[X86][AVX512] Converted the VBROADCAST intrinsics to generic IR

llvm-svn: 274544

4 files changed, 122 insertions, 93 deletions

diff --git a/clang/test/CodeGen/avx512bw-builtins.c b/clang/test/CodeGen/avx512bw-builtins.c
index 75efcbaa754..1cd0a0ccb14 100644
--- a/clang/test/CodeGen/avx512bw-builtins.c
+++ b/clang/test/CodeGen/avx512bw-builtins.c
@@ -1473,41 +1473,44 @@ __m512i test_mm512_movm_epi16(__mmask32 __A) {
   return _mm512_movm_epi16(__A); 
 }
 
-
 __m512i test_mm512_broadcastb_epi8(__m128i __A) {
   // CHECK-LABEL: @test_mm512_broadcastb_epi8
-  // CHECK: @llvm.x86.avx512.pbroadcastb.512
-  return _mm512_broadcastb_epi8(__A); 
+  // CHECK: shufflevector <16 x i8> %{{.*}}, <16 x i8> undef, <64 x i32> zeroinitializer
+  return _mm512_broadcastb_epi8(__A);
 }
 
 __m512i test_mm512_mask_broadcastb_epi8(__m512i __O, __mmask64 __M, __m128i __A) {
   // CHECK-LABEL: @test_mm512_mask_broadcastb_epi8
-  // CHECK: @llvm.x86.avx512.pbroadcastb.512
-  return _mm512_mask_broadcastb_epi8(__O, __M, __A); 
+  // CHECK: shufflevector <16 x i8> %{{.*}}, <16 x i8> undef, <64 x i32> zeroinitializer
+  // CHECK: select <64 x i1> %{{.*}}, <64 x i8> %{{.*}}, <64 x i8> %{{.*}}
+  return _mm512_mask_broadcastb_epi8(__O, __M, __A);
 }
 
 __m512i test_mm512_maskz_broadcastb_epi8(__mmask64 __M, __m128i __A) {
   // CHECK-LABEL: @test_mm512_maskz_broadcastb_epi8
-  // CHECK: @llvm.x86.avx512.pbroadcastb.512
-  return _mm512_maskz_broadcastb_epi8(__M, __A); 
+  // CHECK: shufflevector <16 x i8> %{{.*}}, <16 x i8> undef, <64 x i32> zeroinitializer
+  // CHECK: select <64 x i1> %{{.*}}, <64 x i8> %{{.*}}, <64 x i8> %{{.*}}
+  return _mm512_maskz_broadcastb_epi8(__M, __A);
 }
 
 __m512i test_mm512_broadcastw_epi16(__m128i __A) {
   // CHECK-LABEL: @test_mm512_broadcastw_epi16
-  // CHECK: @llvm.x86.avx512.pbroadcastw.512
-  return _mm512_broadcastw_epi16(__A); 
+  // CHECK: shufflevector <8 x i16> %{{.*}}, <8 x i16> undef, <32 x i32> zeroinitializer
+  return _mm512_broadcastw_epi16(__A);
 }
 
 __m512i test_mm512_mask_broadcastw_epi16(__m512i __O, __mmask32 __M, __m128i __A) {
   // CHECK-LABEL: @test_mm512_mask_broadcastw_epi16
-  // CHECK: @llvm.x86.avx512.pbroadcastw.512
-  return _mm512_mask_broadcastw_epi16(__O, __M, __A); 
+  // CHECK: shufflevector <8 x i16> %{{.*}}, <8 x i16> undef, <32 x i32> zeroinitializer
+  // CHECK: select <32 x i1> %{{.*}}, <32 x i16> %{{.*}}, <32 x i16> %{{.*}}
+  return _mm512_mask_broadcastw_epi16(__O, __M, __A);
 }
 
 __m512i test_mm512_maskz_broadcastw_epi16(__mmask32 __M, __m128i __A) {
   // CHECK-LABEL: @test_mm512_maskz_broadcastw_epi16
-  // CHECK: @llvm.x86.avx512.pbroadcastw.512
-  return _mm512_maskz_broadcastw_epi16(__M, __A); 
+  // CHECK: shufflevector <8 x i16> %{{.*}}, <8 x i16> undef, <32 x i32> zeroinitializer
+  // CHECK: select <32 x i1> %{{.*}}, <32 x i16> %{{.*}}, <32 x i16> %{{.*}}
+  return _mm512_maskz_broadcastw_epi16(__M, __A);
 }
 
