// Sets the four single-precision, floating-point values to the four inputs in reverse order. https://msdn.microsoft.com/en-us/library/vstudio/d2172ct3(v=vs.100).aspx
// Stores four 32-bit integer values as (as a __m128i value) at the address p. https://msdn.microsoft.com/en-us/library/vstudio/edk11s13(v=vs.100).aspx
// Loads a single single-precision, floating-point value, copying it into all four words https://msdn.microsoft.com/en-us/library/vstudio/5cdkf716(v=vs.100).aspx
// for neon, alignment doesn't matter, so _mm_load_ps and _mm_loadu_ps are equivalent for neon
returnvld1q_f32(p);
}
// Loads an single - precision, floating - point value into the low word and clears the upper three words. https://msdn.microsoft.com/en-us/library/548bb9h4%28v=vs.90%29.aspx
// Compares for inequality. https://msdn.microsoft.com/en-us/library/sf44thbx(v=vs.100).aspx
FORCE_INLINE__m128_mm_cmpneq_ps(__m128a,__m128b)
{
return(__m128)vmvnq_s32((__m128i)vceqq_f32(a,b));
}
// Computes the bitwise AND-NOT of the four single-precision, floating-point values of a and b. https://msdn.microsoft.com/en-us/library/vstudio/68h7wd02(v=vs.100).aspx
// Computes the bitwise AND of the 128-bit value in b and the bitwise NOT of the 128-bit value in a. https://msdn.microsoft.com/en-us/library/vstudio/1beaceh8(v=vs.100).aspx
// Computes the bitwise AND of the 128-bit value in a and the 128-bit value in b. https://msdn.microsoft.com/en-us/library/vstudio/6d1txsa8(v=vs.100).aspx
// Computes the bitwise AND of the four single-precision, floating-point values of a and b. https://msdn.microsoft.com/en-us/library/vstudio/73ck1xc5(v=vs.100).aspx
FORCE_INLINE__m128_mm_and_ps(__m128a,__m128b)
{
return(__m128)vandq_s32((__m128i)a,(__m128i)b);
}
// Computes the bitwise OR of the four single-precision, floating-point values of a and b. https://msdn.microsoft.com/en-us/library/vstudio/7ctdsyy0(v=vs.100).aspx
FORCE_INLINE__m128_mm_or_ps(__m128a,__m128b)
{
return(__m128)vorrq_s32((__m128i)a,(__m128i)b);
}
// Computes bitwise EXOR (exclusive-or) of the four single-precision, floating-point values of a and b. https://msdn.microsoft.com/en-us/library/ss6k3wk8(v=vs.100).aspx
FORCE_INLINE__m128_mm_xor_ps(__m128a,__m128b)
{
return(__m128)veorq_s32((__m128i)a,(__m128i)b);
}
// Computes the bitwise OR of the 128-bit value in a and the 128-bit value in b. https://msdn.microsoft.com/en-us/library/vstudio/ew8ty0db(v=vs.100).aspx
// Creates a 4-bit mask from the most significant bits of the four single-precision, floating-point values. https://msdn.microsoft.com/en-us/library/vstudio/4490ys29(v=vs.100).aspx
FORCE_INLINEint_mm_movemask_ps(__m128a)
{
#if ENABLE_CPP_VERSION // I am not yet convinced that the NEON version is faster than the C version of this
// NEON does not support a general purpose permute intrinsic
// Currently I am not sure whether the C implementation is faster or slower than the NEON version.
// Note, this has to be expanded as a template because the shuffle value must be an immediate value.
// The same is true on SSE as well.
