xref: /illumos-gate/usr/src/lib/libm/common/R/atanf.c (revision 97a9db610324e7db4393415018e0e737485a94cd)
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 __atanf = atanf
31 
32 /* INDENT OFF */
33 /*
34  * float atanf(float x);
35  * Table look-up algorithm
36  * By K.C. Ng, March 9, 1989
37  *
38  * Algorithm.
39  *
40  * The algorithm is based on atan(x)=atan(y)+atan((x-y)/(1+x*y)).
41  * We use poly1(x) to approximate atan(x) for x in [0,1/8] with
42  * error (relative)
43  * 	|(atan(x)-poly1(x))/x|<= 2^-115.94 	long double
44  * 	|(atan(x)-poly1(x))/x|<= 2^-58.85	double
45  * 	|(atan(x)-poly1(x))/x|<= 2^-25.53 	float
46  * and use poly2(x) to approximate atan(x) for x in [0,1/65] with
47  * error (absolute)
48  *	|atan(x)-poly2(x)|<= 2^-122.15		long double
49  *	|atan(x)-poly2(x)|<= 2^-64.79		double
50  *	|atan(x)-poly2(x)|<= 2^-35.36		float
51  * and use poly3(x) to approximate atan(x) for x in [1/8,7/16] with
52  * error (relative, on for single precision)
53  * 	|(atan(x)-poly1(x))/x|<= 2^-25.53 	float
54  *
55  * Here poly1-3 are odd polynomial with the following form:
56  *		x + x^3*(a1+x^2*(a2+...))
57  *
58  * (0). Purge off Inf and NaN and 0
59  * (1). Reduce x to positive by atan(x) = -atan(-x).
60  * (2). For x <= 1/8, use
61  *	(2.1) if x < 2^(-prec/2-2), atan(x) =  x  with inexact
62  *	(2.2) Otherwise
63  *		atan(x) = poly1(x)
64  * (3). For x >= 8 then
65  *	(3.1) if x >= 2^(prec+2),   atan(x) = atan(inf) - pio2lo
66  *	(3.2) if x >= 2^(prec/3+2), atan(x) = atan(inf) - 1/x
67  *	(3.3) if x >  65,           atan(x) = atan(inf) - poly2(1/x)
68  *	(3.4) Otherwise,	    atan(x) = atan(inf) - poly1(1/x)
69  *
70  * (4). Now x is in (0.125, 8)
71  *      Find y that match x to 4.5 bit after binary (easy).
72  *	If iy is the high word of y, then
73  *		single : j = (iy - 0x3e000000) >> 19
74  *		(single is modified to (iy-0x3f000000)>>19)
75  *		double : j = (iy - 0x3fc00000) >> 16
76  *		quad   : j = (iy - 0x3ffc0000) >> 12
77  *
78  *	Let s = (x-y)/(1+x*y). Then
79  *	atan(x) = atan(y) + poly1(s)
80  *		= _TBL_r_atan_hi[j] + (_TBL_r_atan_lo[j] + poly2(s) )
81  *
82  *	Note. |s| <= 1.5384615385e-02 = 1/65. Maxium occurs at x = 1.03125
83  *
84  */
85 
86 #include "libm.h"
87 
88 extern const float _TBL_r_atan_hi[], _TBL_r_atan_lo[];
89 static const float
90 	big	=   1.0e37F,
91 	one	=   1.0F,
92 	p1	=  -3.333185951111688247225368498733544672172e-0001F,
93 	p2	=   1.969352894213455405211341983203180636021e-0001F,
94 	q1	=  -3.332921964095646819563419704110132937456e-0001F,
95 	a1	=  -3.333323465223893614063523351509338934592e-0001F,
96 	a2	=   1.999425625935277805494082274808174062403e-0001F,
97 	a3	=  -1.417547090509737780085769846290301788559e-0001F,
98 	a4	=   1.016250813871991983097273733227432685084e-0001F,
99 	a5	=  -5.137023693688358515753093811791755221805e-0002F,
100 	pio2hi	=   1.570796371e+0000F,
101 	pio2lo	=  -4.371139000e-0008F;
102 /* INDENT ON */
103 
104 float
105 atanf(float xx) {
106 	float x, y, z, r, p, s;
107 	volatile double dummy;
108 	int ix, iy, sign, j;
109 
110 	x = xx;
111 	ix = *(int *) &x;
112 	sign = ix & 0x80000000;
113 	ix ^= sign;
114 
115 	/* for |x| < 1/8 */
116 	if (ix < 0x3e000000) {
117 		if (ix < 0x38800000) {	/* if |x| < 2**(-prec/2-2) */
118 			dummy = big + x;	/* get inexact flag if x != 0 */
119 #ifdef lint
120 			dummy = dummy;
121 #endif
122 			return (x);
123 		}
124 		z = x * x;
125 		if (ix < 0x3c000000) {	/* if |x| < 2**(-prec/4-1) */
126 			x = x + (x * z) * p1;
127 			return (x);
128 		} else {
129 			x = x + (x * z) * (p1 + z * p2);
130 			return (x);
131 		}
132 	}
133 
134 	/* for |x| >= 8.0 */
135 	if (ix >= 0x41000000) {
136 		*(int *) &x = ix;
137 		if (ix < 0x42820000) {	/* x <  65 */
138 			r = one / x;
139 			z = r * r;
140 			y = r * (one + z * (p1 + z * p2));	/* poly1 */
141 			y -= pio2lo;
142 		} else if (ix < 0x44800000) {	/* x <  2**(prec/3+2) */
143 			r = one / x;
144 			z = r * r;
145 			y = r * (one + z * q1);	/* poly2 */
146 			y -= pio2lo;
147 		} else if (ix < 0x4c800000) {	/* x <  2**(prec+2) */
148 			y = one / x - pio2lo;
149 		} else if (ix < 0x7f800000) {	/* x <  inf */
150 			y = -pio2lo;
151 		} else {		/* x is inf or NaN */
152 			if (ix > 0x7f800000) {
153 				return (x * x);	/* - -> * for Cheetah */
154 			}
155 			y = -pio2lo;
156 		}
157 
158 		if (sign == 0)
159 			x = pio2hi - y;
160 		else
161 			x = y - pio2hi;
162 		return (x);
163 	}
164 
165 
166 	/* now x is between 1/8 and 8 */
167 	if (ix < 0x3f000000) {	/* between 1/8 and 1/2 */
168 		z = x * x;
169 		x = x + (x * z) * (a1 + z * (a2 + z * (a3 + z * (a4 +
170 			z * a5))));
171 		return (x);
172 	}
173 	*(int *) &x = ix;
174 	iy = (ix + 0x00040000) & 0x7ff80000;
175 	*(int *) &y = iy;
176 	j = (iy - 0x3f000000) >> 19;
177 
178 	if (ix == iy)
179 		p = x - y;	/* p=0.0 */
180 	else {
181 		if (sign == 0)
182 			s = (x - y) / (one + x * y);
183 		else
184 			s = (y - x) / (one + x * y);
185 		z = s * s;
186 		p = s * (one + z * q1);
187 	}
188 	if (sign == 0) {
189 		r = p + _TBL_r_atan_lo[j];
190 		x = r + _TBL_r_atan_hi[j];
191 	} else {
192 		r = p - _TBL_r_atan_lo[j];
193 		x = r - _TBL_r_atan_hi[j];
194 	}
195 	return (x);
196 }
197