 __m512i test_mm512_mask_set1_epi16(__m512i __O, __mmask32 __M, short __A) {
diff --git a/clang/test/CodeGen/avx512f-builtins.c b/clang/test/CodeGen/avx512f-builtins.c
index 018ad93f8df..b8121b2f75a 100644
--- a/clang/test/CodeGen/avx512f-builtins.c
+++ b/clang/test/CodeGen/avx512f-builtins.c
@@ -434,20 +434,6 @@ __m512i test_mm512_maskz_alignr_epi64( __mmask8 u, __m512i a, __m512i b)
   return _mm512_maskz_alignr_epi64(u, a, b, 2);
 }
 
-__m512d test_mm512_broadcastsd_pd(__m128d a)
-{
-  // CHECK-LABEL: @test_mm512_broadcastsd_pd
-  // CHECK: insertelement <8 x double> {{.*}}, i32 0
-  // CHECK: insertelement <8 x double> {{.*}}, i32 1
-  // CHECK: insertelement <8 x double> {{.*}}, i32 2
-  // CHECK: insertelement <8 x double> {{.*}}, i32 3
-  // CHECK: insertelement <8 x double> {{.*}}, i32 4
-  // CHECK: insertelement <8 x double> {{.*}}, i32 5
-  // CHECK: insertelement <8 x double> {{.*}}, i32 6
-  // CHECK: insertelement <8 x double> {{.*}}, i32 7
-  return _mm512_broadcastsd_pd(a);
-}
-
 __m512d test_mm512_fmadd_round_pd(__m512d __A, __m512d __B, __m512d __C) {
   // CHECK-LABEL: @test_mm512_fmadd_round_pd
   // CHECK: @llvm.x86.avx512.mask.vfmadd.pd.512
@@ -4388,64 +4374,84 @@ __m512i test_mm512_maskz_broadcast_i64x4(__mmask8 __M, __m256i __A) {
   return _mm512_maskz_broadcast_i64x4(__M, __A); 
 }
 
+__m512d test_mm512_broadcastsd_pd(__m128d __A) {
+  // CHECK-LABEL: @test_mm512_broadcastsd_pd
+  // CHECK: shufflevector <2 x double> %{{.*}}, <2 x double> undef, <8 x i32> zeroinitializer
+  return _mm512_broadcastsd_pd(__A);
+}
+
 __m512d test_mm512_mask_broadcastsd_pd(__m512d __O, __mmask8 __M, __m128d __A) {
   // CHECK-LABEL: @test_mm512_mask_broadcastsd_pd
-  // CHECK: @llvm.x86.avx512.mask.broadcast.sd.pd.512
-  return _mm512_mask_broadcastsd_pd(__O, __M, __A); 
+  // CHECK: shufflevector <2 x double> %{{.*}}, <2 x double> undef, <8 x i32> zeroinitializer
+  // CHECK: select <8 x i1> %{{.*}}, <8 x double> %{{.*}}, <8 x double> %{{.*}}
+  return _mm512_mask_broadcastsd_pd(__O, __M, __A);
 }
 
 __m512d test_mm512_maskz_broadcastsd_pd(__mmask8 __M, __m128d __A) {
   // CHECK-LABEL: @test_mm512_maskz_broadcastsd_pd
-  // CHECK: @llvm.x86.avx512.mask.broadcast.sd.pd.512
-  return _mm512_maskz_broadcastsd_pd(__M, __A); 
+  // CHECK: shufflevector <2 x double> %{{.*}}, <2 x double> undef, <8 x i32> zeroinitializer
+  // CHECK: select <8 x i1> %{{.*}}, <8 x double> %{{.*}}, <8 x double> %{{.*}}
+  return _mm512_maskz_broadcastsd_pd(__M, __A);
+}
+
+__m512 test_mm512_broadcastss_ps(__m128 __A) {
+  // CHECK-LABEL: @test_mm512_broadcastss_ps
+  // CHECK: shufflevector <4 x float> %{{.*}}, <4 x float> undef, <16 x i32> zeroinitializer
+  return _mm512_broadcastss_ps(__A);
 }
 
