[X86] Avoid passing _mm_undefined* to builtin_shufflevector if we are able to pass the first input a second time.

This is more consistent with other usages of builtin_shufflevector. Later optimization passes or codegen will detect the duplicate vector and replace it with undef. Using _mm_undefined just puts a zeroinitializer that still needs to be optimized out later.

git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@333944 91177308-0d34-0410-b5e6-96231b3b80d8

lib/Headers/avx512bwintrin.h

@@ -1889,8 +1889,7 @@ _mm512_movm_epi16 (__mmask32 __A)
 static __inline__ __m512i __DEFAULT_FN_ATTRS
 _mm512_broadcastb_epi8 (__m128i __A)
 {
-  return (__m512i)__builtin_shufflevector((__v16qi) __A,
-                                          (__v16qi)_mm_undefined_si128(),
+  return (__m512i)__builtin_shufflevector((__v16qi) __A, (__v16qi) __A,
                                           0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
                                           0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
                                           0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
@@ -1932,8 +1931,7 @@ _mm512_maskz_set1_epi16 (__mmask32 __M, short __A)
 static __inline__ __m512i __DEFAULT_FN_ATTRS
 _mm512_broadcastw_epi16 (__m128i __A)
 {
-  return (__m512i)__builtin_shufflevector((__v8hi) __A,
-                                          (__v8hi)_mm_undefined_si128(),
+  return (__m512i)__builtin_shufflevector((__v8hi) __A, (__v8hi) __A,
                                           0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
                                           0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0);
 }

lib/Headers/avx512dqintrin.h

@@ -963,8 +963,7 @@ _mm512_movepi64_mask (__m512i __A)
 static __inline__ __m512 __DEFAULT_FN_ATTRS
 _mm512_broadcast_f32x2 (__m128 __A)
 {
-  return (__m512)__builtin_shufflevector((__v4sf)__A,
-                                         (__v4sf)_mm_undefined_ps(),
+  return (__m512)__builtin_shufflevector((__v4sf)__A, (__v4sf)__A,
                                          0, 1, 0, 1, 0, 1, 0, 1,
                                          0, 1, 0, 1, 0, 1, 0, 1);
 }
@@ -1035,8 +1034,7 @@ _mm512_maskz_broadcast_f64x2(__mmask8 __M, __m128d __A)
 static __inline__ __m512i __DEFAULT_FN_ATTRS
 _mm512_broadcast_i32x2 (__m128i __A)
 {
-  return (__m512i)__builtin_shufflevector((__v4si)__A,
-                                          (__v4si)_mm_undefined_si128(),
+  return (__m512i)__builtin_shufflevector((__v4si)__A, (__v4si)__A,
                                           0, 1, 0, 1, 0, 1, 0, 1,
                                           0, 1, 0, 1, 0, 1, 0, 1);
 }

lib/Headers/avx512fintrin.h

@@ -212,8 +212,7 @@ _mm512_undefined_epi32(void)
 static __inline__ __m512i __DEFAULT_FN_ATTRS
 _mm512_broadcastd_epi32 (__m128i __A)
 {
-  return (__m512i)__builtin_shufflevector((__v4si) __A,
-                                          (__v4si)_mm_undefined_si128(),
+  return (__m512i)__builtin_shufflevector((__v4si) __A, (__v4si) __A,
                                           0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0);
 }
 
@@ -236,8 +235,7 @@ _mm512_maskz_broadcastd_epi32 (__mmask16 __M, __m128i __A)
 static __inline__ __m512i __DEFAULT_FN_ATTRS
 _mm512_broadcastq_epi64 (__m128i __A)
 {
-  return (__m512i)__builtin_shufflevector((__v2di) __A,
-                                          (__v2di) _mm_undefined_si128(),
+  return (__m512i)__builtin_shufflevector((__v2di) __A, (__v2di) __A,
                                           0, 0, 0, 0, 0, 0, 0, 0);
 }
 
@@ -344,8 +342,7 @@ _mm512_maskz_set1_epi64(__mmask8 __M, long long __A)
 static __inline__ __m512 __DEFAULT_FN_ATTRS
 _mm512_broadcastss_ps(__m128 __A)
 {
-  return (__m512)__builtin_shufflevector((__v4sf) __A,
-                                         (__v4sf)_mm_undefined_ps(),
+  return (__m512)__builtin_shufflevector((__v4sf) __A, (__v4sf) __A,
                                          0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0);
 }
 
