sysdeps/ieee754/dbl-64/s_atan.c

   1 /*
   2  * IBM Accurate Mathematical Library
   3  * written by International Business Machines Corp.
   4  * Copyright (C) 2001-2013 Free Software Foundation, Inc.
   5  *
   6  * This program is free software; you can redistribute it and/or modify
   7  * it under the terms of the GNU Lesser General Public License as published by
   8  * the Free Software Foundation; either version 2.1 of the License, or
   9  * (at your option) any later version.
  10  *
  11  * This program is distributed in the hope that it will be useful,
  12  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  13  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  14  * GNU Lesser General Public License for more details.
  15  *
  16  * You should have received a copy of the GNU Lesser General Public License
  17  * along with this program; if not, see <http://www.gnu.org/licenses/>.
  18  */
  19 /************************************************************************/
  20 /*  MODULE_NAME: atnat.c                                                */
  21 /*                                                                      */
  22 /*  FUNCTIONS:  uatan                                                   */
  23 /*              atanMp                                                  */
  24 /*              signArctan                                              */
  25 /*                                                                      */
  26 /*                                                                      */
  27 /*  FILES NEEDED: dla.h endian.h mpa.h mydefs.h atnat.h                 */
  28 /*                mpatan.c mpatan2.c mpsqrt.c                           */
  29 /*                uatan.tbl                                             */
  30 /*                                                                      */
  31 /* An ultimate atan() routine. Given an IEEE double machine number x    */
  32 /* it computes the correctly rounded (to nearest) value of atan(x).     */
  33 /*                                                                      */
  34 /* Assumption: Machine arithmetic operations are performed in           */
  35 /* round to nearest mode of IEEE 754 standard.                          */
  36 /*                                                                      */
  37 /************************************************************************/
  38
  39 #include <dla.h>
  40 #include "mpa.h"
  41 #include "MathLib.h"
  42 #include "uatan.tbl"
  43 #include "atnat.h"
  44 #include <math.h>
  45
  46 void __mpatan (mp_no *, mp_no *, int);  /* see definition in mpatan.c */
  47 static double atanMp (double, const int[]);
  48
  49   /* Fix the sign of y and return */
  50 static double
  51 __signArctan (double x, double y)
  52 {
  53   return __copysign (y, x);
  54 }
  55
  56
  57 /* An ultimate atan() routine. Given an IEEE double machine number x,    */
  58 /* routine computes the correctly rounded (to nearest) value of atan(x). */
  59 double
  60 atan (double x)
  61 {
  62   double cor, s1, ss1, s2, ss2, t1, t2, t3, t7, t8, t9, t10, u, u2, u3,
  63     v, vv, w, ww, y, yy, z, zz;
  64 #ifndef DLA_FMS
  65   double t4, t5, t6;
