[Clang][BuiltIn][AVX512] Adding intrinsics fot align{d|q} and palignr instruction set

[AsafBadouh]

please change from setzero to undefined in the non mask versions.
LGTM after that.

lib/Headers/avx512bwintrin.h

2063

please change it to _mm512_undefined_epi32()

lib/Headers/avx512vlintrin.h

7771

_mm_undefined_ps()

include/clang/Basic/BuiltinsX86.def

 TARGET_BUILTIN(__builtin_ia32_storeups512_mask, "vv*V16fUs", "", "avx512f")
 TARGET_BUILTIN(__builtin_ia32_storeaps512_mask, "vv*V16fUs", "", "avx512f")
 TARGET_BUILTIN(__builtin_ia32_vpermt2vard512_mask, "V16iV16iV16iV16iUs", "", "avx512f")
 TARGET_BUILTIN(__builtin_ia32_vpermt2varq512_mask, "V8LLiV8LLiV8LLiV8LLiUc", "", "avx512f")
 TARGET_BUILTIN(__builtin_ia32_vpermt2varps512_mask, "V16fV16iV16fV16fUs", "", "avx512f")
 TARGET_BUILTIN(__builtin_ia32_vpermt2varpd512_mask, "V8dV8LLiV8dV8dUc", "", "avx512f")
 TARGET_BUILTIN(__builtin_ia32_alignq512_mask, "V8LLiV8LLiV8LLiIiV8LLiUc", "", "avx512f")
 TARGET_BUILTIN(__builtin_ia32_alignd512_mask, "V16iV16iV16iIiV16iUs", "", "avx512f")
+TARGET_BUILTIN(__builtin_ia32_alignd128_mask, "V4iV4iV4iIiV4iUc","","avx512vl")
+TARGET_BUILTIN(__builtin_ia32_alignd256_mask, "V8iV8iV8iIiV8iUc","","avx512vl")
+TARGET_BUILTIN(__builtin_ia32_alignq128_mask, "V2LLiV2LLiV2LLiIiV2LLiUc","","avx512vl")
+TARGET_BUILTIN(__builtin_ia32_alignq256_mask, "V4LLiV4LLiV4LLiIiV4LLiUc","","avx512vl")
 TARGET_BUILTIN(__builtin_ia32_extractf64x4_mask, "V4dV8dIiV4dUc", "", "avx512f")
 TARGET_BUILTIN(__builtin_ia32_extractf32x4_mask, "V4fV16fIiV4fUc", "", "avx512f")
 
 TARGET_BUILTIN(__builtin_ia32_gathersiv8df, "V8dV8dvC*V8iUcIi", "", "avx512f")
 TARGET_BUILTIN(__builtin_ia32_gathersiv16sf, "V16fV16fvC*UsIi", "", "avx512f")
 TARGET_BUILTIN(__builtin_ia32_gatherdiv8df, "V8dV8dvC*V8LLiUcIi", "", "avx512f")
 TARGET_BUILTIN(__builtin_ia32_gatherdiv16sf, "V8fV8fvC*V8LLiUcIi", "", "avx512f")
 TARGET_BUILTIN(__builtin_ia32_gathersiv8di, "V8LLiV8LLivC*V8iUcIi", "", "avx512f")

 TARGET_BUILTIN(__builtin_ia32_shufps128_mask, "V4fV4fV4fIiV4fUc","","avx512vl")
 TARGET_BUILTIN(__builtin_ia32_shufps256_mask, "V8fV8fV8fIiV8fUc","","avx512vl")
 TARGET_BUILTIN(__builtin_ia32_sqrtsd_round_mask, "V2dV2dV2dV2dUcIi","","avx512f")
 TARGET_BUILTIN(__builtin_ia32_sqrtss_round_mask, "V4fV4fV4fV4fUcIi","","avx512f")
 TARGET_BUILTIN(__builtin_ia32_rsqrt14pd128_mask, "V2dV2dV2dUc","","avx512vl")
 TARGET_BUILTIN(__builtin_ia32_rsqrt14pd256_mask, "V4dV4dV4dUc","","avx512vl")
 TARGET_BUILTIN(__builtin_ia32_rsqrt14ps128_mask, "V4fV4fV4fUc","","avx512vl")
 TARGET_BUILTIN(__builtin_ia32_rsqrt14ps256_mask, "V8fV8fV8fUc","","avx512vl")
+TARGET_BUILTIN(__builtin_ia32_palignr512_mask, "V64cV64cV64ciV64cULLi","","avx512bw")
+TARGET_BUILTIN(__builtin_ia32_palignr128_mask, "V16cV16cV16ciV16cUs","","avx512bw,avx512vl")
+TARGET_BUILTIN(__builtin_ia32_palignr256_mask, "V32cV32cV32ciV32cUi","","avx512bw,avx512vl")
+
 