 __m512 test_mm512_mask_broadcastss_ps(__m512 __O, __mmask16 __M, __m128 __A) {
   // CHECK-LABEL: @test_mm512_mask_broadcastss_ps
-  // CHECK: @llvm.x86.avx512.mask.broadcast.ss.ps.512
-  return _mm512_mask_broadcastss_ps(__O, __M, __A); 
+  // CHECK: shufflevector <4 x float> %{{.*}}, <4 x float> undef, <16 x i32> zeroinitializer
+  // CHECK: select <16 x i1> %{{.*}}, <16 x float> %{{.*}}, <16 x float> %{{.*}}
+  return _mm512_mask_broadcastss_ps(__O, __M, __A);
 }
 
 __m512 test_mm512_maskz_broadcastss_ps(__mmask16 __M, __m128 __A) {
   // CHECK-LABEL: @test_mm512_maskz_broadcastss_ps
-  // CHECK: @llvm.x86.avx512.mask.broadcast.ss.ps.512
-  return _mm512_maskz_broadcastss_ps(__M, __A); 
+  // CHECK: shufflevector <4 x float> %{{.*}}, <4 x float> undef, <16 x i32> zeroinitializer
+  // CHECK: select <16 x i1> %{{.*}}, <16 x float> %{{.*}}, <16 x float> %{{.*}}
+  return _mm512_maskz_broadcastss_ps(__M, __A);
 }
 
 __m512i test_mm512_broadcastd_epi32(__m128i __A) {
   // CHECK-LABEL: @test_mm512_broadcastd_epi32
-  // CHECK: @llvm.x86.avx512.pbroadcastd.512
-  return _mm512_broadcastd_epi32(__A); 
+  // CHECK: shufflevector <4 x i32> %{{.*}}, <4 x i32> undef, <16 x i32> zeroinitializer
+  return _mm512_broadcastd_epi32(__A);
 }
 
 __m512i test_mm512_mask_broadcastd_epi32(__m512i __O, __mmask16 __M, __m128i __A) {
   // CHECK-LABEL: @test_mm512_mask_broadcastd_epi32
-  // CHECK: @llvm.x86.avx512.pbroadcastd.512
-  return _mm512_mask_broadcastd_epi32(__O, __M, __A); 
+  // CHECK: shufflevector <4 x i32> %{{.*}}, <4 x i32> undef, <16 x i32> zeroinitializer
+  // CHECK: select <16 x i1> %{{.*}}, <16 x i32> %{{.*}}, <16 x i32> %{{.*}}
+  return _mm512_mask_broadcastd_epi32(__O, __M, __A);
 }
 
 __m512i test_mm512_maskz_broadcastd_epi32(__mmask16 __M, __m128i __A) {
   // CHECK-LABEL: @test_mm512_maskz_broadcastd_epi32
-  // CHECK: @llvm.x86.avx512.pbroadcastd.512
-  return _mm512_maskz_broadcastd_epi32(__M, __A); 
+  // CHECK: shufflevector <4 x i32> %{{.*}}, <4 x i32> undef, <16 x i32> zeroinitializer
+  // CHECK: select <16 x i1> %{{.*}}, <16 x i32> %{{.*}}, <16 x i32> %{{.*}}
+  return _mm512_maskz_broadcastd_epi32(__M, __A);
 }
 
 __m512i test_mm512_broadcastq_epi64(__m128i __A) {
   // CHECK-LABEL: @test_mm512_broadcastq_epi64
-  // CHECK: @llvm.x86.avx512.pbroadcastq.512
-  return _mm512_broadcastq_epi64(__A); 
+  // CHECK: shufflevector <2 x i64> %{{.*}}, <2 x i64> undef, <8 x i32> zeroinitializer
+  return _mm512_broadcastq_epi64(__A);
 }
 
 __m512i test_mm512_mask_broadcastq_epi64(__m512i __O, __mmask8 __M, __m128i __A) {
   // CHECK-LABEL: @test_mm512_mask_broadcastq_epi64
-  // CHECK: @llvm.x86.avx512.pbroadcastq.512
-  return _mm512_mask_broadcastq_epi64(__O, __M, __A); 
+  // CHECK: shufflevector <2 x i64> %{{.*}}, <2 x i64> undef, <8 x i32> zeroinitializer
+  // CHECK: select <8 x i1> %{{.*}}, <8 x i64> %{{.*}}, <8 x i64> %{{.*}}
+  return _mm512_mask_broadcastq_epi64(__O, __M, __A);
 }
 