  66 #endif
  67   int i, ux, dx;
  68   static const int pr[M] = { 6, 8, 10, 32 };
  69   number num;
  70
  71   num.d = x;
  72   ux = num.i[HIGH_HALF];
  73   dx = num.i[LOW_HALF];
  74
  75   /* x=NaN */
  76   if (((ux & 0x7ff00000) == 0x7ff00000)
  77       && (((ux & 0x000fffff) | dx) != 0x00000000))
  78     return x + x;
  79
  80   /* Regular values of x, including denormals +-0 and +-INF */
  81   u = (x < 0) ? -x : x;
  82   if (u < C)
  83     {
  84       if (u < B)
  85         {
  86           if (u < A)
  87             return x;
  88           else
  89             {                   /* A <= u < B */
  90               v = x * x;
  91               yy = d11.d + v * d13.d;
  92               yy = d9.d + v * yy;
  93               yy = d7.d + v * yy;
  94               yy = d5.d + v * yy;
  95               yy = d3.d + v * yy;
  96               yy *= x * v;
  97
  98               if ((y = x + (yy - U1 * x)) == x + (yy + U1 * x))
  99                 return y;
 100
 101               EMULV (x, x, v, vv, t1, t2, t3, t4, t5);  /* v+vv=x^2 */
 102
 103               s1 = f17.d + v * f19.d;
 104               s1 = f15.d + v * s1;
 105               s1 = f13.d + v * s1;
 106               s1 = f11.d + v * s1;
 107               s1 *= v;
 108
 109               ADD2 (f9.d, ff9.d, s1, 0, s2, ss2, t1, t2);
 110               MUL2 (v, vv, s2, ss2, s1, ss1, t1, t2, t3, t4, t5, t6, t7, t8);
 111               ADD2 (f7.d, ff7.d, s1, ss1, s2, ss2, t1, t2);
 112               MUL2 (v, vv, s2, ss2, s1, ss1, t1, t2, t3, t4, t5, t6, t7, t8);
 113               ADD2 (f5.d, ff5.d, s1, ss1, s2, ss2, t1, t2);
 114               MUL2 (v, vv, s2, ss2, s1, ss1, t1, t2, t3, t4, t5, t6, t7, t8);
 115               ADD2 (f3.d, ff3.d, s1, ss1, s2, ss2, t1, t2);
 116               MUL2 (v, vv, s2, ss2, s1, ss1, t1, t2, t3, t4, t5, t6, t7, t8);
 117               MUL2 (x, 0, s1, ss1, s2, ss2, t1, t2, t3, t4, t5, t6, t7,
 118                     t8);
 119               ADD2 (x, 0, s2, ss2, s1, ss1, t1, t2);
 120               if ((y = s1 + (ss1 - U5 * s1)) == s1 + (ss1 + U5 * s1))
 121                 return y;
 122
 123               return atanMp (x, pr);
 124             }
 125         }
 126       else
 127         {                       /* B <= u < C */
 128           i = (TWO52 + TWO8 * u) - TWO52;
 129           i -= 16;
 130           z = u - cij[i][0].d;
 131           yy = cij[i][5].d + z * cij[i][6].d;
 132           yy = cij[i][4].d + z * yy;
 133           yy = cij[i][3].d + z * yy;
 134           yy = cij[i][2].d + z * yy;
 135           yy *= z;
 136
 137           t1 = cij[i][1].d;
 138           if (i < 112)
 139             {
 140               if (i < 48)
 141                 u2 = U21;       /* u < 1/4        */
 142               else
 143                 u2 = U22;
 144             }                   /* 1/4 <= u < 1/2 */
 145           else
 146             {
 147               if (i < 176)
 148                 u2 = U23;       /* 1/2 <= u < 3/4 */
 149               else
 150                 u2 = U24;
 151             }                   /* 3/4 <= u <= 1  */
 152           if ((y = t1 + (yy - u2 * t1)) == t1 + (yy + u2 * t1))
 153             return __signArctan (x, y);
