@cpp-linter/clang-tidy-win32-x64 13.0.0 → 14.0.0

package/lib/clang/{13.0.0 → 14.0.0}/include/avx512vbmi2intrin.h

@@ -129,88 +129,88 @@ _mm512_maskz_expandloadu_epi8(__mmask64 __U, void const *__P)
 }
 
 #define _mm512_shldi_epi64(A, B, I) \
-  (__m512i)__builtin_ia32_vpshldq512((__v8di)(__m512i)(A), \
-                                     (__v8di)(__m512i)(B), (int)(I))
+  ((__m512i)__builtin_ia32_vpshldq512((__v8di)(__m512i)(A), \
+                                      (__v8di)(__m512i)(B), (int)(I)))
 
 #define _mm512_mask_shldi_epi64(S, U, A, B, I) \
-  (__m512i)__builtin_ia32_selectq_512((__mmask8)(U), \
-                                    (__v8di)_mm512_shldi_epi64((A), (B), (I)), \
-                                    (__v8di)(__m512i)(S))
+  ((__m512i)__builtin_ia32_selectq_512((__mmask8)(U), \
+                                     (__v8di)_mm512_shldi_epi64((A), (B), (I)), \
+                                     (__v8di)(__m512i)(S)))
 
 #define _mm512_maskz_shldi_epi64(U, A, B, I) \
-  (__m512i)__builtin_ia32_selectq_512((__mmask8)(U), \
-                                    (__v8di)_mm512_shldi_epi64((A), (B), (I)), \
-                                    (__v8di)_mm512_setzero_si512())
+  ((__m512i)__builtin_ia32_selectq_512((__mmask8)(U), \
+                                     (__v8di)_mm512_shldi_epi64((A), (B), (I)), \
+                                     (__v8di)_mm512_setzero_si512()))
 
 #define _mm512_shldi_epi32(A, B, I) \
-  (__m512i)__builtin_ia32_vpshldd512((__v16si)(__m512i)(A), \
-                                     (__v16si)(__m512i)(B), (int)(I))
+  ((__m512i)__builtin_ia32_vpshldd512((__v16si)(__m512i)(A), \
+                                      (__v16si)(__m512i)(B), (int)(I)))
 
 #define _mm512_mask_shldi_epi32(S, U, A, B, I) \
-  (__m512i)__builtin_ia32_selectd_512((__mmask16)(U), \
-                                   (__v16si)_mm512_shldi_epi32((A), (B), (I)), \
-                                   (__v16si)(__m512i)(S))
+  ((__m512i)__builtin_ia32_selectd_512((__mmask16)(U), \
+                                    (__v16si)_mm512_shldi_epi32((A), (B), (I)), \
+                                    (__v16si)(__m512i)(S)))
 
 #define _mm512_maskz_shldi_epi32(U, A, B, I) \
-  (__m512i)__builtin_ia32_selectd_512((__mmask16)(U), \
-                                   (__v16si)_mm512_shldi_epi32((A), (B), (I)), \
-                                   (__v16si)_mm512_setzero_si512())
+  ((__m512i)__builtin_ia32_selectd_512((__mmask16)(U), \
+                                    (__v16si)_mm512_shldi_epi32((A), (B), (I)), \
+                                    (__v16si)_mm512_setzero_si512()))
 
 #define _mm512_shldi_epi16(A, B, I) \
-  (__m512i)__builtin_ia32_vpshldw512((__v32hi)(__m512i)(A), \
-                                     (__v32hi)(__m512i)(B), (int)(I))
+  ((__m512i)__builtin_ia32_vpshldw512((__v32hi)(__m512i)(A), \
+                                      (__v32hi)(__m512i)(B), (int)(I)))
 
 #define _mm512_mask_shldi_epi16(S, U, A, B, I) \
-  (__m512i)__builtin_ia32_selectw_512((__mmask32)(U), \
-                                   (__v32hi)_mm512_shldi_epi16((A), (B), (I)), \
-                                   (__v32hi)(__m512i)(S))
+  ((__m512i)__builtin_ia32_selectw_512((__mmask32)(U), \
+                                    (__v32hi)_mm512_shldi_epi16((A), (B), (I)), \
+                                    (__v32hi)(__m512i)(S)))
 