@@ -395,8 +392,7 @@ _mm512_set4_ps (float __A, float __B, float __C, float __D)
 static __inline__ __m512d __DEFAULT_FN_ATTRS
 _mm512_broadcastsd_pd(__m128d __A)
 {
-  return (__m512d)__builtin_shufflevector((__v2df) __A,
-                                          (__v2df) _mm_undefined_pd(),
+  return (__m512d)__builtin_shufflevector((__v2df) __A, (__v2df) __A,
                                           0, 0, 0, 0, 0, 0, 0, 0);
 }
 

lib/Headers/avx512vldqintrin.h

@@ -970,8 +970,7 @@ _mm256_movepi64_mask (__m256i __A)
 static __inline__ __m256 __DEFAULT_FN_ATTRS
 _mm256_broadcast_f32x2 (__m128 __A)
 {
-  return (__m256)__builtin_shufflevector((__v4sf)__A,
-                                         (__v4sf)_mm_undefined_ps(),
+  return (__m256)__builtin_shufflevector((__v4sf)__A, (__v4sf)__A,
                                          0, 1, 0, 1, 0, 1, 0, 1);
 }
 
@@ -1017,8 +1016,7 @@ _mm256_maskz_broadcast_f64x2 (__mmask8 __M, __m128d __A)
 static __inline__ __m128i __DEFAULT_FN_ATTRS
 _mm_broadcast_i32x2 (__m128i __A)
 {
-  return (__m128i)__builtin_shufflevector((__v4si)__A,
-                                          (__v4si)_mm_undefined_si128(),
+  return (__m128i)__builtin_shufflevector((__v4si)__A, (__v4si)__A,
                                           0, 1, 0, 1);
 }
 
@@ -1041,8 +1039,7 @@ _mm_maskz_broadcast_i32x2 (__mmask8 __M, __m128i __A)
 static __inline__ __m256i __DEFAULT_FN_ATTRS
 _mm256_broadcast_i32x2 (__m128i __A)
 {
-  return (__m256i)__builtin_shufflevector((__v4si)__A,
-                                          (__v4si)_mm_undefined_si128(),
+  return (__m256i)__builtin_shufflevector((__v4si)__A, (__v4si)__A,
                                           0, 1, 0, 1, 0, 1, 0, 1);
 }
 

test/CodeGen/avx512bw-builtins.c

@@ -1778,40 +1778,40 @@ __m512i test_mm512_movm_epi16(__mmask32 __A) {
 
 __m512i test_mm512_broadcastb_epi8(__m128i __A) {
   // CHECK-LABEL: @test_mm512_broadcastb_epi8
-  // CHECK: shufflevector <16 x i8> %{{.*}}, <16 x i8> zeroinitializer, <64 x i32> zeroinitializer
+  // CHECK: shufflevector <16 x i8> %{{.*}}, <16 x i8> %{{.*}}, <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: shufflevector <16 x i8> %{{.*}}, <16 x i8> zeroinitializer, <64 x i32> zeroinitializer
+  // CHECK: shufflevector <16 x i8> %{{.*}}, <16 x i8> %{{.*}}, <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: shufflevector <16 x i8> %{{.*}}, <16 x i8> zeroinitializer, <64 x i32> zeroinitializer
+  // CHECK: shufflevector <16 x i8> %{{.*}}, <16 x i8> %{{.*}}, <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: shufflevector <8 x i16> %{{.*}}, <8 x i16> zeroinitializer, <32 x i32> zeroinitializer
+  // CHECK: shufflevector <8 x i16> %{{.*}}, <8 x i16> %{{.*}}, <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: shufflevector <8 x i16> %{{.*}}, <8 x i16> zeroinitializer, <32 x i32> zeroinitializer
+  // CHECK: shufflevector <8 x i16> %{{.*}}, <8 x i16> %{{.*}}, <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: shufflevector <8 x i16> %{{.*}}, <8 x i16> zeroinitializer, <32 x i32> zeroinitializer
+  // CHECK: shufflevector <8 x i16> %{{.*}}, <8 x i16> %{{.*}}, <32 x i32> zeroinitializer
   // CHECK: select <32 x i1> %{{.*}}, <32 x i16> %{{.*}}, <32 x i16> %{{.*}}
   return _mm512_maskz_broadcastw_epi16(__M, __A);
 }

