+; CHECK-LABEL: test_int_x86_avx512_mask_shuf_pd_128:
+; CHECK:       ## BB#0:
+; CHECK-NEXT:    movzbl %dil, %eax
+; CHECK-NEXT:    kmovw %eax, %k1
+; CHECK-NEXT:    vshufpd $22, %xmm1, %xmm0, %xmm2 {%k1}
+; CHECK-NEXT:    ## xmm2 = xmm2[0],k1[1]
+; CHECK-NEXT:    vshufpd $22, %xmm1, %xmm0, %xmm3 {%k1} {z}
+; CHECK-NEXT:    ## xmm3 = k1[0],xmm0[1]
+; CHECK-NEXT:    vshufpd $22, %xmm1, %xmm0, %xmm0
+; CHECK-NEXT:    ## xmm0 = xmm0[0],xmm1[1]
+; CHECK-NEXT:    vaddpd %xmm0, %xmm2, %xmm0
+; CHECK-NEXT:    vaddpd %xmm0, %xmm3, %xmm0
+; CHECK-NEXT:    retq
+  %res = call <2 x double> @llvm.x86.avx512.mask.shuf.pd.128(<2 x double> %x0, <2 x double> %x1, i32 22, <2 x double> %x3, i8 %x4)
+  %res1 = call <2 x double> @llvm.x86.avx512.mask.shuf.pd.128(<2 x double> %x0, <2 x double> %x1, i32 22, <2 x double> %x3, i8 -1)
+  %res2 = call <2 x double> @llvm.x86.avx512.mask.shuf.pd.128(<2 x double> %x0, <2 x double> %x1, i32 22, <2 x double> zeroinitializer, i8 %x4)
+  %res3 = fadd <2 x double> %res, %res1
+  %res4 = fadd <2 x double> %res2, %res3
+  ret <2 x double> %res4
+}
+
+declare <4 x double> @llvm.x86.avx512.mask.shuf.pd.256(<4 x double>, <4 x double>, i32, <4 x double>, i8)
+
+define <4 x double>@test_int_x86_avx512_mask_shuf_pd_256(<4 x double> %x0, <4 x double> %x1, <4 x double> %x3, i8 %x4) {
+; CHECK-LABEL: test_int_x86_avx512_mask_shuf_pd_256:
+; CHECK:       ## BB#0:
+; CHECK-NEXT:    movzbl %dil, %eax
+; CHECK-NEXT:    kmovw %eax, %k1
+; CHECK-NEXT:    vshufpd $22, %ymm1, %ymm0, %ymm2 {%k1}
+; CHECK-NEXT:    ## ymm2 = ymm2[0],k1[1],ymm2[3],k1[2]
+; CHECK-NEXT:    vshufpd $22, %ymm1, %ymm0, %ymm0
+; CHECK-NEXT:    ## ymm0 = ymm0[0],ymm1[1],ymm0[3],ymm1[2]
+; CHECK-NEXT:    vaddpd %ymm0, %ymm2, %ymm0
+; CHECK-NEXT:    retq
+  %res = call <4 x double> @llvm.x86.avx512.mask.shuf.pd.256(<4 x double> %x0, <4 x double> %x1, i32 22, <4 x double> %x3, i8 %x4)
+  %res1 = call <4 x double> @llvm.x86.avx512.mask.shuf.pd.256(<4 x double> %x0, <4 x double> %x1, i32 22, <4 x double> %x3, i8 -1)
+  %res2 = fadd <4 x double> %res, %res1
+  ret <4 x double> %res2
+}
+
+declare <4 x float> @llvm.x86.avx512.mask.shuf.ps.128(<4 x float>, <4 x float>, i32, <4 x float>, i8)
+
+define <4 x float>@test_int_x86_avx512_mask_shuf_ps_128(<4 x float> %x0, <4 x float> %x1, <4 x float> %x3, i8 %x4) {
+; CHECK-LABEL: test_int_x86_avx512_mask_shuf_ps_128:
+; CHECK:       ## BB#0:
+; CHECK-NEXT:    movzbl %dil, %eax
+; CHECK-NEXT:    kmovw %eax, %k1
+; CHECK-NEXT:    vshufps $22, %xmm1, %xmm0, %xmm2 {%k1}
+; CHECK-NEXT:    ## xmm2 = xmm2[2,1],k1[1,0]
+; CHECK-NEXT:    vshufps $22, %xmm1, %xmm0, %xmm0
+; CHECK-NEXT:    ## xmm0 = xmm0[2,1],xmm1[1,0]
+; CHECK-NEXT:    vaddps %xmm0, %xmm2, %xmm0
+; CHECK-NEXT:    retq
+  %res = call <4 x float> @llvm.x86.avx512.mask.shuf.ps.128(<4 x float> %x0, <4 x float> %x1, i32 22, <4 x float> %x3, i8 %x4)
+  %res1 = call <4 x float> @llvm.x86.avx512.mask.shuf.ps.128(<4 x float> %x0, <4 x float> %x1, i32 22, <4 x float> %x3, i8 -1)
+  %res2 = fadd <4 x float> %res, %res1
+  ret <4 x float> %res2
+}
+
+declare <8 x float> @llvm.x86.avx512.mask.shuf.ps.256(<8 x float>, <8 x float>, i32, <8 x float>, i8)
+
+define <8 x float>@test_int_x86_avx512_mask_shuf_ps_256(<8 x float> %x0, <8 x float> %x1, <8 x float> %x3, i8 %x4) {
+; CHECK-LABEL: test_int_x86_avx512_mask_shuf_ps_256:
+; CHECK:       ## BB#0:
+; CHECK-NEXT:    movzbl %dil, %eax
+; CHECK-NEXT:    kmovw %eax, %k1
+; CHECK-NEXT:    vshufps $22, %ymm1, %ymm0, %ymm2 {%k1}
+; CHECK-NEXT:    ## ymm2 = ymm2[2,1],k1[1,0],ymm2[6,5],k1[5,4]
+; CHECK-NEXT:    vshufps $22, %ymm1, %ymm0, %ymm0
+; CHECK-NEXT:    ## ymm0 = ymm0[2,1],ymm1[1,0],ymm0[6,5],ymm1[5,4]
+; CHECK-NEXT:    vaddps %ymm0, %ymm2, %ymm0
+; CHECK-NEXT:    retq
+  %res = call <8 x float> @llvm.x86.avx512.mask.shuf.ps.256(<8 x float> %x0, <8 x float> %x1, i32 22, <8 x float> %x3, i8 %x4)
+  %res1 = call <8 x float> @llvm.x86.avx512.mask.shuf.ps.256(<8 x float> %x0, <8 x float> %x1, i32 22, <8 x float> %x3, i8 -1)
+  %res2 = fadd <8 x float> %res, %res1
+  ret <8 x float> %res2
+}
+
+declare <4 x i32> @llvm.x86.avx512.mask.valign.d.128(<4 x i32>, <4 x i32>, i32, <4 x i32>, i8)
+
+define <4 x i32>@test_int_x86_avx512_mask_valign_d_128(<4 x i32> %x0, <4 x i32> %x1,<4 x i32> %x3, i8 %x4) {
+; CHECK-LABEL: test_int_x86_avx512_mask_valign_d_128:
+; CHECK:       ## BB#0:
+; CHECK-NEXT:    movzbl %dil, %eax
+; CHECK-NEXT:    kmovw %eax, %k1
+; CHECK-NEXT:    valignd $22, %xmm1, %xmm0, %xmm2 {%k1}
+; CHECK-NEXT:    valignd $22, %xmm1, %xmm0, %xmm3 {%k1} {z}
+; CHECK-NEXT:    valignd $22, %xmm1, %xmm0, %xmm0
+; CHECK-NEXT:    vpaddd %xmm0, %xmm2, %xmm0
+; CHECK-NEXT:    vpaddd %xmm3, %xmm0, %xmm0
+; CHECK-NEXT:    retq
+  %res = call <4 x i32> @llvm.x86.avx512.mask.valign.d.128(<4 x i32> %x0, <4 x i32> %x1, i32 22, <4 x i32> %x3, i8 %x4)
+  %res1 = call <4 x i32> @llvm.x86.avx512.mask.valign.d.128(<4 x i32> %x0, <4 x i32> %x1, i32 22, <4 x i32> %x3, i8 -1)
+    %res2 = call <4 x i32> @llvm.x86.avx512.mask.valign.d.128(<4 x i32> %x0, <4 x i32> %x1, i32 22, <4 x i32> zeroinitializer,i8 %x4)
+  %res3 = add <4 x i32> %res, %res1
+    %res4 = add <4 x i32> %res3, %res2
+  ret <4 x i32> %res4
+}
+
+declare <8 x i32> @llvm.x86.avx512.mask.valign.d.256(<8 x i32>, <8 x i32>, i32, <8 x i32>, i8)
+
+define <8 x i32>@test_int_x86_avx512_mask_valign_d_256(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x3, i8 %x4) {
+; CHECK-LABEL: test_int_x86_avx512_mask_valign_d_256:
+; CHECK:       ## BB#0:
+; CHECK-NEXT:    movzbl %dil, %eax
+; CHECK-NEXT:    kmovw %eax, %k1
+; CHECK-NEXT:    valignd $22, %ymm1, %ymm0, %ymm2 {%k1}
+; CHECK-NEXT:    valignd $22, %ymm1, %ymm0, %ymm0
+; CHECK-NEXT:    vpaddd %ymm0, %ymm2, %ymm0
+; CHECK-NEXT:    retq
+  %res = call <8 x i32> @llvm.x86.avx512.mask.valign.d.256(<8 x i32> %x0, <8 x i32> %x1, i32 22, <8 x i32> %x3, i8 %x4)
+  %res1 = call <8 x i32> @llvm.x86.avx512.mask.valign.d.256(<8 x i32> %x0, <8 x i32> %x1, i32 22, <8 x i32> %x3, i8 -1)
+  %res2 = add <8 x i32> %res, %res1
+  ret <8 x i32> %res2
+}
+
+declare <2 x i64> @llvm.x86.avx512.mask.valign.q.128(<2 x i64>, <2 x i64>, i32, <2 x i64>, i8)
+
+define <2 x i64>@test_int_x86_avx512_mask_valign_q_128(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> %x3, i8 %x4) {
+; CHECK-LABEL: test_int_x86_avx512_mask_valign_q_128:
+; CHECK:       ## BB#0:
+; CHECK-NEXT:    movzbl %dil, %eax
+; CHECK-NEXT:    kmovw %eax, %k1
+; CHECK-NEXT:    valignq $22, %xmm1, %xmm0, %xmm2 {%k1}
+; CHECK-NEXT:    valignq $22, %xmm1, %xmm0, %xmm0
+; CHECK-NEXT:    vpaddq %xmm0, %xmm2, %xmm0
+; CHECK-NEXT:    retq
+  %res = call <2 x i64> @llvm.x86.avx512.mask.valign.q.128(<2 x i64> %x0, <2 x i64> %x1, i32 22, <2 x i64> %x3, i8 %x4)
+  %res1 = call <2 x i64> @llvm.x86.avx512.mask.valign.q.128(<2 x i64> %x0, <2 x i64> %x1, i32 22, <2 x i64> %x3, i8 -1)
+  %res2 = add <2 x i64> %res, %res1
+  ret <2 x i64> %res2
+}
+
+declare <4 x i64> @llvm.x86.avx512.mask.valign.q.256(<4 x i64>, <4 x i64>, i32, <4 x i64>, i8)
+
+define <4 x i64>@test_int_x86_avx512_mask_valign_q_256(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> %x3, i8 %x4) {
+; CHECK-LABEL: test_int_x86_avx512_mask_valign_q_256:
+; CHECK:       ## BB#0:
+; CHECK-NEXT:    movzbl %dil, %eax
+; CHECK-NEXT:    kmovw %eax, %k1
+; CHECK-NEXT:    valignq $22, %ymm1, %ymm0, %ymm2 {%k1}
+; CHECK-NEXT:    valignq $22, %ymm1, %ymm0, %ymm0
+; CHECK-NEXT:    vpaddq %ymm0, %ymm2, %ymm0
+; CHECK-NEXT:    retq
+  %res = call <4 x i64> @llvm.x86.avx512.mask.valign.q.256(<4 x i64> %x0, <4 x i64> %x1, i32 22, <4 x i64> %x3, i8 %x4)
+  %res1 = call <4 x i64> @llvm.x86.avx512.mask.valign.q.256(<4 x i64> %x0, <4 x i64> %x1, i32 22, <4 x i64> %x3, i8 -1)
+  %res2 = add <4 x i64> %res, %res1
+  ret <4 x i64> %res2
+}
+
+declare <4 x double> @llvm.x86.avx512.mask.vpermil.pd.256(<4 x double>, i32, <4 x double>, i8)
+
+define <4 x double>@test_int_x86_avx512_mask_vpermil_pd_256(<4 x double> %x0, <4 x double> %x2, i8 %x3) {
+; CHECK-LABEL: test_int_x86_avx512_mask_vpermil_pd_256:
+; CHECK:       ## BB#0:
+; CHECK-NEXT:    movzbl %dil, %eax
+; CHECK-NEXT:    kmovw %eax, %k1
+; CHECK-NEXT:    vpermilpd $22, %ymm0, %ymm1 {%k1}
+; CHECK-NEXT:    ## ymm1 = ymm1[0,1,3,2]
+; CHECK-NEXT:    vpermilpd $22, %ymm0, %ymm2 {%k1} {z}
+; CHECK-NEXT:    ## ymm2 = k1[0,1,3,2]
+; CHECK-NEXT:    vpermilpd $22, %ymm0, %ymm0
+; CHECK-NEXT:    ## ymm0 = ymm0[0,1,3,2]
+; CHECK-NEXT:    vaddpd %ymm2, %ymm1, %ymm1
+; CHECK-NEXT:    vaddpd %ymm1, %ymm0, %ymm0
+; CHECK-NEXT:    retq
+  %res = call <4 x double> @llvm.x86.avx512.mask.vpermil.pd.256(<4 x double> %x0, i32 22, <4 x double> %x2, i8 %x3)
+  %res1 = call <4 x double> @llvm.x86.avx512.mask.vpermil.pd.256(<4 x double> %x0, i32 22, <4 x double> zeroinitializer, i8 %x3)
+  %res2 = call <4 x double> @llvm.x86.avx512.mask.vpermil.pd.256(<4 x double> %x0, i32 22, <4 x double> %x2, i8 -1)
+  %res3 = fadd <4 x double> %res, %res1
+  %res4 = fadd <4 x double> %res2, %res3
+  ret <4 x double> %res4
+}
+
+declare <2 x double> @llvm.x86.avx512.mask.vpermil.pd.128(<2 x double>, i32, <2 x double>, i8)
+
+define <2 x double>@test_int_x86_avx512_mask_vpermil_pd_128(<2 x double> %x0, <2 x double> %x2, i8 %x3) {
+; CHECK-LABEL: test_int_x86_avx512_mask_vpermil_pd_128:
+; CHECK:       ## BB#0:
+; CHECK-NEXT:    movzbl %dil, %eax
+; CHECK-NEXT:    kmovw %eax, %k1
+; CHECK-NEXT:    vpermilpd $1, %xmm0, %xmm1 {%k1}
+; CHECK-NEXT:    ## xmm1 = xmm1[1,0]
+; CHECK-NEXT:    vpermilpd $1, %xmm0, %xmm2 {%k1} {z}
+; CHECK-NEXT:    ## xmm2 = k1[1,0]
+; CHECK-NEXT:    vpermilpd $1, %xmm0, %xmm0
+; CHECK-NEXT:    ## xmm0 = xmm0[1,0]
+; CHECK-NEXT:    vaddpd %xmm2, %xmm1, %xmm1
+; CHECK-NEXT:    vaddpd %xmm0, %xmm1, %xmm0
+; CHECK-NEXT:    retq
+  %res = call <2 x double> @llvm.x86.avx512.mask.vpermil.pd.128(<2 x double> %x0, i32 1, <2 x double> %x2, i8 %x3)
+  %res1 = call <2 x double> @llvm.x86.avx512.mask.vpermil.pd.128(<2 x double> %x0, i32 1, <2 x double> zeroinitializer, i8 %x3)
+  %res2 = call <2 x double> @llvm.x86.avx512.mask.vpermil.pd.128(<2 x double> %x0, i32 1, <2 x double> %x2, i8 -1)
+  %res3 = fadd <2 x double> %res, %res1
+  %res4 = fadd <2 x double> %res3, %res2
+  ret <2 x double> %res4
+}
+
+declare <8 x float> @llvm.x86.avx512.mask.vpermil.ps.256(<8 x float>, i32, <8 x float>, i8)
+
+define <8 x float>@test_int_x86_avx512_mask_vpermil_ps_256(<8 x float> %x0, <8 x float> %x2, i8 %x3) {