 #define _mm512_maskz_shldi_epi16(U, A, B, I) \
-  (__m512i)__builtin_ia32_selectw_512((__mmask32)(U), \
-                                   (__v32hi)_mm512_shldi_epi16((A), (B), (I)), \
-                                   (__v32hi)_mm512_setzero_si512())
+  ((__m512i)__builtin_ia32_selectw_512((__mmask32)(U), \
+                                    (__v32hi)_mm512_shldi_epi16((A), (B), (I)), \
+                                    (__v32hi)_mm512_setzero_si512()))
 
 #define _mm512_shrdi_epi64(A, B, I) \
-  (__m512i)__builtin_ia32_vpshrdq512((__v8di)(__m512i)(A), \
-                                     (__v8di)(__m512i)(B), (int)(I))
+  ((__m512i)__builtin_ia32_vpshrdq512((__v8di)(__m512i)(A), \
+                                      (__v8di)(__m512i)(B), (int)(I)))
 
 #define _mm512_mask_shrdi_epi64(S, U, A, B, I) \
-  (__m512i)__builtin_ia32_selectq_512((__mmask8)(U), \
-                                    (__v8di)_mm512_shrdi_epi64((A), (B), (I)), \
-                                    (__v8di)(__m512i)(S))
+  ((__m512i)__builtin_ia32_selectq_512((__mmask8)(U), \
+                                     (__v8di)_mm512_shrdi_epi64((A), (B), (I)), \
+                                     (__v8di)(__m512i)(S)))
 
 #define _mm512_maskz_shrdi_epi64(U, A, B, I) \
-  (__m512i)__builtin_ia32_selectq_512((__mmask8)(U), \
-                                    (__v8di)_mm512_shrdi_epi64((A), (B), (I)), \
-                                    (__v8di)_mm512_setzero_si512())
+  ((__m512i)__builtin_ia32_selectq_512((__mmask8)(U), \
+                                     (__v8di)_mm512_shrdi_epi64((A), (B), (I)), \
+                                     (__v8di)_mm512_setzero_si512()))
 
 #define _mm512_shrdi_epi32(A, B, I) \
-  (__m512i)__builtin_ia32_vpshrdd512((__v16si)(__m512i)(A), \
-                                     (__v16si)(__m512i)(B), (int)(I))
+  ((__m512i)__builtin_ia32_vpshrdd512((__v16si)(__m512i)(A), \
+                                      (__v16si)(__m512i)(B), (int)(I)))
 
 #define _mm512_mask_shrdi_epi32(S, U, A, B, I) \
-  (__m512i)__builtin_ia32_selectd_512((__mmask16)(U), \
-                                   (__v16si)_mm512_shrdi_epi32((A), (B), (I)), \
-                                   (__v16si)(__m512i)(S))
+  ((__m512i)__builtin_ia32_selectd_512((__mmask16)(U), \
+                                    (__v16si)_mm512_shrdi_epi32((A), (B), (I)), \
+                                    (__v16si)(__m512i)(S)))
 
 #define _mm512_maskz_shrdi_epi32(U, A, B, I) \
-  (__m512i)__builtin_ia32_selectd_512((__mmask16)(U), \
-                                   (__v16si)_mm512_shrdi_epi32((A), (B), (I)), \
-                                   (__v16si)_mm512_setzero_si512())
+  ((__m512i)__builtin_ia32_selectd_512((__mmask16)(U), \
+                                    (__v16si)_mm512_shrdi_epi32((A), (B), (I)), \
+                                    (__v16si)_mm512_setzero_si512()))
 
 #define _mm512_shrdi_epi16(A, B, I) \
-  (__m512i)__builtin_ia32_vpshrdw512((__v32hi)(__m512i)(A), \
-                                     (__v32hi)(__m512i)(B), (int)(I))
+  ((__m512i)__builtin_ia32_vpshrdw512((__v32hi)(__m512i)(A), \
+                                      (__v32hi)(__m512i)(B), (int)(I)))
 
