| 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 |
|
|
#ifndef _COS_SIN_H_
|
| 43 |
|
|
#define _COS_SIN_H_ 1
|
| 44 |
|
|
|
| 45 |
|
|
#define M_PI_OVER_4_HI_32 0x3fe921fb
|
| 46 |
|
|
|
| 47 |
|
|
#define M_PI_OVER_4 0.78539816339744827900
|
| 48 |
|
|
#define M_FOUR_OVER_PI 1.27323954478442180616
|
| 49 |
|
|
|
| 50 |
|
|
#define M_PI_OVER_2 1.57079632679489655800
|
| 51 |
|
|
#define M_PI_OVER_2_HI 1.57079632673412561417
|
| 52 |
|
|
#define M_PI_OVER_2_LO 0.0000000000607710050650619224932
|
| 53 |
|
|
|
| 54 |
|
|
#define M_PI_OVER_2F_HI 1.570312500000000000
|
| 55 |
|
|
#define M_PI_OVER_2F_LO 0.000483826794896558
|
| 56 |
|
|
|
| 57 |
|
|
/* The following coefficients correspond to the Taylor series
|
| 58 |
|
|
* coefficients for cos and sin.
|
| 59 |
|
|
*/
|
| 60 |
|
|
#define COS_14 -0.00000000001138218794258068723867
|
| 61 |
|
|
#define COS_12 0.000000002087614008917893178252
|
| 62 |
|
|
#define COS_10 -0.0000002755731724204127572108
|
| 63 |
|
|
#define COS_08 0.00002480158729870839541888
|
| 64 |
|
|
#define COS_06 -0.001388888888888735934799
|
| 65 |
|
|
#define COS_04 0.04166666666666666534980
|
| 66 |
|
|
#define COS_02 -0.5000000000000000000000
|
| 67 |
|
|
#define COS_00 1.0
|
| 68 |
|
|
|
| 69 |
|
|
#define SIN_15 -0.00000000000076471637318198164759
|
| 70 |
|
|
#define SIN_13 0.00000000016059043836821614599
|
| 71 |
|
|
#define SIN_11 -0.000000025052108385441718775
|
| 72 |
|
|
#define SIN_09 0.0000027557319223985890653
|
| 73 |
|
|
#define SIN_07 -0.0001984126984126984127
|
| 74 |
|
|
#define SIN_05 0.008333333333333333333
|
| 75 |
|
|
#define SIN_03 -0.16666666666666666666
|
| 76 |
|
|
#define SIN_01 1.0
|
| 77 |
|
|
|
| 78 |
|
|
|
| 79 |
|
|
/* Compute the following for each floating point element of x.
|
| 80 |
|
|
* x = fmod(x, PI/4);
|
| 81 |
|
|
* ix = (int)x * PI/4;
|
| 82 |
|
|
* This allows one to compute cos / sin over the limited range
|
| 83 |
|
|
* and select the sign and correct result based upon the octant
|
| 84 |
|
|
* of the original angle (as defined by the ix result).