+; CHECK-LABEL: test_int_x86_avx512_mask_vpermil_ps_256:
+; CHECK:       ## BB#0:
+; CHECK-NEXT:    movzbl %dil, %eax
+; CHECK-NEXT:    kmovw %eax, %k1
+; CHECK-NEXT:    vpermilps $22, %ymm0, %ymm1 {%k1}
+; CHECK-NEXT:    ## ymm1 = ymm1[2,1,1,0,6,5,5,4]
+; CHECK-NEXT:    vpermilps $22, %ymm0, %ymm2 {%k1} {z}
+; CHECK-NEXT:    ## ymm2 = k1[2,1,1,0,6,5,5,4]
+; CHECK-NEXT:    vpermilps $22, %ymm0, %ymm0
+; CHECK-NEXT:    ## ymm0 = ymm0[2,1,1,0,6,5,5,4]
+; CHECK-NEXT:    vaddps %ymm2, %ymm1, %ymm1
+; CHECK-NEXT:    vaddps %ymm0, %ymm1, %ymm0
+; CHECK-NEXT:    retq
+  %res = call <8 x float> @llvm.x86.avx512.mask.vpermil.ps.256(<8 x float> %x0, i32 22, <8 x float> %x2, i8 %x3)
+  %res1 = call <8 x float> @llvm.x86.avx512.mask.vpermil.ps.256(<8 x float> %x0, i32 22, <8 x float> zeroinitializer, i8 %x3)
+  %res2 = call <8 x float> @llvm.x86.avx512.mask.vpermil.ps.256(<8 x float> %x0, i32 22, <8 x float> %x2, i8 -1)
+  %res3 = fadd <8 x float> %res, %res1
+  %res4 = fadd <8 x float> %res3, %res2
+  ret <8 x float> %res4
+}
+
+declare <4 x float> @llvm.x86.avx512.mask.vpermil.ps.128(<4 x float>, i32, <4 x float>, i8)
+
+define <4 x float>@test_int_x86_avx512_mask_vpermil_ps_128(<4 x float> %x0, <4 x float> %x2, i8 %x3) {
+; CHECK-LABEL: test_int_x86_avx512_mask_vpermil_ps_128:
+; CHECK:       ## BB#0:
+; CHECK-NEXT:    movzbl %dil, %eax
+; CHECK-NEXT:    kmovw %eax, %k1
+; CHECK-NEXT:    vpermilps $22, %xmm0, %xmm1 {%k1}
+; CHECK-NEXT:    ## xmm1 = xmm1[2,1,1,0]
+; CHECK-NEXT:    vpermilps $22, %xmm0, %xmm2 {%k1} {z}
+; CHECK-NEXT:    ## xmm2 = k1[2,1,1,0]
+; CHECK-NEXT:    vpermilps $22, %xmm0, %xmm0
+; CHECK-NEXT:    ## xmm0 = xmm0[2,1,1,0]
+; CHECK-NEXT:    vaddps %xmm2, %xmm1, %xmm1
+; CHECK-NEXT:    vaddps %xmm1, %xmm0, %xmm0
+; CHECK-NEXT:    retq
+  %res = call <4 x float> @llvm.x86.avx512.mask.vpermil.ps.128(<4 x float> %x0, i32 22, <4 x float> %x2, i8 %x3)
+  %res1 = call <4 x float> @llvm.x86.avx512.mask.vpermil.ps.128(<4 x float> %x0, i32 22, <4 x float> zeroinitializer, i8 %x3)
+  %res2 = call <4 x float> @llvm.x86.avx512.mask.vpermil.ps.128(<4 x float> %x0, i32 22, <4 x float> %x2, i8 -1)
+  %res3 = fadd <4 x float> %res, %res1
+  %res4 = fadd <4 x float> %res2, %res3
+  ret <4 x float> %res4
+}
+
+declare <4 x double> @llvm.x86.avx512.mask.vpermilvar.pd.256(<4 x double>, <4 x i64>, <4 x double>, i8)
+
+define <4 x double>@test_int_x86_avx512_mask_vpermilvar_pd_256(<4 x double> %x0, <4 x i64> %x1, <4 x double> %x2, i8 %x3) {
+; CHECK-LABEL: test_int_x86_avx512_mask_vpermilvar_pd_256:
+; CHECK:       ## BB#0:
+; CHECK-NEXT:    movzbl %dil, %eax
+; CHECK-NEXT:    kmovw %eax, %k1
+; CHECK-NEXT:    vpermilpd %ymm1, %ymm0, %ymm2 {%k1}
+; CHECK-NEXT:    vpermilpd %ymm1, %ymm0, %ymm3 {%k1} {z}
+; CHECK-NEXT:    vpermilpd %ymm1, %ymm0, %ymm0
+; CHECK-NEXT:    vaddpd %ymm3, %ymm2, %ymm1
+; CHECK-NEXT:    vaddpd %ymm1, %ymm0, %ymm0
+; CHECK-NEXT:    retq
+  %res = call <4 x double> @llvm.x86.avx512.mask.vpermilvar.pd.256(<4 x double> %x0, <4 x i64> %x1, <4 x double> %x2, i8 %x3)
+  %res1 = call <4 x double> @llvm.x86.avx512.mask.vpermilvar.pd.256(<4 x double> %x0, <4 x i64> %x1, <4 x double> zeroinitializer, i8 %x3)
+  %res2 = call <4 x double> @llvm.x86.avx512.mask.vpermilvar.pd.256(<4 x double> %x0, <4 x i64> %x1, <4 x double> %x2, i8 -1)
+  %res3 = fadd <4 x double> %res, %res1
+  %res4 = fadd <4 x double> %res2, %res3
+  ret <4 x double> %res4
+}
+
+declare <2 x double> @llvm.x86.avx512.mask.vpermilvar.pd.128(<2 x double>, <2 x i64>, <2 x double>, i8)
+
+define <2 x double>@test_int_x86_avx512_mask_vpermilvar_pd_128(<2 x double> %x0, <2 x i64> %x1, <2 x double> %x2, i8 %x3) {
+; CHECK-LABEL: test_int_x86_avx512_mask_vpermilvar_pd_128:
+; CHECK:       ## BB#0:
+; CHECK-NEXT:    movzbl %dil, %eax
+; CHECK-NEXT:    kmovw %eax, %k1
+; CHECK-NEXT:    vpermilpd %xmm1, %xmm0, %xmm2 {%k1}
+; CHECK-NEXT:    vpermilpd %xmm1, %xmm0, %xmm3 {%k1} {z}
+; CHECK-NEXT:    vpermilpd %xmm1, %xmm0, %xmm0
+; CHECK-NEXT:    vaddpd %xmm3, %xmm2, %xmm1
+; CHECK-NEXT:    vaddpd %xmm0, %xmm1, %xmm0
+; CHECK-NEXT:    retq
+  %res = call <2 x double> @llvm.x86.avx512.mask.vpermilvar.pd.128(<2 x double> %x0, <2 x i64> %x1, <2 x double> %x2, i8 %x3)
+  %res1 = call <2 x double> @llvm.x86.avx512.mask.vpermilvar.pd.128(<2 x double> %x0, <2 x i64> %x1, <2 x double> zeroinitializer, i8 %x3)
+  %res2 = call <2 x double> @llvm.x86.avx512.mask.vpermilvar.pd.128(<2 x double> %x0, <2 x i64> %x1, <2 x double> %x2, i8 -1)
+  %res3 = fadd <2 x double> %res, %res1
+  %res4 = fadd <2 x double> %res3, %res2
+  ret <2 x double> %res4
+}
+
+declare <8 x float> @llvm.x86.avx512.mask.vpermilvar.ps.256(<8 x float>, <8 x i32>, <8 x float>, i8)
+
+define <8 x float>@test_int_x86_avx512_mask_vpermilvar_ps_256(<8 x float> %x0, <8 x i32> %x1, <8 x float> %x2, i8 %x3) {
+; CHECK-LABEL: test_int_x86_avx512_mask_vpermilvar_ps_256:
+; CHECK:       ## BB#0:
+; CHECK-NEXT:    movzbl %dil, %eax
+; CHECK-NEXT:    kmovw %eax, %k1
+; CHECK-NEXT:    vpermilps %ymm1, %ymm0, %ymm2 {%k1}
+; CHECK-NEXT:    vpermilps %ymm1, %ymm0, %ymm3 {%k1} {z}
+; CHECK-NEXT:    vpermilps %ymm1, %ymm0, %ymm0
+; CHECK-NEXT:    vaddps %ymm3, %ymm2, %ymm1
+; CHECK-NEXT:    vaddps %ymm0, %ymm1, %ymm0
+; CHECK-NEXT:    retq
+  %res = call <8 x float> @llvm.x86.avx512.mask.vpermilvar.ps.256(<8 x float> %x0, <8 x i32> %x1, <8 x float> %x2, i8 %x3)
+  %res1 = call <8 x float> @llvm.x86.avx512.mask.vpermilvar.ps.256(<8 x float> %x0, <8 x i32> %x1, <8 x float> zeroinitializer, i8 %x3)
+  %res2 = call <8 x float> @llvm.x86.avx512.mask.vpermilvar.ps.256(<8 x float> %x0, <8 x i32> %x1, <8 x float> %x2, i8 -1)
+  %res3 = fadd <8 x float> %res, %res1
+  %res4 = fadd <8 x float> %res3, %res2
+  ret <8 x float> %res4
+}
+
+declare <4 x float> @llvm.x86.avx512.mask.vpermilvar.ps.128(<4 x float>, <4 x i32>, <4 x float>, i8)
+
+define <4 x float>@test_int_x86_avx512_mask_vpermilvar_ps_128(<4 x float> %x0, <4 x i32> %x1, <4 x float> %x2, i8 %x3) {
+; CHECK-LABEL: test_int_x86_avx512_mask_vpermilvar_ps_128:
+; CHECK:       ## BB#0:
+; CHECK-NEXT:    movzbl %dil, %eax
+; CHECK-NEXT:    kmovw %eax, %k1
+; CHECK-NEXT:    vpermilps %xmm1, %xmm0, %xmm2 {%k1}
+; CHECK-NEXT:    vpermilps %xmm1, %xmm0, %xmm3 {%k1} {z}
+; CHECK-NEXT:    vpermilps %xmm1, %xmm0, %xmm0
+; CHECK-NEXT:    vaddps %xmm3, %xmm2, %xmm1
+; CHECK-NEXT:    vaddps %xmm1, %xmm0, %xmm0
+; CHECK-NEXT:    retq
+  %res = call <4 x float> @llvm.x86.avx512.mask.vpermilvar.ps.128(<4 x float> %x0, <4 x i32> %x1, <4 x float> %x2, i8 %x3)
+  %res1 = call <4 x float> @llvm.x86.avx512.mask.vpermilvar.ps.128(<4 x float> %x0, <4 x i32> %x1, <4 x float> zeroinitializer, i8 %x3)
+  %res2 = call <4 x float> @llvm.x86.avx512.mask.vpermilvar.ps.128(<4 x float> %x0, <4 x i32> %x1, <4 x float> %x2, i8 -1)
+  %res3 = fadd <4 x float> %res, %res1
+  %res4 = fadd <4 x float> %res2, %res3
+  ret <4 x float> %res4
+}
+
+declare <8 x float> @llvm.x86.avx512.mask.insertf32x4.256(<8 x float>, <4 x float>, i32, <8 x float>, i8)
+
+define <8 x float>@test_int_x86_avx512_mask_insertf32x4_256(<8 x float> %x0, <4 x float> %x1, <8 x float> %x3, i8 %x4) {
+; CHECK-LABEL: test_int_x86_avx512_mask_insertf32x4_256:
+; CHECK:       ## BB#0:
+; CHECK-NEXT:    movzbl %dil, %eax
+; CHECK-NEXT:    kmovw %eax, %k1
+; CHECK-NEXT:    vinsertf32x4 $1, %xmm1, %ymm0, %ymm2 {%k1}
+; CHECK-NEXT:    vinsertf32x4 $1, %xmm1, %ymm0, %ymm3 {%k1} {z}
+; CHECK-NEXT:    vinsertf32x4 $1, %xmm1, %ymm0, %ymm0
+; CHECK-NEXT:    vaddps %ymm0, %ymm2, %ymm0
+; CHECK-NEXT:    vaddps %ymm0, %ymm3, %ymm0
+; CHECK-NEXT:    retq
+  %res = call <8 x float> @llvm.x86.avx512.mask.insertf32x4.256(<8 x float> %x0, <4 x float> %x1, i32 1, <8 x float> %x3, i8 %x4)
+  %res1 = call <8 x float> @llvm.x86.avx512.mask.insertf32x4.256(<8 x float> %x0, <4 x float> %x1, i32 1, <8 x float> %x3, i8 -1)
+  %res2 = call <8 x float> @llvm.x86.avx512.mask.insertf32x4.256(<8 x float> %x0, <4 x float> %x1, i32 1, <8 x float> zeroinitializer, i8 %x4)
+  %res3 = fadd <8 x float> %res, %res1
+  %res4 = fadd <8 x float> %res2, %res3
+  ret <8 x float> %res4
+}
+
+declare <8 x i32> @llvm.x86.avx512.mask.inserti32x4.256(<8 x i32>, <4 x i32>, i32, <8 x i32>, i8)
+
+define <8 x i32>@test_int_x86_avx512_mask_inserti32x4_256(<8 x i32> %x0, <4 x i32> %x1, <8 x i32> %x3, i8 %x4) {
+; CHECK-LABEL: test_int_x86_avx512_mask_inserti32x4_256:
+; CHECK:       ## BB#0:
+; CHECK-NEXT:    movzbl %dil, %eax
+; CHECK-NEXT:    kmovw %eax, %k1
+; CHECK-NEXT:    vinserti32x4 $1, %xmm1, %ymm0, %ymm2 {%k1}
+; CHECK-NEXT:    vinserti32x4 $1, %xmm1, %ymm0, %ymm3 {%k1} {z}
+; CHECK-NEXT:    vinserti32x4 $1, %xmm1, %ymm0, %ymm0
+; CHECK-NEXT:    vpaddd %ymm0, %ymm2, %ymm0
+; CHECK-NEXT:    vpaddd %ymm0, %ymm3, %ymm0
+; CHECK-NEXT:    retq
+
+  %res = call <8 x i32> @llvm.x86.avx512.mask.inserti32x4.256(<8 x i32> %x0, <4 x i32> %x1, i32 1, <8 x i32> %x3, i8 %x4)
+  %res1 = call <8 x i32> @llvm.x86.avx512.mask.inserti32x4.256(<8 x i32> %x0, <4 x i32> %x1, i32 1, <8 x i32> %x3, i8 -1)
+  %res2 = call <8 x i32> @llvm.x86.avx512.mask.inserti32x4.256(<8 x i32> %x0, <4 x i32> %x1, i32 1, <8 x i32> zeroinitializer, i8 %x4)
+  %res3 = add <8 x i32> %res, %res1
+  %res4 = add <8 x i32> %res2, %res3
+  ret <8 x i32> %res4
+}
+
+declare <4 x i32> @llvm.x86.avx512.mask.pternlog.d.128(<4 x i32>, <4 x i32>, <4 x i32>, i32, i8)
+
+define <4 x i32>@test_int_x86_avx512_mask_pternlog_d_128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x2, i8 %x4) {
+; CHECK-LABEL: test_int_x86_avx512_mask_pternlog_d_128:
+; CHECK:       ## BB#0:
+; CHECK-NEXT:    movzbl %dil, %eax
+; CHECK-NEXT:    kmovw %eax, %k1
+; CHECK-NEXT:    vmovaps %zmm0, %zmm3
+; CHECK-NEXT:    vpternlogd $33, %xmm2, %xmm1, %xmm3 {%k1}
+; CHECK-NEXT:    vpternlogd $33, %xmm2, %xmm1, %xmm0
+; CHECK-NEXT:    vpaddd %xmm0, %xmm3, %xmm0
+; CHECK-NEXT:    retq
+  %res = call <4 x i32> @llvm.x86.avx512.mask.pternlog.d.128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x2, i32 33, i8 %x4)
+  %res1 = call <4 x i32> @llvm.x86.avx512.mask.pternlog.d.128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x2, i32 33, i8 -1)
+  %res2 = add <4 x i32> %res, %res1
+  ret <4 x i32> %res2
+}
+
+declare <4 x i32> @llvm.x86.avx512.maskz.pternlog.d.128(<4 x i32>, <4 x i32>, <4 x i32>, i32, i8)
+
+define <4 x i32>@test_int_x86_avx512_maskz_pternlog_d_128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x2, i8 %x4) {