 __m512i test_mm512_maskz_broadcastq_epi64(__mmask8 __M, __m128i __A) {
   // CHECK-LABEL: @test_mm512_maskz_broadcastq_epi64
-  // CHECK: @llvm.x86.avx512.pbroadcastq.512
-  return _mm512_maskz_broadcastq_epi64(__M, __A); 
+  // CHECK: shufflevector <2 x i64> %{{.*}}, <2 x i64> undef, <8 x i32> zeroinitializer
+  // CHECK: select <8 x i1> %{{.*}}, <8 x i64> %{{.*}}, <8 x i64> %{{.*}}
+  return _mm512_maskz_broadcastq_epi64(__M, __A);
 }
 
 __m128i test_mm512_cvtsepi32_epi8(__m512i __A) {
diff --git a/clang/test/CodeGen/avx512vl-builtins.c b/clang/test/CodeGen/avx512vl-builtins.c
index f27849a15c0..c4f70fbddcd 100644
--- a/clang/test/CodeGen/avx512vl-builtins.c
+++ b/clang/test/CodeGen/avx512vl-builtins.c
@@ -5353,86 +5353,100 @@ __m256i test_mm256_maskz_broadcast_i32x4(__mmask8 __M, __m128i __A) {
 
 __m256d test_mm256_mask_broadcastsd_pd(__m256d __O, __mmask8 __M, __m128d __A) {
   // CHECK-LABEL: @test_mm256_mask_broadcastsd_pd
-  // CHECK: @llvm.x86.avx512.mask.broadcast.sd.pd.256
-  return _mm256_mask_broadcastsd_pd(__O, __M, __A); 
+  // CHECK: shufflevector <2 x double> %{{.*}}, <2 x double> %{{.*}}, <4 x i32> zeroinitializer
+  // CHECK: select <4 x i1> %{{.*}}, <4 x double> %{{.*}}, <4 x double> %{{.*}}
+  return _mm256_mask_broadcastsd_pd(__O, __M, __A);
 }
 
 __m256d test_mm256_maskz_broadcastsd_pd(__mmask8 __M, __m128d __A) {
   // CHECK-LABEL: @test_mm256_maskz_broadcastsd_pd
-  // CHECK: @llvm.x86.avx512.mask.broadcast.sd.pd.256
-  return _mm256_maskz_broadcastsd_pd(__M, __A); 
+  // CHECK: shufflevector <2 x double> %{{.*}}, <2 x double> %{{.*}}, <4 x i32> zeroinitializer
+  // CHECK: select <4 x i1> %{{.*}}, <4 x double> %{{.*}}, <4 x double> %{{.*}}
+  return _mm256_maskz_broadcastsd_pd(__M, __A);
 }
 
 __m128 test_mm_mask_broadcastss_ps(__m128 __O, __mmask8 __M, __m128 __A) {
   // CHECK-LABEL: @test_mm_mask_broadcastss_ps
-  // CHECK: @llvm.x86.avx512.mask.broadcast.ss.ps.128
-  return _mm_mask_broadcastss_ps(__O, __M, __A); 
+  // CHECK: shufflevector <4 x float> %{{.*}}, <4 x float> %{{.*}}, <4 x i32> zeroinitializer
+  // CHECK: select <4 x i1> %{{.*}}, <4 x float> %{{.*}}, <4 x float> %{{.*}}
+  return _mm_mask_broadcastss_ps(__O, __M, __A);
 }
 
 __m128 test_mm_maskz_broadcastss_ps(__mmask8 __M, __m128 __A) {
   // CHECK-LABEL: @test_mm_maskz_broadcastss_ps
-  // CHECK: @llvm.x86.avx512.mask.broadcast.ss.ps.128
-  return _mm_maskz_broadcastss_ps(__M, __A); 
+  // CHECK: shufflevector <4 x float> %{{.*}}, <4 x float> %{{.*}}, <4 x i32> zeroinitializer
+  // CHECK: select <4 x i1> %{{.*}}, <4 x float> %{{.*}}, <4 x float> %{{.*}}
+  return _mm_maskz_broadcastss_ps(__M, __A);
 }
 
 __m256 test_mm256_mask_broadcastss_ps(__m256 __O, __mmask8 __M, __m128 __A) {
   // CHECK-LABEL: @test_mm256_mask_broadcastss_ps
-  // CHECK: @llvm.x86.avx512.mask.broadcast.ss.ps.256
-  return _mm256_mask_broadcastss_ps(__O, __M, __A); 
+  // CHECK: shufflevector <4 x float> %{{.*}}, <4 x float> %{{.*}}, <8 x i32> zeroinitializer
+  // CHECK: select <8 x i1> %{{.*}}, <8 x float> %{{.*}}, <8 x float> %{{.*}}
+  return _mm256_mask_broadcastss_ps(__O, __M, __A);
 }
 
