diff options
Diffstat (limited to 'lib/Headers/avxintrin.h')
| -rw-r--r-- | lib/Headers/avxintrin.h | 715 |
1 files changed, 361 insertions, 354 deletions
diff --git a/lib/Headers/avxintrin.h b/lib/Headers/avxintrin.h index ee7f835..412d284 100644 --- a/lib/Headers/avxintrin.h +++ b/lib/Headers/avxintrin.h @@ -38,111 +38,111 @@ typedef long long __m256i __attribute__((__vector_size__(32))); /* Arithmetic */ static __inline __m256d __attribute__((__always_inline__, __nodebug__)) -_mm256_add_pd(__m256d a, __m256d b) +_mm256_add_pd(__m256d __a, __m256d __b) { - return a+b; + return __a+__b; } static __inline __m256 __attribute__((__always_inline__, __nodebug__)) -_mm256_add_ps(__m256 a, __m256 b) +_mm256_add_ps(__m256 __a, __m256 __b) { - return a+b; + return __a+__b; } static __inline __m256d __attribute__((__always_inline__, __nodebug__)) -_mm256_sub_pd(__m256d a, __m256d b) +_mm256_sub_pd(__m256d __a, __m256d __b) { - return a-b; + return __a-__b; } static __inline __m256 __attribute__((__always_inline__, __nodebug__)) -_mm256_sub_ps(__m256 a, __m256 b) +_mm256_sub_ps(__m256 __a, __m256 __b) { - return a-b; + return __a-__b; } static __inline __m256d __attribute__((__always_inline__, __nodebug__)) -_mm256_addsub_pd(__m256d a, __m256d b) +_mm256_addsub_pd(__m256d __a, __m256d __b) { - return (__m256d)__builtin_ia32_addsubpd256((__v4df)a, (__v4df)b); + return (__m256d)__builtin_ia32_addsubpd256((__v4df)__a, (__v4df)__b); } static __inline __m256 __attribute__((__always_inline__, __nodebug__)) -_mm256_addsub_ps(__m256 a, __m256 b) +_mm256_addsub_ps(__m256 __a, __m256 __b) { - return (__m256)__builtin_ia32_addsubps256((__v8sf)a, (__v8sf)b); + return (__m256)__builtin_ia32_addsubps256((__v8sf)__a, (__v8sf)__b); } static __inline __m256d __attribute__((__always_inline__, __nodebug__)) -_mm256_div_pd(__m256d a, __m256d b) +_mm256_div_pd(__m256d __a, __m256d __b) { - return a / b; + return __a / __b; } static __inline __m256 __attribute__((__always_inline__, __nodebug__)) -_mm256_div_ps(__m256 a, __m256 b) +_mm256_div_ps(__m256 __a, __m256 __b) { - return a / b; + return __a / __b; } static __inline __m256d __attribute__((__always_inline__, __nodebug__)) -_mm256_max_pd(__m256d a, __m256d b) +_mm256_max_pd(__m256d __a, __m256d __b) { - return (__m256d)__builtin_ia32_maxpd256((__v4df)a, (__v4df)b); + return (__m256d)__builtin_ia32_maxpd256((__v4df)__a, (__v4df)__b); } static __inline __m256 __attribute__((__always_inline__, __nodebug__)) -_mm256_max_ps(__m256 a, __m256 b) +_mm256_max_ps(__m256 __a, __m256 __b) { - return (__m256)__builtin_ia32_maxps256((__v8sf)a, (__v8sf)b); + return (__m256)__builtin_ia32_maxps256((__v8sf)__a, (__v8sf)__b); } static __inline __m256d __attribute__((__always_inline__, __nodebug__)) -_mm256_min_pd(__m256d a, __m256d b) +_mm256_min_pd(__m256d __a, __m256d __b) { - return (__m256d)__builtin_ia32_minpd256((__v4df)a, (__v4df)b); + return (__m256d)__builtin_ia32_minpd256((__v4df)__a, (__v4df)__b); } static __inline __m256 __attribute__((__always_inline__, __nodebug__)) -_mm256_min_ps(__m256 a, __m256 b) +_mm256_min_ps(__m256 __a, __m256 __b) { - return (__m256)__builtin_ia32_minps256((__v8sf)a, (__v8sf)b); + return (__m256)__builtin_ia32_minps256((__v8sf)__a, (__v8sf)__b); } static __inline __m256d __attribute__((__always_inline__, __nodebug__)) -_mm256_mul_pd(__m256d a, __m256d b) +_mm256_mul_pd(__m256d __a, __m256d __b) { - return a * b; + return __a * __b; } static __inline __m256 __attribute__((__always_inline__, __nodebug__)) -_mm256_mul_ps(__m256 a, __m256 b) +_mm256_mul_ps(__m256 __a, __m256 __b) { - return a * b; + return __a * __b; } static __inline __m256d __attribute__((__always_inline__, __nodebug__)) -_mm256_sqrt_pd(__m256d a) +_mm256_sqrt_pd(__m256d __a) { - return (__m256d)__builtin_ia32_sqrtpd256((__v4df)a); + return (__m256d)__builtin_ia32_sqrtpd256((__v4df)__a); } static __inline __m256 __attribute__((__always_inline__, __nodebug__)) -_mm256_sqrt_ps(__m256 a) +_mm256_sqrt_ps(__m256 __a) { - return (__m256)__builtin_ia32_sqrtps256((__v8sf)a); + return (__m256)__builtin_ia32_sqrtps256((__v8sf)__a); } static __inline __m256 __attribute__((__always_inline__, __nodebug__)) -_mm256_rsqrt_ps(__m256 a) +_mm256_rsqrt_ps(__m256 __a) { - return (__m256)__builtin_ia32_rsqrtps256((__v8sf)a); + return (__m256)__builtin_ia32_rsqrtps256((__v8sf)__a); } static __inline __m256 __attribute__((__always_inline__, __nodebug__)) -_mm256_rcp_ps(__m256 a) +_mm256_rcp_ps(__m256 __a) { - return (__m256)__builtin_ia32_rcpps256((__v8sf)a); + return (__m256)__builtin_ia32_rcpps256((__v8sf)__a); } #define _mm256_round_pd(V, M) __extension__ ({ \ @@ -160,102 +160,102 @@ _mm256_rcp_ps(__m256 a) /* Logical */ static __inline __m256d __attribute__((__always_inline__, __nodebug__)) -_mm256_and_pd(__m256d a, __m256d b) +_mm256_and_pd(__m256d __a, __m256d __b) { - return (__m256d)((__v4di)a & (__v4di)b); + return (__m256d)((__v4di)__a & (__v4di)__b); } static __inline __m256 __attribute__((__always_inline__, __nodebug__)) -_mm256_and_ps(__m256 a, __m256 b) +_mm256_and_ps(__m256 __a, __m256 __b) { - return (__m256)((__v8si)a & (__v8si)b); + return (__m256)((__v8si)__a & (__v8si)__b); } static __inline __m256d __attribute__((__always_inline__, __nodebug__)) -_mm256_andnot_pd(__m256d a, __m256d b) +_mm256_andnot_pd(__m256d __a, __m256d __b) { - return (__m256d)(~(__v4di)a & (__v4di)b); + return (__m256d)(~(__v4di)__a & (__v4di)__b); } static __inline __m256 __attribute__((__always_inline__, __nodebug__)) -_mm256_andnot_ps(__m256 a, __m256 b) +_mm256_andnot_ps(__m256 __a, __m256 __b) { - return (__m256)(~(__v8si)a & (__v8si)b); + return (__m256)(~(__v8si)__a & (__v8si)__b); } static __inline __m256d __attribute__((__always_inline__, __nodebug__)) -_mm256_or_pd(__m256d a, __m256d b) +_mm256_or_pd(__m256d __a, __m256d __b) { - return (__m256d)((__v4di)a | (__v4di)b); + return (__m256d)((__v4di)__a | (__v4di)__b); } static __inline __m256 __attribute__((__always_inline__, __nodebug__)) -_mm256_or_ps(__m256 a, __m256 b) +_mm256_or_ps(__m256 __a, __m256 __b) { - return (__m256)((__v8si)a | (__v8si)b); + return (__m256)((__v8si)__a | (__v8si)__b); } static __inline __m256d __attribute__((__always_inline__, __nodebug__)) -_mm256_xor_pd(__m256d a, __m256d b) +_mm256_xor_pd(__m256d __a, __m256d __b) { - return (__m256d)((__v4di)a ^ (__v4di)b); + return (__m256d)((__v4di)__a ^ (__v4di)__b); } static __inline __m256 __attribute__((__always_inline__, __nodebug__)) -_mm256_xor_ps(__m256 a, __m256 b) +_mm256_xor_ps(__m256 __a, __m256 __b) { - return (__m256)((__v8si)a ^ (__v8si)b); + return (__m256)((__v8si)__a ^ (__v8si)__b); } /* Horizontal arithmetic */ static __inline __m256d __attribute__((__always_inline__, __nodebug__)) -_mm256_hadd_pd(__m256d a, __m256d b) +_mm256_hadd_pd(__m256d __a, __m256d __b) { - return (__m256d)__builtin_ia32_haddpd256((__v4df)a, (__v4df)b); + return (__m256d)__builtin_ia32_haddpd256((__v4df)__a, (__v4df)__b); } static __inline __m256 __attribute__((__always_inline__, __nodebug__)) -_mm256_hadd_ps(__m256 a, __m256 b) +_mm256_hadd_ps(__m256 __a, __m256 __b) { - return (__m256)__builtin_ia32_haddps256((__v8sf)a, (__v8sf)b); + return (__m256)__builtin_ia32_haddps256((__v8sf)__a, (__v8sf)__b); } static __inline __m256d __attribute__((__always_inline__, __nodebug__)) -_mm256_hsub_pd(__m256d a, __m256d b) +_mm256_hsub_pd(__m256d __a, __m256d __b) { - return (__m256d)__builtin_ia32_hsubpd256((__v4df)a, (__v4df)b); + return (__m256d)__builtin_ia32_hsubpd256((__v4df)__a, (__v4df)__b); } static __inline __m256 __attribute__((__always_inline__, __nodebug__)) -_mm256_hsub_ps(__m256 a, __m256 b) +_mm256_hsub_ps(__m256 __a, __m256 __b) { - return (__m256)__builtin_ia32_hsubps256((__v8sf)a, (__v8sf)b); + return (__m256)__builtin_ia32_hsubps256((__v8sf)__a, (__v8sf)__b); } /* Vector permutations */ static __inline __m128d __attribute__((__always_inline__, __nodebug__)) -_mm_permutevar_pd(__m128d a, __m128i c) +_mm_permutevar_pd(__m128d __a, __m128i __c) { - return (__m128d)__builtin_ia32_vpermilvarpd((__v2df)a, (__v2di)c); + return (__m128d)__builtin_ia32_vpermilvarpd((__v2df)__a, (__v2di)__c); } static __inline __m256d __attribute__((__always_inline__, __nodebug__)) -_mm256_permutevar_pd(__m256d a, __m256i c) +_mm256_permutevar_pd(__m256d __a, __m256i __c) { - return (__m256d)__builtin_ia32_vpermilvarpd256((__v4df)a, (__v4di)c); + return (__m256d)__builtin_ia32_vpermilvarpd256((__v4df)__a, (__v4di)__c); } static __inline __m128 __attribute__((__always_inline__, __nodebug__)) -_mm_permutevar_ps(__m128 a, __m128i c) +_mm_permutevar_ps(__m128 __a, __m128i __c) { - return (__m128)__builtin_ia32_vpermilvarps((__v4sf)a, (__v4si)c); + return (__m128)__builtin_ia32_vpermilvarps((__v4sf)__a, (__v4si)__c); } static __inline __m256 __attribute__((__always_inline__, __nodebug__)) -_mm256_permutevar_ps(__m256 a, __m256i c) +_mm256_permutevar_ps(__m256 __a, __m256i __c) { - return (__m256)__builtin_ia32_vpermilvarps256((__v8sf)a, - (__v8si)c); + return (__m256)__builtin_ia32_vpermilvarps256((__v8sf)__a, + (__v8si)__c); } #define _mm_permute_pd(A, C) __extension__ ({ \ @@ -313,15 +313,17 @@ _mm256_permutevar_ps(__m256 a, __m256i c) (__m256)__builtin_ia32_blendps256((__v8sf)__V1, (__v8sf)__V2, (M)); }) static __inline __m256d __attribute__((__always_inline__, __nodebug__)) -_mm256_blendv_pd(__m256d a, __m256d b, __m256d c) +_mm256_blendv_pd(__m256d __a, __m256d __b, __m256d __c) { - return (__m256d)__builtin_ia32_blendvpd256((__v4df)a, (__v4df)b, (__v4df)c); + return (__m256d)__builtin_ia32_blendvpd256( + (__v4df)__a, (__v4df)__b, (__v4df)__c); } static __inline __m256 __attribute__((__always_inline__, __nodebug__)) -_mm256_blendv_ps(__m256 a, __m256 b, __m256 c) +_mm256_blendv_ps(__m256 __a, __m256 __b, __m256 __c) { - return (__m256)__builtin_ia32_blendvps256((__v8sf)a, (__v8sf)b, (__v8sf)c); + return (__m256)__builtin_ia32_blendvps256( + (__v8sf)__a, (__v8sf)__b, (__v8sf)__c); } /* Vector Dot Product */ @@ -427,32 +429,32 @@ _mm256_blendv_ps(__m256 a, __m256 b, __m256 c) (__m128i)__builtin_ia32_vextractf128_si256((__v8si)__A, (O)); }) static __inline int __attribute__((__always_inline__, __nodebug__)) -_mm256_extract_epi32(__m256i a, int const imm) +_mm256_extract_epi32(__m256i __a, int const __imm) { - __v8si b = (__v8si)a; - return b[imm]; + __v8si __b = (__v8si)__a; + return __b[__imm]; } static __inline int __attribute__((__always_inline__, __nodebug__)) -_mm256_extract_epi16(__m256i a, int const imm) +_mm256_extract_epi16(__m256i __a, int const __imm) { - __v16hi b = (__v16hi)a; - return b[imm]; + __v16hi __b = (__v16hi)__a; + return __b[__imm]; } static __inline int __attribute__((__always_inline__, __nodebug__)) -_mm256_extract_epi8(__m256i a, int const imm) +_mm256_extract_epi8(__m256i __a, int const __imm) { - __v32qi b = (__v32qi)a; - return b[imm]; + __v32qi __b = (__v32qi)__a; + return __b[__imm]; } #ifdef __x86_64__ static __inline long long __attribute__((__always_inline__, __nodebug__)) -_mm256_extract_epi64(__m256i a, const int imm) +_mm256_extract_epi64(__m256i __a, const int __imm) { - __v4di b = (__v4di)a; - return b[imm]; + __v4di __b = (__v4di)__a; + return __b[__imm]; } #endif @@ -473,237 +475,237 @@ _mm256_extract_epi64(__m256i a, const int