+; CHECK-LABEL: test_int_x86_avx512_maskz_pternlog_d_128:
+; CHECK:       ## BB#0:
+; CHECK-NEXT:    movzbl %dil, %eax
+; CHECK-NEXT:    kmovw %eax, %k1
+; CHECK-NEXT:    vmovaps %zmm0, %zmm3
+; CHECK-NEXT:    vpternlogd $33, %xmm2, %xmm1, %xmm3 {%k1} {z}
+; CHECK-NEXT:    vpternlogd $33, %xmm2, %xmm1, %xmm0
+; CHECK-NEXT:    vpaddd %xmm0, %xmm3, %xmm0
+; CHECK-NEXT:    retq
+  %res = call <4 x i32> @llvm.x86.avx512.maskz.pternlog.d.128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x2, i32 33, i8 %x4)
+  %res1 = call <4 x i32> @llvm.x86.avx512.maskz.pternlog.d.128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x2, i32 33, i8 -1)
+  %res2 = add <4 x i32> %res, %res1
+  ret <4 x i32> %res2
+}
+
+declare <8 x i32> @llvm.x86.avx512.mask.pternlog.d.256(<8 x i32>, <8 x i32>, <8 x i32>, i32, i8)
+
+define <8 x i32>@test_int_x86_avx512_mask_pternlog_d_256(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x2, i8 %x4) {
+; CHECK-LABEL: test_int_x86_avx512_mask_pternlog_d_256:
+; CHECK:       ## BB#0:
+; CHECK-NEXT:    movzbl %dil, %eax
+; CHECK-NEXT:    kmovw %eax, %k1
+; CHECK-NEXT:    vmovaps %zmm0, %zmm3
+; CHECK-NEXT:    vpternlogd $33, %ymm2, %ymm1, %ymm3 {%k1}
+; CHECK-NEXT:    vpternlogd $33, %ymm2, %ymm1, %ymm0
+; CHECK-NEXT:    vpaddd %ymm0, %ymm3, %ymm0
+; CHECK-NEXT:    retq
+  %res = call <8 x i32> @llvm.x86.avx512.mask.pternlog.d.256(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x2, i32 33, i8 %x4)
+  %res1 = call <8 x i32> @llvm.x86.avx512.mask.pternlog.d.256(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x2, i32 33, i8 -1)
+  %res2 = add <8 x i32> %res, %res1
+  ret <8 x i32> %res2
+}
+
+declare <8 x i32> @llvm.x86.avx512.maskz.pternlog.d.256(<8 x i32>, <8 x i32>, <8 x i32>, i32, i8)
+
+define <8 x i32>@test_int_x86_avx512_maskz_pternlog_d_256(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x2, i8 %x4) {
+; CHECK-LABEL: test_int_x86_avx512_maskz_pternlog_d_256:
+; CHECK:       ## BB#0:
+; CHECK-NEXT:    movzbl %dil, %eax
+; CHECK-NEXT:    kmovw %eax, %k1
+; CHECK-NEXT:    vmovaps %zmm0, %zmm3
+; CHECK-NEXT:    vpternlogd $33, %ymm2, %ymm1, %ymm3 {%k1} {z}
+; CHECK-NEXT:    vpternlogd $33, %ymm2, %ymm1, %ymm0
+; CHECK-NEXT:    vpaddd %ymm0, %ymm3, %ymm0
+; CHECK-NEXT:    retq
+  %res = call <8 x i32> @llvm.x86.avx512.maskz.pternlog.d.256(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x2, i32 33, i8 %x4)
+  %res1 = call <8 x i32> @llvm.x86.avx512.maskz.pternlog.d.256(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x2, i32 33, i8 -1)
+  %res2 = add <8 x i32> %res, %res1
+  ret <8 x i32> %res2
+}
+
+declare <2 x i64> @llvm.x86.avx512.mask.pternlog.q.128(<2 x i64>, <2 x i64>, <2 x i64>, i32, i8)
+
+define <2 x i64>@test_int_x86_avx512_mask_pternlog_q_128(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> %x2, i8 %x4) {
+; CHECK-LABEL: test_int_x86_avx512_mask_pternlog_q_128:
+; CHECK:       ## BB#0:
+; CHECK-NEXT:    movzbl %dil, %eax
+; CHECK-NEXT:    kmovw %eax, %k1
+; CHECK-NEXT:    vmovaps %zmm0, %zmm3
+; CHECK-NEXT:    vpternlogq $33, %xmm2, %xmm1, %xmm3 {%k1}
+; CHECK-NEXT:    vpternlogq $33, %xmm2, %xmm1, %xmm0
+; CHECK-NEXT:    vpaddq %xmm0, %xmm3, %xmm0
+; CHECK-NEXT:    retq
+  %res = call <2 x i64> @llvm.x86.avx512.mask.pternlog.q.128(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> %x2, i32 33, i8 %x4)
+  %res1 = call <2 x i64> @llvm.x86.avx512.mask.pternlog.q.128(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> %x2, i32 33, i8 -1)
+  %res2 = add <2 x i64> %res, %res1
+  ret <2 x i64> %res2
+}
+
+declare <2 x i64> @llvm.x86.avx512.maskz.pternlog.q.128(<2 x i64>, <2 x i64>, <2 x i64>, i32, i8)
+
+define <2 x i64>@test_int_x86_avx512_maskz_pternlog_q_128(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> %x2, i8 %x4) {
+; CHECK-LABEL: test_int_x86_avx512_maskz_pternlog_q_128:
+; CHECK:       ## BB#0:
+; CHECK-NEXT:    movzbl %dil, %eax
+; CHECK-NEXT:    kmovw %eax, %k1
+; CHECK-NEXT:    vmovaps %zmm0, %zmm3
+; CHECK-NEXT:    vpternlogq $33, %xmm2, %xmm1, %xmm3 {%k1} {z}
+; CHECK-NEXT:    vpternlogq $33, %xmm2, %xmm1, %xmm0
+; CHECK-NEXT:    vpaddq %xmm0, %xmm3, %xmm0
+; CHECK-NEXT:    retq
+  %res = call <2 x i64> @llvm.x86.avx512.maskz.pternlog.q.128(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> %x2, i32 33, i8 %x4)
+  %res1 = call <2 x i64> @llvm.x86.avx512.maskz.pternlog.q.128(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> %x2, i32 33, i8 -1)
+  %res2 = add <2 x i64> %res, %res1
+  ret <2 x i64> %res2
+}
+
+declare <4 x i64> @llvm.x86.avx512.mask.pternlog.q.256(<4 x i64>, <4 x i64>, <4 x i64>, i32, i8)
+
+define <4 x i64>@test_int_x86_avx512_mask_pternlog_q_256(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> %x2, i8 %x4) {
+; CHECK-LABEL: test_int_x86_avx512_mask_pternlog_q_256:
+; CHECK:       ## BB#0:
+; CHECK-NEXT:    movzbl %dil, %eax
+; CHECK-NEXT:    kmovw %eax, %k1
+; CHECK-NEXT:    vmovaps %zmm0, %zmm3
+; CHECK-NEXT:    vpternlogq $33, %ymm2, %ymm1, %ymm3 {%k1}
+; CHECK-NEXT:    vpternlogq $33, %ymm2, %ymm1, %ymm0
+; CHECK-NEXT:    vpaddq %ymm0, %ymm3, %ymm0
+; CHECK-NEXT:    retq
+  %res = call <4 x i64> @llvm.x86.avx512.mask.pternlog.q.256(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> %x2, i32 33, i8 %x4)
+  %res1 = call <4 x i64> @llvm.x86.avx512.mask.pternlog.q.256(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> %x2, i32 33, i8 -1)
+  %res2 = add <4 x i64> %res, %res1
+  ret <4 x i64> %res2
+}
+
+declare <4 x i64> @llvm.x86.avx512.maskz.pternlog.q.256(<4 x i64>, <4 x i64>, <4 x i64>, i32, i8)
+
+define <4 x i64>@test_int_x86_avx512_maskz_pternlog_q_256(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> %x2, i8 %x4) {
+; CHECK-LABEL: test_int_x86_avx512_maskz_pternlog_q_256:
+; CHECK:       ## BB#0:
+; CHECK-NEXT:    movzbl %dil, %eax
+; CHECK-NEXT:    kmovw %eax, %k1
+; CHECK-NEXT:    vmovaps %zmm0, %zmm3
+; CHECK-NEXT:    vpternlogq $33, %ymm2, %ymm1, %ymm3 {%k1} {z}
+; CHECK-NEXT:    vpternlogq $33, %ymm2, %ymm1, %ymm0
+; CHECK-NEXT:    vpaddq %ymm0, %ymm3, %ymm0
+; CHECK-NEXT:    retq
+  %res = call <4 x i64> @llvm.x86.avx512.maskz.pternlog.q.256(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> %x2, i32 33, i8 %x4)
+  %res1 = call <4 x i64> @llvm.x86.avx512.maskz.pternlog.q.256(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> %x2, i32 33, i8 -1)
+  %res2 = add <4 x i64> %res, %res1
+  ret <4 x i64> %res2
+}
+
+declare <8 x i32> @llvm.x86.avx512.pbroadcastd.256(<4 x i32>, <8 x i32>, i8)
+
+define <8 x i32>@test_int_x86_avx512_pbroadcastd_256(<4 x i32> %x0, <8 x i32> %x1, i8 %mask) {
+; CHECK-LABEL: test_int_x86_avx512_pbroadcastd_256:
+; CHECK:       ## BB#0:
+; CHECK-NEXT:    movzbl %dil, %eax
+; CHECK-NEXT:    kmovw %eax, %k1
+; CHECK-NEXT:    vpbroadcastd %xmm0, %ymm1 {%k1}
+; CHECK-NEXT:    vpbroadcastd %xmm0, %ymm2 {%k1} {z}
+; CHECK-NEXT:    vpbroadcastd %xmm0, %ymm0
+; CHECK-NEXT:    vpaddd %ymm1, %ymm0, %ymm0
+; CHECK-NEXT:    vpaddd %ymm0, %ymm2, %ymm0
+; CHECK-NEXT:    retq
+  %res = call <8 x i32> @llvm.x86.avx512.pbroadcastd.256(<4 x i32> %x0, <8 x i32> %x1, i8 -1)
+  %res1 = call <8 x i32> @llvm.x86.avx512.pbroadcastd.256(<4 x i32> %x0, <8 x i32> %x1, i8 %mask)
+  %res2 = call <8 x i32> @llvm.x86.avx512.pbroadcastd.256(<4 x i32> %x0, <8 x i32> zeroinitializer, i8 %mask)
+  %res3 = add <8 x i32> %res, %res1
+  %res4 = add <8 x i32> %res2, %res3
+  ret <8 x i32> %res4
+}
+
+declare <4 x i32> @llvm.x86.avx512.pbroadcastd.128(<4 x i32>, <4 x i32>, i8)
+
+define <4 x i32>@test_int_x86_avx512_pbroadcastd_128(<4 x i32> %x0, <4 x i32> %x1, i8 %mask) {
+; CHECK-LABEL: test_int_x86_avx512_pbroadcastd_128:
+; CHECK:       ## BB#0:
+; CHECK-NEXT:    movzbl %dil, %eax
+; CHECK-NEXT:    kmovw %eax, %k1
+; CHECK-NEXT:    vpbroadcastd %xmm0, %xmm1 {%k1}
+; CHECK-NEXT:    vpbroadcastd %xmm0, %xmm2 {%k1} {z}
+; CHECK-NEXT:    vpbroadcastd %xmm0, %xmm0
+; CHECK-NEXT:    vpaddd %xmm1, %xmm0, %xmm0
+; CHECK-NEXT:    vpaddd %xmm0, %xmm2, %xmm0
+; CHECK-NEXT:    retq
+  %res = call <4 x i32> @llvm.x86.avx512.pbroadcastd.128(<4 x i32> %x0, <4 x i32> %x1, i8 -1)
+  %res1 = call <4 x i32> @llvm.x86.avx512.pbroadcastd.128(<4 x i32> %x0, <4 x i32> %x1, i8 %mask)
+  %res2 = call <4 x i32> @llvm.x86.avx512.pbroadcastd.128(<4 x i32> %x0, <4 x i32> zeroinitializer, i8 %mask)
+  %res3 = add <4 x i32> %res, %res1
+  %res4 = add <4 x i32> %res2, %res3
+  ret <4 x i32> %res4
+}
+
+declare <4 x i64> @llvm.x86.avx512.pbroadcastq.256(<2 x i64>, <4 x i64>, i8)
+
+define <4 x i64>@test_int_x86_avx512_pbroadcastq_256(<2 x i64> %x0, <4 x i64> %x1, i8 %mask) {
+; CHECK-LABEL: test_int_x86_avx512_pbroadcastq_256:
+; CHECK:       ## BB#0:
+; CHECK-NEXT:    movzbl %dil, %eax
+; CHECK-NEXT:    kmovw %eax, %k1
+; CHECK-NEXT:    vpbroadcastq %xmm0, %ymm1 {%k1}
+; CHECK-NEXT:    vpbroadcastq %xmm0, %ymm2 {%k1} {z}
+; CHECK-NEXT:    vpbroadcastq %xmm0, %ymm0
+; CHECK-NEXT:    vpaddq %ymm1, %ymm0, %ymm0
+; CHECK-NEXT:    vpaddq %ymm0, %ymm2, %ymm0
+; CHECK-NEXT:    retq
+  %res = call <4 x i64> @llvm.x86.avx512.pbroadcastq.256(<2 x i64> %x0, <4 x i64> %x1,i8 -1)
+  %res1 = call <4 x i64> @llvm.x86.avx512.pbroadcastq.256(<2 x i64> %x0, <4 x i64> %x1,i8 %mask)
+  %res2 = call <4 x i64> @llvm.x86.avx512.pbroadcastq.256(<2 x i64> %x0, <4 x i64> zeroinitializer,i8 %mask)
+  %res3 = add <4 x i64> %res, %res1
+  %res4 = add <4 x i64> %res2, %res3
+  ret <4 x i64> %res4
+}
+
+declare <2 x i64> @llvm.x86.avx512.pbroadcastq.128(<2 x i64>, <2 x i64>, i8)
+
+define <2 x i64>@test_int_x86_avx512_pbroadcastq_128(<2 x i64> %x0, <2 x i64> %x1, i8 %mask) {
+; CHECK-LABEL: test_int_x86_avx512_pbroadcastq_128:
+; CHECK:       ## BB#0:
+; CHECK-NEXT:    movzbl %dil, %eax
+; CHECK-NEXT:    kmovw %eax, %k1
+; CHECK-NEXT:    vpbroadcastq %xmm0, %xmm1 {%k1}
+; CHECK-NEXT:    vpbroadcastq %xmm0, %xmm2 {%k1} {z}
+; CHECK-NEXT:    vpbroadcastq %xmm0, %xmm0
+; CHECK-NEXT:    vpaddq %xmm1, %xmm0, %xmm0
+; CHECK-NEXT:    vpaddq %xmm0, %xmm2, %xmm0
+; CHECK-NEXT:    retq
+  %res = call <2 x i64> @llvm.x86.avx512.pbroadcastq.128(<2 x i64> %x0, <2 x i64> %x1,i8 -1)
+  %res1 = call <2 x i64> @llvm.x86.avx512.pbroadcastq.128(<2 x i64> %x0, <2 x i64> %x1,i8 %mask)
+  %res2 = call <2 x i64> @llvm.x86.avx512.pbroadcastq.128(<2 x i64> %x0, <2 x i64> zeroinitializer,i8 %mask)
+  %res3 = add <2 x i64> %res, %res1
+  %res4 = add <2 x i64> %res2, %res3
+  ret <2 x i64> %res4
+}
+
+define <4 x float> @test_x86_vcvtph2ps_128(<8 x i16> %a0) {