 #define _mm512_mask_shrdi_epi16(S, U, A, B, I) \
-  (__m512i)__builtin_ia32_selectw_512((__mmask32)(U), \
-                                   (__v32hi)_mm512_shrdi_epi16((A), (B), (I)), \
-                                   (__v32hi)(__m512i)(S))
+  ((__m512i)__builtin_ia32_selectw_512((__mmask32)(U), \
+                                    (__v32hi)_mm512_shrdi_epi16((A), (B), (I)), \
+                                    (__v32hi)(__m512i)(S)))
 
 #define _mm512_maskz_shrdi_epi16(U, A, B, I) \
-  (__m512i)__builtin_ia32_selectw_512((__mmask32)(U), \
-                                   (__v32hi)_mm512_shrdi_epi16((A), (B), (I)), \
-                                   (__v32hi)_mm512_setzero_si512())
+  ((__m512i)__builtin_ia32_selectw_512((__mmask32)(U), \
+                                    (__v32hi)_mm512_shrdi_epi16((A), (B), (I)), \
+                                    (__v32hi)_mm512_setzero_si512()))
 
 static __inline__ __m512i __DEFAULT_FN_ATTRS
 _mm512_shldv_epi64(__m512i __A, __m512i __B, __m512i __C)

package/lib/clang/{13.0.0 → 14.0.0}/include/avx512vlbf16intrin.h

@@ -420,18 +420,46 @@ static __inline__ __bfloat16 __DEFAULT_FN_ATTRS128 _mm_cvtness_sbh(float __A) {
   return __R[0];
 }
 
+/// Convert Packed BF16 Data to Packed float Data.
+///
+/// \headerfile <x86intrin.h>
+///
+/// \param __A
+///    A 128-bit vector of [4 x bfloat].
+/// \returns A 128-bit vector of [4 x float] come from conversion of __A
+static __inline__ __m128 __DEFAULT_FN_ATTRS128 _mm_cvtpbh_ps(__m128bh __A) {
+  return _mm_castsi128_ps(
+      (__m128i)_mm_slli_epi32((__m128i)_mm_cvtepi16_epi32((__m128i)__A), 16));
+}
+
 /// Convert Packed BF16 Data to Packed float Data.
 ///
 /// \headerfile <x86intrin.h>
 ///
 /// \param __A
 ///    A 128-bit vector of [8 x bfloat].
-/// \returns A 256-bit vector of [8 x float] come from convertion of __A
+/// \returns A 256-bit vector of [8 x float] come from conversion of __A
 static __inline__ __m256 __DEFAULT_FN_ATTRS256 _mm256_cvtpbh_ps(__m128bh __A) {
   return _mm256_castsi256_ps((__m256i)_mm256_slli_epi32(
       (__m256i)_mm256_cvtepi16_epi32((__m128i)__A), 16));
 }
 
+/// Convert Packed BF16 Data to Packed float Data using zeroing mask.
+///
+/// \headerfile <x86intrin.h>
+///
+/// \param __U
+///    A 4-bit mask. Elements are zeroed out when the corresponding mask
+///    bit is not set.
+/// \param __A
+///    A 128-bit vector of [4 x bfloat].
+/// \returns A 128-bit vector of [4 x float] come from conversion of __A
+static __inline__ __m128 __DEFAULT_FN_ATTRS128
+_mm_maskz_cvtpbh_ps(__mmask8 __U, __m128bh __A) {
+  return _mm_castsi128_ps((__m128i)_mm_slli_epi32(
+      (__m128i)_mm_maskz_cvtepi16_epi32((__mmask8)__U, (__m128i)__A), 16));
+}
+
 /// Convert Packed BF16 Data to Packed float Data using zeroing mask.
 ///
 /// \headerfile <x86intrin.h>
@@ -441,13 +469,33 @@ static __inline__ __m256 __DEFAULT_FN_ATTRS256 _mm256_cvtpbh_ps(__m128bh __A) {
 ///    bit is not set.
 /// \param __A
 ///    A 128-bit vector of [8 x bfloat].
-/// \returns A 256-bit vector of [8 x float] come from convertion of __A
+/// \returns A 256-bit vector of [8 x float] come from conversion of __A
 static __inline__ __m256 __DEFAULT_FN_ATTRS256
 _mm256_maskz_cvtpbh_ps(__mmask8 __U, __m128bh __A) {
   return _mm256_castsi256_ps((__m256i)_mm256_slli_epi32(
       (__m256i)_mm256_maskz_cvtepi16_epi32((__mmask8)__U, (__m128i)__A), 16));
 }
 