imm) (__m256i)__builtin_ia32_vinsertf128_si256((__v8si)__V1, (__v4si)__V2, (O)); }) static __inline __m256i __attribute__((__always_inline__, __nodebug__)) -_mm256_insert_epi32(__m256i a, int b, int const imm) +_mm256_insert_epi32(__m256i __a, int __b, int const __imm) { - __v8si c = (__v8si)a; - c[imm & 7] = b; - return (__m256i)c; + __v8si __c = (__v8si)__a; + __c[__imm & 7] = __b; + return (__m256i)__c; } static __inline __m256i __attribute__((__always_inline__, __nodebug__)) -_mm256_insert_epi16(__m256i a, int b, int const imm) +_mm256_insert_epi16(__m256i __a, int __b, int const __imm) { - __v16hi c = (__v16hi)a; - c[imm & 15] = b; - return (__m256i)c; + __v16hi __c = (__v16hi)__a; + __c[__imm & 15] = __b; + return (__m256i)__c; } static __inline __m256i __attribute__((__always_inline__, __nodebug__)) -_mm256_insert_epi8(__m256i a, int b, int const imm) +_mm256_insert_epi8(__m256i __a, int __b, int const __imm) { - __v32qi c = (__v32qi)a; - c[imm & 31] = b; - return (__m256i)c; + __v32qi __c = (__v32qi)__a; + __c[__imm & 31] = __b; + return (__m256i)__c; } #ifdef __x86_64__ static __inline __m256i __attribute__((__always_inline__, __nodebug__)) -_mm256_insert_epi64(__m256i a, int b, int const imm) +_mm256_insert_epi64(__m256i __a, int __b, int const __imm) { - __v4di c = (__v4di)a; - c[imm & 3] = b; - return (__m256i)c; + __v4di __c = (__v4di)__a; + __c[__imm & 3] = __b; + return (__m256i)__c; } #endif /* Conversion */ static __inline __m256d __attribute__((__always_inline__, __nodebug__)) -_mm256_cvtepi32_pd(__m128i a) +_mm256_cvtepi32_pd(__m128i __a) { - return (__m256d)__builtin_ia32_cvtdq2pd256((__v4si) a); + return (__m256d)__builtin_ia32_cvtdq2pd256((__v4si) __a); } static __inline __m256 __attribute__((__always_inline__, __nodebug__)) -_mm256_cvtepi32_ps(__m256i a) +_mm256_cvtepi32_ps(__m256i __a) { - return (__m256)__builtin_ia32_cvtdq2ps256((__v8si) a); + return (__m256)__builtin_ia32_cvtdq2ps256((__v8si) __a); } static __inline __m128 __attribute__((__always_inline__, __nodebug__)) -_mm256_cvtpd_ps(__m256d a) +_mm256_cvtpd_ps(__m256d __a) { - return (__m128)__builtin_ia32_cvtpd2ps256((__v4df) a); + return (__m128)__builtin_ia32_cvtpd2ps256((__v4df) __a); } static __inline __m256i __attribute__((__always_inline__, __nodebug__)) -_mm256_cvtps_epi32(__m256 a) +_mm256_cvtps_epi32(__m256 __a) { - return (__m256i)__builtin_ia32_cvtps2dq256((__v8sf) a); + return (__m256i)__builtin_ia32_cvtps2dq256((__v8sf) __a); } static __inline __m256d __attribute__((__always_inline__, __nodebug__)) -_mm256_cvtps_pd(__m128 a) +_mm256_cvtps_pd(__m128 __a) { - return (__m256d)__builtin_ia32_cvtps2pd256((__v4sf) a); + return (__m256d)__builtin_ia32_cvtps2pd256((__v4sf) __a); } static __inline __m128i __attribute__((__always_inline__, __nodebug__)) -_mm256_cvttpd_epi32(__m256d a) +_mm256_cvttpd_epi32(__m256d __a) { - return (__m128i)__builtin_ia32_cvttpd2dq256((__v4df) a); + return (__m128i)__builtin_ia32_cvttpd2dq256((__v4df) __a); } static __inline __m128i __attribute__((__always_inline__, __nodebug__)) -_mm256_cvtpd_epi32(__m256d a) +_mm256_cvtpd_epi32(__m256d __a) { - return (__m128i)__builtin_ia32_cvtpd2dq256((__v4df) a); + return (__m128i)__builtin_ia32_cvtpd2dq256((__v4df) __a); } static __inline __m256i __attribute__((__always_inline__, __nodebug__)) -_mm256_cvttps_epi32(__m256 a) +_mm256_cvttps_epi32(__m256 __a) { - return (__m256i)__builtin_ia32_cvttps2dq256((__v8sf) a); + return (__m256i)__builtin_ia32_cvttps2dq256((__v8sf) __a); } /* Vector replicate */ static __inline __m256 __attribute__((__always_inline__, __nodebug__)) -_mm256_movehdup_ps(__m256 a) +_mm256_movehdup_ps(__m256 __a) { - return __builtin_shufflevector(a, a, 1, 1, 3, 3, 5, 5, 7, 7); + return __builtin_shufflevector(__a, __a, 1, 1, 3, 3, 5, 5, 7, 7); } static __inline __m256 __attribute__((__always_inline__, __nodebug__)) -_mm256_moveldup_ps(__m256 a) +_mm256_moveldup_ps(__m256 __a) { - return __builtin_shufflevector(a, a, 0, 0, 2, 2, 4, 4, 6, 6); + return __builtin_shufflevector(__a, __a, 0, 0, 2, 2, 4, 4, 6, 6); } static __inline __m256d __attribute__((__always_inline__, __nodebug__)) -_mm256_movedup_pd(__m256d a) +_mm256_movedup_pd(__m256d __a) { - return __builtin_shufflevector(a, a, 0, 0, 2, 2); + return __builtin_shufflevector(__a, __a, 0, 0, 2, 2); } /* Unpack and Interleave */ static __inline __m256d __attribute__((__always_inline__, __nodebug__)) -_mm256_unpackhi_pd(__m256d a, __m256d b) +_mm256_unpackhi_pd(__m256d __a, __m256d __b) { - return __builtin_shufflevector(a, b, 1, 5, 1+2, 5+2); + return __builtin_shufflevector(__a, __b, 1, 5, 1+2, 5+2); } static __inline __m256d __attribute__((__always_inline__, __nodebug__)) -_mm256_unpacklo_pd(__m256d a, __m256d b) +_mm256_unpacklo_pd(__m256d __a, __m256d __b) { - return __builtin_shufflevector(a, b, 0, 4, 0+2, 4+2); + return __builtin_shufflevector(__a, __b, 0, 4, 0+2, 4+2); } static __inline __m256 __attribute__((__always_inline__, __nodebug__)) -_mm256_unpackhi_ps(__m256 a, __m256 b) +_mm256_unpackhi_ps(__m256 __a, __m256 __b) { - return __builtin_shufflevector(a, b, 2, 10, 2+1, 10+1, 6, 14, 6+1, 14+1); + return __builtin_shufflevector(__a, __b, 2, 10, 2+1, 10+1, 6, 14, 6+1, 14+1); } static __inline __m256 __attribute__((__always_inline__, __nodebug__)) -_mm256_unpacklo_ps(__m256 a, __m256 b) +_mm256_unpacklo_ps(__m256 __a, __m256 __b) { - return __builtin_shufflevector(a, b, 0, 8, 0+1, 8+1, 4, 12, 4+1, 12+1); + return __builtin_shufflevector(__a, __b, 0, 8, 0+1, 8+1, 4, 12, 4+1, 12+1); } /* Bit Test */ static __inline int __attribute__((__always_inline__, __nodebug__)) -_mm_testz_pd(__m128d a, __m128d