 154
 155           z = u - hij[i][0].d;
 156
 157           s1 = hij[i][14].d + z * hij[i][15].d;
 158           s1 = hij[i][13].d + z * s1;
 159           s1 = hij[i][12].d + z * s1;
 160           s1 = hij[i][11].d + z * s1;
 161           s1 *= z;
 162
 163           ADD2 (hij[i][9].d, hij[i][10].d, s1, 0, s2, ss2, t1, t2);
 164           MUL2 (z, 0, s2, ss2, s1, ss1, t1, t2, t3, t4, t5, t6, t7, t8);
 165           ADD2 (hij[i][7].d, hij[i][8].d, s1, ss1, s2, ss2, t1, t2);
 166           MUL2 (z, 0, s2, ss2, s1, ss1, t1, t2, t3, t4, t5, t6, t7, t8);
 167           ADD2 (hij[i][5].d, hij[i][6].d, s1, ss1, s2, ss2, t1, t2);
 168           MUL2 (z, 0, s2, ss2, s1, ss1, t1, t2, t3, t4, t5, t6, t7, t8);
 169           ADD2 (hij[i][3].d, hij[i][4].d, s1, ss1, s2, ss2, t1, t2);
 170           MUL2 (z, 0, s2, ss2, s1, ss1, t1, t2, t3, t4, t5, t6, t7, t8);
 171           ADD2 (hij[i][1].d, hij[i][2].d, s1, ss1, s2, ss2, t1, t2);
 172           if ((y = s2 + (ss2 - U6 * s2)) == s2 + (ss2 + U6 * s2))
 173             return __signArctan (x, y);
 174
 175           return atanMp (x, pr);
 176         }
 177     }
 178   else
 179     {
 180       if (u < D)
 181         {                       /* C <= u < D */
 182           w = ONE / u;
 183           EMULV (w, u, t1, t2, t3, t4, t5, t6, t7);
 184           ww = w * ((ONE - t1) - t2);
 185           i = (TWO52 + TWO8 * w) - TWO52;
 186           i -= 16;
 187           z = (w - cij[i][0].d) + ww;
 188
 189           yy = cij[i][5].d + z * cij[i][6].d;
 190           yy = cij[i][4].d + z * yy;
 191           yy = cij[i][3].d + z * yy;
 192           yy = cij[i][2].d + z * yy;
 193           yy  = HPI1 - z * yy;
 194
 195           t1 = HPI - cij[i][1].d;
 196           if (i < 112)
 197             u3 = U31;           /* w <  1/2 */
 198           else
 199             u3 = U32;           /* w >= 1/2 */
 200           if ((y = t1 + (yy - u3)) == t1 + (yy + u3))
 201             return __signArctan (x, y);
 202
 203           DIV2 (ONE, 0, u, 0, w, ww, t1, t2, t3, t4, t5, t6, t7, t8, t9,
 204                 t10);
 205           t1 = w - hij[i][0].d;
 206           EADD (t1, ww, z, zz);
 207
 208           s1 = hij[i][14].d + z * hij[i][15].d;
 209           s1 = hij[i][13].d + z * s1;
 210           s1 = hij[i][12].d + z * s1;
 211           s1 = hij[i][11].d + z * s1;
 212           s1 *= z;
 213
 214           ADD2 (hij[i][9].d, hij[i][10].d, s1, 0, s2, ss2, t1, t2);
 215           MUL2 (z, zz, s2, ss2, s1, ss1, t1, t2, t3, t4, t5, t6, t7, t8);
 216           ADD2 (hij[i][7].d, hij[i][8].d, s1, ss1, s2, ss2, t1, t2);
 217           MUL2 (z, zz, s2, ss2, s1, ss1, t1, t2, t3, t4, t5, t6, t7, t8);
 218           ADD2 (hij[i][5].d, hij[i][6].d, s1, ss1, s2, ss2, t1, t2);
 219           MUL2 (z, zz, s2, ss2, s1, ss1, t1, t2, t3, t4, t5, t6, t7, t8);
 220           ADD2 (hij[i][3].d, hij[i][4].d, s1, ss1, s2, ss2, t1, t2);
 221           MUL2 (z, zz, s2, ss2, s1, ss1, t1, t2, t3, t4, t5, t6, t7, t8);
 222           ADD2 (hij[i][1].d, hij[i][2].d, s1, ss1, s2, ss2, t1, t2);
 223           SUB2 (HPI, HPI1, s2, ss2, s1, ss1, t1, t2);
 224           if ((y = s1 + (ss1 - U7)) == s1 + (ss1 + U7))
 225             return __signArctan (x, y);
