mmintrin.h

Mac OS X 10.4.9 for x86 Source Code gcc 实现源代码

_mm_subs_pu16 (__m64 __m1, __m64 __m2)
{
  return (__m64) __builtin_ia32_psubusw ((__v4hi)__m1, (__v4hi)__m2);
}

static __inline __m64 __attribute__((__always_inline__, __nodebug__))
_m_psubusw (__m64 __m1, __m64 __m2)
{
  return _mm_subs_pu16 (__m1, __m2);
}

/* Multiply four 16-bit values in M1 by four 16-bit values in M2 producing
   four 32-bit intermediate results, which are then summed by pairs to
   produce two 32-bit results.  */
static __inline __m64 __attribute__((__always_inline__, __nodebug__))
_mm_madd_pi16 (__m64 __m1, __m64 __m2)
{
  return (__m64) __builtin_ia32_pmaddwd ((__v4hi)__m1, (__v4hi)__m2);
}

static __inline __m64 __attribute__((__always_inline__, __nodebug__))
_m_pmaddwd (__m64 __m1, __m64 __m2)
{
  return _mm_madd_pi16 (__m1, __m2);
}

/* Multiply four signed 16-bit values in M1 by four signed 16-bit values in
   M2 and produce the high 16 bits of the 32-bit results.  */
static __inline __m64 __attribute__((__always_inline__, __nodebug__))
_mm_mulhi_pi16 (__m64 __m1, __m64 __m2)
{
  return (__m64) __builtin_ia32_pmulhw ((__v4hi)__m1, (__v4hi)__m2);
}

static __inline __m64 __attribute__((__always_inline__, __nodebug__))
_m_pmulhw (__m64 __m1, __m64 __m2)
{
  return _mm_mulhi_pi16 (__m1, __m2);
}

/* Multiply four 16-bit values in M1 by four 16-bit values in M2 and produce
   the low 16 bits of the results.  */
static __inline __m64 __attribute__((__always_inline__, __nodebug__))
_mm_mullo_pi16 (__m64 __m1, __m64 __m2)
{
  return (__m64) __builtin_ia32_pmullw ((__v4hi)__m1, (__v4hi)__m2);
}

static __inline __m64 __attribute__((__always_inline__, __nodebug__))
_m_pmullw (__m64 __m1, __m64 __m2)
{
  return _mm_mullo_pi16 (__m1, __m2);
}

/* Shift four 16-bit values in M left by COUNT.  */
static __inline __m64 __attribute__((__always_inline__, __nodebug__))
_mm_sll_pi16 (__m64 __m, __m64 __count)
{
  return (__m64) __builtin_ia32_psllw ((__v4hi)__m, (long long)__count);
}

static __inline __m64 __attribute__((__always_inline__, __nodebug__))
_m_psllw (__m64 __m, __m64 __count)
{
  return _mm_sll_pi16 (__m, __count);
}

static __inline __m64 __attribute__((__always_inline__, __nodebug__))
_mm_slli_pi16 (__m64 __m, int __count)
{
  return (__m64) __builtin_ia32_psllw ((__v4hi)__m, __count);
}

static __inline __m64 __attribute__((__always_inline__, __nodebug__))
_m_psllwi (__m64 __m, int __count)
{
  return _mm_slli_pi16 (__m, __count);
}

/* Shift two 32-bit values in M left by COUNT.  */
static __inline __m64 __attribute__((__always_inline__, __nodebug__))
_mm_sll_pi32 (__m64 __m, __m64 __count)
{
  return (__m64) __builtin_ia32_pslld ((__v2si)__m, (long long)__count);
}

static __inline __m64 __attribute__((__always_inline__, __nodebug__))
_m_pslld (__m64 __m, __m64 __count)
{
  return _mm_sll_pi32 (__m, __count);
}

static __inline __m64 __attribute__((__always_inline__, __nodebug__))
_mm_slli_pi32 (__m64 __m, int __count)
{
  return (__m64) __builtin_ia32_pslld ((__v2si)__m, __count);
}

static __inline __m64 __attribute__((__always_inline__, __nodebug__))
_m_pslldi (__m64 __m, int __count)
{
  return _mm_slli_pi32 (__m, __count);
}