+  ; CHECK: test_x86_vcvtph2ps_128
+  ; CHECK: vcvtph2ps  %xmm0, %xmm0    
+  %res = call <4 x float> @llvm.x86.avx512.mask.vcvtph2ps.128(<8 x i16> %a0, <4 x float> zeroinitializer, i8 -1)
+  ret <4 x float> %res
+}
+
+define <4 x float> @test_x86_vcvtph2ps_128_rrk(<8 x i16> %a0,<4 x float> %a1, i8 %mask) {
+  ; CHECK: test_x86_vcvtph2ps_128_rrk
+  ; CHECK: vcvtph2ps  %xmm0, %xmm1 {%k1}
+  %res = call <4 x float> @llvm.x86.avx512.mask.vcvtph2ps.128(<8 x i16> %a0, <4 x float> %a1, i8 %mask)
+  ret <4 x float> %res
+}
+
+
+define <4 x float> @test_x86_vcvtph2ps_128_rrkz(<8 x i16> %a0, i8 %mask) {
+  ; CHECK: test_x86_vcvtph2ps_128_rrkz
+  ; CHECK: vcvtph2ps  %xmm0, %xmm0 {%k1} {z}
+  %res = call <4 x float> @llvm.x86.avx512.mask.vcvtph2ps.128(<8 x i16> %a0, <4 x float> zeroinitializer, i8 %mask)
+  ret <4 x float> %res
+}
+
+declare <4 x float> @llvm.x86.avx512.mask.vcvtph2ps.128(<8 x i16>, <4 x float>, i8) nounwind readonly
+
+define <8 x float> @test_x86_vcvtph2ps_256(<8 x i16> %a0) {
+  ; CHECK: test_x86_vcvtph2ps_256
+  ; CHECK: vcvtph2ps  %xmm0, %ymm0
+  %res = call <8 x float> @llvm.x86.avx512.mask.vcvtph2ps.256(<8 x i16> %a0, <8 x float> zeroinitializer, i8 -1)
+  ret <8 x float> %res
+}
+
+define <8 x float> @test_x86_vcvtph2ps_256_rrk(<8 x i16> %a0,<8 x float> %a1, i8 %mask) {
+  ; CHECK: test_x86_vcvtph2ps_256_rrk
+  ; CHECK: vcvtph2ps  %xmm0, %ymm1 {%k1} 
+  %res = call <8 x float> @llvm.x86.avx512.mask.vcvtph2ps.256(<8 x i16> %a0, <8 x float> %a1, i8 %mask)
+  ret <8 x float> %res
+}
+
+define <8 x float> @test_x86_vcvtph2ps_256_rrkz(<8 x i16> %a0, i8 %mask) {
+  ; CHECK: test_x86_vcvtph2ps_256_rrkz
+  ; CHECK: vcvtph2ps  %xmm0, %ymm0 {%k1} {z}
+  %res = call <8 x float> @llvm.x86.avx512.mask.vcvtph2ps.256(<8 x i16> %a0, <8 x float> zeroinitializer, i8 %mask)
+  ret <8 x float> %res
+}
+
+declare <8 x float> @llvm.x86.avx512.mask.vcvtph2ps.256(<8 x i16>, <8 x float>, i8) nounwind readonly
+
+define <8 x i16> @test_x86_vcvtps2ph_128(<4 x float> %a0) {
+  ; CHECK: test_x86_vcvtps2ph_128
+  ; CHECK: vcvtps2ph $2, %xmm0, %xmm0
+  %res = call <8 x i16> @llvm.x86.avx512.mask.vcvtps2ph.128(<4 x float> %a0, i32 2, <8 x i16> zeroinitializer, i8 -1)
+  ret <8 x i16> %res
+}
+
+
+declare <8 x i16> @llvm.x86.avx512.mask.vcvtps2ph.128(<4 x float>, i32, <8 x i16>, i8) nounwind readonly
+
+define <8 x i16> @test_x86_vcvtps2ph_256(<8 x float> %a0) {
+  ; CHECK: test_x86_vcvtps2ph_256
+  ; CHECK: vcvtps2ph $2, %ymm0, %xmm0
+  %res = call <8 x i16> @llvm.x86.avx512.mask.vcvtps2ph.256(<8 x float> %a0, i32 2, <8 x i16> zeroinitializer, i8 -1)
+  ret <8 x i16> %res
+}
+
+declare <8 x i16> @llvm.x86.avx512.mask.vcvtps2ph.256(<8 x float>, i32, <8 x i16>, i8) nounwind readonly
+
+declare <4 x float> @llvm.x86.avx512.mask.movsldup.128(<4 x float>, <4 x float>, i8)
+
+define <4 x float>@test_int_x86_avx512_mask_movsldup_128(<4 x float> %x0, <4 x float> %x1, i8 %x2) {
+; CHECK-LABEL: test_int_x86_avx512_mask_movsldup_128:
+; CHECK:       ## BB#0:
+; CHECK-NEXT:    movzbl %dil, %eax 
+; CHECK-NEXT:    kmovw %eax, %k1 
+; CHECK-NEXT:    vmovsldup %xmm0, %xmm1 {%k1} 
+; CHECK-NEXT:    ## xmm1 = xmm0[0,0,2,2]
+; CHECK-NEXT:    vmovsldup %xmm0, %xmm2 {%k1} {z} 
+; CHECK-NEXT:    ## xmm2 = xmm0[0,0,2,2]
+; CHECK-NEXT:    vmovsldup %xmm0, %xmm0 
+; CHECK-NEXT:    ## xmm0 = xmm0[0,0,2,2]
+; CHECK-NEXT:    vaddps %xmm0, %xmm1, %xmm0 
+; CHECK-NEXT:    vaddps %xmm0, %xmm2, %xmm0 
+; CHECK-NEXT:    retq 
+  %res = call <4 x float> @llvm.x86.avx512.mask.movsldup.128(<4 x float> %x0, <4 x float> %x1, i8 %x2)
+  %res1 = call <4 x float> @llvm.x86.avx512.mask.movsldup.128(<4 x float> %x0, <4 x float> %x1, i8 -1)
+  %res2 = call <4 x float> @llvm.x86.avx512.mask.movsldup.128(<4 x float> %x0, <4 x float> zeroinitializer, i8 %x2)
+  %res3 = fadd <4 x float> %res, %res1
+  %res4 = fadd <4 x float> %res2, %res3
+  ret <4 x float> %res4
+}
+
+declare <8 x float> @llvm.x86.avx512.mask.movsldup.256(<8 x float>, <8 x float>, i8)
+
+define <8 x float>@test_int_x86_avx512_mask_movsldup_256(<8 x float> %x0, <8 x float> %x1, i8 %x2) {
+; CHECK-LABEL: test_int_x86_avx512_mask_movsldup_256:
+; CHECK:       ## BB#0:
+; CHECK-NEXT:    movzbl %dil, %eax 
+; CHECK-NEXT:    kmovw %eax, %k1 
+; CHECK-NEXT:    vmovsldup %ymm0, %ymm1 {%k1} 
+; CHECK-NEXT:    ## ymm1 = ymm0[0,0,2,2,4,4,6,6]
+; CHECK-NEXT:    vmovsldup %ymm0, %ymm2 {%k1} {z} 
+; CHECK-NEXT:    ## ymm2 = ymm0[0,0,2,2,4,4,6,6]
+; CHECK-NEXT:    vmovsldup %ymm0, %ymm0 
+; CHECK-NEXT:    ## ymm0 = ymm0[0,0,2,2,4,4,6,6]
+; CHECK-NEXT:    vaddps %ymm0, %ymm1, %ymm0 
+; CHECK-NEXT:    vaddps %ymm0, %ymm2, %ymm0 
+; CHECK-NEXT:    retq 
+  %res = call <8 x float> @llvm.x86.avx512.mask.movsldup.256(<8 x float> %x0, <8 x float> %x1, i8 %x2)
+  %res1 = call <8 x float> @llvm.x86.avx512.mask.movsldup.256(<8 x float> %x0, <8 x float> %x1, i8 -1)
+  %res2 = call <8 x float> @llvm.x86.avx512.mask.movsldup.256(<8 x float> %x0, <8 x float> zeroinitializer, i8 %x2)
+  %res3 = fadd <8 x float> %res, %res1
+  %res4 = fadd <8 x float> %res2, %res3
+  ret <8 x float> %res4
+}
+
+declare <4 x float> @llvm.x86.avx512.mask.movshdup.128(<4 x float>, <4 x float>, i8)
+
+define <4 x float>@test_int_x86_avx512_mask_movshdup_128(<4 x float> %x0, <4 x float> %x1, i8 %x2) {
+; CHECK-LABEL: test_int_x86_avx512_mask_movshdup_128:
+; CHECK:       ## BB#0:
+; CHECK-NEXT:    movzbl %dil, %eax 
+; CHECK-NEXT:    kmovw %eax, %k1 
+; CHECK-NEXT:    vmovshdup %xmm0, %xmm1 {%k1} 
+; CHECK-NEXT:    ## xmm1 = xmm0[1,1,3,3]
+; CHECK-NEXT:    vmovshdup %xmm0, %xmm2 {%k1} {z} 
+; CHECK-NEXT:    ## xmm2 = xmm0[1,1,3,3]
+; CHECK-NEXT:    vmovshdup %xmm0, %xmm0 
+; CHECK-NEXT:    ## xmm0 = xmm0[1,1,3,3]
+; CHECK-NEXT:    vaddps %xmm0, %xmm1, %xmm0 
+; CHECK-NEXT:    vaddps %xmm0, %xmm2, %xmm0 
+; CHECK-NEXT:    retq 
+  %res = call <4 x float> @llvm.x86.avx512.mask.movshdup.128(<4 x float> %x0, <4 x float> %x1, i8 %x2)
+  %res1 = call <4 x float> @llvm.x86.avx512.mask.movshdup.128(<4 x float> %x0, <4 x float> %x1, i8 -1)
+  %res2 = call <4 x float> @llvm.x86.avx512.mask.movshdup.128(<4 x float> %x0, <4 x float> zeroinitializer, i8 %x2)
+  %res3 = fadd <4 x float> %res, %res1
+  %res4 = fadd <4 x float> %res2, %res3
+  ret <4 x float> %res4
+}
+
+declare <8 x float> @llvm.x86.avx512.mask.movshdup.256(<8 x float>, <8 x float>, i8)
+
+define <8 x float>@test_int_x86_avx512_mask_movshdup_256(<8 x float> %x0, <8 x float> %x1, i8 %x2) {
+; CHECK-LABEL: test_int_x86_avx512_mask_movshdup_256:
+; CHECK:       ## BB#0:
+; CHECK-NEXT:    movzbl %dil, %eax 
+; CHECK-NEXT:    kmovw %eax, %k1 
+; CHECK-NEXT:    vmovshdup %ymm0, %ymm1 {%k1} 
+; CHECK-NEXT:    ## ymm1 = ymm0[1,1,3,3,5,5,7,7]
+; CHECK-NEXT:    vmovshdup %ymm0, %ymm2 {%k1} {z} 
+; CHECK-NEXT:    ## ymm2 = ymm0[1,1,3,3,5,5,7,7]
+; CHECK-NEXT:    vmovshdup %ymm0, %ymm0 
+; CHECK-NEXT:    ## ymm0 = ymm0[1,1,3,3,5,5,7,7]
+; CHECK-NEXT:    vaddps %ymm0, %ymm1, %ymm0 
+; CHECK-NEXT:    vaddps %ymm0, %ymm2, %ymm0 
+; CHECK-NEXT:    retq 
+  %res = call <8 x float> @llvm.x86.avx512.mask.movshdup.256(<8 x float> %x0, <8 x float> %x1, i8 %x2)
+  %res1 = call <8 x float> @llvm.x86.avx512.mask.movshdup.256(<8 x float> %x0, <8 x float> %x1, i8 -1)
+  %res2 = call <8 x float> @llvm.x86.avx512.mask.movshdup.256(<8 x float> %x0, <8 x float> zeroinitializer, i8 %x2)
+  %res3 = fadd <8 x float> %res, %res1
+  %res4 = fadd <8 x float> %res2, %res3
+  ret <8 x float> %res4
+}
+declare <2 x double> @llvm.x86.avx512.mask.movddup.128(<2 x double>, <2 x double>, i8)
+
+define <2 x double>@test_int_x86_avx512_mask_movddup_128(<2 x double> %x0, <2 x double> %x1, i8 %x2) {
+; CHECK-LABEL: test_int_x86_avx512_mask_movddup_128:
+; CHECK:       ## BB#0:
+; CHECK-NEXT:    movzbl %dil, %eax
+; CHECK-NEXT:    kmovw %eax, %k1
+; CHECK-NEXT:    vmovddup %xmm0, %xmm1 {%k1}
+; CHECK-NEXT:    ## xmm1 = xmm0[0,0]
+; CHECK-NEXT:    vmovddup %xmm0, %xmm2 {%k1} {z}
+; CHECK-NEXT:    ## xmm2 = xmm0[0,0]
+; CHECK-NEXT:    vmovddup %xmm0, %xmm0
+; CHECK-NEXT:    ## xmm0 = xmm0[0,0]
+; CHECK-NEXT:    vaddpd %xmm0, %xmm1, %xmm0
+; CHECK-NEXT:    vaddpd %xmm0, %xmm2, %xmm0
+; CHECK-NEXT:    retq
+  %res = call <2 x double> @llvm.x86.avx512.mask.movddup.128(<2 x double> %x0, <2 x double> %x1, i8 %x2)
+  %res1 = call <2 x double> @llvm.x86.avx512.mask.movddup.128(<2 x double> %x0, <2 x double> %x1, i8 -1)
+  %res2 = call <2 x double> @llvm.x86.avx512.mask.movddup.128(<2 x double> %x0, <2 x double> zeroinitializer, i8 %x2)
+  %res3 = fadd <2 x double> %res, %res1
+  %res4 = fadd <2 x double> %res2, %res3
+  ret <2 x double> %res4
+}
+
+declare <4 x double> @llvm.x86.avx512.mask.movddup.256(<4 x double>, <4 x double>, i8)
+
+define <4 x double>@test_int_x86_avx512_mask_movddup_256(<4 x double> %x0, <4 x double> %x1, i8 %x2) {
+; CHECK-LABEL: test_int_x86_avx512_mask_movddup_256:
+; CHECK:       ## BB#0:
+; CHECK-NEXT:    movzbl %dil, %eax
+; CHECK-NEXT:    kmovw %eax, %k1
+; CHECK-NEXT:    vmovddup %ymm0, %ymm1 {%k1}
+; CHECK-NEXT:    ## ymm1 = ymm0[0,0,2,2]
+; CHECK-NEXT:    vmovddup %ymm0, %ymm2 {%k1} {z}
+; CHECK-NEXT:    ## ymm2 = ymm0[0,0,2,2]
+; CHECK-NEXT:    vmovddup %ymm0, %ymm0
+; CHECK-NEXT:    ## ymm0 = ymm0[0,0,2,2]
+; CHECK-NEXT:    vaddpd %ymm0, %ymm1, %ymm0
+; CHECK-NEXT:    vaddpd %ymm0, %ymm2, %ymm0
+; CHECK-NEXT:    retq
+  %res = call <4 x double> @llvm.x86.avx512.mask.movddup.256(<4 x double> %x0, <4 x double> %x1, i8 %x2)
+  %res1 = call <4 x double> @llvm.x86.avx512.mask.movddup.256(<4 x double> %x0, <4 x double> %x1, i8 -1)
+  %res2 = call <4 x double> @llvm.x86.avx512.mask.movddup.256(<4 x double> %x0, <4 x double> zeroinitializer, i8 %x2)
+  %res3 = fadd <4 x double> %res, %res1
+  %res4 = fadd <4 x double> %res2, %res3
+  ret <4 x double> %res4
+}
+
+define <8 x float> @test_rsqrt_ps_256_rr(<8 x float> %a0) {
+; CHECK-LABEL: test_rsqrt_ps_256_rr:
+; CHECK: vrsqrt14ps %ymm0, %ymm0
+  %res = call <8 x float> @llvm.x86.avx512.rsqrt14.ps.256(<8 x float> %a0, <8 x float> zeroinitializer, i8 -1)
+  ret <8 x float> %res
+}
+
+define <8 x float> @test_rsqrt_ps_256_rrkz(<8 x float> %a0, i8 %mask) {
+; CHECK-LABEL: test_rsqrt_ps_256_rrkz:
+; CHECK: vrsqrt14ps %ymm0, %ymm0 {%k1} {z}
+  %res = call <8 x float> @llvm.x86.avx512.rsqrt14.ps.256(<8 x float> %a0, <8 x float> zeroinitializer, i8 %mask)
+  ret <8 x float> %res
+}
+
+define <8 x float> @test_rsqrt_ps_256_rrk(<8 x float> %a0, <8 x float> %a1, i8 %mask) {