b) +_mm_testz_pd(__m128d __a, __m128d __b) { - return __builtin_ia32_vtestzpd((__v2df)a, (__v2df)b); + return __builtin_ia32_vtestzpd((__v2df)__a, (__v2df)__b); } static __inline int __attribute__((__always_inline__, __nodebug__)) -_mm_testc_pd(__m128d a, __m128d b) +_mm_testc_pd(__m128d __a, __m128d __b) { - return __builtin_ia32_vtestcpd((__v2df)a, (__v2df)b); + return __builtin_ia32_vtestcpd((__v2df)__a, (__v2df)__b); } static __inline int __attribute__((__always_inline__, __nodebug__)) -_mm_testnzc_pd(__m128d a, __m128d b) +_mm_testnzc_pd(__m128d __a, __m128d __b) { - return __builtin_ia32_vtestnzcpd((__v2df)a, (__v2df)b); + return __builtin_ia32_vtestnzcpd((__v2df)__a, (__v2df)__b); } static __inline int __attribute__((__always_inline__, __nodebug__)) -_mm_testz_ps(__m128 a, __m128 b) +_mm_testz_ps(__m128 __a, __m128 __b) { - return __builtin_ia32_vtestzps((__v4sf)a, (__v4sf)b); + return __builtin_ia32_vtestzps((__v4sf)__a, (__v4sf)__b); } static __inline int __attribute__((__always_inline__, __nodebug__)) -_mm_testc_ps(__m128 a, __m128 b) +_mm_testc_ps(__m128 __a, __m128 __b) { - return __builtin_ia32_vtestcps((__v4sf)a, (__v4sf)b); + return __builtin_ia32_vtestcps((__v4sf)__a, (__v4sf)__b); } static __inline int __attribute__((__always_inline__, __nodebug__)) -_mm_testnzc_ps(__m128 a, __m128 b) +_mm_testnzc_ps(__m128 __a, __m128 __b) { - return __builtin_ia32_vtestnzcps((__v4sf)a, (__v4sf)b); + return __builtin_ia32_vtestnzcps((__v4sf)__a, (__v4sf)__b); } static __inline int __attribute__((__always_inline__, __nodebug__)) -_mm256_testz_pd(__m256d a, __m256d b) +_mm256_testz_pd(__m256d __a, __m256d __b) { - return __builtin_ia32_vtestzpd256((__v4df)a, (__v4df)b); + return __builtin_ia32_vtestzpd256((__v4df)__a, (__v4df)__b); } static __inline int __attribute__((__always_inline__, __nodebug__)) -_mm256_testc_pd(__m256d a, __m256d b) +_mm256_testc_pd(__m256d __a, __m256d __b) { - return __builtin_ia32_vtestcpd256((__v4df)a, (__v4df)b); + return __builtin_ia32_vtestcpd256((__v4df)__a, (__v4df)__b); } static __inline int __attribute__((__always_inline__, __nodebug__)) -_mm256_testnzc_pd(__m256d a, __m256d b) +_mm256_testnzc_pd(__m256d __a, __m256d __b) { - return __builtin_ia32_vtestnzcpd256((__v4df)a, (__v4df)b); + return __builtin_ia32_vtestnzcpd256((__v4df)__a, (__v4df)__b); } static __inline int __attribute__((__always_inline__, __nodebug__)) -_mm256_testz_ps(__m256 a, __m256 b) +_mm256_testz_ps(__m256 __a, __m256 __b) { - return __builtin_ia32_vtestzps256((__v8sf)a, (__v8sf)b); + return __builtin_ia32_vtestzps256((__v8sf)__a, (__v8sf)__b); } static __inline int __attribute__((__always_inline__, __nodebug__)) -_mm256_testc_ps(__m256 a, __m256 b) +_mm256_testc_ps(__m256 __a, __m256 __b) { - return __builtin_ia32_vtestcps256((__v8sf)a, (__v8sf)b); + return __builtin_ia32_vtestcps256((__v8sf)__a, (__v8sf)__b); } static __inline int __attribute__((__always_inline__, __nodebug__)) -_mm256_testnzc_ps(__m256 a, __m256 b) +_mm256_testnzc_ps(__m256 __a, __m256 __b) { - return __builtin_ia32_vtestnzcps256((__v8sf)a, (__v8sf)b); + return __builtin_ia32_vtestnzcps256((__v8sf)__a, (__v8sf)__b); } static __inline int __attribute__((__always_inline__, __nodebug__)) -_mm256_testz_si256(__m256i a, __m256i b) +_mm256_testz_si256(__m256i __a, __m256i __b) { - return __builtin_ia32_ptestz256((__v4di)a, (__v4di)b); + return __builtin_ia32_ptestz256((__v4di)__a, (__v4di)__b); } static __inline int __attribute__((__always_inline__, __nodebug__)) -_mm256_testc_si256(__m256i a, __m256i b) +_mm256_testc_si256(__m256i __a, __m256i __b) { - return __builtin_ia32_ptestc256((__v4di)a, (__v4di)b); + return __builtin_ia32_ptestc256((__v4di)__a, (__v4di)__b); } static __inline int __attribute__((__always_inline__, __nodebug__)) -_mm256_testnzc_si256(__m256i a, __m256i b) +_mm256_testnzc_si256(__m256i __a, __m256i __b) { - return __builtin_ia32_ptestnzc256((__v4di)a, (__v4di)b); + return __builtin_ia32_ptestnzc256((__v4di)__a, (__v4di)__b); } /* Vector extract sign mask */ static __inline int __attribute__((__always_inline__, __nodebug__)) -_mm256_movemask_pd(__m256d a) +_mm256_movemask_pd(__m256d __a) { - return __builtin_ia32_movmskpd256((__v4df)a); + return __builtin_ia32_movmskpd256((__v4df)__a); } static __inline int __attribute__((__always_inline__, __nodebug__)) -_mm256_movemask_ps(__m256 a) +_mm256_movemask_ps(__m256 __a) { - return __builtin_ia32_movmskps256((__v8sf)a); + return __builtin_ia32_movmskps256((__v8sf)__a); } -/* Vector zero */ +/* Vector __zero */ static __inline void __attribute__((__always_inline__, __nodebug__)) _mm256_zeroall(void) { @@ -718,341 +720,344 @@ _mm256_zeroupper(void) /* Vector load with broadcast */ static __inline __m128 __attribute__((__always_inline__, __nodebug__)) -_mm_broadcast_ss(float const *a) +_mm_broadcast_ss(float const *__a) { - return (__m128)__builtin_ia32_vbroadcastss(a); + return (__m128)__builtin_ia32_vbroadcastss(__a); } static __inline __m256d __attribute__((__always_inline__, __nodebug__)) -_mm256_broadcast_sd(double const *a) +_mm256_broadcast_sd(double const *__a) { - return (__m256d)__builtin_ia32_vbroadcastsd256(a); + return (__m256d)__builtin_ia32_vbroadcastsd256(__a); } static __inline __m256 __attribute__((__always_inline__, __nodebug__)) -_mm256_broadcast_ss(float const *a) +_mm256_broadcast_ss(float const *__a) { - return (__m256)__builtin_ia32_vbroadcastss256(a); + return (__m256)__builtin_ia32_vbroadcastss256(__a); } static __inline __m256d __attribute__((__always_inline__, __nodebug__)) -_mm256_broadcast_pd(__m128d const *a) +_mm256_broadcast_pd(__m128d const *__a) { - return (__m256d)__builtin_ia32_vbroadcastf128_pd256(a); + return (__m256d)__builtin_ia32_vbroadcastf128_pd256(__a); } static __inline __m256 __attribute__((__always_inline__, __nodebug__)) -_mm256_broadcast_ps(__m128 const *a) +_mm256_broadcast_ps(__m128 const *__a) { - return (__m256)__builtin_ia32_vbroadcastf128_ps256(a); + return (__m256)__builtin_ia32_vbroadcastf128_ps256(__a); } /* SIMD load ops */ static __inline __m256d __attribute__((__always_inline__, __nodebug__)) -_mm256_load_pd(double const *p) +_mm256_load_pd(double const *__p) { - return *(__m256d *)p; + return *(__m256d *)__p; } static __inline __m256 __attribute__((__always_inline__, __nodebug__)) -_mm256_load_ps(float const *p) +_mm256_load_ps(float const *__p) { - return *(__m256 *)p; + return *(__m256 *)__p; } static __inline __m256d __attribute__((__always_inline__, __nodebug__)) -_mm256_loadu_pd(double const *p) +_mm256_loadu_pd(double const *__p) { struct __loadu_pd { - __m256d v; + __m256d __v; } __attribute__((packed, may_alias)); - return ((struct __loadu_pd*)p)->v; + return ((struct __loadu_pd*)__p)->__v; } static __inline __m256 __attribute__((__always_inline__, __nodebug__)) -_mm256_loadu_ps(float const *p) +_mm256_loadu_ps(float const *__p) { struct __loadu_ps { - __m256 v; + __m256 __v; } __attribute__((packed, may_alias)); - return ((struct __loadu_ps*)p)->v; + return ((struct __loadu_ps*)__p)->__v; } static __inline __m256i __attribute__((__always_inline__, __nodebug__)) -_mm256_load_si256(__m256i const *p) +_mm256_load_si256(__m256i const *__p) { - return *p; + return *__p; } static __inline __m256i __attribute__((__always_inline__, __nodebug__)) -_mm256_loadu_si256(__m256i const *p) +_mm256_loadu_si256(__m256i const *__p) { struct __loadu_si256 { - __m256i v; + __m256i __v; } __attribute__((packed, may_alias)); - return ((struct __loadu_si256*)p)->v; + return ((struct __loadu_si256*)__p)->__v; } static __inline __m256i __attribute__((__always_inline__, __nodebug__)) -_mm256_lddqu_si256(__m256i const *p) +_mm256_lddqu_si256(__m256i const *__p) { - return (__m256i)__builtin_ia32_lddqu256((char const *)p); + return (__m256i)__builtin_ia32_lddqu256((char const *)__p); } /* SIMD store ops */ static __inline void __attribute__((__always_inline__, __nodebug__)) -_mm256_store_pd(double *p, __m256d a) +_mm256_store_pd(double *__p, __m256d __a) { - *(__m256d *)p = a; + *(__m256d *)__p = __a; } static __inline void __attribute__((__always_inline__, __nodebug__)) -_mm256_store_ps(float *p, __m256 a) +_mm256_store_ps(float *__p, __m256 __a) { - *(__m256 *)p = a; + *(__m256 *)__p = __a; } static __inline void __attribute__((__always_inline__, __nodebug__)) -_mm256_storeu_pd(double *p, __m256d a) +_mm256_storeu_pd(double *__p, __m256d __a) { - __builtin_ia32_storeupd256(p, (__v4df)a); + __builtin_ia32_storeupd256(__p, (__v4df)__a); } static __inline void __attribute__((__always_inline__, __nodebug__)) -_mm256_storeu_ps(float *p, __m256 a) +_mm256_storeu_ps(float *__p, __m256 __a) { - __builtin_ia32_storeups256(p, (__v8sf)a); + __builtin_ia32_storeups256(__p, (__v8sf)__a); } static __inline void __attribute__((__always_inline__, __nodebug__)) -_mm256_store_si256(__m256i *p, __m256i a) +_mm256_store_si256(__m256i *__p, __m256i __a) { - *p = a; + *__p = __a; } static __inline void __attribute__((__always_inline__, __nodebug__)) -_mm256_storeu_si256(__m256i *p, __m256i a) +_mm256_storeu_si256(__m256i *__p, __m256i __a) { - __builtin_ia32_storedqu256((char *)p, (__v32qi)a); + __builtin_ia32_storedqu256((char *)__p, (__v32qi)__a); } /* Conditional load ops */ static __inline __m128d __attribute__((__always_inline__, __nodebug__)) -_mm_maskload_pd(double const *p, __m128d m) +_mm_maskload_pd(double const *__p, __m128d __m) { - return (__m128d)__builtin_ia32_maskloadpd((const __v2df *)p, (__v2df)m); + return (__m128d)__builtin_ia32_maskloadpd((const __v2df *)__p, (__v2df)__m); } static __inline __m256d __attribute__((__always_inline__, __nodebug__)) -_mm256_maskload_pd(double const *p, __m256d m) +_mm256_maskload_pd(double const *__p, __m256d __m) { - return (__m256d)__builtin_ia32_maskloadpd256((const __v4df *)p, (__v4df)m); + return (__m256d)__builtin_ia32_maskloadpd256((const __v4df *)__p, + (__v4df)__m); } static __inline __m128 __attribute__((__always_inline__, __nodebug__)) -_mm_maskload_ps(float const *p, __m128 m) +_mm_maskload_ps(float const *__p, __m128 __m) { - return (__m128)__builtin_ia32_maskloadps((const __v4sf *)p, (__v4sf)m); + return (__m128)__builtin_ia32_maskloadps((const __v4sf *)__p, (__v4sf)__m); } static __inline __m256 __attribute__((__always_inline__, __nodebug__)) -_mm256_maskload_ps(float const *p, __m256 m) +_mm256_maskload_ps(float const *__p, __m256 __m) { - return (__m256)__builtin_ia32_maskloadps256((const __v8sf *)p, (__v8sf)m); + return (__m256)__builtin_ia32_maskloadps256((const __v8sf *)__p, (__v8sf)__m); } /* Conditional store ops */ static __inline void __attribute__((__always_inline__, __nodebug__)) -_mm256_maskstore_ps(float *p, __m256 m, __m256 a) +_mm256_maskstore_ps(float *__p, __m256 __m, __m256 __a) { - __builtin_ia32_maskstoreps256((__v8sf *)p, (__v8sf)m, (__v8sf)a); + __builtin_ia32_maskstoreps256((__v8sf *)__p, (__v8sf)__m, (__v8sf)__a); } static __inline void __attribute__((__always_inline__, __nodebug__)) -_mm_maskstore_pd(double *p, __m128d m, __m128d a) +_mm_maskstore_pd(double *__p, __m128d __m, __m128d __a) { - __builtin_ia32_maskstorepd((__v2df *)p, (__v2df)m, (__v2df)a); + __builtin_ia32_maskstorepd((__v2df *)__p, (__v2df)__m, (__v2df)__a); } static __inline void __attribute__((__always_inline__, __nodebug__)) -_mm256_maskstore_pd(double *p, __m256d m, __m256d a) +_mm256_maskstore_pd(double *__p, __m256d __m, __m256d __a) { - __builtin_ia32_maskstorepd256((__v4df *)p, (__v4df)m, (__v4df)a); + __builtin_ia32_maskstorepd256((__v4df *)__p, (__v4df)__m, (__v4df)__a); } static __inline void __attribute__((__always_inline__, __nodebug__)) -_mm_maskstore_ps(float *p, __m128 m, __m128 a) +_mm_maskstore_ps(float *__p, __m128 __m, __m128 __a) { - __builtin_ia32_maskstoreps((__v4sf *)p, (__v4sf)m, (__v4sf)a); + __builtin_ia32_maskstoreps((__v4sf *)__p, (__v4sf)__m, (__v4sf)__a); } /* Cacheability support ops */ static __inline void __attribute__((__always_inline__, __nodebug__)) -_mm256_stream_si256(__m256i *a, __m256i b) +_mm256_stream_si256(__m256i *__a, __m256i __b) { - __builtin_ia32_movntdq256((__v4di *)a, (__v4di)b); + __builtin_ia32_movntdq256((__v4di *)__a, (__v4di)__b); } static __inline void __attribute__((__always_inline__, __nodebug__)) -_mm256_stream_pd(double *a, __m256d b) +_mm256_stream_pd(double *__a, __m256d __b) { - __builtin_ia32_movntpd256(a, (__v4df)b); + __builtin_ia32_movntpd256(__a, (__v4df)__b); } static __inline void __attribute__((__always_inline__, __nodebug__)) -_mm256_stream_ps(float *p, __m256 a) +_mm256_stream_ps(float *__p, __m256 __a) { - __builtin_ia32_movntps256(p, (__v8sf)a); + __builtin_ia32_movntps256(__p, (__v8sf)__a); } /* Create vectors */ static __inline __m256d __attribute__((__always_inline__, __nodebug__)) -_mm256_set_pd(double a, double b, double c, double d) +_mm256_set_pd(double __a, double __b, double __c, double __d) { - return (__m256d){ d, c, b, a }; + return (__m256d){ __d, __c, __b, __a }; } static __inline __m256 __attribute__((__always_inline__, __nodebug__)) -_mm256_set_ps(float a, float b, float c, float d, - float e, float f, float g, float h) +_mm256_set_ps(float __a, float __b, float __c, float __d, + float __e, float __f, float __g, float __h) { - return (__m256){ h, g, f, e, d, c, b, a }; + return (__m256){ __h, __g, __f, __e, __d, __c, __b, __a }; } static __inline __m256i __attribute__((__always_inline__, __nodebug__)) -_mm256_set_epi32(int i0, int i1, int i2, int i3, - int i4, int i5, int i6, int i7) +_mm256_set_epi32(int __i0, int __i1, int __i2, int __i3, + int __i4, int __i5, int __i6, int __i7) { - return (__m256i)(__v8si){ i7, i6, i5, i4, i3, i2, i1, i0 }; + return (__m256i)(__v8si){ __i7, __i6, __i5, __i4, __i3, __i2, __i1, __i0 }; } static __inline __m256i __attribute__((__always_inline__, __nodebug__)) -_mm256_set_epi16(short w15, short w14, short w13, short w12, - short w11, short w10, short w09, short w08, - short w07, short w06, short w05, short w04, - short w03, short w02, short w01, short w00) +_mm256_set_epi16(short __w15, short __w14, short __w13, short __w12, + short __w11, short __w10, short __w09, short __w08, + short __w07, short __w06, short __w05, short __w04, + short __w03, short __w02, short __w01, short __w00) { - return (__m256i)(__v16hi){ w00, w01, w02, w03, w04, w05, w06, w07, - w08, w09, w10, w11, w12, w13, w14, w15 }; + return (__m256i)(__v16hi){ __w00, __w01, __w02, __w03, __w04, __w05, __w06, + __w07, __w08, __w09, __w10, __w11, __w12, __w13, __w14, __w15 }; } static __inline __m256i __attribute__((__always_inline__, __nodebug__)) -_mm256_set_epi8(char b31, char b30, char b29, char b28, - char b27, char b26, char b25, char b24, - char b23, char b22, char b21, char b20, - char b19, char b18, char b17, char b16, - char b15, char b14, char b13, char b12, - char b11, char b10, char b09, char b08, - char b07, char b06, char b05, char b04, - char b03, char b02, char b01, char b00) +_mm256_set_epi8(char __b31, char __b30, char __b29, char __b28, + char __b27, char __b26, char __b25, char __b24, + char __b23, char __b22, char __b21, char __b20, + char __b19, char __b18, char __b17, char __b16, + char __b15, char __b14, char __b13, char __b12, + char __b11, char __b10, char __b09, char __b08, + char __b07, char __b06, char __b05, char __b04, + char __b03, char __b02, char __b01, char __b00) { return (__m256i)(__v32qi){ - b00, b01, b02, b03, b04, b05, b06, b07, - b08, b09, b10, b11, b12, b13, b14, b15, - b16, b17, b18, b19, b20, b21, b22, b23, - b24, b25, b26, b27, b28, b29, b30, b31 + __b00, __b01, __b02, __b03, __b04, __b05, __b06, __b07, + __b08, __b09, __b10, __b11, __b12, __b13, __b14, __b15, + __b16, __b17, __b18, __b19, __b20, __b21, __b22, __b23, + __b24, __b25, __b26, __b27, __b28, __b29, __b30, __b31 }; } static __inline __m256i __attribute__((__always_inline__, __nodebug__)) -_mm256_set_epi64x(long long a, long long b, long long c, long long d) +_mm256_set_epi64x(long long __a, long long __b, long long __c, long long __d) { - return (__m256i)(__v4di){ d, c, b, a }; + return (__m256i)(__v4di){ __d, __c, __b, __a }; } /* Create vectors with elements in reverse order */ static __inline __m256d __attribute__((__always_inline__, __nodebug__)) -_mm256_setr_pd(double a, double b, double c, double d) +_mm256_setr_pd(double __a, double __b, double __c, double __d) { - return (__m256d){ a, b, c, d }; + return (__m256d){ __a, __b, __c, __d }; } static __inline __m256 __attribute__((__always_inline__, __nodebug__)) -_mm256_setr_ps(float a, float b, float c, float d, - float e, float f, float g, float h) +_mm256_setr_ps(float __a, float __b, float __c, float __d, + float __e, float __f, float __g, float __h) { - return (__m256){ a, b, c, d, e, f, g, h }; + return (__m256){ __a, __b, __c, __d, __e, __f, __g, __h }; } static __inline __m256i __attribute__((__always_inline__, __nodebug__)) -_mm256_setr_epi32(int i0, int i1, int i2, int i3, - int i4, int i5, int i6, int