ldivmod, uldivmod: fix qdivrem calls
[minix.git] / lib / libm / noieee_src / n_asincos.c
blob09da31e15599cf224537f78ea1aee39dc9e34e20
1 /* $NetBSD: n_asincos.c,v 1.7 2003/08/07 16:44:50 agc Exp $ */
2 /*
3 * Copyright (c) 1985, 1993
4 * The Regents of the University of California. All rights reserved.
6 * Redistribution and use in source and binary forms, with or without
7 * modification, are permitted provided that the following conditions
8 * are met:
9 * 1. Redistributions of source code must retain the above copyright
10 * notice, this list of conditions and the following disclaimer.
11 * 2. Redistributions in binary form must reproduce the above copyright
12 * notice, this list of conditions and the following disclaimer in the
13 * documentation and/or other materials provided with the distribution.
14 * 3. Neither the name of the University nor the names of its contributors
15 * may be used to endorse or promote products derived from this software
16 * without specific prior written permission.
18 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
19 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
20 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
21 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
22 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
23 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
24 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
25 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
26 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
27 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
28 * SUCH DAMAGE.
31 #ifndef lint
32 #if 0
33 static char sccsid[] = "@(#)asincos.c 8.1 (Berkeley) 6/4/93";
34 #endif
35 #endif /* not lint */
37 /* ASIN(X)
38 * RETURNS ARC SINE OF X
39 * DOUBLE PRECISION (IEEE DOUBLE 53 bits, VAX D FORMAT 56 bits)
40 * CODED IN C BY K.C. NG, 4/16/85, REVISED ON 6/10/85.
42 * Required system supported functions:
43 * copysign(x,y)
44 * sqrt(x)
46 * Required kernel function:
47 * atan2(y,x)
49 * Method :
50 * asin(x) = atan2(x,sqrt(1-x*x)); for better accuracy, 1-x*x is
51 * computed as follows
52 * 1-x*x if x < 0.5,
53 * 2*(1-|x|)-(1-|x|)*(1-|x|) if x >= 0.5.
55 * Special cases:
56 * if x is NaN, return x itself;
57 * if |x|>1, return NaN.
59 * Accuracy:
60 * 1) If atan2() uses machine PI, then
62 * asin(x) returns (PI/pi) * (the exact arc sine of x) nearly rounded;
63 * and PI is the exact pi rounded to machine precision (see atan2 for
64 * details):
66 * in decimal:
67 * pi = 3.141592653589793 23846264338327 .....
68 * 53 bits PI = 3.141592653589793 115997963 ..... ,
69 * 56 bits PI = 3.141592653589793 227020265 ..... ,
71 * in hexadecimal:
72 * pi = 3.243F6A8885A308D313198A2E....
73 * 53 bits PI = 3.243F6A8885A30 = 2 * 1.921FB54442D18 error=.276ulps
74 * 56 bits PI = 3.243F6A8885A308 = 4 * .C90FDAA22168C2 error=.206ulps
76 * In a test run with more than 200,000 random arguments on a VAX, the
77 * maximum observed error in ulps (units in the last place) was
78 * 2.06 ulps. (comparing against (PI/pi)*(exact asin(x)));
80 * 2) If atan2() uses true pi, then
82 * asin(x) returns the exact asin(x) with error below about 2 ulps.
84 * In a test run with more than 1,024,000 random arguments on a VAX, the
85 * maximum observed error in ulps (units in the last place) was
86 * 1.99 ulps.
89 #include "mathimpl.h"
91 double
92 asin(double x)
94 double s,t,one=1.0;
95 #if !defined(__vax__)&&!defined(tahoe)
96 if(x!=x) return(x); /* x is NaN */
97 #endif /* !defined(__vax__)&&!defined(tahoe) */
98 s=copysign(x,one);
99 if(s <= 0.5)
100 return(atan2(x,sqrt(one-x*x)));
101 else
102 { t=one-s; s=t+t; return(atan2(x,sqrt(s-t*t))); }
106 /* ACOS(X)
107 * RETURNS ARC COS OF X
108 * DOUBLE PRECISION (IEEE DOUBLE 53 bits, VAX D FORMAT 56 bits)
109 * CODED IN C BY K.C. NG, 4/16/85, REVISED ON 6/10/85.
111 * Required system supported functions:
112 * copysign(x,y)
113 * sqrt(x)
115 * Required kernel function:
116 * atan2(y,x)
118 * Method :
119 * ________
120 * / 1 - x
121 * acos(x) = 2*atan2( / -------- , 1 ) .
122 * \/ 1 + x
124 * Special cases:
125 * if x is NaN, return x itself;
126 * if |x|>1, return NaN.
128 * Accuracy:
129 * 1) If atan2() uses machine PI, then
131 * acos(x) returns (PI/pi) * (the exact arc cosine of x) nearly rounded;
132 * and PI is the exact pi rounded to machine precision (see atan2 for
133 * details):
135 * in decimal:
136 * pi = 3.141592653589793 23846264338327 .....
137 * 53 bits PI = 3.141592653589793 115997963 ..... ,
138 * 56 bits PI = 3.141592653589793 227020265 ..... ,
140 * in hexadecimal:
141 * pi = 3.243F6A8885A308D313198A2E....
142 * 53 bits PI = 3.243F6A8885A30 = 2 * 1.921FB54442D18 error=.276ulps
143 * 56 bits PI = 3.243F6A8885A308 = 4 * .C90FDAA22168C2 error=.206ulps
145 * In a test run with more than 200,000 random arguments on a VAX, the
146 * maximum observed error in ulps (units in the last place) was
147 * 2.07 ulps. (comparing against (PI/pi)*(exact acos(x)));
149 * 2) If atan2() uses true pi, then
151 * acos(x) returns the exact acos(x) with error below about 2 ulps.
153 * In a test run with more than 1,024,000 random arguments on a VAX, the
154 * maximum observed error in ulps (units in the last place) was
155 * 2.15 ulps.
158 double
159 acos(double x)
161 double t,one=1.0;
162 #if !defined(__vax__)&&!defined(tahoe)
163 if(x!=x) return(x);
164 #endif /* !defined(__vax__)&&!defined(tahoe) */
165 if( x != -1.0)
166 t=atan2(sqrt((one-x)/(one+x)),one);
167 else
168 t=atan2(one,0.0); /* t = PI/2 */
169 return(t+t);