+; CHECK-LABEL: test_rsqrt_ps_256_rrk:
+; CHECK: vrsqrt14ps %ymm0, %ymm1 {%k1}
+  %res = call <8 x float> @llvm.x86.avx512.rsqrt14.ps.256(<8 x float> %a0, <8 x float> %a1, i8 %mask)
+  ret <8 x float> %res
+}
+
+define <4 x float> @test_rsqrt_ps_128_rr(<4 x float> %a0) {
+; CHECK-LABEL: test_rsqrt_ps_128_rr:
+; CHECK: vrsqrt14ps %xmm0, %xmm0
+  %res = call <4 x float> @llvm.x86.avx512.rsqrt14.ps.128(<4 x float> %a0, <4 x float> zeroinitializer, i8 -1)
+  ret <4 x float> %res
+}
+
+define <4 x float> @test_rsqrt_ps_128_rrkz(<4 x float> %a0, i8 %mask) {
+; CHECK-LABEL: test_rsqrt_ps_128_rrkz:
+; CHECK: vrsqrt14ps %xmm0, %xmm0 {%k1} {z}
+  %res = call <4 x float> @llvm.x86.avx512.rsqrt14.ps.128(<4 x float> %a0, <4 x float> zeroinitializer, i8 %mask)
+  ret <4 x float> %res
+}
+
+define <4 x float> @test_rsqrt_ps_128_rrk(<4 x float> %a0, <4 x float> %a1, i8 %mask) {
+; CHECK-LABEL: test_rsqrt_ps_128_rrk:
+; CHECK: vrsqrt14ps %xmm0, %xmm1 {%k1}
+  %res = call <4 x float> @llvm.x86.avx512.rsqrt14.ps.128(<4 x float> %a0, <4 x float> %a1, i8 %mask)
+  ret <4 x float> %res
+}
+
+declare <8 x float> @llvm.x86.avx512.rsqrt14.ps.256(<8 x float>, <8 x float>, i8) nounwind readnone
+declare <4 x float> @llvm.x86.avx512.rsqrt14.ps.128(<4 x float>, <4 x float>, i8) nounwind readnone
+
+define <8 x float> @test_rcp_ps_256_rr(<8 x float> %a0) {
+; CHECK-LABEL: test_rcp_ps_256_rr:
+; CHECK: vrcp14ps %ymm0, %ymm0
+  %res = call <8 x float> @llvm.x86.avx512.rcp14.ps.256(<8 x float> %a0, <8 x float> zeroinitializer, i8 -1)
+  ret <8 x float> %res
+}
+
+define <8 x float> @test_rcp_ps_256_rrkz(<8 x float> %a0, i8 %mask) {
+; CHECK-LABEL: test_rcp_ps_256_rrkz:
+; CHECK: vrcp14ps %ymm0, %ymm0 {%k1} {z}
+  %res = call <8 x float> @llvm.x86.avx512.rcp14.ps.256(<8 x float> %a0, <8 x float> zeroinitializer, i8 %mask)
+  ret <8 x float> %res
+}
+
+define <8 x float> @test_rcp_ps_256_rrk(<8 x float> %a0, <8 x float> %a1, i8 %mask) {
+; CHECK-LABEL: test_rcp_ps_256_rrk:
+; CHECK: vrcp14ps %ymm0, %ymm1 {%k1}
+  %res = call <8 x float> @llvm.x86.avx512.rcp14.ps.256(<8 x float> %a0, <8 x float> %a1, i8 %mask)
+  ret <8 x float> %res
+}
+
+define <4 x float> @test_rcp_ps_128_rr(<4 x float> %a0) {
+; CHECK-LABEL: test_rcp_ps_128_rr:
+; CHECK: vrcp14ps %xmm0, %xmm0
+  %res = call <4 x float> @llvm.x86.avx512.rcp14.ps.128(<4 x float> %a0, <4 x float> zeroinitializer, i8 -1)
+  ret <4 x float> %res
+}
+
+define <4 x float> @test_rcp_ps_128_rrkz(<4 x float> %a0, i8 %mask) {
+; CHECK-LABEL: test_rcp_ps_128_rrkz:
+; CHECK: vrcp14ps %xmm0, %xmm0 {%k1} {z}
+  %res = call <4 x float> @llvm.x86.avx512.rcp14.ps.128(<4 x float> %a0, <4 x float> zeroinitializer, i8 %mask)
+  ret <4 x float> %res
+}
+
+define <4 x float> @test_rcp_ps_128_rrk(<4 x float> %a0, <4 x float> %a1, i8 %mask) {
+; CHECK-LABEL: test_rcp_ps_128_rrk:
+; CHECK: vrcp14ps %xmm0, %xmm1 {%k1}
+  %res = call <4 x float> @llvm.x86.avx512.rcp14.ps.128(<4 x float> %a0, <4 x float> %a1, i8 %mask)
+  ret <4 x float> %res
+}
+
+declare <8 x float> @llvm.x86.avx512.rcp14.ps.256(<8 x float>, <8 x float>, i8) nounwind readnone
+declare <4 x float> @llvm.x86.avx512.rcp14.ps.128(<4 x float>, <4 x float>, i8) nounwind readnone
+
+
+define <4 x double> @test_rsqrt_pd_256_rr(<4 x double> %a0) {
+; CHECK-LABEL: test_rsqrt_pd_256_rr:
+; CHECK: vrsqrt14pd %ymm0, %ymm0
+  %res = call <4 x double> @llvm.x86.avx512.rsqrt14.pd.256(<4 x double> %a0, <4 x double> zeroinitializer, i8 -1)
+  ret <4 x double> %res
+}
+
+define <4 x double> @test_rsqrt_pd_256_rrkz(<4 x double> %a0, i8 %mask) {
+; CHECK-LABEL: test_rsqrt_pd_256_rrkz:
+; CHECK: vrsqrt14pd %ymm0, %ymm0 {%k1} {z}
+  %res = call <4 x double> @llvm.x86.avx512.rsqrt14.pd.256(<4 x double> %a0, <4 x double> zeroinitializer, i8 %mask)
+  ret <4 x double> %res
+}
+
+define <4 x double> @test_rsqrt_pd_256_rrk(<4 x double> %a0, <4 x double> %a1, i8 %mask) {
+; CHECK-LABEL: test_rsqrt_pd_256_rrk:
+; CHECK: vrsqrt14pd %ymm0, %ymm1 {%k1}
+  %res = call <4 x double> @llvm.x86.avx512.rsqrt14.pd.256(<4 x double> %a0, <4 x double> %a1, i8 %mask)
+  ret <4 x double> %res
+}
+
+define <2 x double> @test_rsqrt_pd_128_rr(<2 x double> %a0) {
+; CHECK-LABEL: test_rsqrt_pd_128_rr:
+; CHECK: vrsqrt14pd %xmm0, %xmm0
+  %res = call <2 x double> @llvm.x86.avx512.rsqrt14.pd.128(<2 x double> %a0, <2 x double> zeroinitializer, i8 -1)
+  ret <2 x double> %res
+}
+
+define <2 x double> @test_rsqrt_pd_128_rrkz(<2 x double> %a0, i8 %mask) {
+; CHECK-LABEL: test_rsqrt_pd_128_rrkz:
+; CHECK: vrsqrt14pd %xmm0, %xmm0 {%k1} {z}
+  %res = call <2 x double> @llvm.x86.avx512.rsqrt14.pd.128(<2 x double> %a0, <2 x double> zeroinitializer, i8 %mask)
+  ret <2 x double> %res
+}
+
+define <2 x double> @test_rsqrt_pd_128_rrk(<2 x double> %a0, <2 x double> %a1, i8 %mask) {
+; CHECK-LABEL: test_rsqrt_pd_128_rrk:
+; CHECK: vrsqrt14pd %xmm0, %xmm1 {%k1}
+  %res = call <2 x double> @llvm.x86.avx512.rsqrt14.pd.128(<2 x double> %a0, <2 x double> %a1, i8 %mask)
+  ret <2 x double> %res
+}
+
+declare <4 x double> @llvm.x86.avx512.rsqrt14.pd.256(<4 x double>, <4 x double>, i8) nounwind readnone
+declare <2 x double> @llvm.x86.avx512.rsqrt14.pd.128(<2 x double>, <2 x double>, i8) nounwind readnone
+
+define <4 x double> @test_rcp_pd_256_rr(<4 x double> %a0) {
+; CHECK-LABEL: test_rcp_pd_256_rr:
+; CHECK: vrcp14pd %ymm0, %ymm0
+  %res = call <4 x double> @llvm.x86.avx512.rcp14.pd.256(<4 x double> %a0, <4 x double> zeroinitializer, i8 -1)
+  ret <4 x double> %res
+}
+
+define <4 x double> @test_rcp_pd_256_rrkz(<4 x double> %a0, i8 %mask) {
+; CHECK-LABEL: test_rcp_pd_256_rrkz:
+; CHECK: vrcp14pd %ymm0, %ymm0 {%k1} {z}
+  %res = call <4 x double> @llvm.x86.avx512.rcp14.pd.256(<4 x double> %a0, <4 x double> zeroinitializer, i8 %mask)
+  ret <4 x double> %res
+}
+
+define <4 x double> @test_rcp_pd_256_rrk(<4 x double> %a0, <4 x double> %a1, i8 %mask) {
+; CHECK-LABEL: test_rcp_pd_256_rrk:
+; CHECK: vrcp14pd %ymm0, %ymm1 {%k1}
+  %res = call <4 x double> @llvm.x86.avx512.rcp14.pd.256(<4 x double> %a0, <4 x double> %a1, i8 %mask)
+  ret <4 x double> %res
+}
+
+define <2 x double> @test_rcp_pd_128_rr(<2 x double> %a0) {
+; CHECK-LABEL: test_rcp_pd_128_rr:
+; CHECK: vrcp14pd %xmm0, %xmm0
+  %res = call <2 x double> @llvm.x86.avx512.rcp14.pd.128(<2 x double> %a0, <2 x double> zeroinitializer, i8 -1)
+  ret <2 x double> %res
+}
+
+define <2 x double> @test_rcp_pd_128_rrkz(<2 x double> %a0, i8 %mask) {
+; CHECK-LABEL: test_rcp_pd_128_rrkz:
+; CHECK: vrcp14pd %xmm0, %xmm0 {%k1} {z}
+  %res = call <2 x double> @llvm.x86.avx512.rcp14.pd.128(<2 x double> %a0, <2 x double> zeroinitializer, i8 %mask)
+  ret <2 x double> %res
+}
+
+define <2 x double> @test_rcp_pd_128_rrk(<2 x double> %a0, <2 x double> %a1, i8 %mask) {
+; CHECK-LABEL: test_rcp_pd_128_rrk:
+; CHECK: vrcp14pd %xmm0, %xmm1 {%k1}
+  %res = call <2 x double> @llvm.x86.avx512.rcp14.pd.128(<2 x double> %a0, <2 x double> %a1, i8 %mask)
+  ret <2 x double> %res
+}
+
+declare <4 x double> @llvm.x86.avx512.rcp14.pd.256(<4 x double>, <4 x double>, i8) nounwind readnone
+declare <2 x double> @llvm.x86.avx512.rcp14.pd.128(<2 x double>, <2 x double>, i8) nounwind readnone
+
+define <4 x double> @test_x86_vbroadcast_sd_pd_256(<2 x double> %a0, <4 x double> %a1, i8 %mask ) {
+; CHECK-LABEL: test_x86_vbroadcast_sd_pd_256:
+; CHECK: kmovw   %eax, %k1
+; CHECK-NEXT: vbroadcastsd %xmm0, %ymm1 {%k1}
+; CHECK-NEXT: vbroadcastsd %xmm0, %ymm2 {%k1} {z}
+; CHECK-NEXT: vbroadcastsd %xmm0, %ymm0
+; CHECK-NEXT: vaddpd %ymm1, %ymm0, %ymm0
+
+  %res = call <4 x double> @llvm.x86.avx512.mask.broadcast.sd.pd.256(<2 x double> %a0, <4 x double> zeroinitializer, i8 -1) 
+  %res1 = call <4 x double> @llvm.x86.avx512.mask.broadcast.sd.pd.256(<2 x double> %a0, <4 x double> %a1, i8 %mask) 
+  %res2 = call <4 x double> @llvm.x86.avx512.mask.broadcast.sd.pd.256(<2 x double> %a0, <4 x double> zeroinitializer, i8 %mask) 
+  %res3 = fadd <4 x double> %res, %res1
+  %res4 = fadd <4 x double> %res2, %res3
+  ret <4 x double> %res4
+}
+declare <4 x double> @llvm.x86.avx512.mask.broadcast.sd.pd.256(<2 x double>, <4 x double>, i8) nounwind readonly
+
+define <8 x float> @test_x86_vbroadcast_ss_ps_256(<4 x float> %a0, <8 x float> %a1, i8 %mask ) {
+; CHECK-LABEL: test_x86_vbroadcast_ss_ps_256:
+; CHECK: kmovw   %eax, %k1
+; CHECK-NEXT: vbroadcastss %xmm0, %ymm1 {%k1}
+; CHECK-NEXT: vbroadcastss %xmm0, %ymm2 {%k1} {z}
+; CHECK-NEXT: vbroadcastss %xmm0, %ymm0
+; CHECK-NEXT: vaddps %ymm1, %ymm0, %ymm0
+
+  %res = call <8 x float> @llvm.x86.avx512.mask.broadcast.ss.ps.256(<4 x float> %a0, <8 x float> zeroinitializer, i8 -1) 
+  %res1 = call <8 x float> @llvm.x86.avx512.mask.broadcast.ss.ps.256(<4 x float> %a0, <8 x float> %a1, i8 %mask) 
+  %res2 = call <8 x float> @llvm.x86.avx512.mask.broadcast.ss.ps.256(<4 x float> %a0, <8 x float> zeroinitializer, i8 %mask) 
+  %res3 = fadd <8 x float> %res, %res1
+  %res4 = fadd <8 x float> %res2, %res3
+  ret <8 x float> %res4
+}
+declare <8 x float> @llvm.x86.avx512.mask.broadcast.ss.ps.256(<4 x float>, <8 x float>, i8) nounwind readonly
+
+define <4 x float> @test_x86_vbroadcast_ss_ps_128(<4 x float> %a0, <4 x float> %a1, i8 %mask ) {
+; CHECK-LABEL: test_x86_vbroadcast_ss_ps_128:
+; CHECK: kmovw   %eax, %k1
+; CHECK-NEXT: vbroadcastss %xmm0, %xmm1 {%k1}
+; CHECK-NEXT: vbroadcastss %xmm0, %xmm2 {%k1} {z}
+; CHECK-NEXT: vbroadcastss %xmm0, %xmm0
+; CHECK-NEXT: vaddps %xmm1, %xmm0, %xmm0
+
+  %res = call <4 x float> @llvm.x86.avx512.mask.broadcast.ss.ps.128(<4 x float> %a0, <4 x float> zeroinitializer, i8 -1) 
+  %res1 = call <4 x float> @llvm.x86.avx512.mask.broadcast.ss.ps.128(<4 x float> %a0, <4 x float> %a1, i8 %mask) 
+  %res2 = call <4 x float> @llvm.x86.avx512.mask.broadcast.ss.ps.128(<4 x float> %a0, <4 x float> zeroinitializer, i8 %mask) 
+  %res3 = fadd <4 x float> %res, %res1
+  %res4 = fadd <4 x float> %res2, %res3
+  ret <4 x float> %res4
+}
+declare <4 x float> @llvm.x86.avx512.mask.broadcast.ss.ps.128(<4 x float>, <4 x float>, i8) nounwind readonly
+
+
+declare <8 x float> @llvm.x86.avx512.mask.broadcastf32x4.256(<4 x float>, <8 x float>, i8)
+
+define <8 x float>@test_int_x86_avx512_mask_broadcastf32x4_256(<4 x float> %x0, <8 x float> %x2, i8 %mask) {
+; CHECK-LABEL: test_int_x86_avx512_mask_broadcastf32x4_256:
+; CHECK: kmovw %eax, %k1
+; CHECK: vshuff32x4 $0, %ymm0, %ymm0, %ymm2 {%k1} {z}
+; CHECK: vshuff32x4 $0, %ymm0, %ymm0, %ymm1 {%k1}