i7) +_mm256_setr_epi32(int __i0, int __i1, int __i2, int __i3, + int __i4, int __i5, int __i6, int __i7) { - return (__m256i)(__v8si){ i0, i1, i2, i3, i4, i5, i6, i7 }; + return (__m256i)(__v8si){ __i0, __i1, __i2, __i3, __i4, __i5, __i6, __i7 }; } static __inline __m256i __attribute__((__always_inline__, __nodebug__)) -_mm256_setr_epi16(short w15, short w14, short w13, short w12, - short w11, short w10, short w09, short w08, - short w07, short w06, short w05, short w04, - short w03, short w02, short w01, short w00) +_mm256_setr_epi16(short __w15, short __w14, short __w13, short __w12, + short __w11, short __w10, short __w09, short __w08, + short __w07, short __w06, short __w05, short __w04, + short __w03, short __w02, short __w01, short __w00) { - return (__m256i)(__v16hi){ w15, w14, w13, w12, w11, w10, w09, w08, - w07, w06, w05, w04, w03, w02, w01, w00 }; + return (__m256i)(__v16hi){ __w15, __w14, __w13, __w12, __w11, __w10, __w09, + __w08, __w07, __w06, __w05, __w04, __w03, __w02, __w01, __w00 }; } static __inline __m256i __attribute__((__always_inline__, __nodebug__)) -_mm256_setr_epi8(char b31, char b30, char b29, char b28, - char b27, char b26, char b25, char b24, - char b23, char b22, char b21, char b20, - char b19, char b18, char b17, char b16, - char b15, char b14, char b13, char b12, - char b11, char b10, char b09, char b08, - char b07, char b06, char b05, char b04, - char b03, char b02, char b01, char b00) +_mm256_setr_epi8(char __b31, char __b30, char __b29, char __b28, + char __b27, char __b26, char __b25, char __b24, + char __b23, char __b22, char __b21, char __b20, + char __b19, char __b18, char __b17, char __b16, + char __b15, char __b14, char __b13, char __b12, + char __b11, char __b10, char __b09, char __b08, + char __b07, char __b06, char __b05, char __b04, + char __b03, char __b02, char __b01, char __b00) { return (__m256i)(__v32qi){ - b31, b30, b29, b28, b27, b26, b25, b24, - b23, b22, b21, b20, b19, b18, b17, b16, - b15, b14, b13, b12, b11, b10, b09, b08, - b07, b06, b05, b04, b03, b02, b01, b00 }; + __b31, __b30, __b29, __b28, __b27, __b26, __b25, __b24, + __b23, __b22, __b21, __b20, __b19, __b18, __b17, __b16, + __b15, __b14, __b13, __b12, __b11, __b10, __b09, __b08, + __b07, __b06, __b05, __b04, __b03, __b02, __b01, __b00 }; } static __inline __m256i __attribute__((__always_inline__, __nodebug__)) -_mm256_setr_epi64x(long long a, long long b, long long c, long long d) +_mm256_setr_epi64x(long long __a, long long __b, long long __c, long long __d) { - return (__m256i)(__v4di){ a, b, c, d }; + return (__m256i)(__v4di){ __a, __b, __c, __d }; } /* Create vectors with repeated elements */ static __inline __m256d __attribute__((__always_inline__, __nodebug__)) -_mm256_set1_pd(double w) +_mm256_set1_pd(double __w) { - return (__m256d){ w, w, w, w }; + return (__m256d){ __w, __w, __w, __w }; } static __inline __m256 __attribute__((__always_inline__, __nodebug__)) -_mm256_set1_ps(float w) +_mm256_set1_ps(float __w) { - return (__m256){ w, w, w, w, w, w, w, w }; + return (__m256){ __w, __w, __w, __w, __w, __w, __w, __w }; } static __inline __m256i __attribute__((__always_inline__, __nodebug__)) -_mm256_set1_epi32(int i) +_mm256_set1_epi32(int __i) { - return (__m256i)(__v8si){ i, i, i, i, i, i, i, i }; + return (__m256i)(__v8si){ __i, __i, __i, __i, __i, __i, __i, __i }; } static __inline __m256i __attribute__((__always_inline__, __nodebug__)) -_mm256_set1_epi16(short w) +_mm256_set1_epi16(short __w) { - return (__m256i)(__v16hi){ w, w, w, w, w, w, w, w, w, w, w, w, w, w, w, w }; + return (__m256i)(__v16hi){ __w, __w, __w, __w, __w, __w, __w, __w, __w, __w, + __w, __w, __w, __w, __w, __w }; } static __inline __m256i __attribute__((__always_inline__, __nodebug__)) -_mm256_set1_epi8(char b) +_mm256_set1_epi8(char __b) { - return (__m256i)(__v32qi){ b, b, b, b, b, b, b, b, b, b, b, b, b, b, b, b, - b, b, b, b, b, b, b, b, b, b, b, b, b, b, b, b }; + return (__m256i)(__v32qi){ __b, __b, __b, __b, __b, __b, __b, __b, __b, __b, + __b, __b, __b, __b, __b, __b, __b, __b, __b, __b, __b, __b, __b, __b, __b, + __b, __b, __b, __b, __b, __b, __b }; } static __inline __m256i __attribute__((__always_inline__, __nodebug__)) -_mm256_set1_epi64x(long long q) +_mm256_set1_epi64x(long long __q) { - return (__m256i)(__v4di){ q, q, q, q }; + return (__m256i)(__v4di){ __q, __q, __q, __q }; } -/* Create zeroed vectors */ +/* Create __zeroed vectors */ static __inline __m256d __attribute__((__always_inline__, __nodebug__)) _mm256_setzero_pd(void) { @@ -1073,143 +1078,145 @@ _mm256_setzero_si256(void) /* Cast between vector types */ static __inline __m256 __attribute__((__always_inline__, __nodebug__)) -_mm256_castpd_ps(__m256d in) +_mm256_castpd_ps(__m256d __in) { - return (__m256)in; + return (__m256)__in; } static __inline __m256i __attribute__((__always_inline__, __nodebug__)) -_mm256_castpd_si256(__m256d in) +_mm256_castpd_si256(__m256d __in) { - return (__m256i)in; + return (__m256i)__in; } static __inline __m256d __attribute__((__always_inline__, __nodebug__)) -_mm256_castps_pd(__m256 in) +_mm256_castps_pd(__m256 __in) { - return (__m256d)in; + return (__m256d)__in; } static __inline __m256i __attribute__((__always_inline__, __nodebug__)) -_mm256_castps_si256(__m256 in) +_mm256_castps_si256(__m256 __in) { - return (__m256i)in; + return (__m256i)__in; } static __inline __m256 __attribute__((__always_inline__, __nodebug__)) -_mm256_castsi256_ps(__m256i in) +_mm256_castsi256_ps(__m256i __in) { - return (__m256)in; + return (__m256)__in; } static __inline __m256d __attribute__((__always_inline__, __nodebug__)) -_mm256_castsi256_pd(__m256i