1 |
148 |
jeremybenn |
/* -------------------------------------------------------------- */
|
2 |
|
|
/* (C)Copyright 2007,2008, */
|
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 |
|
|
|
39 |
|
|
#ifdef __SPU__
|
40 |
|
|
#ifndef _LOG1PF4_H_
|
41 |
|
|
#define _LOG1PF4_H_ 1
|
42 |
|
|
|
43 |
|
|
#include <spu_intrinsics.h>
|
44 |
|
|
#include "simdmath.h"
|
45 |
|
|
|
46 |
|
|
#include "logf4.h"
|
47 |
|
|
#include "divf4.h"
|
48 |
|
|
|
49 |
|
|
/*
|
50 |
|
|
* FUNCTION
|
51 |
|
|
* vector float _log1pf4(vector float x)
|
52 |
|
|
*
|
53 |
|
|
* DESCRIPTION
|
54 |
|
|
* The function _log1pf4 computes the natural logarithm of x + 1
|
55 |
|
|
* for each of the float word elements of x.
|
56 |
|
|
*
|
57 |
|
|
*
|
58 |
|
|
*/
|
59 |
|
|
|
60 |
|
|
#define LOG1PF4_P0 0.0000000000000000000000000e+00f
|
61 |
|
|
#define LOG1PF4_P1 1.0000000000000000000000000e+00f
|
62 |
|
|
#define LOG1PF4_P2 1.4220868022897381610647471e+00f
|
63 |
|
|
#define LOG1PF4_P3 5.4254553902256308361984338e-01f
|
64 |
|
|
#define LOG1PF4_P4 4.5971908823142115796400731e-02f
|
65 |
|
|
|
66 |
|
|
#define LOG1PF4_Q0 1.0000000000000000000000000e+00f
|
67 |
|
|
#define LOG1PF4_Q1 1.9220868007537357247116461e+00f
|
68 |
|
|
#define LOG1PF4_Q2 1.1702556461286610645089468e+00f
|
69 |
|
|
#define LOG1PF4_Q3 2.4040413392943396631018516e-01f
|
70 |
|
|
#define LOG1PF4_Q4 1.0637426466449625625521058e-02f
|
71 |
|
|
|
72 |
|
|
|
73 |
|
|
static __inline vector float _log1pf4(vector float x)
|
74 |
|
|
{
|
75 |
|
|
vector float onef = spu_splats(1.0f);
|
76 |
|
|
vector float range = spu_splats(0.35f);
|
77 |
|
|
vector unsigned int use_log;
|
78 |
|
|
vector float pr, qr;
|
79 |
|
|
vector float eresult;
|
80 |
|
|
vector float rresult;
|
81 |
|
|
vector float result;
|
82 |
|
|
|
83 |
|
|
use_log = spu_cmpabsgt(x, range);
|
84 |
|
|
|
85 |
|
|
/*
|
86 |
|
|
* Calculate directly using log(x+1)
|
87 |
|
|
*/
|
88 |
|
|
eresult = _logf4(spu_add(x, onef));
|
89 |
|
|
|
90 |
|
|
/*
|
91 |
|
|
* For x in [-0.35,0.35], use a rational approximation.
|
92 |
|
|
*/
|
93 |
|
|
pr = spu_madd(x, spu_splats((float)LOG1PF4_P4), spu_splats((float)LOG1PF4_P3));
|
94 |
|
|
qr = spu_madd(x, spu_splats((float)LOG1PF4_Q4), spu_splats((float)LOG1PF4_Q3));
|
95 |
|
|
pr = spu_madd(pr, x, spu_splats((float)LOG1PF4_P2));
|
96 |
|
|
qr = spu_madd(qr, x, spu_splats((float)LOG1PF4_Q2));
|
97 |
|
|
pr = spu_madd(pr, x, spu_splats((float)LOG1PF4_P1));
|
98 |
|
|
qr = spu_madd(qr, x, spu_splats((float)LOG1PF4_Q1));
|
99 |
|
|
pr = spu_madd(pr, x, spu_splats((float)LOG1PF4_P0));
|
100 |
|
|
qr = spu_madd(qr, x, spu_splats((float)LOG1PF4_Q0));
|
101 |
|
|
rresult = _divf4(pr, qr);
|
102 |
|
|
|
103 |
|
|
/*
|
104 |
|
|
* Select either direct calculation or rational approximation.
|
105 |
|
|
*/
|
106 |
|
|
result = spu_sel(rresult, eresult, use_log);
|
107 |
|
|
|
108 |
|
|
return result;
|
109 |
|
|
}
|
110 |
|
|
|
111 |
|
|
#endif /* _LOG1PF4_H_ */
|
112 |
|
|
#endif /* __SPU__ */
|