 __m256 test_mm256_maskz_broadcastss_ps(__mmask8 __M, __m128 __A) {
   // CHECK-LABEL: @test_mm256_maskz_broadcastss_ps
-  // CHECK: @llvm.x86.avx512.mask.broadcast.ss.ps.256
-  return _mm256_maskz_broadcastss_ps(__M, __A); 
+  // CHECK: shufflevector <4 x float> %{{.*}}, <4 x float> %{{.*}}, <8 x i32> zeroinitializer
+  // CHECK: select <8 x i1> %{{.*}}, <8 x float> %{{.*}}, <8 x float> %{{.*}}
+  return _mm256_maskz_broadcastss_ps(__M, __A);
 }
 
 __m128i test_mm_mask_broadcastd_epi32(__m128i __O, __mmask8 __M, __m128i __A) {
   // CHECK-LABEL: @test_mm_mask_broadcastd_epi32
-  // CHECK: @llvm.x86.avx512.pbroadcastd.128
-  return _mm_mask_broadcastd_epi32(__O, __M, __A); 
+  // CHECK: shufflevector <4 x i32> %{{.*}}, <4 x i32> %{{.*}}, <4 x i32> zeroinitializer
+  // CHECK: select <4 x i1> %{{.*}}, <4 x i32> %{{.*}}, <4 x i32> %{{.*}}
+  return _mm_mask_broadcastd_epi32(__O, __M, __A);
 }
 
 __m128i test_mm_maskz_broadcastd_epi32(__mmask8 __M, __m128i __A) {
   // CHECK-LABEL: @test_mm_maskz_broadcastd_epi32
-  // CHECK: @llvm.x86.avx512.pbroadcastd.128
-  return _mm_maskz_broadcastd_epi32(__M, __A); 
+  // CHECK: shufflevector <4 x i32> %{{.*}}, <4 x i32> %{{.*}}, <4 x i32> zeroinitializer
+  // CHECK: select <4 x i1> %{{.*}}, <4 x i32> %{{.*}}, <4 x i32> %{{.*}}
+  return _mm_maskz_broadcastd_epi32(__M, __A);
 }
 
 __m256i test_mm256_mask_broadcastd_epi32(__m256i __O, __mmask8 __M, __m128i __A) {
   // CHECK-LABEL: @test_mm256_mask_broadcastd_epi32
-  // CHECK: @llvm.x86.avx512.pbroadcastd.256
-  return _mm256_mask_broadcastd_epi32(__O, __M, __A); 
+  // CHECK: shufflevector <4 x i32> %{{.*}}, <4 x i32> %{{.*}}, <8 x i32> zeroinitializer
+  // CHECK: select <8 x i1> %{{.*}}, <8 x i32> %{{.*}}, <8 x i32> %{{.*}}
+  return _mm256_mask_broadcastd_epi32(__O, __M, __A);
 }
 
 __m256i test_mm256_maskz_broadcastd_epi32(__mmask8 __M, __m128i __A) {
   // CHECK-LABEL: @test_mm256_maskz_broadcastd_epi32
-  // CHECK: @llvm.x86.avx512.pbroadcastd.256
-  return _mm256_maskz_broadcastd_epi32(__M, __A); 
+  // CHECK: shufflevector <4 x i32> %{{.*}}, <4 x i32> %{{.*}}, <8 x i32> zeroinitializer
+  // CHECK: select <8 x i1> %{{.*}}, <8 x i32> %{{.*}}, <8 x i32> %{{.*}}
+  return _mm256_maskz_broadcastd_epi32(__M, __A);
 }
 
 __m128i test_mm_mask_broadcastq_epi64(__m128i __O, __mmask8 __M, __m128i __A) {
   // CHECK-LABEL: @test_mm_mask_broadcastq_epi64
-  // CHECK: @llvm.x86.avx512.pbroadcastq.128
-  return _mm_mask_broadcastq_epi64(__O, __M, __A); 
+  // CHECK: shufflevector <2 x i64> %{{.*}}, <2 x i64> %{{.*}}, <2 x i32> zeroinitializer
+  // CHECK: select <2 x i1> %{{.*}}, <2 x i64> %{{.*}}, <2 x i64> %{{.*}}
+  return _mm_mask_broadcastq_epi64(__O, __M, __A);
 }
 