+/// Convert Packed BF16 Data to Packed float Data using merging mask.
+///
+/// \headerfile <x86intrin.h>
+///
+/// \param __S
+///    A 128-bit vector of [4 x float]. Elements are copied from __S when
+///     the corresponding mask bit is not set.
+/// \param __U
+///    A 4-bit mask. Elements are zeroed out when the corresponding mask
+///    bit is not set.
+/// \param __A
+///    A 128-bit vector of [4 x bfloat].
+/// \returns A 128-bit vector of [4 x float] come from conversion of __A
+static __inline__ __m128 __DEFAULT_FN_ATTRS128
+_mm_mask_cvtpbh_ps(__m128 __S, __mmask8 __U, __m128bh __A) {
+  return _mm_castsi128_ps((__m128i)_mm_mask_slli_epi32(
+      (__m128i)__S, (__mmask8)__U, (__m128i)_mm_cvtepi16_epi32((__m128i)__A),
+      16));
+}
+
 /// Convert Packed BF16 Data to Packed float Data using merging mask.
 ///
 /// \headerfile <x86intrin.h>
@@ -460,7 +508,7 @@ _mm256_maskz_cvtpbh_ps(__mmask8 __U, __m128bh __A) {
 ///    bit is not set.
 /// \param __A
 ///    A 128-bit vector of [8 x bfloat].
-/// \returns A 256-bit vector of [8 x float] come from convertion of __A
+/// \returns A 256-bit vector of [8 x float] come from conversion of __A
 static __inline__ __m256 __DEFAULT_FN_ATTRS256
 _mm256_mask_cvtpbh_ps(__m256 __S, __mmask8 __U, __m128bh __A) {
   return _mm256_castsi256_ps((__m256i)_mm256_mask_slli_epi32(

package/lib/clang/{13.0.0 → 14.0.0}/include/avx512vlbwintrin.h

@@ -21,84 +21,84 @@
 /* Integer compare */
 
 #define _mm_cmp_epi8_mask(a, b, p) \
-  (__mmask16)__builtin_ia32_cmpb128_mask((__v16qi)(__m128i)(a), \
-                                         (__v16qi)(__m128i)(b), (int)(p), \
-                                         (__mmask16)-1)
+  ((__mmask16)__builtin_ia32_cmpb128_mask((__v16qi)(__m128i)(a), \
+                                          (__v16qi)(__m128i)(b), (int)(p), \
+                                          (__mmask16)-1))
 
 #define _mm_mask_cmp_epi8_mask(m, a, b, p) \
-  (__mmask16)__builtin_ia32_cmpb128_mask((__v16qi)(__m128i)(a), \
-                                         (__v16qi)(__m128i)(b), (int)(p), \
-                                         (__mmask16)(m))
+  ((__mmask16)__builtin_ia32_cmpb128_mask((__v16qi)(__m128i)(a), \
+                                          (__v16qi)(__m128i)(b), (int)(p), \
+                                          (__mmask16)(m)))
 
 #define _mm_cmp_epu8_mask(a, b, p) \
-  (__mmask16)__builtin_ia32_ucmpb128_mask((__v16qi)(__m128i)(a), \
-                                          (__v16qi)(__m128i)(b), (int)(p), \
-                                          (__mmask16)-1)
+  ((__mmask16)__builtin_ia32_ucmpb128_mask((__v16qi)(__m128i)(a), \
+                                           (__v16qi)(__m128i)(b), (int)(p), \
+                                           (__mmask16)-1))
 
