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 _ATANF4_H_
|
44 |
|
|
#define _ATANF4_H_ 1
|
45 |
|
|
|
46 |
|
|
#include <spu_intrinsics.h>
|
47 |
|
|
|
48 |
|
|
#include "simdmath.h"
|
49 |
|
|
#include "recipf4.h"
|
50 |
|
|
|
51 |
|
|
/*
|
52 |
|
|
* FUNCTION
|
53 |
|
|
* vector float _atanf4(vector float x)
|
54 |
|
|
*
|
55 |
|
|
* DESCRIPTION
|
56 |
|
|
* The _atanf4 function computes the arc tangent of a vector of values x;
|
57 |
|
|
* that is the values whose tangent is x.
|
58 |
|
|
*
|
59 |
|
|
* The _atanf4 function returns the arc tangents in radians and the value
|
60 |
|
|
* is mathematically defined to be in the range -PI/2 to PI/2.
|
61 |
|
|
*
|
62 |
|
|
* The arc tangent function is computed using a polynomial approximation
|
63 |
|
|
* (B. Carlson, M. Goldstein, Los Alamos Scientific Laboratory, 1955).
|
64 |
|
|
* __8__
|
65 |
|
|
* \
|
66 |
|
|
* \
|
67 |
|
|
* atan(x) = / Ci*x^(2*i+1)
|
68 |
|
|
* /____
|
69 |
|
|
* i=0
|
70 |
|
|
*
|
71 |
|
|
* for x in the range -1 to 1. The remaining regions are defined to be:
|
72 |
|
|
*
|
73 |
|
|
* [1, infinity] : PI/2 + atanf(-1/x)
|
74 |
|
|
* [-infinity, -1] : -PI/2 + atanf(-1/x)
|
75 |
|
|
*
|
76 |
|
|
*/
|
77 |
|
|
static __inline vector float _atanf4(vector float x)
|
78 |
|
|
{
|
79 |
|
|
vector float bias;
|
80 |
|
|
vector float x2, x3, x4, x8, x9;
|
81 |
|
|
vector float hi, lo;
|
82 |
|
|
vector float result;
|
83 |
|
|
vector float inv_x;
|
84 |
|
|
vector unsigned int sign;
|
85 |
|
|
vector unsigned int select;
|
86 |
|
|
|
87 |
|
|
sign = spu_sl(spu_rlmask((vector unsigned int)x, -31), 31);
|
88 |
|
|
inv_x = _recipf4(x);
|
89 |
|
|
inv_x = (vector float)spu_xor((vector unsigned int)inv_x, spu_splats(0x80000000));
|
90 |
|
|
|
91 |
|
|
select = (vector unsigned int)spu_cmpabsgt(x, spu_splats(1.0f));
|
92 |
|
|
bias = (vector float)spu_or(sign, (vector unsigned int)(spu_splats((float)SM_PI_2)));
|
93 |
|
|
bias = (vector float)spu_and((vector unsigned int)bias, select);
|
94 |
|
|
|
95 |
|
|
x = spu_sel(x, inv_x, select);
|
96 |
|
|
|
97 |
|
|
/* Instruction counts can be reduced if the polynomial was
|
98 |
|
|
* computed entirely from nested (dependent) fma's. However,
|
99 |
|
|
* to reduce the number of pipeline stalls, the polygon is evaluated
|
100 |
|
|
* in two halves(hi and lo).
|
101 |
|
|
*/
|
102 |
|
|
bias = spu_add(bias, x);
|
103 |
|
|
x2 = spu_mul(x, x);
|
104 |
|
|
x3 = spu_mul(x2, x);
|
105 |
|
|
x4 = spu_mul(x2, x2);
|
106 |
|
|
x8 = spu_mul(x4, x4);
|
107 |
|
|
x9 = spu_mul(x8, x);
|
108 |
|
|
hi = spu_madd(spu_splats(0.0028662257f), x2, spu_splats(-0.0161657367f));
|
109 |
|
|
hi = spu_madd(hi, x2, spu_splats(0.0429096138f));
|
110 |
|
|
hi = spu_madd(hi, x2, spu_splats(-0.0752896400f));
|
111 |
|
|
hi = spu_madd(hi, x2, spu_splats(0.1065626393f));
|
112 |
|
|
lo = spu_madd(spu_splats(-0.1420889944f), x2, spu_splats(0.1999355085f));
|
113 |
|
|
lo = spu_madd(lo, x2, spu_splats(-0.3333314528f));
|
114 |
|
|
lo = spu_madd(lo, x3, bias);
|
115 |
|
|
|
116 |
|
|
result = spu_madd(hi, x9, lo);
|
117 |
|
|
|
118 |
|
|
return (result);
|
119 |
|
|
}
|
120 |
|
|
|
121 |
|
|
#endif /* _ATANF4_H_ */
|
122 |
|
|
#endif /* __SPU__ */
|
123 |
|
|
|
124 |
|
|
|