 #undef BUILTIN
 #undef TARGET_BUILTIN

lib/Headers/avx512bwintrin.h

 
 static __inline__ __mmask32 __DEFAULT_FN_ATTRS
 _mm512_mask_testn_epi16_mask (__mmask32 __U, __m512i __A, __m512i __B)
 {
   return (__mmask32) __builtin_ia32_ptestnmw512 ((__v32hi) __A,
              (__v32hi) __B, __U);
 }
 
+#define _mm512_alignr_epi8( __A, __B, __N) __extension__ ({\
+__builtin_ia32_palignr512_mask ((__v8di) __A,\
+                 (__v8di) __B ,__N * 8,\
+                 (__v8di) _mm512_setzero_si512 (),\
+                 (__mmask64) -1);\
+})
+
+#define _mm512_mask_alignr_epi8( __W, __U, __A, __B, __N) __extension__({\
+__builtin_ia32_palignr512_mask ((__v8di) __A,\
+                 (__v8di) __B,\
+                 __N * 8,\
+                 (__v8di) __W,\
+                 (__mmask64) __U);\
+})
+
+#define _mm512_maskz_alignr_epi8 ( __U, __A, __B, __N) __extension__({\
+__builtin_ia32_palignr512_mask ((__v8di) __A,\
+                 (__v8di) __B,\
+                 __N * 8,\
+                 (__v8di) _mm512_setzero_si512 (),\
+                 (__mmask64) __U);\
+})
 
 #undef __DEFAULT_FN_ATTRS
 
 #endif

lib/Headers/avx512fintrin.h

 }
 
 #define _mm512_alignr_epi64(A, B, I) __extension__ ({ \
   (__m512i)__builtin_ia32_alignq512_mask((__v8di)(__m512i)(A), \
                                          (__v8di)(__m512i)(B), \
                                          (I), (__v8di)_mm512_setzero_si512(), \
                                          (__mmask8)-1); })
 
+#define _mm512_mask_alignr_epi64( __W,  __U,  __A, __B, __imm) __extension__({\
+  (__m512i)__builtin_ia32_alignq512_mask ((__v8di) __A,\
+                                         (__v8di) __B, __imm,\
+                                         (__v8di) __W,\
+                                         (__mmask8) __U);\
+})
+
+#define _mm512_maskz_alignr_epi64( __U,  __A,  __B, __imm) __extension__({\
+  (__m512i)__builtin_ia32_alignq512_mask ((__v8di) __A,\
+                                         (__v8di) __B, __imm,\
+                                         (__v8di) _mm512_setzero_si512 (),\
+                                         (__mmask8) __U);\
+})
+
 #define _mm512_alignr_epi32(A, B, I) __extension__ ({ \
     (__m512i)__builtin_ia32_alignd512_mask((__v16si)(__m512i)(A), \
                                          (__v16si)(__m512i)(B), \
                                          (I), (__v16si)_mm512_setzero_si512(), \
-                                         (__mmask16)-1); })
+                                         (__mmask16)-1);\
+})
+                                         
+#define _mm512_mask_alignr_epi32( __W, __U, __A, __B,  __imm) __extension__ ({\
+    (__m512i) __builtin_ia32_alignd512_mask((__v16si) __A,\
+                                         (__v16si) __B, __imm,\
+                                         (__v16si) __W,\
+                                         (__mmask16) __U);\
+})
 
+#define _mm512_maskz_alignr_epi32( __U, __A, __B, __imm) __extension__({\
+    (__m512i) __builtin_ia32_alignd512_mask ((__v16si) __A,\
+                                         (__v16si) __B, __imm,\
+                                         (__v16si) _mm512_setzero_si512 (),\
+                                         (__mmask16) __U);\
+}
 /* Vector Extract */
 
