xref: /original-bsd/old/libm/libm/log10.c (revision aba77441)
1 /*
2  * Copyright (c) 1985 Regents of the University of California.
3  *
4  * Use and reproduction of this software are granted  in  accordance  with
5  * the terms and conditions specified in  the  Berkeley  Software  License
6  * Agreement (in particular, this entails acknowledgement of the programs'
7  * source, and inclusion of this notice) with the additional understanding
8  * that  all  recipients  should regard themselves as participants  in  an
9  * ongoing  research  project and hence should  feel  obligated  to report
10  * their  experiences (good or bad) with these elementary function  codes,
11  * using "sendbug 4bsd-bugs@BERKELEY", to the authors.
12  */
13 
14 #ifndef lint
15 static char sccsid[] = "@(#)log10.c	1.1 (Berkeley) 05/23/85";
16 #endif not lint
17 
18 /* LOG10(X)
19  * RETURN THE BASE 10 LOGARITHM OF x
20  * DOUBLE PRECISION (VAX D format 56 bits, IEEE DOUBLE 53 BITS)
21  * CODED IN C BY K.C. NG, 1/20/85;
22  * REVISED BY K.C. NG on 1/23/85, 3/7/85, 4/16/85.
23  *
24  * Required kernel function:
25  *	log(x)
26  *
27  * Method :
28  *			     log(x)
29  *		log10(x) = ---------  or  [1/log(10)]*log(x)
30  *			    log(10)
31  *
32  *    Note:
33  *	  [log(10)]   rounded to 56 bits has error  .0895  ulps,
34  *	  [1/log(10)] rounded to 53 bits has error  .198   ulps;
35  *	  therefore, for better accuracy, in VAX D format, we divide
36  *	  log(x) by log(10), but in IEEE Double format, we multiply
37  *	  log(x) by [1/log(10)].
38  *
39  * Special cases:
40  *	log10(x) is NAN with signal if x < 0;
41  *	log10(+INF) is +INF with no signal; log10(0) is -INF with signal;
42  *	log10(NAN) is that NAN with no signal.
43  *
44  * Accuracy:
45  *	log10(X) returns the exact log10(x) nearly rounded. In a test run
46  *	with 1,536,000 random arguments on a VAX, the maximum observed
47  *	error was 1.74 ulps (units in the last place).
48  *
49  * Constants:
50  * The hexadecimal values are the intended ones for the following constants.
51  * The decimal values may be used, provided that the compiler will convert
52  * from decimal to binary accurately enough to produce the hexadecimal values
53  * shown.
54  */
55 
56 #ifdef VAX	/* VAX D format (56 bits) */
57 /* static double */
58 /* ln10hi =  2.3025850929940456790E0     ; Hex   2^  2   *  .935D8DDDAAA8AC */
59 static long    ln10hix[] = { 0x5d8d4113, 0xa8acddaa};
60 #define   ln10hi    (*(double*)ln10hix)
61 #else  		/* IEEE double precision (53 bits) */
62 static double
63 ivln10 =  4.3429448190325181667E-1    ; /*Hex   2^ -2   *  1.BCB7B1526E50E */
64 #endif
65 
66 double log10(x)
67 double x;
68 {
69 	double log();
70 
71 #ifdef VAX
72 	return(log(x)/ln10hi);
73 #else
74 	return(ivln10*log(x));
75 #endif
76 }
77