in) +_mm256_castsi256_pd(__m256i __in) { - return (__m256d)in; + return (__m256d)__in; } static __inline __m128d __attribute__((__always_inline__, __nodebug__)) -_mm256_castpd256_pd128(__m256d in) +_mm256_castpd256_pd128(__m256d __in) { - return __builtin_shufflevector(in, in, 0, 1); + return __builtin_shufflevector(__in, __in, 0, 1); } static __inline __m128 __attribute__((__always_inline__, __nodebug__)) -_mm256_castps256_ps128(__m256 in) +_mm256_castps256_ps128(__m256 __in) { - return __builtin_shufflevector(in, in, 0, 1, 2, 3); + return __builtin_shufflevector(__in, __in, 0, 1, 2, 3); } static __inline __m128i __attribute__((__always_inline__, __nodebug__)) -_mm256_castsi256_si128(__m256i in) +_mm256_castsi256_si128(__m256i __in) { - return __builtin_shufflevector(in, in, 0, 1); + return __builtin_shufflevector(__in, __in, 0, 1); } static __inline __m256d __attribute__((__always_inline__, __nodebug__)) -_mm256_castpd128_pd256(__m128d in) +_mm256_castpd128_pd256(__m128d __in) { - __m128d zero = _mm_setzero_pd(); - return __builtin_shufflevector(in, zero, 0, 1, 2, 2); + __m128d __zero = _mm_setzero_pd(); + return __builtin_shufflevector(__in, __zero, 0, 1, 2, 2); } static __inline __m256 __attribute__((__always_inline__, __nodebug__)) -_mm256_castps128_ps256(__m128 in) +_mm256_castps128_ps256(__m128 __in) { - __m128 zero = _mm_setzero_ps(); - return __builtin_shufflevector(in, zero, 0, 1, 2, 3, 4, 4, 4, 4); + __m128 __zero = _mm_setzero_ps(); + return __builtin_shufflevector(__in, __zero, 0, 1, 2, 3, 4, 4, 4, 4); } static __inline __m256i __attribute__((__always_inline__, __nodebug__)) -_mm256_castsi128_si256(__m128i in) +_mm256_castsi128_si256(__m128i __in) { - __m128i zero = _mm_setzero_si128(); - return __builtin_shufflevector(in, zero, 0, 1, 2, 2); + __m128i __zero = _mm_setzero_si128(); + return __builtin_shufflevector(__in, __zero, 0, 1, 2, 2); } /* SIMD load ops (unaligned) */ static __inline __m256 __attribute__((__always_inline__, __nodebug__)) -_mm256_loadu2_m128(float const *addr_hi, float const *addr_lo) +_mm256_loadu2_m128(float const *__addr_hi, float const *__addr_lo) { struct __loadu_ps { - __m128 v; + __m128 __v; } __attribute__((__packed__, __may_alias__)); - __m256 v256 = _mm256_castps128_ps256(((struct __loadu_ps*)addr_lo)->v); - return _mm256_insertf128_ps(v256, ((struct __loadu_ps*)addr_hi)->v, 1); + __m256 __v256 = _mm256_castps128_ps256(((struct __loadu_ps*)__addr_lo)->__v); + return _mm256_insertf128_ps(__v256, ((struct __loadu_ps*)__addr_hi)->__v, 1); } static __inline __m256d __attribute__((__always_inline__, __nodebug__)) -_mm256_loadu2_m128d(double const *addr_hi, double const *addr_lo) +_mm256_loadu2_m128d(double const *__addr_hi, double const *__addr_lo) { struct __loadu_pd { - __m128d v; + __m128d __v; } __attribute__((__packed__, __may_alias__)); - __m256d v256 = _mm256_castpd128_pd256(((struct __loadu_pd*)addr_lo)->v); - return _mm256_insertf128_pd(v256, ((struct __loadu_pd*)addr_hi)->v, 1); + __m256d __v256 = _mm256_castpd128_pd256(((struct __loadu_pd*)__addr_lo)->__v); + return _mm256_insertf128_pd(__v256, ((struct __loadu_pd*)__addr_hi)->__v, 1); } static __inline __m256i __attribute__((__always_inline__, __nodebug__)) -_mm256_loadu2_m128i(__m128i const *addr_hi, __m128i const *addr_lo) +_mm256_loadu2_m128i(__m128i const *__addr_hi, __m128i const *__addr_lo) { struct __loadu_si128 { - __m128i v; + __m128i __v; } __attribute__((packed, may_alias)); - __m256i v256 = _mm256_castsi128_si256(((struct __loadu_si128*)addr_lo)->v); - return _mm256_insertf128_si256(v256, ((struct __loadu_si128*)addr_hi)->v, 1); + __m256i __v256 = _mm256_castsi128_si256( + ((struct __loadu_si128*)__addr_lo)->__v); + return _mm256_insertf128_si256(__v256, + ((struct __loadu_si128*)__addr_hi)->__v, 1); } /* SIMD store ops (unaligned) */ static __inline void __attribute__((__always_inline__, __nodebug__)) -_mm256_storeu2_m128(float *addr_hi, float *addr_lo, __m256 a) +_mm256_storeu2_m128(float *__addr_hi, float *__addr_lo, __m256 __a) { - __m128 v128; + __m128 __v128; - v128 = _mm256_castps256_ps128(a); - __builtin_ia32_storeups(addr_lo, v128); - v128 = _mm256_extractf128_ps(a, 1); - __builtin_ia32_storeups(addr_hi, v128); + __v128 = _mm256_castps256_ps128(__a); + __builtin_ia32_storeups(__addr_lo, __v128); + __v128 = _mm256_extractf128_ps(__a, 1); + __builtin_ia32_storeups(__addr_hi, __v128); } static __inline void __attribute__((__always_inline__, __nodebug__)) -_mm256_storeu2_m128d(double *addr_hi, double *addr_lo, __m256d a) +_mm256_storeu2_m128d(double *__addr_hi, double *__addr_lo, __m256d __a) { - __m128d v128; + __m128d __v128; - v128 = _mm256_castpd256_pd128(a); - __builtin_ia32_storeupd(addr_lo, v128); - v128 = _mm256_extractf128_pd(a, 1); - __builtin_ia32_storeupd(addr_hi, v128); + __v128 = _mm256_castpd256_pd128(__a); + __builtin_ia32_storeupd(__addr_lo, __v128); + __v128 = _mm256_extractf128_pd(__a, 1); + __builtin_ia32_storeupd(__addr_hi, __v128); } static __inline void __attribute__((__always_inline__, __nodebug__)) -_mm256_storeu2_m128i(__m128i *addr_hi, __m128i *addr_lo, __m256i a) +_mm256_storeu2_m128i(__m128i *__addr_hi, __m128i *__addr_lo, __m256i __a) { - __m128i v128; + __m128i __v128; - v128 = _mm256_castsi256_si128(a); - __builtin_ia32_storedqu((char *)addr_lo, (__v16qi)v128); - v128 = _mm256_extractf128_si256(a, 1); - __builtin_ia32_storedqu((char *)addr_hi, (__v16qi)v128); + __v128 = _mm256_castsi256_si128(__a); + __builtin_ia32_storedqu((char *)__addr_lo, (__v16qi)__v128); + __v128 = _mm256_extractf128_si256(__a, 1); + __builtin_ia32_storedqu((char *)__addr_hi, (__v16qi)__v128); } |