|
| 85 |
|
|
*
|
| 86 |
|
|
* Expected Inputs Types:
|
| 87 |
|
|
* x = vec_float4
|
| 88 |
|
|
* ix = vec_int4
|
| 89 |
|
|
*/
|
| 90 |
|
|
#define MOD_PI_OVER_FOUR_F(_x, _ix) { \
|
| 91 |
|
|
vec_float4 fx; \
|
| 92 |
|
|
\
|
| 93 |
|
|
_ix = spu_convts(spu_mul(_x, spu_splats((float)M_FOUR_OVER_PI)), 0); \
|
| 94 |
|
|
_ix = spu_add(_ix, spu_add(spu_rlmaska((vec_int4)_x, -31), 1)); \
|
| 95 |
|
|
\
|
| 96 |
|
|
fx = spu_convtf(spu_rlmaska(_ix, -1), 0); \
|
| 97 |
|
|
_x = spu_nmsub(fx, spu_splats((float)M_PI_OVER_2F_HI), _x); \
|
| 98 |
|
|
_x = spu_nmsub(fx, spu_splats((float)M_PI_OVER_2F_LO), _x); \
|
| 99 |
|
|
}
|
| 100 |
|
|
|
| 101 |
|
|
/* Double precision MOD_PI_OVER_FOUR
|
| 102 |
|
|
*
|
| 103 |
|
|
* Expected Inputs Types:
|
| 104 |
|
|
* x = vec_double2
|
| 105 |
|
|
* ix = vec_int4
|
| 106 |
|
|
*/
|
| 107 |
|
|
#define MOD_PI_OVER_FOUR(_x, _ix) { \
|
| 108 |
|
|
vec_float4 fx; \
|
| 109 |
|
|
vec_double2 dix; \
|
| 110 |
|
|
\
|
| 111 |
|
|
fx = spu_roundtf(spu_mul(_x, spu_splats(M_FOUR_OVER_PI))); \
|
| 112 |
|
|
_ix = spu_convts(fx, 0); \
|
| 113 |
|
|
_ix = spu_add(_ix, spu_add(spu_rlmaska((vec_int4)fx, -31), 1)); \
|
| 114 |
|
|
\
|
| 115 |
|
|
dix = spu_extend(spu_convtf(spu_rlmaska(_ix, -1), 0)); \
|
| 116 |
|
|
_x = spu_nmsub(spu_splats(M_PI_OVER_2_HI), dix, _x); \
|
| 117 |
|
|
_x = spu_nmsub(spu_splats(M_PI_OVER_2_LO), dix, _x); \
|
| 118 |
|
|
}
|
| 119 |
|
|
|
| 120 |
|
|
|
| 121 |
|
|
/* Compute the cos(x) and sin(x) for the range reduced angle x.
|
| 122 |
|
|
* In order to compute these trig functions to full single precision
|
| 123 |
|
|
* accuracy, we solve the Taylor series.
|
| 124 |
|
|
*
|
| 125 |
|
|
* c = cos(x) = 1 - x^2/2! + x^4/4! - x^6/6! + x^8/8! - x^10/10!
|
| 126 |
|
|
* s = sin(x) = x - x^3/4! + x^5/5! - x^7/7! + x^9/9! - x^11/11!
|
| 127 |
|
|
*
|
| 128 |
|
|
* Expected Inputs Types:
|
| 129 |
|
|
* x = vec_float4
|
| 130 |
|
|
* c = vec_float4
|
| 131 |
|
|
* s = vec_float4
|
| 132 |
|
|
*/
|
| 133 |
|
|
|
| 134 |
|
|
#define COMPUTE_COS_SIN_F(_x, _c, _s) { \
|
| 135 |
|
|
vec_float4 x2, x4, x6; \
|
| 136 |
|
|
vec_float4 cos_hi, cos_lo; \
|
| 137 |
|
|
vec_float4 sin_hi, sin_lo; \
|
| 138 |
|
|
\
|
| 139 |
|
|
x2 = spu_mul(_x, _x); \
|
| 140 |
|
|
x4 = spu_mul(x2, x2); \
|
| 141 |
|
|
x6 = spu_mul(x2, x4); \
|
| 142 |
|
|
\
|
| 143 |
|
|
cos_hi = spu_madd(spu_splats((float)COS_10), x2, spu_splats((float)COS_08)); \
|
| 144 |
|
|
cos_lo = spu_madd(spu_splats((float)COS_04), x2, spu_splats((float)COS_02)); \
|
| 145 |
|
|
cos_hi = spu_madd(cos_hi, x2, spu_splats((float)COS_06)); \
|
| 146 |
|
|
cos_lo = spu_madd(cos_lo, x2, spu_splats((float)COS_00)); \
|
| 147 |
|
|
_c = spu_madd(cos_hi, x6, cos_lo); \
|
| 148 |
|
|
\
|
| 149 |
|
|
sin_hi = spu_madd(spu_splats((float)SIN_11), x2, spu_splats((float)SIN_09)); \
|
| 150 |
|
|
sin_lo = spu_madd(spu_splats((float)SIN_05), x2, spu_splats((float)SIN_03)); \
|
| 151 |
|
|
sin_hi = spu_madd(sin_hi, x2, spu_splats((float)SIN_07)); \
|
| 152 |
|
|
sin_lo = spu_madd(sin_lo, x2, spu_splats((float)SIN_01)); \
|
| 153 |
|
|
_s = spu_madd(sin_hi, x6, sin_lo); \
|
| 154 |
|
|
_s = spu_mul(_s, _x); \
|
| 155 |
|
|
}
|
| 156 |
|
|
|
| 157 |
|
|
|
| 158 |
|
|
/* Compute the cos(x) and sin(x) for the range reduced angle x.