test/CodeGen/avx512dq-builtins.c

@@ -955,20 +955,20 @@ __mmask8 test_mm512_movepi64_mask(__m512i __A) {
 
 __m512 test_mm512_broadcast_f32x2(__m128 __A) {
   // CHECK-LABEL: @test_mm512_broadcast_f32x2
-  // CHECK: shufflevector <4 x float> %{{.*}}, <4 x float> zeroinitializer, <16 x i32> <i32 0, i32 1, i32 0, i32 1, i32 0, i32 1, i32 0, i32 1, i32 0, i32 1, i32 0, i32 1, i32 0, i32 1, i32 0, i32 1>
+  // CHECK: shufflevector <4 x float> %{{.*}}, <4 x float> %{{.*}}, <16 x i32> <i32 0, i32 1, i32 0, i32 1, i32 0, i32 1, i32 0, i32 1, i32 0, i32 1, i32 0, i32 1, i32 0, i32 1, i32 0, i32 1>
   return _mm512_broadcast_f32x2(__A); 
 }
 
 __m512 test_mm512_mask_broadcast_f32x2(__m512 __O, __mmask16 __M, __m128 __A) {
   // CHECK-LABEL: @test_mm512_mask_broadcast_f32x2
-  // CHECK: shufflevector <4 x float> %{{.*}}, <4 x float> zeroinitializer, <16 x i32> <i32 0, i32 1, i32 0, i32 1, i32 0, i32 1, i32 0, i32 1, i32 0, i32 1, i32 0, i32 1, i32 0, i32 1, i32 0, i32 1>
+  // CHECK: shufflevector <4 x float> %{{.*}}, <4 x float> %{{.*}}, <16 x i32> <i32 0, i32 1, i32 0, i32 1, i32 0, i32 1, i32 0, i32 1, i32 0, i32 1, i32 0, i32 1, i32 0, i32 1, i32 0, i32 1>
   // CHECK: select <16 x i1> %{{.*}}, <16 x float> %{{.*}}, <16 x float> %{{.*}}
   return _mm512_mask_broadcast_f32x2(__O, __M, __A); 
 }
 
 __m512 test_mm512_maskz_broadcast_f32x2(__mmask16 __M, __m128 __A) {
   // CHECK-LABEL: @test_mm512_maskz_broadcast_f32x2
-  // CHECK: shufflevector <4 x float> %{{.*}}, <4 x float> zeroinitializer, <16 x i32> <i32 0, i32 1, i32 0, i32 1, i32 0, i32 1, i32 0, i32 1, i32 0, i32 1, i32 0, i32 1, i32 0, i32 1, i32 0, i32 1>
+  // CHECK: shufflevector <4 x float> %{{.*}}, <4 x float> %{{.*}}, <16 x i32> <i32 0, i32 1, i32 0, i32 1, i32 0, i32 1, i32 0, i32 1, i32 0, i32 1, i32 0, i32 1, i32 0, i32 1, i32 0, i32 1>
   // CHECK: select <16 x i1> %{{.*}}, <16 x float> %{{.*}}, <16 x float> %{{.*}}
   return _mm512_maskz_broadcast_f32x2(__M, __A); 
 }
@@ -1015,20 +1015,20 @@ __m512d test_mm512_maskz_broadcast_f64x2(__mmask8 __M, double const* __A) {
 
 __m512i test_mm512_broadcast_i32x2(__m128i __A) {
   // CHECK-LABEL: @test_mm512_broadcast_i32x2
-  // CHECK: shufflevector <4 x i32> %{{.*}}, <4 x i32> zeroinitializer, <16 x i32> <i32 0, i32 1, i32 0, i32 1, i32 0, i32 1, i32 0, i32 1, i32 0, i32 1, i32 0, i32 1, i32 0, i32 1, i32 0, i32 1>
+  // CHECK: shufflevector <4 x i32> %{{.*}}, <4 x i32> %{{.*}}, <16 x i32> <i32 0, i32 1, i32 0, i32 1, i32 0, i32 1, i32 0, i32 1, i32 0, i32 1, i32 0, i32 1, i32 0, i32 1, i32 0, i32 1>
   return _mm512_broadcast_i32x2(__A); 
 }
 