// Selects four specific single-precision, floating-point values from a and b, based on the mask i. https://msdn.microsoft.com/en-us/library/vstudio/5f0858x0(v=vs.100).aspx
// Shuffles the upper 4 signed or unsigned 16 - bit integers in a as specified by imm. https://msdn.microsoft.com/en-us/library/13ywktbs(v=vs.100).aspx
// Shifts the 4 signed or unsigned 32-bit integers in a left by count bits while shifting in zeros. : https://msdn.microsoft.com/en-us/library/z2k3bbtb%28v=vs.90%29.aspx
//Shifts the 4 signed or unsigned 32-bit integers in a right by count bits while shifting in zeros. https://msdn.microsoft.com/en-us/library/w486zcfa(v=vs.100).aspx
// Shifts the 4 signed 32 - bit integers in a right by count bits while shifting in the sign bit. https://msdn.microsoft.com/en-us/library/z1939387(v=vs.100).aspx
// Shifts the 128 - bit value in a right by imm bytes while shifting in zeros.imm must be an immediate. https://msdn.microsoft.com/en-us/library/305w28yz(v=vs.100).aspx
//#define _mm_srli_si128( a, imm ) (__m128i)vmaxq_s8((int8x16_t)a, vextq_s8((int8x16_t)a, vdupq_n_s8(0), imm))
// Shifts the 128-bit value in a left by imm bytes while shifting in zeros. imm must be an immediate. https://msdn.microsoft.com/en-us/library/34d3k2kt(v=vs.100).aspx
// Creates a 16-bit mask from the most significant bits of the 16 signed or unsigned 8-bit integers in a and zero extends the upper bits. https://msdn.microsoft.com/en-us/library/vstudio/s090c8fk(v=vs.100).aspx
// Subtracts the four single-precision, floating-point values of a and b. https://msdn.microsoft.com/en-us/library/vstudio/1zad2k61(v=vs.100).aspx
FORCE_INLINE__m128_mm_sub_ps(__m128a,__m128b)
{
returnvsubq_f32(a,b);
}
// Subtracts the 4 signed or unsigned 32-bit integers of b from the 4 signed or unsigned 32-bit integers of a. https://msdn.microsoft.com/en-us/library/vstudio/fhh866h0(v=vs.100).aspx
// Adds the four single-precision, floating-point values of a and b. https://msdn.microsoft.com/en-us/library/vstudio/c9848chc(v=vs.100).aspx
FORCE_INLINE__m128_mm_add_ps(__m128a,__m128b)
{
returnvaddq_f32(a,b);
}
// Adds the 4 signed or unsigned 32-bit integers in a to the 4 signed or unsigned 32-bit integers in b. https://msdn.microsoft.com/en-us/library/vstudio/09xs4fkk(v=vs.100).aspx
// Adds the 8 signed or unsigned 16-bit integers in a to the 8 signed or unsigned 16-bit integers in b. https://msdn.microsoft.com/en-us/library/fceha5k4(v=vs.100).aspx
// Multiplies the 8 signed or unsigned 16-bit integers from a by the 8 signed or unsigned 16-bit integers from b. https://msdn.microsoft.com/en-us/library/vstudio/9ks1472s(v=vs.100).aspx
// Multiplies the 4 signed or unsigned 32-bit integers from a by the 4 signed or unsigned 32-bit integers from b. https://msdn.microsoft.com/en-us/library/vstudio/bb531409(v=vs.100).aspx
// Multiplies the four single-precision, floating-point values of a and b. https://msdn.microsoft.com/en-us/library/vstudio/22kbk6t9(v=vs.100).aspx
FORCE_INLINE__m128_mm_mul_ps(__m128a,__m128b)
{
returnvmulq_f32(a,b);
}
// This version does additional iterations to improve accuracy. Between 1 and 4 recommended.
// Computes the approximations of reciprocals of the four single-precision, floating-point values of a. https://msdn.microsoft.com/en-us/library/vstudio/796k1tty(v=vs.100).aspx
// Computes the approximations of reciprocals of the four single-precision, floating-point values of a. https://msdn.microsoft.com/en-us/library/vstudio/796k1tty(v=vs.100).aspx
FORCE_INLINE__m128_mm_rcp_ps(__m128in)
{
__m128recip=vrecpeq_f32(in);
recip=vmulq_f32(recip,vrecpsq_f32(recip,in));
returnrecip;
}
// Computes the approximations of square roots of the four single-precision, floating-point values of a. First computes reciprocal square roots and then reciprocals of the four values. https://msdn.microsoft.com/en-us/library/vstudio/8z67bwwk(v=vs.100).aspx
FORCE_INLINE__m128_mm_sqrt_ps(__m128in)
{
__m128recipsq=vrsqrteq_f32(in);
__m128sq=vrecpeq_f32(recipsq);
// ??? use step versions of both sqrt and recip for better accuracy?