 226
 227           return atanMp (x, pr);
 228         }
 229       else
 230         {
 231           if (u < E)
 232             {                   /* D <= u < E */
 233               w = ONE / u;
 234               v = w * w;
 235               EMULV (w, u, t1, t2, t3, t4, t5, t6, t7);
 236
 237               yy = d11.d + v * d13.d;
 238               yy = d9.d + v * yy;
 239               yy = d7.d + v * yy;
 240               yy = d5.d + v * yy;
 241               yy = d3.d + v * yy;
 242               yy *= w * v;
 243
 244               ww = w * ((ONE - t1) - t2);
 245               ESUB (HPI, w, t3, cor);
 246               yy = ((HPI1 + cor) - ww) - yy;
 247               if ((y = t3 + (yy - U4)) == t3 + (yy + U4))
 248                 return __signArctan (x, y);
 249
 250               DIV2 (ONE, 0, u, 0, w, ww, t1, t2, t3, t4, t5, t6, t7, t8,
 251                     t9, t10);
 252               MUL2 (w, ww, w, ww, v, vv, t1, t2, t3, t4, t5, t6, t7, t8);
 253
 254               s1 = f17.d + v * f19.d;
 255               s1 = f15.d + v * s1;
 256               s1 = f13.d + v * s1;
 257               s1 = f11.d + v * s1;
 258               s1 *= v;
 259
 260               ADD2 (f9.d, ff9.d, s1, 0, s2, ss2, t1, t2);
 261               MUL2 (v, vv, s2, ss2, s1, ss1, t1, t2, t3, t4, t5, t6, t7, t8);
 262               ADD2 (f7.d, ff7.d, s1, ss1, s2, ss2, t1, t2);
 263               MUL2 (v, vv, s2, ss2, s1, ss1, t1, t2, t3, t4, t5, t6, t7, t8);
 264               ADD2 (f5.d, ff5.d, s1, ss1, s2, ss2, t1, t2);
 265               MUL2 (v, vv, s2, ss2, s1, ss1, t1, t2, t3, t4, t5, t6, t7, t8);
 266               ADD2 (f3.d, ff3.d, s1, ss1, s2, ss2, t1, t2);
 267               MUL2 (v, vv, s2, ss2, s1, ss1, t1, t2, t3, t4, t5, t6, t7, t8);
 268               MUL2 (w, ww, s1, ss1, s2, ss2, t1, t2, t3, t4, t5, t6, t7, t8);
 269               ADD2 (w, ww, s2, ss2, s1, ss1, t1, t2);
 270               SUB2 (HPI, HPI1, s1, ss1, s2, ss2, t1, t2);
 271
 272               if ((y = s2 + (ss2 - U8)) == s2 + (ss2 + U8))
 273                 return __signArctan (x, y);
 274
 275               return atanMp (x, pr);
 276             }
 277           else
 278             {
 279               /* u >= E */
 280               if (x > 0)
 281                 return HPI;
 282               else
 283                 return MHPI;
 284             }
 285         }
 286     }
 287 }
 288
 289  /* Final stages. Compute atan(x) by multiple precision arithmetic */
 290 static double
 291 atanMp (double x, const int pr[])
 292 {
 293   mp_no mpx, mpy, mpy2, mperr, mpt1, mpy1;
 294   double y1, y2;
 295   int i, p;
 296
 297   for (i = 0; i < M; i++)
 298     {
 299       p = pr[i];
 300       __dbl_mp (x, &mpx, p);
 301       __mpatan (&mpx, &mpy, p);
 302       __dbl_mp (u9[i].d, &mpt1, p);
 303       __mul (&mpy, &mpt1, &mperr, p);
 304       __add (&mpy, &mperr, &mpy1, p);
 305       __sub (&mpy, &mperr, &mpy2, p);
 306       __mp_dbl (&mpy1, &y1, p);
 307       __mp_dbl (&mpy2, &y2, p);
 308       if (y1 == y2)
 309         return y1;
 310     }
 311   return y1;                    /*if impossible to do exact computing */
 312 }
 313
 314 #ifdef NO_LONG_DOUBLE
 315 weak_alias (atan, atanl)
 316 #endif