 __m512i test_mm512_mask_broadcast_i32x2(__m512i __O, __mmask16 __M, __m128i __A) {
   // CHECK-LABEL: @test_mm512_mask_broadcast_i32x2
-  // CHECK: shufflevector <4 x i32> %{{.*}}, <4 x i32> zeroinitializer, <16 x i32> <i32 0, i32 1, i32 0, i32 1, i32 0, i32 1, i32 0, i32 1, i32 0, i32 1, i32 0, i32 1, i32 0, i32 1, i32 0, i32 1>
+  // CHECK: shufflevector <4 x i32> %{{.*}}, <4 x i32> %{{.*}}, <16 x i32> <i32 0, i32 1, i32 0, i32 1, i32 0, i32 1, i32 0, i32 1, i32 0, i32 1, i32 0, i32 1, i32 0, i32 1, i32 0, i32 1>
   // CHECK: select <16 x i1> %{{.*}}, <16 x i32> %{{.*}}, <16 x i32> %{{.*}}
   return _mm512_mask_broadcast_i32x2(__O, __M, __A); 
 }
 
 __m512i test_mm512_maskz_broadcast_i32x2(__mmask16 __M, __m128i __A) {
   // CHECK-LABEL: @test_mm512_maskz_broadcast_i32x2
-  // CHECK: shufflevector <4 x i32> %{{.*}}, <4 x i32> zeroinitializer, <16 x i32> <i32 0, i32 1, i32 0, i32 1, i32 0, i32 1, i32 0, i32 1, i32 0, i32 1, i32 0, i32 1, i32 0, i32 1, i32 0, i32 1>
+  // CHECK: shufflevector <4 x i32> %{{.*}}, <4 x i32> %{{.*}}, <16 x i32> <i32 0, i32 1, i32 0, i32 1, i32 0, i32 1, i32 0, i32 1, i32 0, i32 1, i32 0, i32 1, i32 0, i32 1, i32 0, i32 1>
   // CHECK: select <16 x i1> %{{.*}}, <16 x i32> %{{.*}}, <16 x i32> %{{.*}}
   return _mm512_maskz_broadcast_i32x2(__M, __A); 
 }

test/CodeGen/avx512f-builtins.c

@@ -5072,80 +5072,80 @@ __m512i test_mm512_maskz_broadcast_i64x4(__mmask8 __M, __m256i const* __A) {
 
 __m512d test_mm512_broadcastsd_pd(__m128d __A) {
   // CHECK-LABEL: @test_mm512_broadcastsd_pd
-  // CHECK: shufflevector <2 x double> %{{.*}}, <2 x double> zeroinitializer, <8 x i32> zeroinitializer
+  // CHECK: shufflevector <2 x double> %{{.*}}, <2 x double> %{{.*}}, <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: shufflevector <2 x double> %{{.*}}, <2 x double> zeroinitializer, <8 x i32> zeroinitializer
+  // CHECK: shufflevector <2 x double> %{{.*}}, <2 x double> %{{.*}}, <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: shufflevector <2 x double> %{{.*}}, <2 x double> zeroinitializer, <8 x i32> zeroinitializer
+  // CHECK: shufflevector <2 x double> %{{.*}}, <2 x double> %{{.*}}, <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> zeroinitializer, <16 x i32> zeroinitializer
+  // CHECK: shufflevector <4 x float> %{{.*}}, <4 x float> %{{.*}}, <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: shufflevector <4 x float> %{{.*}}, <4 x float> zeroinitializer, <16 x i32> zeroinitializer
+  // CHECK: shufflevector <4 x float> %{{.*}}, <4 x float> %{{.*}}, <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: shufflevector <4 x float> %{{.*}}, <4 x float> zeroinitializer, <16 x i32> zeroinitializer
+  // CHECK: shufflevector <4 x float> %{{.*}}, <4 x float> %{{.*}}, <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: shufflevector <4 x i32> %{{.*}}, <4 x i32> zeroinitializer, <16 x i32> zeroinitializer
+  // CHECK: shufflevector <4 x i32> %{{.*}}, <4 x i32> %{{.*}}, <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: shufflevector <4 x i32> %{{.*}}, <4 x i32> zeroinitializer, <16 x i32> zeroinitializer
+  // CHECK: shufflevector <4 x i32> %{{.*}}, <4 x i32> %{{.*}}, <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: shufflevector <4 x i32> %{{.*}}, <4 x i32> zeroinitializer, <16 x i32> zeroinitializer
+  // CHECK: shufflevector <4 x i32> %{{.*}}, <4 x i32> %{{.*}}, <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: shufflevector <2 x i64> %{{.*}}, <2 x i64> zeroinitializer, <8 x i32> zeroinitializer
+  // CHECK: shufflevector <2 x i64> %{{.*}}, <2 x i64> %{{.*}}, <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: shufflevector <2 x i64> %{{.*}}, <2 x i64> zeroinitializer, <8 x i32> zeroinitializer
+  // CHECK: shufflevector <2 x i64> %{{.*}}, <2 x i64> %{{.*}}, <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: shufflevector <2 x i64> %{{.*}}, <2 x i64> zeroinitializer, <8 x i32> zeroinitializer
+  // CHECK: shufflevector <2 x i64> %{{.*}}, <2 x i64> %{{.*}}, <8 x i32> zeroinitializer
   // CHECK: select <8 x i1> %{{.*}}, <8 x i64> %{{.*}}, <8 x i64> %{{.*}}
   return _mm512_maskz_broadcastq_epi64(__M, __A);
 }