 __m128i test_mm_maskz_broadcastq_epi64(__mmask8 __M, __m128i __A) {
   // CHECK-LABEL: @test_mm_maskz_broadcastq_epi64
-  // CHECK: @llvm.x86.avx512.pbroadcastq.128
-  return _mm_maskz_broadcastq_epi64(__M, __A); 
+  // CHECK: shufflevector <2 x i64> %{{.*}}, <2 x i64> %{{.*}}, <2 x i32> zeroinitializer
+  // CHECK: select <2 x i1> %{{.*}}, <2 x i64> %{{.*}}, <2 x i64> %{{.*}}
+  return _mm_maskz_broadcastq_epi64(__M, __A);
 }
 
 __m256i test_mm256_mask_broadcastq_epi64(__m256i __O, __mmask8 __M, __m128i __A) {
   // CHECK-LABEL: @test_mm256_mask_broadcastq_epi64
-  // CHECK: @llvm.x86.avx512.pbroadcastq.256
-  return _mm256_mask_broadcastq_epi64(__O, __M, __A); 
+  // CHECK: shufflevector <2 x i64> %{{.*}}, <2 x i64> %{{.*}}, <4 x i32> zeroinitializer
+  // CHECK: select <4 x i1> %{{.*}}, <4 x i64> %{{.*}}, <4 x i64> %{{.*}}
+  return _mm256_mask_broadcastq_epi64(__O, __M, __A);
 }
 
 __m256i test_mm256_maskz_broadcastq_epi64(__mmask8 __M, __m128i __A) {
   // CHECK-LABEL: @test_mm256_maskz_broadcastq_epi64
-  // CHECK: @llvm.x86.avx512.pbroadcastq.256
-  return _mm256_maskz_broadcastq_epi64(__M, __A); 
+  // CHECK: shufflevector <2 x i64> %{{.*}}, <2 x i64> %{{.*}}, <4 x i32> zeroinitializer
+  // CHECK: select <4 x i1> %{{.*}}, <4 x i64> %{{.*}}, <4 x i64> %{{.*}}
+  return _mm256_maskz_broadcastq_epi64(__M, __A);
 }
 
 __m128i test_mm_cvtsepi32_epi8(__m128i __A) {
diff --git a/clang/test/CodeGen/avx512vlbw-builtins.c b/clang/test/CodeGen/avx512vlbw-builtins.c
index 6779b7275d3..6bfa09f5d1c 100644
--- a/clang/test/CodeGen/avx512vlbw-builtins.c
+++ b/clang/test/CodeGen/avx512vlbw-builtins.c
@@ -2280,54 +2280,60 @@ __m256i test_mm256_movm_epi16(__mmask16 __A) {
   return _mm256_movm_epi16(__A); 
 }
 
-
-
 __m128i test_mm_mask_broadcastb_epi8(__m128i __O, __mmask16 __M, __m128i __A) {
   // CHECK-LABEL: @test_mm_mask_broadcastb_epi8
-  // CHECK: @llvm.x86.avx512.pbroadcastb.128
-  return _mm_mask_broadcastb_epi8(__O, __M, __A); 
+  // CHECK: shufflevector <16 x i8> %{{.*}}, <16 x i8> %{{.*}}, <16 x i32> zeroinitializer
+  // CHECK: select <16 x i1> %{{.*}}, <16 x i8> %{{.*}}, <16 x i8> %{{.*}}
+  return _mm_mask_broadcastb_epi8(__O, __M, __A);
 }
 
 __m128i test_mm_maskz_broadcastb_epi8(__mmask16 __M, __m128i __A) {
   // CHECK-LABEL: @test_mm_maskz_broadcastb_epi8
-  // CHECK: @llvm.x86.avx512.pbroadcastb.128
-  return _mm_maskz_broadcastb_epi8(__M, __A); 
+  // CHECK: shufflevector <16 x i8> %{{.*}}, <16 x i8> %{{.*}}, <16 x i32> zeroinitializer
+  // CHECK: select <16 x i1> %{{.*}}, <16 x i8> %{{.*}}, <16 x i8> %{{.*}}
+  return _mm_maskz_broadcastb_epi8(__M, __A);
 }
 