 #define _mm_mask_cmp_epu8_mask(m, a, b, p) \
-  (__mmask16)__builtin_ia32_ucmpb128_mask((__v16qi)(__m128i)(a), \
-                                          (__v16qi)(__m128i)(b), (int)(p), \
-                                          (__mmask16)(m))
+  ((__mmask16)__builtin_ia32_ucmpb128_mask((__v16qi)(__m128i)(a), \
+                                           (__v16qi)(__m128i)(b), (int)(p), \
+                                           (__mmask16)(m)))
 
 #define _mm256_cmp_epi8_mask(a, b, p) \
-  (__mmask32)__builtin_ia32_cmpb256_mask((__v32qi)(__m256i)(a), \
-                                         (__v32qi)(__m256i)(b), (int)(p), \
-                                         (__mmask32)-1)
+  ((__mmask32)__builtin_ia32_cmpb256_mask((__v32qi)(__m256i)(a), \
+                                          (__v32qi)(__m256i)(b), (int)(p), \
+                                          (__mmask32)-1))
 
 #define _mm256_mask_cmp_epi8_mask(m, a, b, p) \
-  (__mmask32)__builtin_ia32_cmpb256_mask((__v32qi)(__m256i)(a), \
-                                         (__v32qi)(__m256i)(b), (int)(p), \
-                                         (__mmask32)(m))
+  ((__mmask32)__builtin_ia32_cmpb256_mask((__v32qi)(__m256i)(a), \
+                                          (__v32qi)(__m256i)(b), (int)(p), \
+                                          (__mmask32)(m)))
 
 #define _mm256_cmp_epu8_mask(a, b, p) \
-  (__mmask32)__builtin_ia32_ucmpb256_mask((__v32qi)(__m256i)(a), \
-                                          (__v32qi)(__m256i)(b), (int)(p), \
-                                          (__mmask32)-1)
+  ((__mmask32)__builtin_ia32_ucmpb256_mask((__v32qi)(__m256i)(a), \
+                                           (__v32qi)(__m256i)(b), (int)(p), \
+                                           (__mmask32)-1))
 
 #define _mm256_mask_cmp_epu8_mask(m, a, b, p) \
-  (__mmask32)__builtin_ia32_ucmpb256_mask((__v32qi)(__m256i)(a), \
-                                          (__v32qi)(__m256i)(b), (int)(p), \
-                                          (__mmask32)(m))
+  ((__mmask32)__builtin_ia32_ucmpb256_mask((__v32qi)(__m256i)(a), \
+                                           (__v32qi)(__m256i)(b), (int)(p), \
+                                           (__mmask32)(m)))
 
 #define _mm_cmp_epi16_mask(a, b, p) \
-  (__mmask8)__builtin_ia32_cmpw128_mask((__v8hi)(__m128i)(a), \
-                                        (__v8hi)(__m128i)(b), (int)(p), \
-                                        (__mmask8)-1)
+  ((__mmask8)__builtin_ia32_cmpw128_mask((__v8hi)(__m128i)(a), \
+                                         (__v8hi)(__m128i)(b), (int)(p), \
+                                         (__mmask8)-1))
 
 #define _mm_mask_cmp_epi16_mask(m, a, b, p) \
-  (__mmask8)__builtin_ia32_cmpw128_mask((__v8hi)(__m128i)(a), \
-                                        (__v8hi)(__m128i)(b), (int)(p), \
-                                        (__mmask8)(m))
+  ((__mmask8)__builtin_ia32_cmpw128_mask((__v8hi)(__m128i)(a), \
+                                         (__v8hi)(__m128i)(b), (int)(p), \
+                                         (__mmask8)(m)))
 
 #define _mm_cmp_epu16_mask(a, b, p) \
-  (__mmask8)__builtin_ia32_ucmpw128_mask((__v8hi)(__m128i)(a), \
-                                         (__v8hi)(__m128i)(b), (int)(p), \
-                                         (__mmask8)-1)
+  ((__mmask8)__builtin_ia32_ucmpw128_mask((__v8hi)(__m128i)(a), \
+                                          (__v8hi)(__m128i)(b), (int)(p), \
+                                          (__mmask8)-1))
 
