static __inline __m256d __attribute__((__always_inline__, __nodebug__))
_mm256_and_pd(__m256d a, __m256d b)
{
- return (__m256d)__builtin_ia32_andpd256((__v4df)a, (__v4df)b);
+ return (__m256d)((__v4df)a & (__v4df)b);
}
static __inline __m256 __attribute__((__always_inline__, __nodebug__))
_mm256_and_ps(__m256 a, __m256 b)
{
- return (__m256)__builtin_ia32_andps256((__v8sf)a, (__v8sf)b);
+ return (__m256)((__v8sf)a & (__v8sf)b);
}
static __inline __m256d __attribute__((__always_inline__, __nodebug__))
_mm256_andnot_pd(__m256d a, __m256d b)
{
- return (__m256d)__builtin_ia32_andnpd256((__v4df)a, (__v4df)b);
+ return (__m256d)(~(__v4df)a & (__v4df)b);
}
static __inline __m256 __attribute__((__always_inline__, __nodebug__))
_mm256_andnot_ps(__m256 a, __m256 b)
{
- return (__m256)__builtin_ia32_andnps256((__v8sf)a, (__v8sf)b);
+ return (__m256)(~(__v8sf)a & (__v8sf)b);
}
static __inline __m256d __attribute__((__always_inline__, __nodebug__))
_mm256_or_pd(__m256d a, __m256d b)
{
- return (__m256d)__builtin_ia32_orpd256((__v4df)a, (__v4df)b);
+ return (__m256d)((__v4df)a | (__v4df)b);
}
static __inline __m256 __attribute__((__always_inline__, __nodebug__))
_mm256_or_ps(__m256 a, __m256 b)
{
- return (__m256)__builtin_ia32_orps256((__v8sf)a, (__v8sf)b);
+ return (__m256)((__v8sf)a | (__v8sf)b);
}
static __inline __m256d __attribute__((__always_inline__, __nodebug__))
_mm256_xor_pd(__m256d a, __m256d b)
{
- return (__m256d)__builtin_ia32_xorpd256((__v4df)a, (__v4df)b);
+ return (__m256d)((__v4df)a ^ (__v4df)b);
}
static __inline __m256 __attribute__((__always_inline__, __nodebug__))
_mm256_xor_ps(__m256 a, __m256 b)
{
- return (__m256)__builtin_ia32_xorps256((__v8sf)a, (__v8sf)b);
+ return (__m256)((__v8sf)a ^ (__v8sf)b);
}
/* Horizontal arithmetic */