7529cdb3f6
The upcoming vectormath that will used to speed up the SPU version of Extras/BulletMultiThreaded depends on this.
96 lines
4.6 KiB
C
96 lines
4.6 KiB
C
/* frexpd2 - for each element of vector x, return the normalized fraction and store the exponent of x'
|
|
Copyright (C) 2006, 2007 Sony Computer Entertainment Inc.
|
|
All rights reserved.
|
|
|
|
Redistribution and use in source and binary forms,
|
|
with or without modification, are permitted provided that the
|
|
following conditions are met:
|
|
* Redistributions of source code must retain the above copyright
|
|
notice, this list of conditions and the following disclaimer.
|
|
* Redistributions in binary form must reproduce the above copyright
|
|
notice, this list of conditions and the following disclaimer in the
|
|
documentation and/or other materials provided with the distribution.
|
|
* Neither the name of the Sony Computer Entertainment Inc nor the names
|
|
of its contributors may be used to endorse or promote products derived
|
|
from this software without specific prior written permission.
|
|
|
|
THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
|
|
AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
|
|
IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
|
|
ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE
|
|
LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
|
|
CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
|
|
SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
|
|
INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
|
|
CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
|
|
ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
|
|
POSSIBILITY OF SUCH DAMAGE.
|
|
*/
|
|
|
|
#include <simdmath.h>
|
|
#include <spu_intrinsics.h>
|
|
#include <math.h>
|
|
|
|
#ifndef DBL_NAN
|
|
#define DBL_NAN ((long long)0x7FF8000000000000ull)
|
|
#endif
|
|
|
|
vector double
|
|
frexpd2 (vector double x, vector signed long long *pexp)
|
|
{
|
|
vec_uchar16 even = (vec_uchar16)(vec_uint4){ 0x00010203, 0x00010203, 0x08090a0b, 0x08090a0b };
|
|
vec_uchar16 swapEvenOdd = (vec_uchar16)(vec_uint4){ 0x04050607, 0x00010203, 0x0c0d0e0f, 0x08090a0b };
|
|
|
|
vec_ullong2 maskdw = (vec_ullong2){0xffffffffffffffffull, 0ull};
|
|
|
|
vec_ullong2 sign = spu_splats(0x8000000000000000ull);
|
|
vec_ullong2 expn = spu_splats(0x7ff0000000000000ull);
|
|
vec_ullong2 zero = spu_splats(0x0000000000000000ull);
|
|
|
|
vec_ullong2 isnan, isinf, iszero;
|
|
vec_ullong2 e0, x0, x1;
|
|
vec_uint4 cmpgt, cmpeq, cmpzr;
|
|
vec_int4 lz, lz0, sh, ex;
|
|
vec_double2 fr, frac = (vec_double2)zero;
|
|
|
|
//NAN: x is NaN (all-ones exponent and non-zero mantissa)
|
|
cmpgt = spu_cmpgt( (vec_uint4)spu_or( (vec_ullong2)x, sign ), (vec_uint4)spu_or(sign, expn) );
|
|
cmpeq = spu_cmpeq( (vec_uint4)spu_or( (vec_ullong2)x, sign ), (vec_uint4)spu_or(sign, expn) );
|
|
isnan = (vec_ullong2)spu_or( cmpgt, spu_and( cmpeq, spu_rlqwbyte( cmpgt, -4 ) ) );
|
|
isnan = (vec_ullong2)spu_shuffle( isnan, isnan, even );
|
|
frac = spu_sel( frac, (vec_double2)spu_splats((long long)DBL_NAN), isnan );
|
|
|
|
//INF: x is infinite (all-ones exponent and zero mantissa)
|
|
isinf = (vec_ullong2)spu_and( cmpeq, spu_shuffle( cmpeq, cmpeq, swapEvenOdd ) );
|
|
frac = spu_sel( frac, x , isinf );
|
|
|
|
//x is zero (zero exponent and zero mantissa)
|
|
cmpzr = spu_cmpeq( (vec_uint4)spu_andc( (vec_ullong2)x, sign ), (vec_uint4)zero );
|
|
iszero = (vec_ullong2)spu_and( cmpzr, spu_shuffle( cmpzr, cmpzr, swapEvenOdd ) );
|
|
|
|
frac = spu_sel( frac, (vec_double2)zero , iszero );
|
|
*pexp = spu_sel( *pexp, (vec_llong2)zero , iszero );
|
|
|
|
//Integer Exponent: if x is normal or subnormal
|
|
|
|
//...shift left to normalize fraction, zero shift if normal
|
|
lz = (vec_int4)spu_cntlz( (vec_uint4)spu_andc( (vec_ullong2)x, sign) );
|
|
lz0 = (vec_int4)spu_shuffle( lz, lz, even );
|
|
sh = spu_sel( (vec_int4)zero, spu_sub( lz0, spu_splats((int)11) ), spu_cmpgt( lz0, (int)11 ) );
|
|
sh = spu_sel( sh, spu_add( sh, lz ), spu_cmpeq( lz0, (int)32 ) );
|
|
|
|
x0 = spu_slqw( spu_slqwbytebc( spu_and( (vec_ullong2)x, maskdw ), spu_extract(sh, 1) ), spu_extract(sh, 1) );
|
|
x1 = spu_slqw( spu_slqwbytebc( (vec_ullong2)x, spu_extract(sh, 3) ), spu_extract(sh, 3) );
|
|
fr = (vec_double2)spu_sel( x1, x0, maskdw );
|
|
fr = spu_sel( fr, (vec_double2)spu_splats(0x3FE0000000000000ull), expn );
|
|
fr = spu_sel( fr, x, sign );
|
|
|
|
e0 = spu_rlmaskqw( spu_rlmaskqwbyte(spu_and( (vec_ullong2)x, expn ),-6), -4 );
|
|
ex = spu_sel( spu_sub( (vec_int4)e0, spu_splats((int)1022) ), spu_sub( spu_splats((int)-1021), sh ), spu_cmpgt( sh, (int)0 ) );
|
|
|
|
frac = spu_sel( frac, fr, spu_nor( isnan, spu_or( isinf, iszero ) ) );
|
|
*pexp = spu_sel( *pexp, spu_extend( ex ), spu_nor( isnan, spu_or( isinf, iszero ) ) );
|
|
|
|
return frac;
|
|
}
|