[Clang][BuiltIn][AVX512]Adding intrinsics for cmp{ss|sd} instruction set

include/clang/Basic/BuiltinsX86.def

 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_cmpsd_mask, "UcV2dV2dIiUcIi","","avx512f")
+TARGET_BUILTIN(__builtin_ia32_cmpss_mask, "UcV4fV4fIiUcIi","","avx512f")
 
 #undef BUILTIN
 #undef TARGET_BUILTIN

lib/Headers/avx512fintrin.h

 
 #define _mm_maskz_sqrt_round_ss( __U, __A, __B, __R) __extension__ ({ \
 __builtin_ia32_sqrtss_round_mask ((__v4sf)( __B),\
                  (__v4sf)( __A),(__v4sf) _mm_setzero_ps(),\
                  (__mmask8) __U,\
                  __R);\
 })
 
+#define _mm_cmp_round_ss_mask( __X, __Y, __P, __R) __extension__ ({ \
+__builtin_ia32_cmpss_mask ((__v4sf)( __X),\
+                 (__v4sf)( __Y), __P,\
+                 (__mmask8) -1, __R);\
+})
+
+#define _mm_mask_cmp_round_ss_mask( __M, __X, __Y, __P, __R) __extension__ ({ \
+__builtin_ia32_cmpss_mask ((__v4sf)( __X),\
+                 (__v4sf)( __Y), __P,\
+                 (__mmask8)( __M), __R);\
+})
+
+#define _mm_cmp_ss_mask( __X, __Y, __P) __extension__ ({ \
+__builtin_ia32_cmpss_mask ((__v4sf)( __X),\
+                 (__v4sf)( __Y),( __P),\
+                 (__mmask8) -1,\
+                 _MM_FROUND_CUR_DIRECTION);\
+})
+
+#define _mm_mask_cmp_ss_mask( __M, __X, __Y, __P) __extension__ ({ \
+__builtin_ia32_cmpss_mask ((__v4sf)( __X),\
+                 (__v4sf)( __Y),( __P),\
+                 (__mmask8)( __M),\
+                 _MM_FROUND_CUR_DIRECTION);\
+})
+
+#define _mm_cmp_round_sd_mask( __X, __Y, __P,__R) __extension__ ({ \
+__builtin_ia32_cmpsd_mask ((__v2df)( __X),\
+                 (__v2df)( __Y), __P,\
+                 (__mmask8) -1, __R);\
+})
+
+#define _mm_mask_cmp_round_sd_mask( __M, __X, __Y, __P, __R) __extension__ ({ \
+__builtin_ia32_cmpsd_mask ((__v2df)( __X),\
+                 (__v2df)( __Y), __P,\
+                 (__mmask8)( __M), __R);\
+})
+
+#define _mm_cmp_sd_mask( __X, __Y, __P) __extension__ ({ \
+__builtin_ia32_cmpsd_mask ((__v2df)( __X),\
+                 (__v2df)( __Y),( __P),\
+                 (__mmask8) -1,\
+                 _MM_FROUND_CUR_DIRECTION);\
+})
+
+#define _mm_mask_cmp_sd_mask( __M, __X, __Y, __P) __extension__ ({ \
+__builtin_ia32_cmpsd_mask ((__v2df)( __X),\
+                 (__v2df)( __Y),( __P),\
+                 (__mmask8)( __M),\
+                 _MM_FROUND_CUR_DIRECTION);\
+})
+
 #undef __DEFAULT_FN_ATTRS
 
 #endif // __AVX512FINTRIN_H

test/CodeGen/avx512f-builtins.c

// CHECK: @llvm.x86.avx512.mask.sqrt.ss
     return _mm_maskz_sqrt_ss(__U,__A,__B);
 }
 
 __m128 test_mm_maskz_sqrt_round_ss(__mmask8 __U, __m128 __A, __m128 __B){
   // CHECK: @llvm.x86.avx512.mask.sqrt.ss
     return _mm_maskz_sqrt_round_ss(__U,__A,__B,_MM_FROUND_CUR_DIRECTION);
 }
 
+__mmask8 test_mm_cmp_round_ss_mask(__m128 __X, __m128 __Y) {
+  // CHECK-LABEL: @test_mm_cmp_round_ss_mask
+  // CHECK: @llvm.x86.avx512.mask.cmp
+  return _mm_cmp_round_ss_mask(__X, __Y, 5, _MM_FROUND_CUR_DIRECTION); 
+}
+
+__mmask8 test_mm_mask_cmp_round_ss_mask(__mmask8 __M, __m128 __X, __m128 __Y) {
+  // CHECK-LABEL: @test_mm_mask_cmp_round_ss_mask
+  // CHECK: @llvm.x86.avx512.mask.cmp
+  return _mm_mask_cmp_round_ss_mask(__M, __X, __Y, 5, _MM_FROUND_CUR_DIRECTION); 
+}
+
+__mmask8 test_mm_cmp_ss_mask(__m128 __X, __m128 __Y) {
+  // CHECK-LABEL: @test_mm_cmp_ss_mask
+  // CHECK: @llvm.x86.avx512.mask.cmp
+  return _mm_cmp_ss_mask(__X, __Y, 5); 
+}
+
+__mmask8 test_mm_mask_cmp_ss_mask(__mmask8 __M, __m128 __X, __m128 __Y) {
+  // CHECK-LABEL: @test_mm_mask_cmp_ss_mask
+  // CHECK: @llvm.x86.avx512.mask.cmp
+  return _mm_mask_cmp_ss_mask(__M, __X, __Y, 5); 
+}
 
+__mmask8 test_mm_cmp_round_sd_mask(__m128d __X, __m128d __Y) {
+  // CHECK-LABEL: @test_mm_cmp_round_sd_mask
+  // CHECK: @llvm.x86.avx512.mask.cmp
+  return _mm_cmp_round_sd_mask(__X, __Y, 5, _MM_FROUND_CUR_DIRECTION); 
+}
+
+__mmask8 test_mm_mask_cmp_round_sd_mask(__mmask8 __M, __m128d __X, __m128d __Y) {
+  // CHECK-LABEL: @test_mm_mask_cmp_round_sd_mask
+  // CHECK: @llvm.x86.avx512.mask.cmp
+  return _mm_mask_cmp_round_sd_mask(__M, __X, __Y, 5, _MM_FROUND_CUR_DIRECTION); 
+}
+
+__mmask8 test_mm_cmp_sd_mask(__m128d __X, __m128d __Y) {
+  // CHECK-LABEL: @test_mm_cmp_sd_mask
+  // CHECK: @llvm.x86.avx512.mask.cmp
+  return _mm_cmp_sd_mask(__X, __Y, 5); 
+}
+
+__mmask8 test_mm_mask_cmp_sd_mask(__mmask8 __M, __m128d __X, __m128d __Y) {
+  // CHECK-LABEL: @test_mm_mask_cmp_sd_mask
+  // CHECK: @llvm.x86.avx512.mask.cmp
+  return _mm_mask_cmp_sd_mask(__M, __X, __Y, 5); 
+}