6 files changed, 132 insertions, 29 deletions
diff --git a/llvm/test/CodeGen/X86/avx512-cvt.ll b/llvm/test/CodeGen/X86/avx512-cvt.ll
index aa34076c910..586a2954501 100644
--- a/llvm/test/CodeGen/X86/avx512-cvt.ll
+++ b/llvm/test/CodeGen/X86/avx512-cvt.ll
@@ -52,14 +52,14 @@ define <4 x i64> @f32tosl(<4 x float> %a) {
 }
 
 ; CHECK-LABEL: sltof432
-; CHECK: vcvtqq2ps 
+; CHECK: vcvtqq2ps
 define <4 x float> @sltof432(<4 x i64> %a) {
   %b = sitofp <4 x i64> %a to <4 x float>
   ret <4 x float> %b
 }
 
 ; CHECK-LABEL: ultof432
-; CHECK: vcvtuqq2ps 
+; CHECK: vcvtuqq2ps
 define <4 x float> @ultof432(<4 x i64> %a) {
   %b = uitofp <4 x i64> %a to <4 x float>
   ret <4 x float> %b
@@ -279,12 +279,14 @@ define i32 @float_to_int(float %x) {
    ret i32 %res
 }
 
-; CHECK-LABEL: uitof64
-; CHECK: vcvtudq2pd
-; CHECK: vextracti64x4
-; CHECK: vcvtudq2pd
-; CHECK: ret
 define <16 x double> @uitof64(<16 x i32> %a) nounwind {
+; CHECK-LABEL: uitof64:
+; CHECK:       ## BB#0:
+; CHECK-NEXT:    vcvtudq2pd %ymm0, %zmm2
+; CHECK-NEXT:    vextracti32x8 $1, %zmm0, %ymm0
+; CHECK-NEXT:    vcvtudq2pd %ymm0, %zmm1
+; CHECK-NEXT:    vmovaps %zmm2, %zmm0
+; CHECK-NEXT:    retq 
   %b = uitofp <16 x i32> %a to <16 x double>
   ret <16 x double> %b
 }
@@ -407,7 +409,7 @@ define <8 x double> @sitofp_8i1_double(<8 x double> %a) {
 }
 
 ; CHECK-LABEL: @uitofp_16i8
-; CHECK:  vpmovzxbd  
+; CHECK:  vpmovzxbd
 ; CHECK: vcvtudq2ps
 define <16 x float> @uitofp_16i8(<16 x i8>%a) {
   %b = uitofp <16 x i8> %a to <16 x float>
diff --git a/llvm/test/CodeGen/X86/avx512-insert-extract.ll b/llvm/test/CodeGen/X86/avx512-insert-extract.ll
index 6f985f0bf3a..52398acd48f 100644
--- a/llvm/test/CodeGen/X86/avx512-insert-extract.ll
+++ b/llvm/test/CodeGen/X86/avx512-insert-extract.ll
@@ -12,14 +12,24 @@ define <16 x float> @test1(<16 x float> %x, float* %br, float %y) nounwind {
   ret <16 x float> %rrr3
 }
 
-;CHECK-LABEL: test2:
-;KNL: vinsertf32x4 $0
-;SKX: vinsertf64x2 $0
-;CHECK: vextractf32x4 $3
-;KNL: vinsertf32x4 $3
-;SKX: vinsertf64x2 $3
-;CHECK: ret
 define <8 x double> @test2(<8 x double> %x, double* %br, double %y) nounwind {
+; KNL-LABEL: test2:
+; KNL:       ## BB#0:
+; KNL-NEXT:    vmovhpd (%rdi), %xmm0, %xmm2
+; KNL-NEXT:    vinsertf32x4 $0, %xmm2, %zmm0, %zmm0
+; KNL-NEXT:    vextractf32x4 $3, %zmm0, %xmm2
+; KNL-NEXT:    vmovsd %xmm1, %xmm2, %xmm1
+; KNL-NEXT:    vinsertf32x4 $3, %xmm1, %zmm0, %zmm0
+; KNL-NEXT:    retq
+;
+; SKX-LABEL: test2:
+; SKX:       ## BB#0:
+; SKX-NEXT:    vmovhpd (%rdi), %xmm0, %xmm2
+; SKX-NEXT:    vinsertf64x2 $0, %xmm2, %zmm0, %zmm0
+; SKX-NEXT:    vextractf64x2 $3, %zmm0, %xmm2
+; SKX-NEXT:    vmovsd %xmm1, %xmm2, %xmm1
+; SKX-NEXT:    vinsertf64x2 $3, %xmm1, %zmm0, %zmm0
+; SKX-NEXT:    retq
   %rrr = load double, double* %br
   %rrr2 = insertelement <8 x double> %x, double %rrr, i32 1
   %rrr3 = insertelement <8 x double> %rrr2, double %y, i32 6
@@ -36,12 +46,22 @@ define <16 x float> @test3(<16 x float> %x) nounwind {
   ret <16 x float> %rrr2
 }
 