 #define _mm512_extractf64x4_pd(A, I) __extension__ ({                    \
       (__m256d)                                                          \
         __builtin_ia32_extractf64x4_mask((__v8df)(__m512d)(A),           \
                                          (I),                            \
                                          (__v4df)_mm256_setzero_si256(), \
                                          (__mmask8) -1); })

lib/Headers/avx512vlbwintrin.h

 
 static __inline__ __mmask16 __DEFAULT_FN_ATTRS
 _mm256_mask_testn_epi16_mask (__mmask16 __U, __m256i __A, __m256i __B)
 {
   return (__mmask16) __builtin_ia32_ptestnmw256 ((__v16hi) __A,
              (__v16hi) __B, __U);
 }
 
+#define _mm_mask_alignr_epi8( __W, __U, __A, __B, __N) __extension__ ({ \
+__builtin_ia32_palignr128_mask ((__v2di)( __A),\
+               (__v2di)( __B),\
+               ( __N) * 8,\
+               (__v2di)( __W),\
+               (__mmask16)( __U));\
+})
+
+#define _mm_maskz_alignr_epi8( __U, __A, __B, __N) __extension__ ({ \
+__builtin_ia32_palignr128_mask ((__v2di)( __A),\
+               (__v2di)( __B),\
+               ( __N) * 8,\
+               (__v2di)\
+               _mm_setzero_si128 (),\
+               (__mmask16)( __U));\
+})
+
+#define _mm256_mask_alignr_epi8( __W, __U, __A, __B, __N) __extension__ ({ \
+__builtin_ia32_palignr256_mask ((__v4di)( __A),\
+               (__v4di)( __B),\
+               ( __N) * 8,\
+               (__v4di)( __W),\
+               (__mmask32)( __U));\
+})
 
+#define _mm256_maskz_alignr_epi8( __U, __A, __B, __N) __extension__ ({ \
+__builtin_ia32_palignr256_mask ((__v4di)( __A),\
+               (__v4di)( __B),\
+               ( __N) * 8,\
+               (__v4di)\
+               _mm256_setzero_si256 (),\
+               (__mmask32)( __U));\
+})
 
 #undef __DEFAULT_FN_ATTRS
 
 #endif /* __AVX512VLBWINTRIN_H */

lib/Headers/avx512vlintrin.h

 _mm256_maskz_rsqrt14_ps (__mmask8 __U, __m256 __A)
 {
   return (__m256) __builtin_ia32_rsqrt14ps256_mask ((__v8sf) __A,
                 (__v8sf)
                 _mm256_setzero_ps (),
                 (__mmask8) __U);
 }
 