test/CodeGen/avx512vldq-builtins.c

@@ -924,20 +924,20 @@ __mmask8 test_mm256_movepi64_mask(__m256i __A) {
 
 __m256 test_mm256_broadcast_f32x2(__m128 __A) {
   // CHECK-LABEL: @test_mm256_broadcast_f32x2
-  // CHECK: shufflevector <4 x float> %{{.*}}, <4 x float> zeroinitializer, <8 x i32> <i32 0, i32 1, i32 0, i32 1, i32 0, i32 1, i32 0, i32 1>
+  // CHECK: shufflevector <4 x float> %{{.*}}, <4 x float> %{{.*}}, <8 x i32> <i32 0, i32 1, i32 0, i32 1, i32 0, i32 1, i32 0, i32 1>
   return _mm256_broadcast_f32x2(__A); 
 }
 
 __m256 test_mm256_mask_broadcast_f32x2(__m256 __O, __mmask8 __M, __m128 __A) {
   // CHECK-LABEL: @test_mm256_mask_broadcast_f32x2
-  // CHECK: shufflevector <4 x float> %{{.*}}, <4 x float> zeroinitializer, <8 x i32> <i32 0, i32 1, i32 0, i32 1, i32 0, i32 1, i32 0, i32 1>
+  // CHECK: shufflevector <4 x float> %{{.*}}, <4 x float> %{{.*}}, <8 x i32> <i32 0, i32 1, i32 0, i32 1, i32 0, i32 1, i32 0, i32 1>
   // CHECK: select <8 x i1> %{{.*}}, <8 x float> %{{.*}}, <8 x float> %{{.*}}
   return _mm256_mask_broadcast_f32x2(__O, __M, __A); 
 }
 
 __m256 test_mm256_maskz_broadcast_f32x2(__mmask8 __M, __m128 __A) {
   // CHECK-LABEL: @test_mm256_maskz_broadcast_f32x2
-  // CHECK: shufflevector <4 x float> %{{.*}}, <4 x float> zeroinitializer, <8 x i32> <i32 0, i32 1, i32 0, i32 1, i32 0, i32 1, i32 0, i32 1>
+  // CHECK: shufflevector <4 x float> %{{.*}}, <4 x float> %{{.*}}, <8 x i32> <i32 0, i32 1, i32 0, i32 1, i32 0, i32 1, i32 0, i32 1>
   // CHECK: select <8 x i1> %{{.*}}, <8 x float> %{{.*}}, <8 x float> %{{.*}}
   return _mm256_maskz_broadcast_f32x2(__M, __A); 
 }
@@ -964,40 +964,40 @@ __m256d test_mm256_maskz_broadcast_f64x2(__mmask8 __M, double const* __A) {
 