-;CHECK-LABEL: test4:
-;CHECK: vextracti32x4 $2
-;KNL: vinserti32x4 $0
-;SKX: vinserti64x2 $0
-;CHECK: ret
 define <8 x i64> @test4(<8 x i64> %x) nounwind {
+; KNL-LABEL: test4:
+; KNL:       ## BB#0:
+; KNL-NEXT:    vextracti32x4 $2, %zmm0, %xmm1
+; KNL-NEXT:    vmovq %xmm1, %rax
+; KNL-NEXT:    vpinsrq $1, %rax, %xmm0, %xmm1
+; KNL-NEXT:    vinserti32x4 $0, %xmm1, %zmm0, %zmm0
+; KNL-NEXT:    retq
+;
+; SKX-LABEL: test4:
+; SKX:       ## BB#0:
+; SKX-NEXT:    vextracti64x2 $2, %zmm0, %xmm1
+; SKX-NEXT:    vmovq %xmm1, %rax
+; SKX-NEXT:    vpinsrq $1, %rax, %xmm0, %xmm1
+; SKX-NEXT:    vinserti64x2 $0, %xmm1, %zmm0, %zmm0
+; SKX-NEXT:    retq
   %eee = extractelement <8 x i64> %x, i32 4
   %rrr2 = insertelement <8 x i64> %x, i64 %eee, i32 1
   ret <8 x i64> %rrr2
@@ -142,7 +162,7 @@ define i64 @test12(<16 x i64>%a, <16 x i64>%b, i64 %a1, i64 %b1) {
 ;CHECK: andl    $1, %eax
 ;CHECK: kmovw   %eax, %k0
 ;CHECK: movw    $-4
-;CHECK: korw    
+;CHECK: korw
 define i16 @test13(i32 %a, i32 %b) {
   %cmp_res = icmp ult i32 %a, %b
   %maskv = insertelement <16 x i1> <i1 true, i1 false, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true>, i1 %cmp_res, i32 0
diff --git a/llvm/test/CodeGen/X86/avx512-intrinsics.ll b/llvm/test/CodeGen/X86/avx512-intrinsics.ll
index 6d18b25b770..4846f760cb7 100644
--- a/llvm/test/CodeGen/X86/avx512-intrinsics.ll
+++ b/llvm/test/CodeGen/X86/avx512-intrinsics.ll
@@ -911,38 +911,38 @@ declare i8 @llvm.x86.avx512.mask.ucmp.q.512(<8 x i64>, <8 x i64>, i32, i8) nounw
 define <4 x float> @test_mask_vextractf32x4(<4 x float> %b, <16 x float> %a, i8 %mask) {
 ; CHECK-LABEL: test_mask_vextractf32x4:
 ; CHECK: vextractf32x4 $2, %zmm1, %xmm0 {%k1}
-  %res = call <4 x float> @llvm.x86.avx512.mask.vextractf32x4.512(<16 x float> %a, i8 2, <4 x float> %b, i8 %mask)
+  %res = call <4 x float> @llvm.x86.avx512.mask.vextractf32x4.512(<16 x float> %a, i32 2, <4 x float> %b, i8 %mask)
   ret <4 x float> %res
 }
 
-declare <4 x float> @llvm.x86.avx512.mask.vextractf32x4.512(<16 x float>, i8, <4 x float>, i8)
+declare <4 x float> @llvm.x86.avx512.mask.vextractf32x4.512(<16 x float>, i32, <4 x float>, i8)
 
 define <4 x i64> @test_mask_vextracti64x4(<4 x i64> %b, <8 x i64> %a, i8 %mask) {
 ; CHECK-LABEL: test_mask_vextracti64x4:
 ; CHECK: vextracti64x4 $2, %zmm1, %ymm0 {%k1}
-  %res = call <4 x i64> @llvm.x86.avx512.mask.vextracti64x4.512(<8 x i64> %a, i8 2, <4 x i64> %b, i8 %mask)
+  %res = call <4 x i64> @llvm.x86.avx512.mask.vextracti64x4.512(<8 x i64> %a, i32 2, <4 x i64> %b, i8 %mask)
   ret <4 x i64> %res
 }
 