+; CHECK: vshuff32x4 $0, %ymm0, %ymm0, %ymm0
+; CHECK: vaddps %ymm1, %ymm0, %ymm0
+; CHECK: vaddps %ymm0, %ymm2, %ymm0
+
+  %res1 = call <8 x float> @llvm.x86.avx512.mask.broadcastf32x4.256(<4 x float> %x0, <8 x float> %x2, i8 -1)
+  %res2 = call <8 x float> @llvm.x86.avx512.mask.broadcastf32x4.256(<4 x float> %x0, <8 x float> %x2, i8 %mask)
+  %res3 = call <8 x float> @llvm.x86.avx512.mask.broadcastf32x4.256(<4 x float> %x0, <8 x float> zeroinitializer, i8 %mask)
+  %res4 = fadd <8 x float> %res1, %res2
+  %res5 = fadd <8 x float> %res3, %res4
+  ret <8 x float> %res5
+}
+
+declare <8 x i32> @llvm.x86.avx512.mask.broadcasti32x4.256(<4 x i32>, <8 x i32>, i8)
+
+define <8 x i32>@test_int_x86_avx512_mask_broadcasti32x4_256(<4 x i32> %x0, <8 x i32> %x2, i8 %mask) {
+; CHECK-LABEL: test_int_x86_avx512_mask_broadcasti32x4_256:
+; CHECK: kmovw %eax, %k1
+; CHECK: vshufi32x4 $0, %ymm0, %ymm0, %ymm2 {%k1} {z}
+; CHECK: vshufi32x4 $0, %ymm0, %ymm0, %ymm1 {%k1}
+; CHECK: vshufi32x4 $0, %ymm0, %ymm0, %ymm0
+; CHECK: vpaddd %ymm1, %ymm0, %ymm0
+; CHECK: vpaddd %ymm0, %ymm2, %ymm0
+
+  %res1 = call <8 x i32> @llvm.x86.avx512.mask.broadcasti32x4.256(<4 x i32> %x0, <8 x i32> %x2, i8 -1)
+  %res2 = call <8 x i32> @llvm.x86.avx512.mask.broadcasti32x4.256(<4 x i32> %x0, <8 x i32> %x2, i8 %mask)
+  %res3 = call <8 x i32> @llvm.x86.avx512.mask.broadcasti32x4.256(<4 x i32> %x0, <8 x i32> zeroinitializer, i8 %mask)
+  %res4 = add <8 x i32> %res1, %res2
+  %res5 = add <8 x i32> %res3, %res4
+  ret <8 x i32> %res5
+}
+
+declare <2 x i64> @llvm.x86.avx512.mask.psrl.q.128(<2 x i64>, <2 x i64>, <2 x i64>, i8)
+
+define <2 x i64>@test_int_x86_avx512_mask_psrl_q_128(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> %x2, i8 %x3) {
+; CHECK-LABEL: test_int_x86_avx512_mask_psrl_q_128:
+; CHECK:       ## BB#0:
+; CHECK-NEXT:    movzbl %dil, %eax
+; CHECK-NEXT:    kmovw %eax, %k1
+; CHECK-NEXT:    vpsrlq %xmm1, %xmm0, %xmm2 {%k1}
+; CHECK-NEXT:    vpsrlq %xmm1, %xmm0, %xmm3 {%k1} {z}
+; CHECK-NEXT:    vpsrlq %xmm1, %xmm0, %xmm0
+; CHECK-NEXT:    vpaddq %xmm0, %xmm2, %xmm0
+; CHECK-NEXT:    vpaddq %xmm3, %xmm0, %xmm0
+; CHECK-NEXT:    retq
+  %res = call <2 x i64> @llvm.x86.avx512.mask.psrl.q.128(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> %x2, i8 %x3)
+  %res1 = call <2 x i64> @llvm.x86.avx512.mask.psrl.q.128(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> %x2, i8 -1)
+  %res2 = call <2 x i64> @llvm.x86.avx512.mask.psrl.q.128(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> zeroinitializer, i8 %x3)
+  %res3 = add <2 x i64> %res, %res1
+  %res4 = add <2 x i64> %res3, %res2
+  ret <2 x i64> %res4
+}
+
+declare <4 x i64> @llvm.x86.avx512.mask.psrl.q.256(<4 x i64>, <2 x i64>, <4 x i64>, i8)
+
+define <4 x i64>@test_int_x86_avx512_mask_psrl_q_256(<4 x i64> %x0, <2 x i64> %x1, <4 x i64> %x2, i8 %x3) {
+; CHECK-LABEL: test_int_x86_avx512_mask_psrl_q_256:
+; CHECK:       ## BB#0:
+; CHECK-NEXT:    movzbl %dil, %eax
+; CHECK-NEXT:    kmovw %eax, %k1
+; CHECK-NEXT:    vpsrlq %xmm1, %ymm0, %ymm2 {%k1}
+; CHECK-NEXT:    vpsrlq %xmm1, %ymm0, %ymm3 {%k1} {z}
+; CHECK-NEXT:    vpsrlq %xmm1, %ymm0, %ymm0
+; CHECK-NEXT:    vpaddq %ymm0, %ymm2, %ymm0
+; CHECK-NEXT:    vpaddq %ymm3, %ymm0, %ymm0
+; CHECK-NEXT:    retq
+  %res = call <4 x i64> @llvm.x86.avx512.mask.psrl.q.256(<4 x i64> %x0, <2 x i64> %x1, <4 x i64> %x2, i8 %x3)
+  %res1 = call <4 x i64> @llvm.x86.avx512.mask.psrl.q.256(<4 x i64> %x0, <2 x i64> %x1, <4 x i64> %x2, i8 -1)
+  %res2 = call <4 x i64> @llvm.x86.avx512.mask.psrl.q.256(<4 x i64> %x0, <2 x i64> %x1, <4 x i64> zeroinitializer, i8 %x3)
+  %res3 = add <4 x i64> %res, %res1
+  %res4 = add <4 x i64> %res3, %res2
+  ret <4 x i64> %res4
+}
+
+declare <2 x i64> @llvm.x86.avx512.mask.psrl.qi.128(<2 x i64>, i8, <2 x i64>, i8)
+
+define <2 x i64>@test_int_x86_avx512_mask_psrl_qi_128(<2 x i64> %x0, i8 %x1, <2 x i64> %x2, i8 %x3) {
+; CHECK-LABEL: test_int_x86_avx512_mask_psrl_qi_128:
+; CHECK:       ## BB#0:
+; CHECK-NEXT:    movzbl %sil, %eax
+; CHECK-NEXT:    kmovw %eax, %k1
+; CHECK-NEXT:    vpsrlq $255, %xmm0, %xmm1 {%k1}
+; CHECK-NEXT:    vpsrlq $255, %xmm0, %xmm2 {%k1} {z}
+; CHECK-NEXT:    vpsrlq $255, %xmm0, %xmm0
+; CHECK-NEXT:    vpaddq %xmm0, %xmm1, %xmm0
+; CHECK-NEXT:    vpaddq %xmm0, %xmm2, %xmm0
+; CHECK-NEXT:    retq
+  %res = call <2 x i64> @llvm.x86.avx512.mask.psrl.qi.128(<2 x i64> %x0, i8 255, <2 x i64> %x2, i8 %x3)
+  %res1 = call <2 x i64> @llvm.x86.avx512.mask.psrl.qi.128(<2 x i64> %x0, i8 255, <2 x i64> %x2, i8 -1)
+  %res2 = call <2 x i64> @llvm.x86.avx512.mask.psrl.qi.128(<2 x i64> %x0, i8 255, <2 x i64> zeroinitializer, i8 %x3)
+  %res3 = add <2 x i64> %res, %res1
+  %res4 = add <2 x i64> %res2, %res3
+  ret <2 x i64> %res4
+}
+
+declare <4 x i64> @llvm.x86.avx512.mask.psrl.qi.256(<4 x i64>, i8, <4 x i64>, i8)
+
+define <4 x i64>@test_int_x86_avx512_mask_psrl_qi_256(<4 x i64> %x0, i8 %x1, <4 x i64> %x2, i8 %x3) {
+; CHECK-LABEL: test_int_x86_avx512_mask_psrl_qi_256:
+; CHECK:       ## BB#0:
+; CHECK-NEXT:    movzbl %sil, %eax
+; CHECK-NEXT:    kmovw %eax, %k1
+; CHECK-NEXT:    vpsrlq $255, %ymm0, %ymm1 {%k1}
+; CHECK-NEXT:    vpsrlq $255, %ymm0, %ymm2 {%k1} {z}
+; CHECK-NEXT:    vpsrlq $255, %ymm0, %ymm0
+; CHECK-NEXT:    vpaddq %ymm0, %ymm1, %ymm0
+; CHECK-NEXT:    vpaddq %ymm0, %ymm2, %ymm0
+; CHECK-NEXT:    retq
+  %res = call <4 x i64> @llvm.x86.avx512.mask.psrl.qi.256(<4 x i64> %x0, i8 255, <4 x i64> %x2, i8 %x3)
+  %res1 = call <4 x i64> @llvm.x86.avx512.mask.psrl.qi.256(<4 x i64> %x0, i8 255, <4 x i64> %x2, i8 -1)
+  %res2 = call <4 x i64> @llvm.x86.avx512.mask.psrl.qi.256(<4 x i64> %x0, i8 255, <4 x i64> zeroinitializer, i8 %x3)
+  %res3 = add <4 x i64> %res, %res1
+  %res4 = add <4 x i64> %res2, %res3
+  ret <4 x i64> %res4
+}
+declare <4 x i32> @llvm.x86.avx512.mask.psrl.d.128(<4 x i32>, <4 x i32>, <4 x i32>, i8)
+define <4 x i32>@test_int_x86_avx512_mask_psrl_d_128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x2, i8 %x3) {
+; CHECK-LABEL: test_int_x86_avx512_mask_psrl_d_128:
+; CHECK:       ## BB#0:
+; CHECK-NEXT:    movzbl %dil, %eax
+; CHECK-NEXT:    kmovw %eax, %k1
+; CHECK-NEXT:    vpsrld %xmm1, %xmm0, %xmm2 {%k1}
+; CHECK-NEXT:    vpsrld %xmm1, %xmm0, %xmm3 {%k1} {z}
+; CHECK-NEXT:    vpsrld %xmm1, %xmm0, %xmm0
+; CHECK-NEXT:    vpaddd %xmm0, %xmm2, %xmm0
+; CHECK-NEXT:    vpaddd %xmm3, %xmm0, %xmm0
+; CHECK-NEXT:    retq
+  %res = call <4 x i32> @llvm.x86.avx512.mask.psrl.d.128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x2, i8 %x3)
+  %res1 = call <4 x i32> @llvm.x86.avx512.mask.psrl.d.128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x2, i8 -1)
+  %res2 = call <4 x i32> @llvm.x86.avx512.mask.psrl.d.128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> zeroinitializer, i8 %x3)
+  %res3 = add <4 x i32> %res, %res1
+  %res4 = add <4 x i32> %res3, %res2
+  ret <4 x i32> %res4
+}
+
+declare <8 x i32> @llvm.x86.avx512.mask.psrl.d.256(<8 x i32>, <4 x i32>, <8 x i32>, i8)
+
+define <8 x i32>@test_int_x86_avx512_mask_psrl_d_256(<8 x i32> %x0, <4 x i32> %x1, <8 x i32> %x2, i8 %x3) {
+; CHECK-LABEL: test_int_x86_avx512_mask_psrl_d_256:
+; CHECK:       ## BB#0:
+; CHECK-NEXT:    movzbl %dil, %eax
+; CHECK-NEXT:    kmovw %eax, %k1
+; CHECK-NEXT:    vpsrld %xmm1, %ymm0, %ymm2 {%k1}
+; CHECK-NEXT:    vpsrld %xmm1, %ymm0, %ymm3 {%k1} {z}
+; CHECK-NEXT:    vpsrld %xmm1, %ymm0, %ymm0
+; CHECK-NEXT:    vpaddd %ymm0, %ymm2, %ymm0
+; CHECK-NEXT:    vpaddd %ymm0, %ymm3, %ymm0
+; CHECK-NEXT:    retq
+  %res = call <8 x i32> @llvm.x86.avx512.mask.psrl.d.256(<8 x i32> %x0, <4 x i32> %x1, <8 x i32> %x2, i8 %x3)
+  %res1 = call <8 x i32> @llvm.x86.avx512.mask.psrl.d.256(<8 x i32> %x0, <4 x i32> %x1, <8 x i32> %x2, i8 -1)
+  %res2 = call <8 x i32> @llvm.x86.avx512.mask.psrl.d.256(<8 x i32> %x0, <4 x i32> %x1, <8 x i32> zeroinitializer, i8 %x3)
+  %res3 = add <8 x i32> %res, %res1
+  %res4 = add <8 x i32> %res2, %res3
+  ret <8 x i32> %res4
+}
+
+declare <4 x i32> @llvm.x86.avx512.mask.psrl.di.128(<4 x i32>, i8, <4 x i32>, i8)
+
+define <4 x i32>@test_int_x86_avx512_mask_psrl_di_128(<4 x i32> %x0, i8 %x1, <4 x i32> %x2, i8 %x3) {
+; CHECK-LABEL: test_int_x86_avx512_mask_psrl_di_128:
+; CHECK:       ## BB#0:
+; CHECK-NEXT:    movzbl %sil, %eax
+; CHECK-NEXT:    kmovw %eax, %k1
+; CHECK-NEXT:    vpsrld $255, %xmm0, %xmm1 {%k1}
+; CHECK-NEXT:    vpsrld $255, %xmm0, %xmm2 {%k1} {z}
+; CHECK-NEXT:    vpsrld $255, %xmm0, %xmm0
+; CHECK-NEXT:    vpaddd %xmm0, %xmm1, %xmm0
+; CHECK-NEXT:    vpaddd %xmm0, %xmm2, %xmm0
+; CHECK-NEXT:    retq
+  %res = call <4 x i32> @llvm.x86.avx512.mask.psrl.di.128(<4 x i32> %x0, i8 255, <4 x i32> %x2, i8 %x3)
+  %res1 = call <4 x i32> @llvm.x86.avx512.mask.psrl.di.128(<4 x i32> %x0, i8 255, <4 x i32> %x2, i8 -1)
+  %res2 = call <4 x i32> @llvm.x86.avx512.mask.psrl.di.128(<4 x i32> %x0, i8 255, <4 x i32> zeroinitializer, i8 %x3)
+  %res3 = add <4 x i32> %res, %res1
+  %res4 = add <4 x i32> %res2, %res3
+  ret <4 x i32> %res4
+}
+
+declare <8 x i32> @llvm.x86.avx512.mask.psrl.di.256(<8 x i32>, i8, <8 x i32>, i8)
+
+define <8 x i32>@test_int_x86_avx512_mask_psrl_di_256(<8 x i32> %x0, i8 %x1, <8 x i32> %x2, i8 %x3) {
+; CHECK-LABEL: test_int_x86_avx512_mask_psrl_di_256:
+; CHECK:       ## BB#0:
+; CHECK-NEXT:    movzbl %sil, %eax
+; CHECK-NEXT:    kmovw %eax, %k1
+; CHECK-NEXT:    vpsrld $255, %ymm0, %ymm1 {%k1}
+; CHECK-NEXT:    vpsrld $255, %ymm0, %ymm2 {%k1} {z}
+; CHECK-NEXT:    vpsrld $255, %ymm0, %ymm0
+; CHECK-NEXT:    vpaddd %ymm0, %ymm1, %ymm0
+; CHECK-NEXT:    vpaddd %ymm0, %ymm2, %ymm0
+; CHECK-NEXT:    retq
+  %res = call <8 x i32> @llvm.x86.avx512.mask.psrl.di.256(<8 x i32> %x0, i8 255, <8 x i32> %x2, i8 %x3)
+  %res1 = call <8 x i32> @llvm.x86.avx512.mask.psrl.di.256(<8 x i32> %x0, i8 255, <8 x i32> %x2, i8 -1)
+  %res2 = call <8 x i32> @llvm.x86.avx512.mask.psrl.di.256(<8 x i32> %x0, i8 255, <8 x i32> zeroinitializer, i8 %x3)
+  %res3 = add <8 x i32> %res, %res1
+  %res4 = add <8 x i32> %res2, %res3
+  ret <8 x i32> %res4
+}
+
+declare <16 x i32> @llvm.x86.avx512.mask.psrl.di.512(<16 x i32>, i8, <16 x i32>, i16)
+
+define <16 x i32>@test_int_x86_avx512_mask_psrl_di_512(<16 x i32> %x0, i8 %x1, <16 x i32> %x2, i16 %x3) {
+; CHECK-LABEL: test_int_x86_avx512_mask_psrl_di_512:
+; CHECK:       ## BB#0:
+; CHECK-NEXT:    kmovw %esi, %k1
+; CHECK-NEXT:    vpsrld $255, %zmm0, %zmm1 {%k1}
+; CHECK-NEXT:    vpsrld $255, %zmm0, %zmm2 {%k1} {z}