+#define _mm_alignr_epi32( __A, __B, __imm) __extension__ ({ \
+__builtin_ia32_alignd128_mask ((__v4si)( __A),\
+              (__v4si)( __B),( __imm),\
+              (__v4si) _mm_setzero_si128 (),\
+              (__mmask8) -1);\
+})
+
+#define _mm_mask_alignr_epi32( __W, __U, __A, __B, __imm) __extension__ ({ \
+__builtin_ia32_alignd128_mask ((__v4si)( __A),\
+              (__v4si)( __B),( __imm),\
+              (__v4si)( __W),\
+              (__mmask8)( __U));\
+})
+
+#define _mm_maskz_alignr_epi32( __U, __A, __B, __imm) __extension__ ({ \
+__builtin_ia32_alignd128_mask ((__v4si)( __A),\
+              (__v4si)( __B),( __imm),\
+              (__v4si) _mm_setzero_si128 (),\
+              (__mmask8)( __U));\
+})
+
+#define _mm256_alignr_epi32( __A, __B, __imm) __extension__ ({ \
+__builtin_ia32_alignd256_mask ((__v8si)( __A),\
+              (__v8si)( __B),( __imm),\
+              (__v8si) _mm256_setzero_si256 (),\
+              (__mmask8) -1);\
+})
+
+#define _mm256_mask_alignr_epi32( __W, __U, __A, __B, __imm) __extension__ ({ \
+__builtin_ia32_alignd256_mask ((__v8si)( __A),\
+              (__v8si)( __B),( __imm),\
+              (__v8si)( __W),\
+              (__mmask8)( __U));\
+})
+
+#define _mm256_maskz_alignr_epi32( __U, __A, __B, __imm) __extension__ ({ \
+__builtin_ia32_alignd256_mask ((__v8si)( __A),\
+              (__v8si)( __B),( __imm),\
+              (__v8si) _mm256_setzero_si256 (),\
+              (__mmask8)( __U));\
+})
+
+#define _mm_alignr_epi64( __A, __B, __imm) __extension__ ({ \
+__builtin_ia32_alignq128_mask ((__v2di)( __A),\
+              (__v2di)( __B),( __imm),\
+              (__v2di) _mm_setzero_di (),\
+              (__mmask8) -1);\
+})
+
+#define _mm_mask_alignr_epi64( __W, __U, __A, __B, __imm) __extension__ ({ \
+__builtin_ia32_alignq128_mask ((__v2di)( __A),\
+              (__v2di)( __B),( __imm),\
+              (__v2di)( __W),\
+              (__mmask8)( __U));\
+})
+
+#define _mm_maskz_alignr_epi64( __U, __A, __B, __imm) __extension__ ({ \
+__builtin_ia32_alignq128_mask ((__v2di)( __A),\
+              (__v2di)( __B),( __imm),\
+              (__v2di) _mm_setzero_di (),\
+              (__mmask8)( __U));\
+})
+
+#define _mm256_alignr_epi64( __A, __B, __imm) __extension__ ({ \
+__builtin_ia32_alignq256_mask ((__v4di)( __A),\
+              (__v4di)( __B),( __imm),\
+              (__v4di) _mm256_setzero_si256 (),\
+              (__mmask8) -1);\
+})
+
+#define _mm256_mask_alignr_epi64( __W, __U, __A, __B, __imm) __extension__ ({ \
+__builtin_ia32_alignq256_mask ((__v4di)( __A),\
+              (__v4di)( __B),( __imm),\
+              (__v4di)( __W),\
+              (__mmask8)( __U));\
+})
+
+#define _mm256_maskz_alignr_epi64( __U, __A, __B, __imm) __extension__ ({ \
+__builtin_ia32_alignq256_mask ((__v4di)( __A),\
+              (__v4di)( __B),( __imm),\
+              (__v4di) _mm256_setzero_si256 (),\
+              (__mmask8)( __U));\
+})
+
 #undef __DEFAULT_FN_ATTRS
 #undef __DEFAULT_FN_ATTRS_BOTH
 
 #endif /* __AVX512VLINTRIN_H */

test/CodeGen/avx512bw-builtins.c

 }
 
 __mmask32 test_mm512_mask_testn_epi16_mask(__mmask32 __U, __m512i __A, __m512i __B) {
   // CHECK-LABEL: @test_mm512_mask_testn_epi16_mask
   // CHECK: @llvm.x86.avx512.ptestnm.w.
   return _mm512_mask_testn_epi16_mask(__U, __A, __B); 
 }
 
+__m512i test_mm512_alignr_epi8(__m512i __A,__m512i __B){
+    // CHECK-LABEL: @test_mm512_alignr_epi8
+    // CHECK: @llvm.x86.avx512.mask.palignr.512
+    _mm512_alignr_epi8(__A, __B, 2); 
+}
+
+__m512i test_mm512_mask_alignr_epi8(__m512i __W, __mmask64 __U, __m512i __A,__m512i __B){
+    // CHECK-LABEL: @test_mm512_mask_alignr_epi8
+    // CHECK: @llvm.x86.avx512.mask.palignr.512
+    _mm512_mask_alignr_epi8(__W, __U, __A, __B, 2); 
+}
+
+__m512i test_mm512_maskz_alignr_epi8(__mmask64 __U, __m512i __A,__m512i __B){
+    // CHECK-LABEL: @test_mm512_maskz_alignr_epi8
+    // CHECK: @llvm.x86.avx512.mask.palignr.512
+   _mm512_maskz_alignr_epi8(__U, __A, __B, 2); 
+}
+
+
+

test/CodeGen/avx512f-builtins.c

 
 __m512i test_mm512_alignr_epi32(__m512i a, __m512i b)
 {
   // CHECK-LABEL: @test_mm512_alignr_epi32
   // CHECK: @llvm.x86.avx512.mask.valign.d.512
   return _mm512_alignr_epi32(a, b, 2);
 }
 