-declare <4 x i64> @llvm.x86.avx512.mask.vextracti64x4.512(<8 x i64>, i8, <4 x i64>, i8)
+declare <4 x i64> @llvm.x86.avx512.mask.vextracti64x4.512(<8 x i64>, i32, <4 x i64>, i8)
 
 define <4 x i32> @test_maskz_vextracti32x4(<16 x i32> %a, i8 %mask) {
 ; CHECK-LABEL: test_maskz_vextracti32x4:
 ; CHECK: vextracti32x4 $2, %zmm0, %xmm0 {%k1} {z}
-  %res = call <4 x i32> @llvm.x86.avx512.mask.vextracti32x4.512(<16 x i32> %a, i8 2, <4 x i32> zeroinitializer, i8 %mask)
+  %res = call <4 x i32> @llvm.x86.avx512.mask.vextracti32x4.512(<16 x i32> %a, i32 2, <4 x i32> zeroinitializer, i8 %mask)
   ret <4 x i32> %res
 }
 
-declare <4 x i32> @llvm.x86.avx512.mask.vextracti32x4.512(<16 x i32>, i8, <4 x i32>, i8)
+declare <4 x i32> @llvm.x86.avx512.mask.vextracti32x4.512(<16 x i32>, i32, <4 x i32>, i8)
 
 define <4 x double> @test_vextractf64x4(<8 x double> %a) {
 ; CHECK-LABEL: test_vextractf64x4:
 ; CHECK: vextractf64x4 $2, %zmm0, %ymm0 ##
-  %res = call <4 x double> @llvm.x86.avx512.mask.vextractf64x4.512(<8 x double> %a, i8 2, <4 x double> zeroinitializer, i8 -1)
+  %res = call <4 x double> @llvm.x86.avx512.mask.vextractf64x4.512(<8 x double> %a, i32 2, <4 x double> zeroinitializer, i8 -1)
   ret <4 x double> %res
 }
 
-declare <4 x double> @llvm.x86.avx512.mask.vextractf64x4.512(<8 x double>, i8, <4 x double>, i8)
+declare <4 x double> @llvm.x86.avx512.mask.vextractf64x4.512(<8 x double>, i32, <4 x double>, i8)
 