/* Shift the 64-bit value in M left by COUNT.  */
static __inline __m64 __attribute__((__always_inline__, __nodebug__))
_mm_sll_si64 (__m64 __m, __m64 __count)
{
  return (__m64) __builtin_ia32_psllq ((long long)__m, (long long)__count);
}

static __inline __m64 __attribute__((__always_inline__, __nodebug__))
_m_psllq (__m64 __m, __m64 __count)
{
  return _mm_sll_si64 (__m, __count);
}

static __inline __m64 __attribute__((__always_inline__, __nodebug__))
_mm_slli_si64 (__m64 __m, int __count)
{
  return (__m64) __builtin_ia32_psllq ((long long)__m, (long long)__count);
}

static __inline __m64 __attribute__((__always_inline__, __nodebug__))
_m_psllqi (__m64 __m, int __count)
{
  return _mm_slli_si64 (__m, __count);
}

/* Shift four 16-bit values in M right by COUNT; shift in the sign bit.  */
static __inline __m64 __attribute__((__always_inline__, __nodebug__))
_mm_sra_pi16 (__m64 __m, __m64 __count)
{
  return (__m64) __builtin_ia32_psraw ((__v4hi)__m, (long long)__count);
}

static __inline __m64 __attribute__((__always_inline__, __nodebug__))
_m_psraw (__m64 __m, __m64 __count)
{
  return _mm_sra_pi16 (__m, __count);
}

static __inline __m64 __attribute__((__always_inline__, __nodebug__))
_mm_srai_pi16 (__m64 __m, int __count)
{
  return (__m64) __builtin_ia32_psraw ((__v4hi)__m, __count);
}

static __inline __m64 __attribute__((__always_inline__, __nodebug__))
_m_psrawi (__m64 __m, int __count)
{
  return _mm_srai_pi16 (__m, __count);
}

/* Shift two 32-bit values in M right by COUNT; shift in the sign bit.  */
static __inline __m64 __attribute__((__always_inline__, __nodebug__))
_mm_sra_pi32 (__m64 __m, __m64 __count)
{
  return (__m64) __builtin_ia32_psrad ((__v2si)__m, (long long)__count);
}

static __inline __m64 __attribute__((__always_inline__, __nodebug__))
_m_psrad (__m64 __m, __m64 __count)
{
  return _mm_sra_pi32 (__m, __count);
}

static __inline __m64 __attribute__((__always_inline__, __nodebug__))
_mm_srai_pi32 (__m64 __m, int __count)
{
  return (__m64) __builtin_ia32_psrad ((__v2si)__m, __count);
}

static __inline __m64 __attribute__((__always_inline__, __nodebug__))
_m_psradi (__m64 __m, int __count)
{
  return _mm_srai_pi32 (__m, __count);
}

/* Shift four 16-bit values in M right by COUNT; shift in zeros.  */
static __inline __m64 __attribute__((__always_inline__, __nodebug__))
_mm_srl_pi16 (__m64 __m, __m64 __count)
{
  return (__m64) __builtin_ia32_psrlw ((__v4hi)__m, (long long)__count);
}

static __inline __m64 __attribute__((__always_inline__, __nodebug__))
_m_psrlw (__m64 __m, __m64 __count)
{
  return _mm_srl_pi16 (__m, __count);
}

static __inline __m64 __attribute__((__always_inline__, __nodebug__))
_mm_srli_pi16 (__m64 __m, int __count)
{
  return (__m64) __builtin_ia32_psrlw ((__v4hi)__m, __count);
}

static __inline __m64 __attribute__((__always_inline__, __nodebug__))
_m_psrlwi (__m64 __m, int __count)
{
  return _mm_srli_pi16 (__m, __count);
}

/* Shift two 32-bit values in M right by COUNT; shift in zeros.  */
static __inline __m64 __attribute__((__always_inline__, __nodebug__))
_mm_srl_pi32 (__m64 __m, __m64 __count)
{
  return (__m64) __builtin_ia32_psrld ((__v2si)__m, (long long)__count);
}

static __inline __m64 __attribute__((__always_inline__, __nodebug__))
_m_psrld (__m64 __m, __m64 __count)
{
  return _mm_srl_pi32 (__m, __count);
}

static __inline __m64 __attribute__((__always_inline__, __nodebug__))
_mm_srli_pi32 (__m64 __m, int __count)
{
  return (__m64) __builtin_ia32_psrld ((__v2si)__m, __count);
}

static __inline __m64 __attribute__((__always_inline__, __nodebug__))
_m_psrldi (__m64 __m, int __count)
{
  return _mm_srli_pi32 (__m, __count);
}

