xref: /illumos-gate/usr/src/lib/libm/common/Q/logl.c (revision ddc0e0b5)
1 /*
2  * CDDL HEADER START
3  *
4  * The contents of this file are subject to the terms of the
5  * Common Development and Distribution License (the "License").
6  * You may not use this file except in compliance with the License.
7  *
8  * You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
9  * or http://www.opensolaris.org/os/licensing.
10  * See the License for the specific language governing permissions
11  * and limitations under the License.
12  *
13  * When distributing Covered Code, include this CDDL HEADER in each
14  * file and include the License file at usr/src/OPENSOLARIS.LICENSE.
15  * If applicable, add the following below this CDDL HEADER, with the
16  * fields enclosed by brackets "[]" replaced with your own identifying
17  * information: Portions Copyright [yyyy] [name of copyright owner]
18  *
19  * CDDL HEADER END
20  */
21 
22 /*
23  * Copyright 2011 Nexenta Systems, Inc.  All rights reserved.
24  */
25 /*
26  * Copyright 2006 Sun Microsystems, Inc.  All rights reserved.
27  * Use is subject to license terms.
28  */
29 
30 #pragma weak __logl = logl
31 
32 /*
33  * logl(x)
34  * Table look-up algorithm
35  * By K.C. Ng, March 6, 1989
36  *
37  * (a). For x in [31/33,33/31], using a special approximation:
38  *	f = x - 1;
39  *	s = f/(2.0+f);	... here |s| <= 0.03125
40  *	z = s*s;
41  *	return f-s*(f-z*(B1+z*(B2+z*(B3+z*(B4+...+z*B9)...))));
42  *
43  * (b). Otherwise, normalize x = 2^n * 1.f.
44  *	Use a 6-bit table look-up: find a 6 bit g that match f to 6.5 bits,
45  *	then
46  *	    log(x) = n*ln2 + log(1.g) + log(1.f/1.g).
47  *	Here the leading and trailing values of log(1.g) are obtained from
48  *	a size-64 table.
49  *	For log(1.f/1.g), let s = (1.f-1.g)/(1.f+1.g), then
50  *	    log(1.f/1.g) = log((1+s)/(1-s)) = 2s + 2/3 s^3 + 2/5 s^5 +...
51  *	Note that |s|<2**-8=0.00390625. We use an odd s-polynomial
52  *	approximation to compute log(1.f/1.g):
53  *		s*(A1+s^2*(A2+s^2*(A3+s^2*(A4+s^2*(A5+s^2*(A6+s^2*A7))))))
54  *	(Precision is 2**-136.91 bits, absolute error)
55  *
56  * (c). The final result is computed by
57  *		(n*ln2_hi+_TBL_logl_hi[j]) +
58  *			( (n*ln2_lo+_TBL_logl_lo[j]) + s*(A1+...) )
59  *
60  * Note.
61  *	For ln2_hi and _TBL_logl_hi[j], we force their last 32 bit to be zero
62  *	so that n*ln2_hi + _TBL_logl_hi[j] is exact. Here
63  *	_TBL_logl_hi[j] + _TBL_logl_lo[j] match log(1+j*2**-6) to 194 bits
64  *
65  *
66  * Special cases:
67  *	log(x) is NaN with signal if x < 0 (including -INF) ;
68  *	log(+INF) is +INF; log(0) is -INF with signal;
69  *	log(NaN) is that NaN with no signal.
70  *
71  * Constants:
72  * The hexadecimal values are the intended ones for the following constants.
73  * The decimal values may be used, provided that the compiler will convert
74  * from decimal to binary accurately enough to produce the hexadecimal values
75  * shown.
