| 1 |
207 |
jeremybenn |
/* -------------------------------------------------------------- */
|
| 2 |
|
|
/* (C)Copyright 2001,2008, */
|
| 3 |
|
|
/* International Business Machines Corporation, */
|
| 4 |
|
|
/* Sony Computer Entertainment, Incorporated, */
|
| 5 |
|
|
/* Toshiba Corporation, */
|
| 6 |
|
|
/* */
|
| 7 |
|
|
/* All Rights Reserved. */
|
| 8 |
|
|
/* */
|
| 9 |
|
|
/* Redistribution and use in source and binary forms, with or */
|
| 10 |
|
|
/* without modification, are permitted provided that the */
|
| 11 |
|
|
/* following conditions are met: */
|
| 12 |
|
|
/* */
|
| 13 |
|
|
/* - Redistributions of source code must retain the above copyright*/
|
| 14 |
|
|
/* notice, this list of conditions and the following disclaimer. */
|
| 15 |
|
|
/* */
|
| 16 |
|
|
/* - Redistributions in binary form must reproduce the above */
|
| 17 |
|
|
/* copyright notice, this list of conditions and the following */
|
| 18 |
|
|
/* disclaimer in the documentation and/or other materials */
|
| 19 |
|
|
/* provided with the distribution. */
|
| 20 |
|
|
/* */
|
| 21 |
|
|
/* - Neither the name of IBM Corporation nor the names of its */
|
| 22 |
|
|
/* contributors may be used to endorse or promote products */
|
| 23 |
|
|
/* derived from this software without specific prior written */
|
| 24 |
|
|
/* permission. */
|
| 25 |
|
|
/* */
|
| 26 |
|
|
/* THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND */
|
| 27 |
|
|
/* CONTRIBUTORS "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, */
|
| 28 |
|
|
/* INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF */
|
| 29 |
|
|
/* MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE */
|
| 30 |
|
|
/* DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR */
|
| 31 |
|
|
/* CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, */
|
| 32 |
|
|
/* SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT */
|
| 33 |
|
|
/* NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; */
|
| 34 |
|
|
/* LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) */
|
| 35 |
|
|
/* HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN */
|
| 36 |
|
|
/* CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR */
|
| 37 |
|
|
/* OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, */
|
| 38 |
|
|
/* EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. */
|
| 39 |
|
|
/* -------------------------------------------------------------- */
|
| 40 |
|
|
/* PROLOG END TAG zYx */
|
| 41 |
|
|
#ifdef __SPU__
|
| 42 |
|
|
|
| 43 |
|
|
#ifndef _LOG2F4_H_
|
| 44 |
|
|
#define _LOG2F4_H_ 1
|
| 45 |
|
|
|
| 46 |
|
|
#include <spu_intrinsics.h>
|
| 47 |
|
|
|
| 48 |
|
|
/*
|
| 49 |
|
|
* FUNCTION
|
| 50 |
|
|
* vector float _log2f4(vector float x)
|
| 51 |
|
|
*
|
| 52 |
|
|
* DESCRIPTION
|
| 53 |
|
|
* The _log2f4 function computes log (base 2) on a vector if inputs
|
| 54 |
|
|
* values x. The _log2f4 function is approximated as a polynomial of
|
| 55 |
|
|
* order 8 (C. Hastings, Jr, 1955).
|
| 56 |
|
|
*
|
| 57 |
|
|
* __8__
|
| 58 |
|
|
* \
|
| 59 |
|
|
* \
|
| 60 |
|
|
* log2f(1+x) = / Ci*x^i
|
| 61 |
|
|
* /____
|
| 62 |
|
|
* i=1
|
| 63 |
|
|
*
|
| 64 |
|
|
* for x in the range 0.0 to 1.0
|
| 65 |
|
|
*
|
| 66 |
|
|
* C1 = 1.4426898816672
|
| 67 |
|
|
* C2 = -0.72116591947498
|
| 68 |
|
|
* C3 = 0.47868480909345
|
| 69 |
|
|
* C4 = -0.34730547155299
|
| 70 |
|
|
* C5 = 0.24187369696082
|
| 71 |
|
|
* C6 = -0.13753123777116
|
| 72 |
|
|
* C7 = 0.052064690894143
|
| 73 |
|
|
* C8 = -0.0093104962134977
|
| 74 |
|
|
*
|
| 75 |
|
|
* This function assumes that x is a non-zero positive value.
|
| 76 |
|
|
*
|
| 77 |
|
|
*/
|
| 78 |
|
|
static __inline vector float _log2f4(vector float x)
|
| 79 |
|
|
{
|
| 80 |
|
|
vector signed int exponent;
|
| 81 |
|
|
vector float result;
|
| 82 |
|
|
vector float x2, x4;
|
| 83 |
|
|
vector float hi, lo;
|
| 84 |
|
|
|
| 85 |
|
|
/* Extract the exponent from the input X.
|
| 86 |
|
|
*/
|
| 87 |
|
|
exponent = (vector signed int)spu_and(spu_rlmask((vector unsigned int)(x), -23), 0xFF);
|
| 88 |
|
|
exponent = spu_add(exponent, -127);
|
| 89 |
|
|
|
| 90 |
|
|
/* Compute the remainder after removing the exponent.
|
| 91 |
|
|
*/
|
| 92 |
|
|
x = (vector float)spu_sub((vector signed int)(x), spu_sl(exponent, 23));
|
| 93 |
|
|
|
| 94 |
|
|
/* Calculate the log2 of the remainder using the polynomial
|
| 95 |
|
|
* approximation.
|
| 96 |
|
|
*/
|
| 97 |
|
|
x = spu_sub(x, spu_splats(1.0f));
|
| 98 |
|
|
|
| 99 |
|
|
/* Instruction counts can be reduced if the polynomial was
|
| 100 |
|
|
* computed entirely from nested (dependent) fma's. However,
|
| 101 |
|
|
* to reduce the number of pipeline stalls, the polygon is evaluated
|
| 102 |
|
|
* in two halves (hi amd lo).
|
| 103 |
|
|
*/
|
| 104 |
|
|
x2 = spu_mul(x, x);
|
| 105 |
|
|
x4 = spu_mul(x2, x2);
|
| 106 |
|
|
|
| 107 |
|
|
hi = spu_madd(x, spu_splats(-0.0093104962134977f), spu_splats(0.052064690894143f));
|
| 108 |
|
|
hi = spu_madd(x, hi, spu_splats(-0.13753123777116f));
|
| 109 |
|
|
hi = spu_madd(x, hi, spu_splats( 0.24187369696082f));
|
| 110 |
|
|
hi = spu_madd(x, hi, spu_splats(-0.34730547155299f));
|
| 111 |
|
|
lo = spu_madd(x, spu_splats(0.47868480909345f), spu_splats(-0.72116591947498f));
|
| 112 |
|
|
lo = spu_madd(x, lo, spu_splats(1.4426898816672f));
|
| 113 |
|
|
lo = spu_mul(x, lo);
|
| 114 |
|
|
result = spu_madd(x4, hi, lo);
|
| 115 |
|
|
|
| 116 |
|
|
/* Add the exponent back into the result.
|
| 117 |
|
|
*/
|
| 118 |
|
|
result = spu_add(result, spu_convtf(exponent, 0));
|
| 119 |
|
|
|
| 120 |
|
|
return (result);
|
| 121 |
|
|
}
|
| 122 |
|
|
|
| 123 |
|
|
#endif /* _LOG2F4_H_ */
|
| 124 |
|
|
#endif /* __SPU__ */
|