 #define _mm_mask_cmp_epu16_mask(m, a, b, p) \
-  (__mmask8)__builtin_ia32_ucmpw128_mask((__v8hi)(__m128i)(a), \
-                                         (__v8hi)(__m128i)(b), (int)(p), \
-                                         (__mmask8)(m))
+  ((__mmask8)__builtin_ia32_ucmpw128_mask((__v8hi)(__m128i)(a), \
+                                          (__v8hi)(__m128i)(b), (int)(p), \
+                                          (__mmask8)(m)))
 
 #define _mm256_cmp_epi16_mask(a, b, p) \
-  (__mmask16)__builtin_ia32_cmpw256_mask((__v16hi)(__m256i)(a), \
-                                         (__v16hi)(__m256i)(b), (int)(p), \
-                                         (__mmask16)-1)
+  ((__mmask16)__builtin_ia32_cmpw256_mask((__v16hi)(__m256i)(a), \
+                                          (__v16hi)(__m256i)(b), (int)(p), \
+                                          (__mmask16)-1))
 
 #define _mm256_mask_cmp_epi16_mask(m, a, b, p) \
-  (__mmask16)__builtin_ia32_cmpw256_mask((__v16hi)(__m256i)(a), \
-                                         (__v16hi)(__m256i)(b), (int)(p), \
-                                         (__mmask16)(m))
+  ((__mmask16)__builtin_ia32_cmpw256_mask((__v16hi)(__m256i)(a), \
+                                          (__v16hi)(__m256i)(b), (int)(p), \
+                                          (__mmask16)(m)))
 
 #define _mm256_cmp_epu16_mask(a, b, p) \
-  (__mmask16)__builtin_ia32_ucmpw256_mask((__v16hi)(__m256i)(a), \
-                                          (__v16hi)(__m256i)(b), (int)(p), \
-                                          (__mmask16)-1)
+  ((__mmask16)__builtin_ia32_ucmpw256_mask((__v16hi)(__m256i)(a), \
+                                           (__v16hi)(__m256i)(b), (int)(p), \
+                                           (__mmask16)-1))
 
 #define _mm256_mask_cmp_epu16_mask(m, a, b, p) \
-  (__mmask16)__builtin_ia32_ucmpw256_mask((__v16hi)(__m256i)(a), \
-                                          (__v16hi)(__m256i)(b), (int)(p), \
-                                          (__mmask16)(m))
+  ((__mmask16)__builtin_ia32_ucmpw256_mask((__v16hi)(__m256i)(a), \
+                                           (__v16hi)(__m256i)(b), (int)(p), \
+                                           (__mmask16)(m)))
 
 #define _mm_cmpeq_epi8_mask(A, B) \
     _mm_cmp_epi8_mask((A), (B), _MM_CMPINT_EQ)
@@ -1821,46 +1821,46 @@ _mm256_maskz_cvtepu8_epi16 (__mmask16 __U, __m128i __A)
 
 
 #define _mm_mask_shufflehi_epi16(W, U, A, imm) \
-  (__m128i)__builtin_ia32_selectw_128((__mmask8)(U), \
-                                      (__v8hi)_mm_shufflehi_epi16((A), (imm)), \
-                                      (__v8hi)(__m128i)(W))
+  ((__m128i)__builtin_ia32_selectw_128((__mmask8)(U), \
+                                       (__v8hi)_mm_shufflehi_epi16((A), (imm)), \
+                                       (__v8hi)(__m128i)(W)))
 
 #define _mm_maskz_shufflehi_epi16(U, A, imm) \
-  (__m128i)__builtin_ia32_selectw_128((__mmask8)(U), \
-                                      (__v8hi)_mm_shufflehi_epi16((A), (imm)), \
-                                      (__v8hi)_mm_setzero_si128())
+  ((__m128i)__builtin_ia32_selectw_128((__mmask8)(U), \
+                                       (__v8hi)_mm_shufflehi_epi16((A), (imm)), \
+                                       (__v8hi)_mm_setzero_si128()))
 