/* Shift the 64-bit value in M left by COUNT; shift in zeros.  */
static __inline __m64 __attribute__((__always_inline__, __nodebug__))
_mm_srl_si64 (__m64 __m, __m64 __count)
{
  return (__m64) __builtin_ia32_psrlq ((long long)__m, (long long)__count);
}

static __inline __m64 __attribute__((__always_inline__, __nodebug__))
_m_psrlq (__m64 __m, __m64 __count)
{
  return _mm_srl_si64 (__m, __count);
}

static __inline __m64 __attribute__((__always_inline__, __nodebug__))
_mm_srli_si64 (__m64 __m, int __count)
{
  return (__m64) __builtin_ia32_psrlq ((long long)__m, (long long)__count);
}

static __inline __m64 __attribute__((__always_inline__, __nodebug__))
_m_psrlqi (__m64 __m, int __count)
{
  return _mm_srli_si64 (__m, __count);
}

/* Bit-wise AND the 64-bit values in M1 and M2.  */
static __inline __m64 __attribute__((__always_inline__, __nodebug__))
_mm_and_si64 (__m64 __m1, __m64 __m2)
{
  return __builtin_ia32_pand (__m1, __m2);
}

static __inline __m64 __attribute__((__always_inline__, __nodebug__))
_m_pand (__m64 __m1, __m64 __m2)
{
  return _mm_and_si64 (__m1, __m2);
}

/* Bit-wise complement the 64-bit value in M1 and bit-wise AND it with the
   64-bit value in M2.  */
static __inline __m64 __attribute__((__always_inline__, __nodebug__))
_mm_andnot_si64 (__m64 __m1, __m64 __m2)
{
  return __builtin_ia32_pandn (__m1, __m2);
}

static __inline __m64 __attribute__((__always_inline__, __nodebug__))
_m_pandn (__m64 __m1, __m64 __m2)
{
  return _mm_andnot_si64 (__m1, __m2);
}

/* Bit-wise inclusive OR the 64-bit values in M1 and M2.  */
static __inline __m64 __attribute__((__always_inline__, __nodebug__))
_mm_or_si64 (__m64 __m1, __m64 __m2)
{
  return __builtin_ia32_por (__m1, __m2);
}

static __inline __m64 __attribute__((__always_inline__, __nodebug__))
_m_por (__m64 __m1, __m64 __m2)
{
  return _mm_or_si64 (__m1, __m2);
}

/* Bit-wise exclusive OR the 64-bit values in M1 and M2.  */
static __inline __m64 __attribute__((__always_inline__, __nodebug__))
_mm_xor_si64 (__m64 __m1, __m64 __m2)
{
  return __builtin_ia32_pxor (__m1, __m2);
}

static __inline __m64 __attribute__((__always_inline__, __nodebug__))
_m_pxor (__m64 __m1, __m64 __m2)
{
  return _mm_xor_si64 (__m1, __m2);
}

/* Compare eight 8-bit values.  The result of the comparison is 0xFF if the
   test is true and zero if false.  */
static __inline __m64 __attribute__((__always_inline__, __nodebug__))
_mm_cmpeq_pi8 (__m64 __m1, __m64 __m2)
{
  return (__m64) __builtin_ia32_pcmpeqb ((__v8qi)__m1, (__v8qi)__m2);
}

static __inline __m64 __attribute__((__always_inline__, __nodebug__))
_m_pcmpeqb (__m64 __m1, __m64 __m2)
{
  return _mm_cmpeq_pi8 (__m1, __m2);
}

static __inline __m64 __attribute__((__always_inline__, __nodebug__))
_mm_cmpgt_pi8 (__m64 __m1, __m64 __m2)
{
  return (__m64) __builtin_ia32_pcmpgtb ((__v8qi)__m1, (__v8qi)__m2);
}

static __inline __m64 __attribute__((__always_inline__, __nodebug__))
_m_pcmpgtb (__m64 __m1, __m64 __m2)
{
  return _mm_cmpgt_pi8 (__m1, __m2);
}

/* Compare four 16-bit values.  The result of the comparison is 0xFFFF if
   the test is true and zero if false.  */
static __inline __m64 __attribute__((__always_inline__, __nodebug__))
_mm_cmpeq_pi16 (__m64 __m1, __m64 __m2)
{
  return (__m64) __builtin_ia32_pcmpeqw ((__v4hi)__m1, (__v4hi)__m2);
}

static __inline __m64 __attribute__((__always_inline__, __nodebug__))
_m_pcmpeqw (__m64 __m1, __m64 __m2)
{
  return _mm_cmpeq_pi16 (__m1, __m2);
}

static __inline __m64 __attribute__((__always_inline__, __nodebug__))
_mm_cmpgt_pi16 (__m64 __m1, __m64 __m2)
{
  return (__m64) __builtin_ia32_pcmpgtw ((__v4hi)__m1, (__v4hi)__m2);
}

static __inline __m64 __attribute__((__always_inline__, __nodebug__))
_m_pcmpgtw (__m64 __m1, __m64 __m2)
{
  return _mm_cmpgt_pi16 (__m1, __m2);
}