returnsq;
}
// Computes the maximums of the four single-precision, floating-point values of a and b. https://msdn.microsoft.com/en-us/library/vstudio/ff5d607a(v=vs.100).aspx
FORCE_INLINE__m128_mm_max_ps(__m128a,__m128b)
{
returnvmaxq_f32(a,b);
}
// Computes the minima of the four single-precision, floating-point values of a and b. https://msdn.microsoft.com/en-us/library/vstudio/wh13kadz(v=vs.100).aspx
FORCE_INLINE__m128_mm_min_ps(__m128a,__m128b)
{
returnvminq_f32(a,b);
}
// Computes the pairwise minima of the 8 signed 16-bit integers from a and the 8 signed 16-bit integers from b. https://msdn.microsoft.com/en-us/library/vstudio/6te997ew(v=vs.100).aspx
// Computes the pariwise maximums of the four signed 32-bit integer values of a and b. https://msdn.microsoft.com/en-us/library/vstudio/bb514055(v=vs.100).aspx
// Computes the pariwise minima of the four signed 32-bit integer values of a and b. https://msdn.microsoft.com/en-us/library/vstudio/bb531476(v=vs.100).aspx
// Multiplies the 8 signed 16-bit integers from a by the 8 signed 16-bit integers from b. https://msdn.microsoft.com/en-us/library/vstudio/59hddw1d(v=vs.100).aspx
// Compares for less than https://msdn.microsoft.com/en-us/library/vstudio/f330yhc8(v=vs.100).aspx
FORCE_INLINE__m128_mm_cmplt_ps(__m128a,__m128b)
{
return(__m128)vcltq_f32(a,b);
}
// Compares for greater than. https://msdn.microsoft.com/en-us/library/vstudio/11dy102s(v=vs.100).aspx
FORCE_INLINE__m128_mm_cmpgt_ps(__m128a,__m128b)
{
return(__m128)vcgtq_f32(a,b);
}
// Compares for greater than or equal. https://msdn.microsoft.com/en-us/library/vstudio/fs813y2t(v=vs.100).aspx
FORCE_INLINE__m128_mm_cmpge_ps(__m128a,__m128b)
{
return(__m128)vcgeq_f32(a,b);
}
// Compares for less than or equal. https://msdn.microsoft.com/en-us/library/vstudio/1s75w83z(v=vs.100).aspx
FORCE_INLINE__m128_mm_cmple_ps(__m128a,__m128b)
{
return(__m128)vcleq_f32(a,b);
}
// Compares for equality. https://msdn.microsoft.com/en-us/library/vstudio/36aectz5(v=vs.100).aspx
FORCE_INLINE__m128_mm_cmpeq_ps(__m128a,__m128b)
{
return(__m128)vceqq_f32(a,b);
}
// Compares the 4 signed 32-bit integers in a and the 4 signed 32-bit integers in b for less than. https://msdn.microsoft.com/en-us/library/vstudio/4ak0bf5d(v=vs.100).aspx
// Compares the 4 signed 32-bit integers in a and the 4 signed 32-bit integers in b for greater than. https://msdn.microsoft.com/en-us/library/vstudio/1s9f2z0y(v=vs.100).aspx
// Compares the four 32-bit floats in a and b to check if any values are NaN. Ordered compare between each value returns true for "orderable" and false for "not orderable" (NaN). https://msdn.microsoft.com/en-us/library/vstudio/0h9w00fx(v=vs.100).aspx
// Converts the four single-precision, floating-point values of a to signed 32-bit integer values using truncate. https://msdn.microsoft.com/en-us/library/vstudio/1h005y6x(v=vs.100).aspx
FORCE_INLINE__m128i_mm_cvttps_epi32(__m128a)
{
returnvcvtq_s32_f32(a);
}
// Converts the four signed 32-bit integer values of a to single-precision, floating-point values https://msdn.microsoft.com/en-us/library/vstudio/36bwxcx5(v=vs.100).aspx
FORCE_INLINE__m128_mm_cvtepi32_ps(__m128ia)
{
returnvcvtq_f32_s32(a);
}