 define <16 x i32> @test_x86_avx512_pslli_d(<16 x i32> %a0) {
   ; CHECK-LABEL: test_x86_avx512_pslli_d
diff --git a/llvm/test/CodeGen/X86/avx512dq-intrinsics.ll b/llvm/test/CodeGen/X86/avx512dq-intrinsics.ll
index 67a88e15570..c1335096ab2 100644
--- a/llvm/test/CodeGen/X86/avx512dq-intrinsics.ll
+++ b/llvm/test/CodeGen/X86/avx512dq-intrinsics.ll
@@ -315,3 +315,44 @@ define <2 x double>@test_int_x86_avx512_mask_range_sd(<2 x double> %x0, <2 x dou
   %res2 = fadd <2 x double> %res, %res1
   ret <2 x double> %res2
 }
+
+
+declare <2 x double> @llvm.x86.avx512.mask.vextractf64x2.512(<8 x double>, i32, <2 x double>, i8)
+
+define <2 x double>@test_int_x86_avx512_mask_vextractf64x2_512(<8 x double> %x0, <2 x double> %x2, i8 %x3) {
+; CHECK-LABEL: test_int_x86_avx512_mask_vextractf64x2_512:
+; CHECK:       ## BB#0:
+; CHECK-NEXT:    kmovw %edi, %k1
+; CHECK-NEXT:    vextractf64x2 $1, %zmm0, %xmm1 {%k1}
+; CHECK-NEXT:    vextractf64x2 $1, %zmm0, %xmm2 {%k1} {z}
+; CHECK-NEXT:    vextractf64x2 $1, %zmm0, %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.vextractf64x2.512(<8 x double> %x0,i32 1, <2 x double> %x2, i8 %x3)
+  %res2 = call <2 x double> @llvm.x86.avx512.mask.vextractf64x2.512(<8 x double> %x0,i32 1, <2 x double> zeroinitializer, i8 %x3)
+  %res1 = call <2 x double> @llvm.x86.avx512.mask.vextractf64x2.512(<8 x double> %x0,i32 1, <2 x double> zeroinitializer, i8 -1)
+  %res3 = fadd <2 x double> %res, %res1
+  %res4 = fadd <2 x double> %res2, %res3
+  ret <2 x double> %res4
+}
+
+declare <8 x float> @llvm.x86.avx512.mask.vextractf32x8.512(<16 x float>, i32, <8 x float>, i8)
+
+define <8 x float>@test_int_x86_avx512_mask_vextractf32x8(<16 x float> %x0, <8 x float> %x2, i8 %x3) {
+; CHECK-LABEL: test_int_x86_avx512_mask_vextractf32x8:
+; CHECK:       ## BB#0:
+; CHECK-NEXT:    kmovw %edi, %k1
+; CHECK-NEXT:    vextractf32x8 $1, %zmm0, %ymm1 {%k1}
+; CHECK-NEXT:    vextractf32x8 $1, %zmm0, %ymm2 {%k1} {z}
+; CHECK-NEXT:    vextractf32x8 $1, %zmm0, %ymm0
+; CHECK-NEXT:    vaddps %ymm0, %ymm1, %ymm0
+; CHECK-NEXT:    vaddps %ymm0, %ymm2, %ymm0
+; CHECK-NEXT:    retq
+  %res  = call <8 x float> @llvm.x86.avx512.mask.vextractf32x8.512(<16 x float> %x0,i32 1, <8 x float> %x2, i8 %x3)
+  %res2 = call <8 x float> @llvm.x86.avx512.mask.vextractf32x8.512(<16 x float> %x0,i32 1, <8 x float> zeroinitializer, i8 %x3)
+  %res1 = call <8 x float> @llvm.x86.avx512.mask.vextractf32x8.512(<16 x float> %x0,i32 1, <8 x float> zeroinitializer, i8 -1)
+  %res3 = fadd <8 x float> %res, %res1
+  %res4 = fadd <8 x float> %res2, %res3
+  ret <8 x float> %res4
+}
diff --git a/llvm/test/CodeGen/X86/avx512dqvl-intrinsics.ll b/llvm/test/CodeGen/X86/avx512dqvl-intrinsics.ll
index 2fcfac0f1bb..8956f279acb 100644
--- a/llvm/test/CodeGen/X86/avx512dqvl-intrinsics.ll
+++ b/llvm/test/CodeGen/X86/avx512dqvl-intrinsics.ll
@@ -1648,3 +1648,23 @@ define <8 x float>@test_int_x86_avx512_mask_range_ps_256(<8 x float> %x0, <8 x f
   %res2 = fadd <8 x float> %res, %res1
   ret <8 x float> %res2
 }
+
+declare <2 x double> @llvm.x86.avx512.mask.vextractf64x2.256(<4 x double>, i32, <2 x double>, i8)
+
+define <2 x double>@test_int_x86_avx512_mask_vextractf64x2_256(<4 x double> %x0, <2 x double> %x2, i8 %x3) {
+; CHECK-LABEL: test_int_x86_avx512_mask_vextractf64x2_256:
+; CHECK:       ## BB#0:
+; CHECK-NEXT:    kmovw %edi, %k1
+; CHECK-NEXT:    vextractf64x2 $1, %ymm0, %xmm1 {%k1}
+; CHECK-NEXT:    vextractf64x2 $1, %ymm0, %xmm2 {%k1} {z}
+; CHECK-NEXT:    vextractf64x2 $1, %ymm0, %xmm0
+; CHECK-NEXT:    vaddpd %xmm0, %xmm1, %xmm0
+; CHECK-NEXT:    vaddpd %xmm2, %xmm0, %xmm0
+; CHECK-NEXT:    retq
+  %res = call <2 x double> @llvm.x86.avx512.mask.vextractf64x2.256(<4 x double> %x0,i32 1, <2 x double> %x2, i8 %x3)
+  %res2 = call <2 x double> @llvm.x86.avx512.mask.vextractf64x2.256(<4 x double> %x0,i32 1, <2 x double> zeroinitializer, i8 %x3)
+  %res1 = call <2 x double> @llvm.x86.avx512.mask.vextractf64x2.256(<4 x double> %x0,i32 1, <2 x double> zeroinitializer, i8 -1)
+  %res3 = fadd <2 x double> %res, %res1
+  %res4 = fadd <2 x double> %res3, %res2
+  ret <2 x double> %res4
+}
diff --git a/llvm/test/CodeGen/X86/avx512vl-intrinsics.ll b/llvm/test/CodeGen/X86/avx512vl-intrinsics.ll
index b148db68092..1f0bacb0186 100644
--- a/llvm/test/CodeGen/X86/avx512vl-intrinsics.ll
+++ b/llvm/test/CodeGen/X86/avx512vl-intrinsics.ll
@@ -4508,6 +4508,26 @@ define <8 x float>@test_int_x86_avx512_mask_rndscale_ps_256(<8 x float> %x0, <8
   ret <8 x float> %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) {