|
| 159 |
|
|
* This version computes the cosine and sine to double precision
|
| 160 |
|
|
* accuracy using the Taylor series:
|
| 161 |
|
|
*
|
| 162 |
|
|
* c = cos(x) = 1 - x^2/2! + x^4/4! - x^6/6! + x^8/8! - x^10/10! + x^12/12! - x^14/14!
|
| 163 |
|
|
* s = sin(x) = x - x^3/4! + x^5/5! - x^7/7! + x^9/9! - x^11/11! + x^13/13! - x^15/15!
|
| 164 |
|
|
*
|
| 165 |
|
|
* Expected Inputs Types:
|
| 166 |
|
|
* x = vec_double2
|
| 167 |
|
|
* c = vec_double2
|
| 168 |
|
|
* s = vec_double2
|
| 169 |
|
|
*/
|
| 170 |
|
|
|
| 171 |
|
|
#define COMPUTE_COS_SIN(_x, _c, _s) { \
|
| 172 |
|
|
vec_double2 x2, x4, x8; \
|
| 173 |
|
|
vec_double2 cos_hi, cos_lo; \
|
| 174 |
|
|
vec_double2 sin_hi, sin_lo; \
|
| 175 |
|
|
\
|
| 176 |
|
|
x2 = spu_mul(_x, _x); \
|
| 177 |
|
|
x4 = spu_mul(x2, x2); \
|
| 178 |
|
|
x8 = spu_mul(x4, x4); \
|
| 179 |
|
|
\
|
| 180 |
|
|
cos_hi = spu_madd(spu_splats(COS_14), x2, spu_splats(COS_12)); \
|
| 181 |
|
|
cos_lo = spu_madd(spu_splats(COS_06), x2, spu_splats(COS_04)); \
|
| 182 |
|
|
cos_hi = spu_madd(cos_hi, x2, spu_splats(COS_10)); \
|
| 183 |
|
|
cos_lo = spu_madd(cos_lo, x2, spu_splats(COS_02)); \
|
| 184 |
|
|
cos_hi = spu_madd(cos_hi, x2, spu_splats(COS_08)); \
|
| 185 |
|
|
cos_lo = spu_madd(cos_lo, x2, spu_splats(COS_00)); \
|
| 186 |
|
|
_c = spu_madd(cos_hi, x8, cos_lo); \
|
| 187 |
|
|
\
|
| 188 |
|
|
sin_hi = spu_madd(spu_splats(SIN_15), x2, spu_splats(SIN_13)); \
|
| 189 |
|
|
sin_lo = spu_madd(spu_splats(SIN_07), x2, spu_splats(SIN_05)); \
|
| 190 |
|
|
sin_hi = spu_madd(sin_hi, x2, spu_splats(SIN_11)); \
|
| 191 |
|
|
sin_lo = spu_madd(sin_lo, x2, spu_splats(SIN_03)); \
|
| 192 |
|
|
sin_hi = spu_madd(sin_hi, x2, spu_splats(SIN_09)); \
|
| 193 |
|
|
sin_lo = spu_madd(sin_lo, x2, spu_splats(SIN_01)); \
|
| 194 |
|
|
_s = spu_madd(sin_hi, x8, sin_lo); \
|
| 195 |
|
|
_s = spu_mul(_s, _x); \
|
| 196 |
|
|
}
|
| 197 |
|
|
|
| 198 |
|
|
|
| 199 |
|
|
|
| 200 |
|
|
|
| 201 |
|
|
#endif /* _COS_SIN_H_ */
|
| 202 |
|
|
#endif /* __SPU__ */
|
| 203 |
|
|
|
| 204 |
|
|
|