76  */
77 
78 #include "libm.h"
79 
80 extern const long double _TBL_logl_hi[], _TBL_logl_lo[];
81 
82 static const long double
83 	zero	=   0.0L,
84 	one	=   1.0L,
85 	two	=   2.0L,
86 	two113  =   10384593717069655257060992658440192.0L,
87 	ln2hi	=   6.931471805599453094172319547495844850203e-0001L,
88 	ln2lo	=   1.667085920830552208890449330400379754169e-0025L,
89 	A1	=   2.000000000000000000000000000000000000024e+0000L,
90 	A2	=   6.666666666666666666666666666666091393804e-0001L,
91 	A3	=   4.000000000000000000000000407167070220671e-0001L,
92 	A4	=   2.857142857142857142730077490612903681164e-0001L,
93 	A5	=   2.222222222222242577702836920812882605099e-0001L,
94 	A6	=   1.818181816435493395985912667105885828356e-0001L,
95 	A7	=   1.538537835211839751112067512805496931725e-0001L,
96 	B1	=   6.666666666666666666666666666666961498329e-0001L,
97 	B2	=   3.999999999999999999999999990037655042358e-0001L,
98 	B3	=   2.857142857142857142857273426428347457918e-0001L,
99 	B4	=   2.222222222222222221353229049747910109566e-0001L,
100 	B5	=   1.818181818181821503532559306309070138046e-0001L,
101 	B6	=   1.538461538453809210486356084587356788556e-0001L,
102 	B7	=   1.333333344463358756121456892645178795480e-0001L,
103 	B8	=   1.176460904783899064854645174603360383792e-0001L,
104 	B9	=   1.057293869956598995326368602518056990746e-0001L;
105 
106 long double
logl(long double x)107 logl(long double x) {
108 	long double f, s, z, qn, h, t;
109 	int *px = (int *) &x;
110 	int *pz = (int *) &z;
111 	int i, j, ix, i0, i1, n;
112 
113 	/* get long double precision word ordering */
114 	if (*(int *) &one == 0) {
115 		i0 = 3;
116 		i1 = 0;
117 	} else {
118 		i0 = 0;
119 		i1 = 3;
120 	}
121 
122 	n = 0;
123 	ix = px[i0];
124 	if (ix > 0x3ffee0f8) {	/* if x >  31/33 */
125 		if (ix < 0x3fff1084) {	/* if x < 33/31 */
126 			f = x - one;
127 			z = f * f;
128 			if (((ix - 0x3fff0000) | px[i1] | px[2] | px[1]) == 0) {
129 				return (zero);	/* log(1)= +0 */
130 			}
131 			s = f / (two + f);	/* |s|<2**-8 */
132 			z = s * s;
133 			return (f - s * (f - z * (B1 + z * (B2 + z * (B3 +
134 				z * (B4 + z * (B5 + z * (B6 + z * (B7 +
135 				z * (B8 + z * B9))))))))));
136 		}
137 		if (ix >= 0x7fff0000)
138 			return (x + x);	/* x is +inf or NaN */
139 		goto LARGE_N;
140 	}
141 	if (ix >= 0x00010000)
142 		goto LARGE_N;
143 	i = ix & 0x7fffffff;
144 	if ((i | px[i1] | px[2] | px[1]) == 0) {
145 		px[i0] |= 0x80000000;
146 		return (one / x);	/* log(0.0) = -inf */
147 	}
148 	if (ix < 0) {
149 		if ((unsigned) ix >= 0xffff0000)
150 			return (x - x);	/* x is -inf or NaN */
151 		return (zero / zero);	/* log(x<0) is NaN  */
152 	}
153 	/* subnormal x */
154 	x *= two113;
155 	n = -113;
156 	ix = px[i0];
157 LARGE_N:
158 	n += ((ix + 0x200) >> 16) - 0x3fff;
159 	ix = (ix & 0x0000ffff) | 0x3fff0000;	/* scale x to [1,2] */
160 	px[i0] = ix;
161 	i = ix + 0x200;
162 	pz[i0] = i & 0xfffffc00;
163 	pz[i1] = pz[1] = pz[2] = 0;
164 	s = (x - z) / (x + z);
165 	j = (i >> 10) & 0x3f;
166 	z = s * s;
167 	qn = (long double) n;
168 	t = qn * ln2lo + _TBL_logl_lo[j];
169 	h = qn * ln2hi + _TBL_logl_hi[j];
170 	f = t + s * (A1 + z * (A2 + z * (A3 + z * (A4 + z * (A5 +
171 		z * (A6 + z * A7))))));
172 	return (h + f);
173 }
174