 __m128i test_mm_broadcast_i32x2(__m128i __A) {
   // CHECK-LABEL: @test_mm_broadcast_i32x2
-  // CHECK: shufflevector <4 x i32> %{{.*}}, <4 x i32> zeroinitializer, <4 x i32> <i32 0, i32 1, i32 0, i32 1>
+  // CHECK: shufflevector <4 x i32> %{{.*}}, <4 x i32> %{{.*}}, <4 x i32> <i32 0, i32 1, i32 0, i32 1>
   return _mm_broadcast_i32x2(__A); 
 }
 
 __m128i test_mm_mask_broadcast_i32x2(__m128i __O, __mmask8 __M, __m128i __A) {
   // CHECK-LABEL: @test_mm_mask_broadcast_i32x2
-  // CHECK: shufflevector <4 x i32> %{{.*}}, <4 x i32> zeroinitializer, <4 x i32> <i32 0, i32 1, i32 0, i32 1>
+  // CHECK: shufflevector <4 x i32> %{{.*}}, <4 x i32> %{{.*}}, <4 x i32> <i32 0, i32 1, i32 0, i32 1>
   // CHECK: select <4 x i1> %{{.*}}, <4 x i32> %{{.*}}, <4 x i32> %{{.*}}
   return _mm_mask_broadcast_i32x2(__O, __M, __A); 
 }
 
 __m128i test_mm_maskz_broadcast_i32x2(__mmask8 __M, __m128i __A) {
   // CHECK-LABEL: @test_mm_maskz_broadcast_i32x2
-  // CHECK: shufflevector <4 x i32> %{{.*}}, <4 x i32> zeroinitializer, <4 x i32> <i32 0, i32 1, i32 0, i32 1>
+  // CHECK: shufflevector <4 x i32> %{{.*}}, <4 x i32> %{{.*}}, <4 x i32> <i32 0, i32 1, i32 0, i32 1>
   // CHECK: select <4 x i1> %{{.*}}, <4 x i32> %{{.*}}, <4 x i32> %{{.*}}
   return _mm_maskz_broadcast_i32x2(__M, __A); 
 }
 
 __m256i test_mm256_broadcast_i32x2(__m128i __A) {
   // CHECK-LABEL: @test_mm256_broadcast_i32x2
-  // CHECK: shufflevector <4 x i32> %{{.*}}, <4 x i32> zeroinitializer, <8 x i32> <i32 0, i32 1, i32 0, i32 1, i32 0, i32 1, i32 0, i32 1>
+  // CHECK: shufflevector <4 x i32> %{{.*}}, <4 x i32> %{{.*}}, <8 x i32> <i32 0, i32 1, i32 0, i32 1, i32 0, i32 1, i32 0, i32 1>
   return _mm256_broadcast_i32x2(__A); 
 }
 
 __m256i test_mm256_mask_broadcast_i32x2(__m256i __O, __mmask8 __M, __m128i __A) {
   // CHECK-LABEL: @test_mm256_mask_broadcast_i32x2
-  // CHECK: shufflevector <4 x i32> %{{.*}}, <4 x i32> zeroinitializer, <8 x i32> <i32 0, i32 1, i32 0, i32 1, i32 0, i32 1, i32 0, i32 1>
+  // CHECK: shufflevector <4 x i32> %{{.*}}, <4 x i32> %{{.*}}, <8 x i32> <i32 0, i32 1, i32 0, i32 1, i32 0, i32 1, i32 0, i32 1>
   // CHECK: select <8 x i1> %{{.*}}, <8 x i32> %{{.*}}, <8 x i32> %{{.*}}
   return _mm256_mask_broadcast_i32x2(__O, __M, __A); 
 }
 
 __m256i test_mm256_maskz_broadcast_i32x2(__mmask8 __M, __m128i __A) {
   // CHECK-LABEL: @test_mm256_maskz_broadcast_i32x2
-  // CHECK: shufflevector <4 x i32> %{{.*}}, <4 x i32> zeroinitializer, <8 x i32> <i32 0, i32 1, i32 0, i32 1, i32 0, i32 1, i32 0, i32 1>
+  // CHECK: shufflevector <4 x i32> %{{.*}}, <4 x i32> %{{.*}}, <8 x i32> <i32 0, i32 1, i32 0, i32 1, i32 0, i32 1, i32 0, i32 1>
   // CHECK: select <8 x i1> %{{.*}}, <8 x i32> %{{.*}}, <8 x i32> %{{.*}}
   return _mm256_maskz_broadcast_i32x2(__M, __A); 
 }