/* Compare two 32-bit values.  The result of the comparison is 0xFFFFFFFF if
   the test is true and zero if false.  */
static __inline __m64 __attribute__((__always_inline__, __nodebug__))
_mm_cmpeq_pi32 (__m64 __m1, __m64 __m2)
{
  return (__m64) __builtin_ia32_pcmpeqd ((__v2si)__m1, (__v2si)__m2);
}

static __inline __m64 __attribute__((__always_inline__, __nodebug__))
_m_pcmpeqd (__m64 __m1, __m64 __m2)
{
  return _mm_cmpeq_pi32 (__m1, __m2);
}

static __inline __m64 __attribute__((__always_inline__, __nodebug__))
_mm_cmpgt_pi32 (__m64 __m1, __m64 __m2)
{
  return (__m64) __builtin_ia32_pcmpgtd ((__v2si)__m1, (__v2si)__m2);
}

static __inline __m64 __attribute__((__always_inline__, __nodebug__))
_m_pcmpgtd (__m64 __m1, __m64 __m2)
{
  return _mm_cmpgt_pi32 (__m1, __m2);
}

/* Creates a 64-bit zero.  */
static __inline __m64 __attribute__((__always_inline__, __nodebug__))
_mm_setzero_si64 (void)
{
  return (__m64)0LL;
}

/* Creates a vector of two 32-bit values; I0 is least significant.  */
static __inline __m64 __attribute__((__always_inline__, __nodebug__))
_mm_set_pi32 (int __i1, int __i0)
{
  return (__m64) __builtin_ia32_vec_init_v2si (__i0, __i1);
}

/* Creates a vector of four 16-bit values; W0 is least significant.  */
static __inline __m64 __attribute__((__always_inline__, __nodebug__))
_mm_set_pi16 (short __w3, short __w2, short __w1, short __w0)
{
  return (__m64) __builtin_ia32_vec_init_v4hi (__w0, __w1, __w2, __w3);
}

/* Creates a vector of eight 8-bit values; B0 is least significant.  */
static __inline __m64 __attribute__((__always_inline__, __nodebug__))
_mm_set_pi8 (char __b7, char __b6, char __b5, char __b4,
	     char __b3, char __b2, char __b1, char __b0)
{
  return (__m64) __builtin_ia32_vec_init_v8qi (__b0, __b1, __b2, __b3,
					       __b4, __b5, __b6, __b7);
}

/* Similar, but with the arguments in reverse order.  */
static __inline __m64 __attribute__((__always_inline__, __nodebug__))
_mm_setr_pi32 (int __i0, int __i1)
{
  return _mm_set_pi32 (__i1, __i0);
}

static __inline __m64 __attribute__((__always_inline__, __nodebug__))
_mm_setr_pi16 (short __w0, short __w1, short __w2, short __w3)
{
  return _mm_set_pi16 (__w3, __w2, __w1, __w0);
}

static __inline __m64 __attribute__((__always_inline__, __nodebug__))
_mm_setr_pi8 (char __b0, char __b1, char __b2, char __b3,
	      char __b4, char __b5, char __b6, char __b7)
{
  return _mm_set_pi8 (__b7, __b6, __b5, __b4, __b3, __b2, __b1, __b0);
}

/* Creates a vector of two 32-bit values, both elements containing I.  */
static __inline __m64 __attribute__((__always_inline__, __nodebug__))
_mm_set1_pi32 (int __i)
{
  return _mm_set_pi32 (__i, __i);
}

/* Creates a vector of four 16-bit values, all elements containing W.  */
static __inline __m64 __attribute__((__always_inline__, __nodebug__))
_mm_set1_pi16 (short __w)
{
  return _mm_set_pi16 (__w, __w, __w, __w);
}

/* Creates a vector of eight 8-bit values, all elements containing B.  */
static __inline __m64 __attribute__((__always_inline__, __nodebug__))
_mm_set1_pi8 (char __b)
{
  return _mm_set_pi8 (__b, __b, __b, __b, __b, __b, __b, __b);
}
/* APPLE LOCAL end radar 4152603 */

#endif /* __MMX__ */
#endif /* _MMINTRIN_H_INCLUDED */