| 1 |
9 |
jlechner |
/* -------------------------------------------------------------- */
|
| 2 |
|
|
/* (C)Copyright 2006,2008, */
|
| 3 |
|
|
/* International Business Machines Corporation */
|
| 4 |
|
|
/* All Rights Reserved. */
|
| 5 |
|
|
/* */
|
| 6 |
|
|
/* Redistribution and use in source and binary forms, with or */
|
| 7 |
|
|
/* without modification, are permitted provided that the */
|
| 8 |
|
|
/* following conditions are met: */
|
| 9 |
|
|
/* */
|
| 10 |
|
|
/* - Redistributions of source code must retain the above copyright*/
|
| 11 |
|
|
/* notice, this list of conditions and the following disclaimer. */
|
| 12 |
|
|
/* */
|
| 13 |
|
|
/* - Redistributions in binary form must reproduce the above */
|
| 14 |
|
|
/* copyright notice, this list of conditions and the following */
|
| 15 |
|
|
/* disclaimer in the documentation and/or other materials */
|
| 16 |
|
|
/* provided with the distribution. */
|
| 17 |
|
|
/* */
|
| 18 |
|
|
/* - Neither the name of IBM Corporation nor the names of its */
|
| 19 |
|
|
/* contributors may be used to endorse or promote products */
|
| 20 |
|
|
/* derived from this software without specific prior written */
|
| 21 |
|
|
/* permission. */
|
| 22 |
|
|
/* */
|
| 23 |
|
|
/* THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND */
|
| 24 |
|
|
/* CONTRIBUTORS "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, */
|
| 25 |
|
|
/* INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF */
|
| 26 |
|
|
/* MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE */
|
| 27 |
|
|
/* DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR */
|
| 28 |
|
|
/* CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, */
|
| 29 |
|
|
/* SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT */
|
| 30 |
|
|
/* NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; */
|
| 31 |
|
|
/* LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) */
|
| 32 |
|
|
/* HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN */
|
| 33 |
|
|
/* CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR */
|
| 34 |
|
|
/* OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, */
|
| 35 |
|
|
/* EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. */
|
| 36 |
|
|
/* -------------------------------------------------------------- */
|
| 37 |
|
|
/* PROLOG END TAG zYx */
|
| 38 |
|
|
#ifdef __SPU__
|
| 39 |
|
|
#ifndef _COSHF4_H_
|
| 40 |
|
|
#define _COSHF4_H_ 1
|
| 41 |
|
|
|
| 42 |
|
|
#include <spu_intrinsics.h>
|
| 43 |
|
|
|
| 44 |
|
|
#include "expf4.h"
|
| 45 |
|
|
#include "recipf4.h"
|
| 46 |
|
|
|
| 47 |
|
|
|
| 48 |
|
|
/*
|
| 49 |
|
|
* FUNCTION
|
| 50 |
|
|
* vector float _coshf4(vector float x)
|
| 51 |
|
|
*
|
| 52 |
|
|
* DESCRIPTION
|
| 53 |
|
|
* The _coshf4 function computes the hyperbolic cosines of a vector of
|
| 54 |
|
|
* angles (expressed in radians) to an accuracy of a single precision
|
| 55 |
|
|
* floating point.
|
| 56 |
|
|
*
|
| 57 |
|
|
*/
|
| 58 |
|
|
static __inline vector float _coshf4(vector float x)
|
| 59 |
|
|
{
|
| 60 |
|
|
// 1.0000 (above this number, use sinh(x) = 0.5 * (e^x - e^-x)
|
| 61 |
|
|
vec_uint4 threshold = (vec_uint4)spu_splats(0x3F800000);
|
| 62 |
|
|
|
| 63 |
|
|
vec_uint4 sign_mask = (vec_uint4)spu_splats(0x80000000);
|
| 64 |
|
|
|
| 65 |
|
|
// Coefficents for the Taylor series
|
| 66 |
|
|
vec_float4 f02 = spu_splats(5.0000000000000000E-1f); // 1/2!
|
| 67 |
|
|
vec_float4 f04 = spu_splats(4.1666666666666667E-2f); // 1/4!
|
| 68 |
|
|
vec_float4 f06 = spu_splats(1.3888888888888889E-3f); // 1/6!
|
| 69 |
|
|
vec_float4 f08 = spu_splats(2.4801587301587302E-5f); // 1/8!
|
| 70 |
|
|
vec_float4 f10 = spu_splats(2.7557319223985891E-7f); // 1/10!
|
| 71 |
|
|
vec_float4 f12 = spu_splats(2.0876756987868099E-9f); // 1/12!
|
| 72 |
|
|
|
| 73 |
|
|
// Perform the calculation as a Taylor series
|
| 74 |
|
|
vec_float4 result;
|
| 75 |
|
|
vector float x2 = spu_mul(x,x);
|
| 76 |
|
|
result = spu_madd(x2,f12,f10);
|
| 77 |
|
|
result = spu_madd(x2,result,f08);
|
| 78 |
|
|
result = spu_madd(x2,result,f06);
|
| 79 |
|
|
result = spu_madd(x2,result,f04);
|
| 80 |
|
|
result = spu_madd(x2,result,f02);
|
| 81 |
|
|
result = spu_madd(x2,result,spu_splats(1.0f));
|
| 82 |
|
|
|
| 83 |
|
|
|
| 84 |
|
|
// Perform calculation as a function of 0.5 * (e^x - e^-x)
|
| 85 |
|
|
vec_float4 ex = _expf4(x);
|
| 86 |
|
|
vec_float4 ex_inv = _recipf4(ex);
|
| 87 |
|
|
|
| 88 |
|
|
vec_float4 r2= spu_add(ex,ex_inv);
|
| 89 |
|
|
r2 = spu_mul(r2,f02); // we can reused f02 here
|
| 90 |
|
|
|
| 91 |
|
|
vec_uint4 xabs = spu_andc((vec_uint4)x,sign_mask);
|
| 92 |
|
|
vec_uint4 use_exp = spu_cmpgt(xabs,threshold);
|
| 93 |
|
|
|
| 94 |
|
|
// Select either the Taylor or exp version
|
| 95 |
|
|
result = spu_sel(result,r2,use_exp);
|
| 96 |
|
|
|
| 97 |
|
|
return result;
|
| 98 |
|
|
}
|
| 99 |
|
|
#endif /* _COSHF4_H_ */
|
| 100 |
|
|
#endif /* __SPU__ */
|