 #define _mm256_mask_shufflehi_epi16(W, U, A, imm) \
-  (__m256i)__builtin_ia32_selectw_256((__mmask16)(U), \
-                                      (__v16hi)_mm256_shufflehi_epi16((A), (imm)), \
-                                      (__v16hi)(__m256i)(W))
+  ((__m256i)__builtin_ia32_selectw_256((__mmask16)(U), \
+                                       (__v16hi)_mm256_shufflehi_epi16((A), (imm)), \
+                                       (__v16hi)(__m256i)(W)))
 
 #define _mm256_maskz_shufflehi_epi16(U, A, imm) \
-  (__m256i)__builtin_ia32_selectw_256((__mmask16)(U), \
-                                      (__v16hi)_mm256_shufflehi_epi16((A), (imm)), \
-                                      (__v16hi)_mm256_setzero_si256())
+  ((__m256i)__builtin_ia32_selectw_256((__mmask16)(U), \
+                                       (__v16hi)_mm256_shufflehi_epi16((A), (imm)), \
+                                       (__v16hi)_mm256_setzero_si256()))
 
 #define _mm_mask_shufflelo_epi16(W, U, A, imm) \
-  (__m128i)__builtin_ia32_selectw_128((__mmask8)(U), \
-                                      (__v8hi)_mm_shufflelo_epi16((A), (imm)), \
-                                      (__v8hi)(__m128i)(W))
+  ((__m128i)__builtin_ia32_selectw_128((__mmask8)(U), \
+                                       (__v8hi)_mm_shufflelo_epi16((A), (imm)), \
+                                       (__v8hi)(__m128i)(W)))
 
 #define _mm_maskz_shufflelo_epi16(U, A, imm) \
-  (__m128i)__builtin_ia32_selectw_128((__mmask8)(U), \
-                                      (__v8hi)_mm_shufflelo_epi16((A), (imm)), \
-                                      (__v8hi)_mm_setzero_si128())
+  ((__m128i)__builtin_ia32_selectw_128((__mmask8)(U), \
+                                       (__v8hi)_mm_shufflelo_epi16((A), (imm)), \
+                                       (__v8hi)_mm_setzero_si128()))
 
 #define _mm256_mask_shufflelo_epi16(W, U, A, imm) \
-  (__m256i)__builtin_ia32_selectw_256((__mmask16)(U), \
-                                      (__v16hi)_mm256_shufflelo_epi16((A), \
-                                                                      (imm)), \
-                                      (__v16hi)(__m256i)(W))
+  ((__m256i)__builtin_ia32_selectw_256((__mmask16)(U), \
+                                       (__v16hi)_mm256_shufflelo_epi16((A), \
+                                                                       (imm)), \
+                                       (__v16hi)(__m256i)(W)))
 
 #define _mm256_maskz_shufflelo_epi16(U, A, imm) \
-  (__m256i)__builtin_ia32_selectw_256((__mmask16)(U), \
-                                      (__v16hi)_mm256_shufflelo_epi16((A), \
-                                                                      (imm)), \
-                                      (__v16hi)_mm256_setzero_si256())
+  ((__m256i)__builtin_ia32_selectw_256((__mmask16)(U), \
+                                       (__v16hi)_mm256_shufflelo_epi16((A), \
+                                                                       (imm)), \
+                                       (__v16hi)_mm256_setzero_si256()))
 
 static __inline__ __m256i __DEFAULT_FN_ATTRS256
 _mm256_sllv_epi16(__m256i __A, __m256i __B)
@@ -2756,52 +2756,52 @@ _mm256_mask_permutexvar_epi16 (__m256i __W, __mmask16 __M, __m256i __A,
 }
 
 #define _mm_mask_alignr_epi8(W, U, A, B, N) \
-  (__m128i)__builtin_ia32_selectb_128((__mmask16)(U), \
+  ((__m128i)__builtin_ia32_selectb_128((__mmask16)(U), \
                                  (__v16qi)_mm_alignr_epi8((A), (B), (int)(N)), \
-                                 (__v16qi)(__m128i)(W))
+                                 (__v16qi)(__m128i)(W)))
 
