log1p.cl
5.81 KB
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
/*
* Copyright (c) 2014 Advanced Micro Devices, Inc.
*
* Permission is hereby granted, free of charge, to any person obtaining a copy
* of this software and associated documentation files (the "Software"), to deal
* in the Software without restriction, including without limitation the rights
* to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
* copies of the Software, and to permit persons to whom the Software is
* furnished to do so, subject to the following conditions:
*
* The above copyright notice and this permission notice shall be included in
* all copies or substantial portions of the Software.
*
* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
* IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
* FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
* AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
* LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
* OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
* THE SOFTWARE.
*/
#include <clc/clc.h>
#include "math.h"
#include "tables.h"
#include "../clcmacro.h"
_CLC_OVERLOAD _CLC_DEF float log1p(float x)
{
float w = x;
uint ux = as_uint(x);
uint ax = ux & EXSIGNBIT_SP32;
// |x| < 2^-4
float u2 = MATH_DIVIDE(x, 2.0f + x);
float u = u2 + u2;
float v = u * u;
// 2/(5 * 2^5), 2/(3 * 2^3)
float zsmall = mad(-u2, x, mad(v, 0x1.99999ap-7f, 0x1.555556p-4f) * v * u) + x;
// |x| >= 2^-4
ux = as_uint(x + 1.0f);
int m = (int)((ux >> EXPSHIFTBITS_SP32) & 0xff) - EXPBIAS_SP32;
float mf = (float)m;
uint indx = (ux & 0x007f0000) + ((ux & 0x00008000) << 1);
float F = as_float(indx | 0x3f000000);
// x > 2^24
float fg24 = F - as_float(0x3f000000 | (ux & MANTBITS_SP32));
// x <= 2^24
uint xhi = ux & 0xffff8000;
float xh = as_float(xhi);
float xt = (1.0f - xh) + w;
uint xnm = ((~(xhi & 0x7f800000)) - 0x00800000) & 0x7f800000;
xt = xt * as_float(xnm) * 0.5f;
float fl24 = F - as_float(0x3f000000 | (xhi & MANTBITS_SP32)) - xt;
float f = mf > 24.0f ? fg24 : fl24;
indx = indx >> 16;
float r = f * USE_TABLE(log_inv_tbl, indx);
// 1/3, 1/2
float poly = mad(mad(r, 0x1.555556p-2f, 0x1.0p-1f), r*r, r);
const float LOG2_HEAD = 0x1.62e000p-1f; // 0.693115234
const float LOG2_TAIL = 0x1.0bfbe8p-15f; // 0.0000319461833
float2 tv = USE_TABLE(loge_tbl, indx);
float z1 = mad(mf, LOG2_HEAD, tv.s0);
float z2 = mad(mf, LOG2_TAIL, -poly) + tv.s1;
float z = z1 + z2;
z = ax < 0x3d800000U ? zsmall : z;
// Edge cases
z = ax >= PINFBITPATT_SP32 ? w : z;
z = w < -1.0f ? as_float(QNANBITPATT_SP32) : z;
z = w == -1.0f ? as_float(NINFBITPATT_SP32) : z;
//fix subnormals
z = ax < 0x33800000 ? x : z;
return z;
}
_CLC_UNARY_VECTORIZE(_CLC_OVERLOAD _CLC_DEF, float, log1p, float);
#ifdef cl_khr_fp64
#pragma OPENCL EXTENSION cl_khr_fp64 : enable
_CLC_OVERLOAD _CLC_DEF double log1p(double x)
{
// Computes natural log(1+x). Algorithm based on:
// Ping-Tak Peter Tang
// "Table-driven implementation of the logarithm function in IEEE
// floating-point arithmetic"
// ACM Transactions on Mathematical Software (TOMS)
// Volume 16, Issue 4 (December 1990)
// Note that we use a lookup table of size 64 rather than 128,
// and compensate by having extra terms in the minimax polynomial
// for the kernel approximation.
// Process Inside the threshold now
ulong ux = as_ulong(1.0 + x);
int xexp = ((as_int2(ux).hi >> 20) & 0x7ff) - EXPBIAS_DP64;
double f = as_double(ONEEXPBITS_DP64 | (ux & MANTBITS_DP64));
int j = as_int2(ux).hi >> 13;
j = ((0x80 | (j & 0x7e)) >> 1) + (j & 0x1);
double f1 = (double)j * 0x1.0p-6;
j -= 64;
double f2temp = f - f1;
double m2 = as_double(convert_ulong(0x3ff - xexp) << EXPSHIFTBITS_DP64);
double f2l = fma(m2, x, m2 - f1);
double f2g = fma(m2, x, -f1) + m2;
double f2 = xexp <= MANTLENGTH_DP64-1 ? f2l : f2g;
f2 = (xexp <= -2) | (xexp >= MANTLENGTH_DP64+8) ? f2temp : f2;
double2 tv = USE_TABLE(ln_tbl, j);
double z1 = tv.s0;
double q = tv.s1;
double u = MATH_DIVIDE(f2, fma(0.5, f2, f1));
double v = u * u;
double poly = v * fma(v,
fma(v, 2.23219810758559851206e-03, 1.24999999978138668903e-02),
8.33333333333333593622e-02);
// log2_lead and log2_tail sum to an extra-precise version of log(2)
const double log2_lead = 6.93147122859954833984e-01; /* 0x3fe62e42e0000000 */
const double log2_tail = 5.76999904754328540596e-08; /* 0x3e6efa39ef35793c */
double z2 = q + fma(u, poly, u);
double dxexp = (double)xexp;
double r1 = fma(dxexp, log2_lead, z1);
double r2 = fma(dxexp, log2_tail, z2);
double result1 = r1 + r2;
// Process Outside the threshold now
double r = x;
u = r / (2.0 + r);
double correction = r * u;
u = u + u;
v = u * u;
r1 = r;
poly = fma(v,
fma(v,
fma(v, 4.34887777707614552256e-04, 2.23213998791944806202e-03),
1.25000000037717509602e-02),
8.33333333333317923934e-02);
r2 = fma(u*v, poly, -correction);
// The values exp(-1/16)-1 and exp(1/16)-1
const double log1p_thresh1 = -0x1.f0540438fd5c3p-5;
const double log1p_thresh2 = 0x1.082b577d34ed8p-4;
double result2 = r1 + r2;
result2 = x < log1p_thresh1 | x > log1p_thresh2 ? result1 : result2;
result2 = isinf(x) ? x : result2;
result2 = x < -1.0 ? as_double(QNANBITPATT_DP64) : result2;
result2 = x == -1.0 ? as_double(NINFBITPATT_DP64) : result2;
return result2;
}
_CLC_UNARY_VECTORIZE(_CLC_OVERLOAD _CLC_DEF, double, log1p, double);
#endif // cl_khr_fp64