OSDN Git Service

PR c++/22494
[pf3gnuchains/gcc-fork.git] / libgcc-math / dbl-64 / mplog.c
1
2 /*
3  * IBM Accurate Mathematical Library
4  * written by International Business Machines Corp.
5  * Copyright (C) 2001 Free Software Foundation
6  *
7  * This program is free software; you can redistribute it and/or modify
8  * it under the terms of the GNU Lesser General Public License as published by
9  * the Free Software Foundation; either version 2.1 of the License, or
10  * (at your option) any later version.
11  *
12  * This program is distributed in the hope that it will be useful,
13  * but WITHOUT ANY WARRANTY; without even the implied warranty of
14  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
15  * GNU Lesser General Public License for more details.
16  *
17  * You should have received a copy of the GNU Lesser General Public License
18  * along with this program; if not, write to the Free Software
19  * Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.
20  */
21 /************************************************************************/
22 /*                                                                      */
23 /*     MODULE_NAME:mplog.c                                              */
24 /*                                                                      */
25 /*     FUNCTIONS:  mplog                                                */
26 /*                                                                      */
27 /*     FILES NEEDED: endian.h mpa.h  mplog.h                            */
28 /*                    mpexp.c                                           */
29 /*                                                                      */
30 /* Multi-Precision logarithm function subroutine (for precision p >= 4, */
31 /* 2**(-1024) < x < 2**1024) and x is outside of the interval           */
32 /* [1-2**(-54),1+2**(-54)]. Upon entry, x should be set to the          */
33 /* multi-precision value of the input and y should be set into a multi- */
34 /* precision value of an approximation of log(x) with relative error    */
35 /* bound of at most 2**(-52). The routine improves the accuracy of y.   */
36 /*                                                                      */
37 /************************************************************************/
38 #include "endian.h"
39 #include "mpa.h"
40
41 void __mpexp(mp_no *, mp_no *, int);
42
43 void __mplog(mp_no *x, mp_no *y, int p) {
44 #include "mplog.h"
45   int i,m;
46 #if 0
47   int j,k,m1,m2,n;
48   double a,b;
49 #endif
50   static const int mp[33] = {0,0,0,0,0,1,1,2,2,2,2,3,3,3,3,3,3,3,3,
51                              4,4,4,4,4,4,4,4,4,4,4,4,4,4};
52   mp_no mpone = {0,{0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,
53                     0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,
54                     0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0}};
55   mp_no mpt1,mpt2;
56
57   /* Choose m and initiate mpone */
58   m = mp[p];  mpone.e = 1;  mpone.d[0]=mpone.d[1]=ONE;
59
60   /* Perform m newton iterations to solve for y: exp(y)-x=0.     */
61   /* The iterations formula is:  y(n+1)=y(n)+(x*exp(-y(n))-1).   */
62   __cpy(y,&mpt1,p);
63   for (i=0; i<m; i++) {
64     mpt1.d[0]=-mpt1.d[0];
65     __mpexp(&mpt1,&mpt2,p);
66     __mul(x,&mpt2,&mpt1,p);
67     __sub(&mpt1,&mpone,&mpt2,p);
68     __add(y,&mpt2,&mpt1,p);
69     __cpy(&mpt1,y,p);
70   }
71   return;
72 }