+; CHECK-NEXT:    vpsrld $255, %zmm0, %zmm0
+; CHECK-NEXT:    vpaddd %zmm0, %zmm1, %zmm0
+; CHECK-NEXT:    vpaddd %zmm0, %zmm2, %zmm0
+; CHECK-NEXT:    retq
+  %res = call <16 x i32> @llvm.x86.avx512.mask.psrl.di.512(<16 x i32> %x0, i8 255, <16 x i32> %x2, i16 %x3)
+  %res1 = call <16 x i32> @llvm.x86.avx512.mask.psrl.di.512(<16 x i32> %x0, i8 255, <16 x i32> %x2, i16 -1)
+  %res2 = call <16 x i32> @llvm.x86.avx512.mask.psrl.di.512(<16 x i32> %x0, i8 255, <16 x i32> zeroinitializer, i16 %x3)
+  %res3 = add <16 x i32> %res, %res1
+  %res4 = add <16 x i32> %res2, %res3
+  ret <16 x i32> %res4
+}
+
+declare <2 x i64> @llvm.x86.avx512.mask.psrlv2.di(<2 x i64>, <2 x i64>, <2 x i64>, i8)
+
+define <2 x i64>@test_int_x86_avx512_mask_psrlv2_di(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> %x2, i8 %x3) {
+; CHECK-LABEL: test_int_x86_avx512_mask_psrlv2_di:
+; CHECK:       ## BB#0:
+; CHECK-NEXT:    movzbl %dil, %eax
+; CHECK-NEXT:    kmovw %eax, %k1
+; CHECK-NEXT:    vpsrlvq %xmm1, %xmm0, %xmm2 {%k1}
+; CHECK-NEXT:    vpsrlvq %xmm1, %xmm0, %xmm3 {%k1} {z}
+; CHECK-NEXT:    vpsrlvq %xmm1, %xmm0, %xmm0
+; CHECK-NEXT:    vpaddq %xmm3, %xmm2, %xmm1
+; CHECK-NEXT:    vpaddq %xmm0, %xmm1, %xmm0
+; CHECK-NEXT:    retq
+  %res = call <2 x i64> @llvm.x86.avx512.mask.psrlv2.di(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> %x2, i8 %x3)
+  %res1 = call <2 x i64> @llvm.x86.avx512.mask.psrlv2.di(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> zeroinitializer, i8 %x3)
+  %res2 = call <2 x i64> @llvm.x86.avx512.mask.psrlv2.di(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> %x2, i8 -1)
+  %res3 = add <2 x i64> %res, %res1
+  %res4 = add <2 x i64> %res3, %res2
+  ret <2 x i64> %res4
+}
+
+declare <4 x i64> @llvm.x86.avx512.mask.psrlv4.di(<4 x i64>, <4 x i64>, <4 x i64>, i8)
+
+define <4 x i64>@test_int_x86_avx512_mask_psrlv4_di(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> %x2, i8 %x3) {
+; CHECK-LABEL: test_int_x86_avx512_mask_psrlv4_di:
+; CHECK:       ## BB#0:
+; CHECK-NEXT:    movzbl %dil, %eax
+; CHECK-NEXT:    kmovw %eax, %k1
+; CHECK-NEXT:    vpsrlvq %ymm1, %ymm0, %ymm2 {%k1}
+; CHECK-NEXT:    vpsrlvq %ymm1, %ymm0, %ymm3 {%k1} {z}
+; CHECK-NEXT:    vpsrlvq %ymm1, %ymm0, %ymm0
+; CHECK-NEXT:    vpaddq %ymm3, %ymm2, %ymm1
+; CHECK-NEXT:    vpaddq %ymm0, %ymm1, %ymm0
+; CHECK-NEXT:    retq
+  %res = call <4 x i64> @llvm.x86.avx512.mask.psrlv4.di(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> %x2, i8 %x3)
+  %res1 = call <4 x i64> @llvm.x86.avx512.mask.psrlv4.di(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> zeroinitializer, i8 %x3)
+  %res2 = call <4 x i64> @llvm.x86.avx512.mask.psrlv4.di(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> %x2, i8 -1)
+  %res3 = add <4 x i64> %res, %res1
+  %res4 = add <4 x i64> %res3, %res2
+  ret <4 x i64> %res4
+}
+
+declare <4 x i32> @llvm.x86.avx512.mask.psrlv4.si(<4 x i32>, <4 x i32>, <4 x i32>, i8)
+
+define <4 x i32>@test_int_x86_avx512_mask_psrlv4_si(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x2, i8 %x3) {
+; CHECK-LABEL: test_int_x86_avx512_mask_psrlv4_si:
+; CHECK:       ## BB#0:
+; CHECK-NEXT:    movzbl %dil, %eax
+; CHECK-NEXT:    kmovw %eax, %k1
+; CHECK-NEXT:    vpsrlvd %xmm1, %xmm0, %xmm2 {%k1}
+; CHECK-NEXT:    vpsrlvd %xmm1, %xmm0, %xmm3 {%k1} {z}
+; CHECK-NEXT:    vpsrlvd %xmm1, %xmm0, %xmm0
+; CHECK-NEXT:    vpaddd %xmm3, %xmm2, %xmm1
+; CHECK-NEXT:    vpaddd %xmm0, %xmm1, %xmm0
+; CHECK-NEXT:    retq
+  %res = call <4 x i32> @llvm.x86.avx512.mask.psrlv4.si(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x2, i8 %x3)
+  %res1 = call <4 x i32> @llvm.x86.avx512.mask.psrlv4.si(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> zeroinitializer, i8 %x3)
+  %res2 = call <4 x i32> @llvm.x86.avx512.mask.psrlv4.si(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x2, i8 -1)
+  %res3 = add <4 x i32> %res, %res1
+  %res4 = add <4 x i32> %res3, %res2
+  ret <4 x i32> %res4
+}
+
+declare <8 x i32> @llvm.x86.avx512.mask.psrlv8.si(<8 x i32>, <8 x i32>, <8 x i32>, i8)
+
+define <8 x i32>@test_int_x86_avx512_mask_psrlv8_si(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x2, i8 %x3) {
+; CHECK-LABEL: test_int_x86_avx512_mask_psrlv8_si:
+; CHECK:       ## BB#0:
+; CHECK-NEXT:    movzbl %dil, %eax
+; CHECK-NEXT:    kmovw %eax, %k1
+; CHECK-NEXT:    vpsrlvd %ymm1, %ymm0, %ymm2 {%k1}
+; CHECK-NEXT:    vpsrlvd %ymm1, %ymm0, %ymm3 {%k1} {z}
+; CHECK-NEXT:    vpsrlvd %ymm1, %ymm0, %ymm0
+; CHECK-NEXT:    vpaddd %ymm3, %ymm2, %ymm1
+; CHECK-NEXT:    vpaddd %ymm0, %ymm1, %ymm0
+; CHECK-NEXT:    retq
+  %res = call <8 x i32> @llvm.x86.avx512.mask.psrlv8.si(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x2, i8 %x3)
+  %res1 = call <8 x i32> @llvm.x86.avx512.mask.psrlv8.si(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> zeroinitializer, i8 %x3)
+  %res2 = call <8 x i32> @llvm.x86.avx512.mask.psrlv8.si(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x2, i8 -1)
+  %res3 = add <8 x i32> %res, %res1
+  %res4 = add <8 x i32> %res3, %res2
+  ret <8 x i32> %res4
+}
+
+declare <4 x i32> @llvm.x86.avx512.mask.psra.d.128(<4 x i32>, <4 x i32>, <4 x i32>, i8)
+
+define <4 x i32>@test_int_x86_avx512_mask_psra_d_128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x2, i8 %x3) {
+; CHECK-LABEL: test_int_x86_avx512_mask_psra_d_128:
+; CHECK:       ## BB#0:
+; CHECK-NEXT:    movzbl %dil, %eax
+; CHECK-NEXT:    kmovw %eax, %k1
+; CHECK-NEXT:    vpsrad %xmm1, %xmm0, %xmm2 {%k1}
+; CHECK-NEXT:    vpsrad %xmm1, %xmm0, %xmm3 {%k1} {z}
+; CHECK-NEXT:    vpsrad %xmm1, %xmm0, %xmm0
+; CHECK-NEXT:    vpaddd %xmm3, %xmm2, %xmm1
+; CHECK-NEXT:    vpaddd %xmm0, %xmm1, %xmm0
+; CHECK-NEXT:    retq
+  %res = call <4 x i32> @llvm.x86.avx512.mask.psra.d.128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x2, i8 %x3)
+  %res1 = call <4 x i32> @llvm.x86.avx512.mask.psra.d.128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> zeroinitializer, i8 %x3)
+  %res2 = call <4 x i32> @llvm.x86.avx512.mask.psra.d.128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x2, i8 -1)
+  %res3 = add <4 x i32> %res, %res1
+  %res4 = add <4 x i32> %res3, %res2
+  ret <4 x i32> %res4
+}
+
+declare <8 x i32> @llvm.x86.avx512.mask.psra.d.256(<8 x i32>, <4 x i32>, <8 x i32>, i8)
+
+define <8 x i32>@test_int_x86_avx512_mask_psra_d_256(<8 x i32> %x0, <4 x i32> %x1, <8 x i32> %x2, i8 %x3) {
+; CHECK-LABEL: test_int_x86_avx512_mask_psra_d_256:
+; CHECK:       ## BB#0:
+; CHECK-NEXT:    movzbl %dil, %eax
+; CHECK-NEXT:    kmovw %eax, %k1
+; CHECK-NEXT:    vpsrad %xmm1, %ymm0, %ymm2 {%k1}
+; CHECK-NEXT:    vpsrad %xmm1, %ymm0, %ymm3 {%k1} {z}
+; CHECK-NEXT:    vpsrad %xmm1, %ymm0, %ymm0
+; CHECK-NEXT:    vpaddd %ymm3, %ymm2, %ymm1
+; CHECK-NEXT:    vpaddd %ymm0, %ymm1, %ymm0
+; CHECK-NEXT:    retq
+  %res = call <8 x i32> @llvm.x86.avx512.mask.psra.d.256(<8 x i32> %x0, <4 x i32> %x1, <8 x i32> %x2, i8 %x3)
+  %res1 = call <8 x i32> @llvm.x86.avx512.mask.psra.d.256(<8 x i32> %x0, <4 x i32> %x1, <8 x i32> zeroinitializer, i8 %x3)
+  %res2 = call <8 x i32> @llvm.x86.avx512.mask.psra.d.256(<8 x i32> %x0, <4 x i32> %x1, <8 x i32> %x2, i8 -1)
+  %res3 = add <8 x i32> %res, %res1
+  %res4 = add <8 x i32> %res3, %res2
+  ret <8 x i32> %res4
+}
+
+declare <4 x i32> @llvm.x86.avx512.mask.psra.di.128(<4 x i32>, i8, <4 x i32>, i8)
+
+define <4 x i32>@test_int_x86_avx512_mask_psra_di_128(<4 x i32> %x0, i8 %x1, <4 x i32> %x2, i8 %x3) {
+; CHECK-LABEL: test_int_x86_avx512_mask_psra_di_128:
+; CHECK:       ## BB#0:
+; CHECK-NEXT:    movzbl %sil, %eax
+; CHECK-NEXT:    kmovw %eax, %k1
+; CHECK-NEXT:    vpsrad $3, %xmm0, %xmm1 {%k1}
+; CHECK-NEXT:    vpsrad $3, %xmm0, %xmm2 {%k1} {z}
+; CHECK-NEXT:    vpsrad $3, %xmm0, %xmm0
+; CHECK-NEXT:    vpaddd %xmm2, %xmm1, %xmm1
+; CHECK-NEXT:    vpaddd %xmm0, %xmm1, %xmm0
+; CHECK-NEXT:    retq
+  %res = call <4 x i32> @llvm.x86.avx512.mask.psra.di.128(<4 x i32> %x0, i8 3, <4 x i32> %x2, i8 %x3)
+  %res1 = call <4 x i32> @llvm.x86.avx512.mask.psra.di.128(<4 x i32> %x0, i8 3, <4 x i32> zeroinitializer, i8 %x3)
+  %res2 = call <4 x i32> @llvm.x86.avx512.mask.psra.di.128(<4 x i32> %x0, i8 3, <4 x i32> %x2, i8 -1)
+  %res3 = add <4 x i32> %res, %res1
+  %res4 = add <4 x i32> %res3, %res2
+  ret <4 x i32> %res4
+}
+
+declare <8 x i32> @llvm.x86.avx512.mask.psra.di.256(<8 x i32>, i8, <8 x i32>, i8)
+
+define <8 x i32>@test_int_x86_avx512_mask_psra_di_256(<8 x i32> %x0, i8 %x1, <8 x i32> %x2, i8 %x3) {
+; CHECK-LABEL: test_int_x86_avx512_mask_psra_di_256:
+; CHECK:       ## BB#0:
+; CHECK-NEXT:    movzbl %sil, %eax
+; CHECK-NEXT:    kmovw %eax, %k1
+; CHECK-NEXT:    vpsrad $3, %ymm0, %ymm1 {%k1}
+; CHECK-NEXT:    vpsrad $3, %ymm0, %ymm2 {%k1} {z}
+; CHECK-NEXT:    vpsrad $3, %ymm0, %ymm0
+; CHECK-NEXT:    vpaddd %ymm2, %ymm1, %ymm1
+; CHECK-NEXT:    vpaddd %ymm0, %ymm1, %ymm0
+; CHECK-NEXT:    retq
+  %res = call <8 x i32> @llvm.x86.avx512.mask.psra.di.256(<8 x i32> %x0, i8 3, <8 x i32> %x2, i8 %x3)
+  %res1 = call <8 x i32> @llvm.x86.avx512.mask.psra.di.256(<8 x i32> %x0, i8 3, <8 x i32> zeroinitializer, i8 %x3)
+  %res2 = call <8 x i32> @llvm.x86.avx512.mask.psra.di.256(<8 x i32> %x0, i8 3, <8 x i32> %x2, i8 -1)
+  %res3 = add <8 x i32> %res, %res1
+  %res4 = add <8 x i32> %res3, %res2
+  ret <8 x i32> %res4
+}
+
+declare <2 x i64> @llvm.x86.avx512.mask.psra.q.128(<2 x i64>, <2 x i64>, <2 x i64>, i8)
+
+define <2 x i64>@test_int_x86_avx512_mask_psra_q_128(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> %x2, i8 %x3) {
+; CHECK-LABEL: test_int_x86_avx512_mask_psra_q_128:
+; CHECK:       ## BB#0:
+; CHECK-NEXT:    movzbl %dil, %eax
+; CHECK-NEXT:    kmovw %eax, %k1
+; CHECK-NEXT:    vpsraq %xmm1, %xmm0, %xmm2 {%k1}
+; CHECK-NEXT:    vpsraq %xmm1, %xmm0, %xmm3 {%k1} {z}
+; CHECK-NEXT:    vpsraq %xmm1, %xmm0, %xmm0
+; CHECK-NEXT:    vpaddq %xmm3, %xmm2, %xmm1
+; CHECK-NEXT:    vpaddq %xmm0, %xmm1, %xmm0
+; CHECK-NEXT:    retq
+  %res = call <2 x i64> @llvm.x86.avx512.mask.psra.q.128(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> %x2, i8 %x3)
+  %res1 = call <2 x i64> @llvm.x86.avx512.mask.psra.q.128(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> zeroinitializer, i8 %x3)
+  %res2 = call <2 x i64> @llvm.x86.avx512.mask.psra.q.128(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> %x2, i8 -1)
+  %res3 = add <2 x i64> %res, %res1
+  %res4 = add <2 x i64> %res3, %res2
+  ret <2 x i64> %res4
+}
+
