#ifndef _AMMINTRIN_H_INCLUDED
#define _AMMINTRIN_H_INCLUDED
#ifndef __SSE4A__
# error "SSE4A instruction set not enabled"
#else
#include <pmmintrin.h>
#define __always_inline__ __always_inline__, __nodebug__
#undef __STATIC_INLINE
#ifdef __GNUC_STDC_INLINE__
#define __STATIC_INLINE __inline
#else
#define __STATIC_INLINE static __inline
#endif
__STATIC_INLINE void __attribute__((__always_inline__))
_mm_stream_sd (double * __P, __m128d __Y)
{
__builtin_ia32_movntsd (__P, (__v2df) __Y);
}
__STATIC_INLINE void __attribute__((__always_inline__))
_mm_stream_ss (float * __P, __m128 __Y)
{
__builtin_ia32_movntss (__P, (__v4sf) __Y);
}
__STATIC_INLINE __m128i __attribute__((__always_inline__))
_mm_extract_si64 (__m128i __X, __m128i __Y)
{
return (__m128i) __builtin_ia32_extrq ((__v2di) __X, (__v16qi) __Y);
}
#ifdef __OPTIMIZE__
__STATIC_INLINE __m128i __attribute__((__always_inline__))
_mm_extracti_si64 (__m128i __X, unsigned const int __I, unsigned const int __L)
{
return (__m128i) __builtin_ia32_extrqi ((__v2di) __X, __I, __L);
}
#else
#define _mm_extracti_si64(X, I, L) \
((__m128i) __builtin_ia32_extrqi ((__v2di)(X), I, L))
#endif
__STATIC_INLINE __m128i __attribute__((__always_inline__))
_mm_insert_si64 (__m128i __X,__m128i __Y)
{
return (__m128i) __builtin_ia32_insertq ((__v2di)__X, (__v2di)__Y);
}
#ifdef __OPTIMIZE__
__STATIC_INLINE __m128i __attribute__((__always_inline__))
_mm_inserti_si64(__m128i __X, __m128i __Y, unsigned const int __I, unsigned const int __L)
{
return (__m128i) __builtin_ia32_insertqi ((__v2di)__X, (__v2di)__Y, __I, __L);
}
#else
#define _mm_inserti_si64(X, Y, I, L) \
((__m128i) __builtin_ia32_insertqi ((__v2di)(X), (__v2di)(Y), I, L))
#endif
#endif
#undef __always_inline__
#endif