 #define _mm_maskz_alignr_epi8(U, A, B, N) \
-  (__m128i)__builtin_ia32_selectb_128((__mmask16)(U), \
+  ((__m128i)__builtin_ia32_selectb_128((__mmask16)(U), \
                                  (__v16qi)_mm_alignr_epi8((A), (B), (int)(N)), \
-                                 (__v16qi)_mm_setzero_si128())
+                                 (__v16qi)_mm_setzero_si128()))
 
 #define _mm256_mask_alignr_epi8(W, U, A, B, N) \
-  (__m256i)__builtin_ia32_selectb_256((__mmask32)(U), \
+  ((__m256i)__builtin_ia32_selectb_256((__mmask32)(U), \
                               (__v32qi)_mm256_alignr_epi8((A), (B), (int)(N)), \
-                              (__v32qi)(__m256i)(W))
+                              (__v32qi)(__m256i)(W)))
 
 #define _mm256_maskz_alignr_epi8(U, A, B, N) \
-  (__m256i)__builtin_ia32_selectb_256((__mmask32)(U), \
+  ((__m256i)__builtin_ia32_selectb_256((__mmask32)(U), \
                               (__v32qi)_mm256_alignr_epi8((A), (B), (int)(N)), \
-                              (__v32qi)_mm256_setzero_si256())
+                              (__v32qi)_mm256_setzero_si256()))
 
 #define _mm_dbsad_epu8(A, B, imm) \
-  (__m128i)__builtin_ia32_dbpsadbw128((__v16qi)(__m128i)(A), \
-                                      (__v16qi)(__m128i)(B), (int)(imm))
+  ((__m128i)__builtin_ia32_dbpsadbw128((__v16qi)(__m128i)(A), \
+                                       (__v16qi)(__m128i)(B), (int)(imm)))
 
 #define _mm_mask_dbsad_epu8(W, U, A, B, imm) \
-  (__m128i)__builtin_ia32_selectw_128((__mmask8)(U), \
+  ((__m128i)__builtin_ia32_selectw_128((__mmask8)(U), \
                                       (__v8hi)_mm_dbsad_epu8((A), (B), (imm)), \
-                                      (__v8hi)(__m128i)(W))
+                                      (__v8hi)(__m128i)(W)))
 
 #define _mm_maskz_dbsad_epu8(U, A, B, imm) \
-  (__m128i)__builtin_ia32_selectw_128((__mmask8)(U), \
+  ((__m128i)__builtin_ia32_selectw_128((__mmask8)(U), \
                                       (__v8hi)_mm_dbsad_epu8((A), (B), (imm)), \
-                                      (__v8hi)_mm_setzero_si128())
+                                      (__v8hi)_mm_setzero_si128()))
 
 #define _mm256_dbsad_epu8(A, B, imm) \
-  (__m256i)__builtin_ia32_dbpsadbw256((__v32qi)(__m256i)(A), \
-                                      (__v32qi)(__m256i)(B), (int)(imm))
+  ((__m256i)__builtin_ia32_dbpsadbw256((__v32qi)(__m256i)(A), \
+                                       (__v32qi)(__m256i)(B), (int)(imm)))
 
 #define _mm256_mask_dbsad_epu8(W, U, A, B, imm) \
-  (__m256i)__builtin_ia32_selectw_256((__mmask16)(U), \
+  ((__m256i)__builtin_ia32_selectw_256((__mmask16)(U), \
                                   (__v16hi)_mm256_dbsad_epu8((A), (B), (imm)), \
-                                  (__v16hi)(__m256i)(W))
+                                  (__v16hi)(__m256i)(W)))
 
 #define _mm256_maskz_dbsad_epu8(U, A, B, imm) \
-  (__m256i)__builtin_ia32_selectw_256((__mmask16)(U), \
+  ((__m256i)__builtin_ia32_selectw_256((__mmask16)(U), \
                                   (__v16hi)_mm256_dbsad_epu8((A), (B), (imm)), \
-                                  (__v16hi)_mm256_setzero_si256())
+                                  (__v16hi)_mm256_setzero_si256()))
 
 #undef __DEFAULT_FN_ATTRS128
 #undef __DEFAULT_FN_ATTRS256