 __m256i test_mm256_mask_broadcastb_epi8(__m256i __O, __mmask32 __M, __m128i __A) {
   // CHECK-LABEL: @test_mm256_mask_broadcastb_epi8
-  // CHECK: @llvm.x86.avx512.pbroadcastb.256
-  return _mm256_mask_broadcastb_epi8(__O, __M, __A); 
+  // CHECK: shufflevector <16 x i8> %{{.*}}, <16 x i8> %{{.*}}, <32 x i32> zeroinitializer
+  // CHECK: select <32 x i1> %{{.*}}, <32 x i8> %{{.*}}, <32 x i8> %{{.*}}
+  return _mm256_mask_broadcastb_epi8(__O, __M, __A);
 }
 
 __m256i test_mm256_maskz_broadcastb_epi8(__mmask32 __M, __m128i __A) {
   // CHECK-LABEL: @test_mm256_maskz_broadcastb_epi8
-  // CHECK: @llvm.x86.avx512.pbroadcastb.256
-  return _mm256_maskz_broadcastb_epi8(__M, __A); 
+  // CHECK: shufflevector <16 x i8> %{{.*}}, <16 x i8> %{{.*}}, <32 x i32> zeroinitializer
+  // CHECK: select <32 x i1> %{{.*}}, <32 x i8> %{{.*}}, <32 x i8> %{{.*}}
+  return _mm256_maskz_broadcastb_epi8(__M, __A);
 }
 
 __m128i test_mm_mask_broadcastw_epi16(__m128i __O, __mmask8 __M, __m128i __A) {
   // CHECK-LABEL: @test_mm_mask_broadcastw_epi16
-  // CHECK: @llvm.x86.avx512.pbroadcastw.128
-  return _mm_mask_broadcastw_epi16(__O, __M, __A); 
+  // CHECK: shufflevector <8 x i16> %{{.*}}, <8 x i16> %{{.*}}, <8 x i32> zeroinitializer
+  // CHECK: select <8 x i1> %{{.*}}, <8 x i16> %{{.*}}, <8 x i16> %{{.*}}
+  return _mm_mask_broadcastw_epi16(__O, __M, __A);
 }
 
 __m128i test_mm_maskz_broadcastw_epi16(__mmask8 __M, __m128i __A) {
   // CHECK-LABEL: @test_mm_maskz_broadcastw_epi16
-  // CHECK: @llvm.x86.avx512.pbroadcastw.128
-  return _mm_maskz_broadcastw_epi16(__M, __A); 
+  // CHECK: shufflevector <8 x i16> %{{.*}}, <8 x i16> %{{.*}}, <8 x i32> zeroinitializer
+  // CHECK: select <8 x i1> %{{.*}}, <8 x i16> %{{.*}}, <8 x i16> %{{.*}}
+  return _mm_maskz_broadcastw_epi16(__M, __A);
 }
 
 __m256i test_mm256_mask_broadcastw_epi16(__m256i __O, __mmask16 __M, __m128i __A) {
   // CHECK-LABEL: @test_mm256_mask_broadcastw_epi16
-  // CHECK: @llvm.x86.avx512.pbroadcastw.256
-  return _mm256_mask_broadcastw_epi16(__O, __M, __A); 
+  // CHECK: shufflevector <8 x i16> %{{.*}}, <8 x i16> %{{.*}}, <16 x i32> zeroinitializer
+  // CHECK: select <16 x i1> %{{.*}}, <16 x i16> %{{.*}}, <16 x i16> %{{.*}}
+  return _mm256_mask_broadcastw_epi16(__O, __M, __A);
 }
 
 __m256i test_mm256_maskz_broadcastw_epi16(__mmask16 __M, __m128i __A) {
   // CHECK-LABEL: @test_mm256_maskz_broadcastw_epi16
-  // CHECK: @llvm.x86.avx512.pbroadcastw.256
-  return _mm256_maskz_broadcastw_epi16(__M, __A); 
+  // CHECK: shufflevector <8 x i16> %{{.*}}, <8 x i16> %{{.*}}, <16 x i32> zeroinitializer
+  // CHECK: select <16 x i1> %{{.*}}, <16 x i16> %{{.*}}, <16 x i16> %{{.*}}
+  return _mm256_maskz_broadcastw_epi16(__M, __A);
 }
 
 __m256i test_mm256_mask_set1_epi16(__m256i __O, __mmask16 __M, short __A) {