// Converts the four single-precision, floating-point values of a to signed 32-bit integer values. https://msdn.microsoft.com/en-us/library/vstudio/xdc42k5e(v=vs.100).aspx
// Moves the least significant 32 bits of a to a 32-bit integer. https://msdn.microsoft.com/en-us/library/5z7a9642%28v=vs.90%29.aspx
FORCE_INLINEint_mm_cvtsi128_si32(__m128ia)
{
returnvgetq_lane_s32(a,0);
}
// Moves 32-bit integer a to the least significant 32 bits of an __m128 object, zero extending the upper bits. https://msdn.microsoft.com/en-us/library/ct3539ha%28v=vs.90%29.aspx
FORCE_INLINE__m128i_mm_cvtsi32_si128(inta)
{
__m128iresult=vdupq_n_s32(0);
returnvsetq_lane_s32(a,result,0);
}
// Applies a type cast to reinterpret four 32-bit floating point values passed in as a 128-bit parameter as packed 32-bit integers. https://msdn.microsoft.com/en-us/library/bb514099.aspx
FORCE_INLINE__m128i_mm_castps_si128(__m128a)
{
returnvcvtq_s32_f32(a);
//return *((const __m128i *) &a);
}
// Applies a type cast to reinterpret four 32-bit integers passed in as a 128-bit parameter as packed 32-bit floating point values. https://msdn.microsoft.com/en-us/library/bb514029.aspx
// Packs the 16 signed 16-bit integers from a and b into 8-bit integers and saturates. https://msdn.microsoft.com/en-us/library/k4y4f7w5%28v=vs.90%29.aspx
// Packs the 16 signed 16 - bit integers from a and b into 8 - bit unsigned integers and saturates. https://msdn.microsoft.com/en-us/library/07ad1wx4(v=vs.100).aspx
// Packs the 8 signed 32-bit integers from a and b into signed 16-bit integers and saturates. https://msdn.microsoft.com/en-us/library/393t56f9%28v=vs.90%29.aspx
// Interleaves the lower 8 signed or unsigned 8-bit integers in a with the lower 8 signed or unsigned 8-bit integers in b. https://msdn.microsoft.com/en-us/library/xf7k860c%28v=vs.90%29.aspx
// Interleaves the lower 4 signed or unsigned 16-bit integers in a with the lower 4 signed or unsigned 16-bit integers in b. https://msdn.microsoft.com/en-us/library/btxb17bw%28v=vs.90%29.aspx
// Interleaves the lower 2 signed or unsigned 32 - bit integers in a with the lower 2 signed or unsigned 32 - bit integers in b. https://msdn.microsoft.com/en-us/library/x8atst9d(v=vs.100).aspx
// Selects and interleaves the lower two single-precision, floating-point values from a and b. https://msdn.microsoft.com/en-us/library/25st103b%28v=vs.90%29.aspx
// Selects and interleaves the upper two single-precision, floating-point values from a and b. https://msdn.microsoft.com/en-us/library/skccxx7d%28v=vs.90%29.aspx
// Interleaves the upper 2 signed or unsigned 32-bit integers in a with the upper 2 signed or unsigned 32-bit integers in b. https://msdn.microsoft.com/en-us/library/65sa7cbs(v=vs.100).aspx
// Guarantees that every preceding store is globally visible before any subsequent store. https://msdn.microsoft.com/en-us/library/5h2w73d1%28v=vs.90%29.aspx
FORCE_INLINEvoid_mm_sfence(void)
{
__sync_synchronize();
}
// Stores the data in a to the address p without polluting the caches. If the cache line containing address p is already in the cache, the cache will be updated.Address p must be 16 - byte aligned. https://msdn.microsoft.com/en-us/library/ba08y07y%28v=vs.90%29.aspx