+__m512i test_mm512_mask_alignr_epi32(__m512i w, __mmask16 u, __m512i a, __m512i b)
+{
+  // CHECK-LABEL: @test_mm512_mask_alignr_epi32
+  // CHECK: @llvm.x86.avx512.mask.valign.d.512
+  return _mm512_mask_alignr_epi32(w, u, a, b, 2);
+}
+
+__m512i test_mm512_maskz_alignr_epi32( __mmask16 u, __m512i a, __m512i b)
+{
+  // CHECK-LABEL: @test_mm512_alignr_epi32
+  // CHECK: @llvm.x86.avx512.mask.valign.d.512
+  return _mm512_maskz_alignr_epi32(u, a, b, 2);
+}
+
 __m512i test_mm512_alignr_epi64(__m512i a, __m512i b)
 {
   // CHECK-LABEL: @test_mm512_alignr_epi64
   // CHECK: @llvm.x86.avx512.mask.valign.q.512
   return _mm512_alignr_epi64(a, b, 2);
 }
 
+__m512i test_mm512_mask_alignr_epi64(__m512i w, __mmask8 u, __m512i a, __m512i b)
+{
+  // CHECK-LABEL: @test_mm512_mask_alignr_epi64
+  // CHECK: @llvm.x86.avx512.mask.valign.q.512
+  return _mm512_mask_alignr_epi64(w, u, a, b, 2);
+}
+
+__m512i test_mm512_maskz_alignr_epi64( __mmask8 u, __m512i a, __m512i b)
+{
+  // CHECK-LABEL: @test_mm512_maskz_alignr_epi64
+  // CHECK: @llvm.x86.avx512.mask.valign.q.512
+  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

test/CodeGen/avx512vl-builtins.c

__m256 test_mm256_mask_rsqrt14_ps(__m256 __W, __mmask8 __U, __m256 __A) {
   return _mm256_mask_rsqrt14_ps(__W, __U, __A); 
 }
 