+declare <4 x i64> @llvm.x86.avx512.mask.psra.q.256(<4 x i64>, <2 x i64>, <4 x i64>, i8)
+
+define <4 x i64>@test_int_x86_avx512_mask_psra_q_256(<4 x i64> %x0, <2 x i64> %x1, <4 x i64> %x2, i8 %x3) {
+; CHECK-LABEL: test_int_x86_avx512_mask_psra_q_256:
+; CHECK:       ## BB#0:
+; CHECK-NEXT:    movzbl %dil, %eax
+; CHECK-NEXT:    kmovw %eax, %k1
+; CHECK-NEXT:    vpsraq %xmm1, %ymm0, %ymm2 {%k1}
+; CHECK-NEXT:    vpsraq %xmm1, %ymm0, %ymm3 {%k1} {z}
+; CHECK-NEXT:    vpsraq %xmm1, %ymm0, %ymm0
+; CHECK-NEXT:    vpaddq %ymm3, %ymm2, %ymm1
+; CHECK-NEXT:    vpaddq %ymm0, %ymm1, %ymm0
+; CHECK-NEXT:    retq
+  %res = call <4 x i64> @llvm.x86.avx512.mask.psra.q.256(<4 x i64> %x0, <2 x i64> %x1, <4 x i64> %x2, i8 %x3)
+  %res1 = call <4 x i64> @llvm.x86.avx512.mask.psra.q.256(<4 x i64> %x0, <2 x i64> %x1, <4 x i64> zeroinitializer, i8 %x3)
+  %res2 = call <4 x i64> @llvm.x86.avx512.mask.psra.q.256(<4 x i64> %x0, <2 x i64> %x1, <4 x i64> %x2, i8 -1)
+  %res3 = add <4 x i64> %res, %res1
+  %res4 = add <4 x i64> %res3, %res2
+  ret <4 x i64> %res4
+}
+
+declare <2 x i64> @llvm.x86.avx512.mask.psra.qi.128(<2 x i64>, i8, <2 x i64>, i8)
+
+define <2 x i64>@test_int_x86_avx512_mask_psra_qi_128(<2 x i64> %x0, i8 %x1, <2 x i64> %x2, i8 %x3) {
+; CHECK-LABEL: test_int_x86_avx512_mask_psra_qi_128:
+; CHECK:       ## BB#0:
+; CHECK-NEXT:    movzbl %sil, %eax
+; CHECK-NEXT:    kmovw %eax, %k1
+; CHECK-NEXT:    vpsraq $3, %xmm0, %xmm1 {%k1}
+; CHECK-NEXT:    vpsraq $3, %xmm0, %xmm2 {%k1} {z}
+; CHECK-NEXT:    vpsraq $3, %xmm0, %xmm0
+; CHECK-NEXT:    vpaddq %xmm2, %xmm1, %xmm1
+; CHECK-NEXT:    vpaddq %xmm0, %xmm1, %xmm0
+; CHECK-NEXT:    retq
+  %res = call <2 x i64> @llvm.x86.avx512.mask.psra.qi.128(<2 x i64> %x0, i8 3, <2 x i64> %x2, i8 %x3)
+  %res1 = call <2 x i64> @llvm.x86.avx512.mask.psra.qi.128(<2 x i64> %x0, i8 3, <2 x i64> zeroinitializer, i8 %x3)
+  %res2 = call <2 x i64> @llvm.x86.avx512.mask.psra.qi.128(<2 x i64> %x0, i8 3, <2 x i64> %x2, i8 -1)
+  %res3 = add <2 x i64> %res, %res1
+  %res4 = add <2 x i64> %res3, %res2
+  ret <2 x i64> %res4
+}
+
+declare <4 x i64> @llvm.x86.avx512.mask.psra.qi.256(<4 x i64>, i8, <4 x i64>, i8)
+
+define <4 x i64>@test_int_x86_avx512_mask_psra_qi_256(<4 x i64> %x0, i8 %x1, <4 x i64> %x2, i8 %x3) {
+; CHECK-LABEL: test_int_x86_avx512_mask_psra_qi_256:
+; CHECK:       ## BB#0:
+; CHECK-NEXT:    movzbl %sil, %eax
+; CHECK-NEXT:    kmovw %eax, %k1
+; CHECK-NEXT:    vpsraq $3, %ymm0, %ymm1 {%k1}
+; CHECK-NEXT:    vpsraq $3, %ymm0, %ymm2 {%k1} {z}
+; CHECK-NEXT:    vpsraq $3, %ymm0, %ymm0
+; CHECK-NEXT:    vpaddq %ymm2, %ymm1, %ymm1
+; CHECK-NEXT:    vpaddq %ymm0, %ymm1, %ymm0
+; CHECK-NEXT:    retq
+  %res = call <4 x i64> @llvm.x86.avx512.mask.psra.qi.256(<4 x i64> %x0, i8 3, <4 x i64> %x2, i8 %x3)
+  %res1 = call <4 x i64> @llvm.x86.avx512.mask.psra.qi.256(<4 x i64> %x0, i8 3, <4 x i64> zeroinitializer, i8 %x3)
+  %res2 = call <4 x i64> @llvm.x86.avx512.mask.psra.qi.256(<4 x i64> %x0, i8 3, <4 x i64> %x2, i8 -1)
+  %res3 = add <4 x i64> %res, %res1
+  %res4 = add <4 x i64> %res3, %res2
+  ret <4 x i64> %res4
+}
+
+
+declare <4 x i32> @llvm.x86.avx512.mask.psll.d.128(<4 x i32>, <4 x i32>, <4 x i32>, i8)
+
+define <4 x i32>@test_int_x86_avx512_mask_psll_d_128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x2, i8 %x3) {
+; CHECK-LABEL: test_int_x86_avx512_mask_psll_d_128:
+; CHECK:       ## BB#0:
+; CHECK-NEXT:    movzbl %dil, %eax
+; CHECK-NEXT:    kmovw %eax, %k1
+; CHECK-NEXT:    vpslld %xmm1, %xmm0, %xmm2 {%k1}
+; CHECK-NEXT:    vpslld %xmm1, %xmm0, %xmm3 {%k1} {z}
+; CHECK-NEXT:    vpslld %xmm1, %xmm0, %xmm0
+; CHECK-NEXT:    vpaddd %xmm3, %xmm2, %xmm1
+; CHECK-NEXT:    vpaddd %xmm0, %xmm1, %xmm0
+; CHECK-NEXT:    retq
+  %res = call <4 x i32> @llvm.x86.avx512.mask.psll.d.128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x2, i8 %x3)
+  %res1 = call <4 x i32> @llvm.x86.avx512.mask.psll.d.128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> zeroinitializer, i8 %x3)
+  %res2 = call <4 x i32> @llvm.x86.avx512.mask.psll.d.128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x2, i8 -1)
+  %res3 = add <4 x i32> %res, %res1
+  %res4 = add <4 x i32> %res3, %res2
+  ret <4 x i32> %res4
+}
+
+declare <8 x i32> @llvm.x86.avx512.mask.psll.d.256(<8 x i32>, <4 x i32>, <8 x i32>, i8)
+
+define <8 x i32>@test_int_x86_avx512_mask_psll_d_256(<8 x i32> %x0, <4 x i32> %x1, <8 x i32> %x2, i8 %x3) {
+; CHECK-LABEL: test_int_x86_avx512_mask_psll_d_256:
+; CHECK:       ## BB#0:
+; CHECK-NEXT:    movzbl %dil, %eax
+; CHECK-NEXT:    kmovw %eax, %k1
+; CHECK-NEXT:    vpslld %xmm1, %ymm0, %ymm2 {%k1}
+; CHECK-NEXT:    vpslld %xmm1, %ymm0, %ymm3 {%k1} {z}
+; CHECK-NEXT:    vpslld %xmm1, %ymm0, %ymm0
+; CHECK-NEXT:    vpaddd %ymm3, %ymm2, %ymm1
+; CHECK-NEXT:    vpaddd %ymm0, %ymm1, %ymm0
+; CHECK-NEXT:    retq
+  %res = call <8 x i32> @llvm.x86.avx512.mask.psll.d.256(<8 x i32> %x0, <4 x i32> %x1, <8 x i32> %x2, i8 %x3)
+  %res1 = call <8 x i32> @llvm.x86.avx512.mask.psll.d.256(<8 x i32> %x0, <4 x i32> %x1, <8 x i32> zeroinitializer, i8 %x3)
+  %res2 = call <8 x i32> @llvm.x86.avx512.mask.psll.d.256(<8 x i32> %x0, <4 x i32> %x1, <8 x i32> %x2, i8 -1)
+  %res3 = add <8 x i32> %res, %res1
+  %res4 = add <8 x i32> %res3, %res2
+  ret <8 x i32> %res4
+}
+
+declare <4 x i32> @llvm.x86.avx512.mask.psll.di.128(<4 x i32>, i8, <4 x i32>, i8)
+
+define <4 x i32>@test_int_x86_avx512_mask_psll_di_128(<4 x i32> %x0, i8 %x1, <4 x i32> %x2, i8 %x3) {
+; CHECK-LABEL: test_int_x86_avx512_mask_psll_di_128:
+; CHECK:       ## BB#0:
+; CHECK-NEXT:    movzbl %sil, %eax
+; CHECK-NEXT:    kmovw %eax, %k1
+; CHECK-NEXT:    vpslld $3, %xmm0, %xmm1 {%k1}
+; CHECK-NEXT:    vpslld $3, %xmm0, %xmm2 {%k1} {z}
+; CHECK-NEXT:    vpslld $3, %xmm0, %xmm0
+; CHECK-NEXT:    vpaddd %xmm2, %xmm1, %xmm1
+; CHECK-NEXT:    vpaddd %xmm0, %xmm1, %xmm0
+; CHECK-NEXT:    retq
+  %res = call <4 x i32> @llvm.x86.avx512.mask.psll.di.128(<4 x i32> %x0, i8 3, <4 x i32> %x2, i8 %x3)
+  %res1 = call <4 x i32> @llvm.x86.avx512.mask.psll.di.128(<4 x i32> %x0, i8 3, <4 x i32> zeroinitializer, i8 %x3)
+  %res2 = call <4 x i32> @llvm.x86.avx512.mask.psll.di.128(<4 x i32> %x0, i8 3, <4 x i32> %x2, i8 -1)
+  %res3 = add <4 x i32> %res, %res1
+  %res4 = add <4 x i32> %res3, %res2
+  ret <4 x i32> %res4
+}
+
+declare <8 x i32> @llvm.x86.avx512.mask.psll.di.256(<8 x i32>, i8, <8 x i32>, i8)
+
+define <8 x i32>@test_int_x86_avx512_mask_psll_di_256(<8 x i32> %x0, i8 %x1, <8 x i32> %x2, i8 %x3) {
+; CHECK-LABEL: test_int_x86_avx512_mask_psll_di_256:
+; CHECK:       ## BB#0:
+; CHECK-NEXT:    movzbl %sil, %eax
+; CHECK-NEXT:    kmovw %eax, %k1
+; CHECK-NEXT:    vpslld $3, %ymm0, %ymm1 {%k1}
+; CHECK-NEXT:    vpslld $3, %ymm0, %ymm2 {%k1} {z}
+; CHECK-NEXT:    vpslld $3, %ymm0, %ymm0
+; CHECK-NEXT:    vpaddd %ymm2, %ymm1, %ymm1
+; CHECK-NEXT:    vpaddd %ymm0, %ymm1, %ymm0
+; CHECK-NEXT:    retq
+  %res = call <8 x i32> @llvm.x86.avx512.mask.psll.di.256(<8 x i32> %x0, i8 3, <8 x i32> %x2, i8 %x3)
+  %res1 = call <8 x i32> @llvm.x86.avx512.mask.psll.di.256(<8 x i32> %x0, i8 3, <8 x i32> zeroinitializer, i8 %x3)
+  %res2 = call <8 x i32> @llvm.x86.avx512.mask.psll.di.256(<8 x i32> %x0, i8 3, <8 x i32> %x2, i8 -1)
+  %res3 = add <8 x i32> %res, %res1
+  %res4 = add <8 x i32> %res3, %res2
+  ret <8 x i32> %res4
+}
+
+declare <4 x i64> @llvm.x86.avx512.mask.psll.q.256(<4 x i64>, <2 x i64>, <4 x i64>, i8)
+
+define <4 x i64>@test_int_x86_avx512_mask_psll_q_256(<4 x i64> %x0, <2 x i64> %x1, <4 x i64> %x2, i8 %x3) {
+; CHECK-LABEL: test_int_x86_avx512_mask_psll_q_256:
+; CHECK:       ## BB#0:
+; CHECK-NEXT:    movzbl %dil, %eax
+; CHECK-NEXT:    kmovw %eax, %k1
+; CHECK-NEXT:    vpsllq %xmm1, %ymm0, %ymm2 {%k1}
+; CHECK-NEXT:    vpsllq %xmm1, %ymm0, %ymm3 {%k1} {z}
+; CHECK-NEXT:    vpsllq %xmm1, %ymm0, %ymm0
+; CHECK-NEXT:    vpaddq %ymm3, %ymm2, %ymm1
+; CHECK-NEXT:    vpaddq %ymm0, %ymm1, %ymm0
+; CHECK-NEXT:    retq
+  %res = call <4 x i64> @llvm.x86.avx512.mask.psll.q.256(<4 x i64> %x0, <2 x i64> %x1, <4 x i64> %x2, i8 %x3)
+  %res1 = call <4 x i64> @llvm.x86.avx512.mask.psll.q.256(<4 x i64> %x0, <2 x i64> %x1, <4 x i64> zeroinitializer, i8 %x3)
+  %res2 = call <4 x i64> @llvm.x86.avx512.mask.psll.q.256(<4 x i64> %x0, <2 x i64> %x1, <4 x i64> %x2, i8 -1)
+  %res3 = add <4 x i64> %res, %res1
+  %res4 = add <4 x i64> %res3, %res2
+  ret <4 x i64> %res4
+}
+
+declare <2 x i64> @llvm.x86.avx512.mask.psll.qi.128(<2 x i64>, i8, <2 x i64>, i8)
+
+define <2 x i64>@test_int_x86_avx512_mask_psll_qi_128(<2 x i64> %x0, i8 %x1, <2 x i64> %x2, i8 %x3) {
+; CHECK-LABEL: test_int_x86_avx512_mask_psll_qi_128:
+; CHECK:       ## BB#0:
+; CHECK-NEXT:    movzbl %sil, %eax
+; CHECK-NEXT:    kmovw %eax, %k1
+; CHECK-NEXT:    vpsllq $3, %xmm0, %xmm1 {%k1}
+; CHECK-NEXT:    vpsllq $3, %xmm0, %xmm2 {%k1} {z}
+; CHECK-NEXT:    vpsllq $3, %xmm0, %xmm0
+; CHECK-NEXT:    vpaddq %xmm2, %xmm1, %xmm1
+; CHECK-NEXT:    vpaddq %xmm0, %xmm1, %xmm0
+; CHECK-NEXT:    retq
+  %res = call <2 x i64> @llvm.x86.avx512.mask.psll.qi.128(<2 x i64> %x0, i8 3, <2 x i64> %x2, i8 %x3)
+  %res1 = call <2 x i64> @llvm.x86.avx512.mask.psll.qi.128(<2 x i64> %x0, i8 3, <2 x i64> zeroinitializer, i8 %x3)
+  %res2 = call <2 x i64> @llvm.x86.avx512.mask.psll.qi.128(<2 x i64> %x0, i8 3, <2 x i64> %x2, i8 -1)
+  %res3 = add <2 x i64> %res, %res1
+  %res4 = add <2 x i64> %res3, %res2
+  ret <2 x i64> %res4
+}
+
+declare <4 x i64> @llvm.x86.avx512.mask.psll.qi.256(<4 x i64>, i8, <4 x i64>, i8)
+
+define <4 x i64>@test_int_x86_avx512_mask_psll_qi_256(<4 x i64> %x0, i8 %x1, <4 x i64> %x2, i8 %x3) {
+; CHECK-LABEL: test_int_x86_avx512_mask_psll_qi_256:
+; CHECK:       ## BB#0:
+; CHECK-NEXT:    movzbl %sil, %eax
+; CHECK-NEXT:    kmovw %eax, %k1
+; CHECK-NEXT:    vpsllq $3, %ymm0, %ymm1 {%k1}
+; CHECK-NEXT:    vpsllq $3, %ymm0, %ymm2 {%k1} {z}
+; CHECK-NEXT:    vpsllq $3, %ymm0, %ymm0
+; CHECK-NEXT:    vpaddq %ymm2, %ymm1, %ymm1
+; CHECK-NEXT:    vpaddq %ymm0, %ymm1, %ymm0
+; CHECK-NEXT:    retq
+  %res = call <4 x i64> @llvm.x86.avx512.mask.psll.qi.256(<4 x i64> %x0, i8 3, <4 x i64> %x2, i8 %x3)
+  %res1 = call <4 x i64> @llvm.x86.avx512.mask.psll.qi.256(<4 x i64> %x0, i8 3, <4 x i64> zeroinitializer, i8 %x3)
+  %res2 = call <4 x i64> @llvm.x86.avx512.mask.psll.qi.256(<4 x i64> %x0, i8 3, <4 x i64> %x2, i8 -1)
+  %res3 = add <4 x i64> %res, %res1
+  %res4 = add <4 x i64> %res3, %res2
+ ret <4 x i64> %res4
+}