 __m256 test_mm256_maskz_rsqrt14_ps(__mmask8 __U, __m256 __A) {
   // CHECK-LABEL: @test_mm256_maskz_rsqrt14_ps
   // CHECK: @llvm.x86.avx512.rsqrt14.ps.256
   return _mm256_maskz_rsqrt14_ps(__U, __A); 
 }
+
+__m128i test_mm_alignr_epi32(__m128i __A, __m128i __B) {
+  // CHECK-LABEL: @test_mm_alignr_epi32
+  // CHECK: @llvm.x86.avx512.mask.valign.d.128
+  return _mm_alignr_epi32(__A, __B, 1); 
+}
+
+__m128i test_mm_mask_alignr_epi32(__m128i __W, __mmask8 __U, __m128i __A, __m128i __B) {
+  // CHECK-LABEL: @test_mm_mask_alignr_epi32
+  // CHECK: @llvm.x86.avx512.mask.valign.d.128
+  return _mm_mask_alignr_epi32(__W, __U, __A, __B, 1); 
+}
+
+__m128i test_mm_maskz_alignr_epi32(__mmask8 __U, __m128i __A, __m128i __B) {
+  // CHECK-LABEL: @test_mm_maskz_alignr_epi32
+  // CHECK: @llvm.x86.avx512.mask.valign.d.128
+  return _mm_maskz_alignr_epi32(__U, __A, __B, 1); 
+}
+
+__m256i test_mm256_alignr_epi32(__m256i __A, __m256i __B) {
+  // CHECK-LABEL: @test_mm256_alignr_epi32
+  // CHECK: @llvm.x86.avx512.mask.valign.d.256
+  return _mm256_alignr_epi32(__A, __B, 1); 
+}
+
+__m256i test_mm256_mask_alignr_epi32(__m256i __W, __mmask8 __U, __m256i __A, __m256i __B) {
+  // CHECK-LABEL: @test_mm256_mask_alignr_epi32
+  // CHECK: @llvm.x86.avx512.mask.valign.d.256
+  return _mm256_mask_alignr_epi32(__W, __U, __A, __B, 1); 
+}
+
+__m256i test_mm256_maskz_alignr_epi32(__mmask8 __U, __m256i __A, __m256i __B) {
+  // CHECK-LABEL: @test_mm256_maskz_alignr_epi32
+  // CHECK: @llvm.x86.avx512.mask.valign.d.256
+  return _mm256_maskz_alignr_epi32(__U, __A, __B, 1); 
+}
+
+__m128i test_mm_alignr_epi64(__m128i __A, __m128i __B) {
+  // CHECK-LABEL: @test_mm_alignr_epi64
+  // CHECK: @llvm.x86.avx512.mask.valign.q.128
+  return _mm_alignr_epi64(__A, __B, 1); 
+}
+
+__m128i test_mm_mask_alignr_epi64(__m128i __W, __mmask8 __U, __m128i __A, __m128i __B) {
+  // CHECK-LABEL: @test_mm_mask_alignr_epi64
+  // CHECK: @llvm.x86.avx512.mask.valign.q.128
+  return _mm_mask_alignr_epi64(__W, __U, __A, __B, 1); 
+}
+
+__m128i test_mm_maskz_alignr_epi64(__mmask8 __U, __m128i __A, __m128i __B) {
+  // CHECK-LABEL: @test_mm_maskz_alignr_epi64
+  // CHECK: @llvm.x86.avx512.mask.valign.q.128
+  return _mm_maskz_alignr_epi64(__U, __A, __B, 1); 
+}
+
+__m256i test_mm256_alignr_epi64(__m256i __A, __m256i __B) {
+  // CHECK-LABEL: @test_mm256_alignr_epi64
+  // CHECK: @llvm.x86.avx512.mask.valign.q.256
+  return _mm256_alignr_epi64(__A, __B, 1); 
+}
+
+__m256i test_mm256_mask_alignr_epi64(__m256i __W, __mmask8 __U, __m256i __A, __m256i __B) {
+  // CHECK-LABEL: @test_mm256_mask_alignr_epi64
+  // CHECK: @llvm.x86.avx512.mask.valign.q.256
+  return _mm256_mask_alignr_epi64(__W, __U, __A, __B, 1); 
+}
+
+__m256i test_mm256_maskz_alignr_epi64(__mmask8 __U, __m256i __A, __m256i __B) {
+  // CHECK-LABEL: @test_mm256_maskz_alignr_epi64
+  // CHECK: @llvm.x86.avx512.mask.valign.q.256
+  return _mm256_maskz_alignr_epi64(__U, __A, __B, 1); 
+}

test/CodeGen/avx512vlbw-builtins.c

 }
 
 __mmask16 test_mm256_mask_testn_epi16_mask(__mmask16 __U, __m256i __A, __m256i __B) {
   // CHECK-LABEL: @test_mm256_mask_testn_epi16_mask
   // CHECK: @llvm.x86.avx512.ptestnm.w.256
   return _mm256_mask_testn_epi16_mask(__U, __A, __B); 
 }
 
+__m128i test_mm_mask_alignr_epi8(__m128i __W, __mmask16 __U, __m128i __A, __m128i __B) {
+  // CHECK-LABEL: @test_mm_mask_alignr_epi8
+  // CHECK: @llvm.x86.avx512.mask.palignr.128
+  return _mm_mask_alignr_epi8(__W, __U, __A, __B, 2); 
+}
+
+__m128i test_mm_maskz_alignr_epi8(__mmask16 __U, __m128i __A, __m128i __B) {
+  // CHECK-LABEL: @test_mm_maskz_alignr_epi8
+  // CHECK: @llvm.x86.avx512.mask.palignr.128
+  return _mm_maskz_alignr_epi8(__U, __A, __B, 2); 
+}
+
+__m256i test_mm256_mask_alignr_epi8(__m256i __W, __mmask32 __U, __m256i __A, __m256i __B) {
+  // CHECK-LABEL: @test_mm256_mask_alignr_epi8
+  // CHECK: @llvm.x86.avx512.mask.palignr.256
+  return _mm256_mask_alignr_epi8(__W, __U, __A, __B, 2); 
+}
+
+__m256i test_mm256_maskz_alignr_epi8(__mmask32 __U, __m256i __A, __m256i __B) {
+  // CHECK-LABEL: @test_mm256_maskz_alignr_epi8
+  // CHECK: @llvm.x86.avx512.mask.palignr.256
+  return _mm256_maskz_alignr_epi8(__U, __A, __B, 2); 
+}
+