xref: /netbsd-src/external/lgpl3/mpfr/dist/src/cos.c (revision e5d758f832e07a177fa24707c434b7ce26d0f762)
1 /* mpfr_cos -- cosine of a floating-point number
2 
3 Copyright 2001-2020 Free Software Foundation, Inc.
4 Contributed by the AriC and Caramba projects, INRIA.
5 
6 This file is part of the GNU MPFR Library.
7 
8 The GNU MPFR Library is free software; you can redistribute it and/or modify
9 it under the terms of the GNU Lesser General Public License as published by
10 the Free Software Foundation; either version 3 of the License, or (at your
11 option) any later version.
12 
13 The GNU MPFR Library is distributed in the hope that it will be useful, but
14 WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY
15 or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU Lesser General Public
16 License for more details.
17 
18 You should have received a copy of the GNU Lesser General Public License
19 along with the GNU MPFR Library; see the file COPYING.LESSER.  If not, see
20 https://www.gnu.org/licenses/ or write to the Free Software Foundation, Inc.,
21 51 Franklin St, Fifth Floor, Boston, MA 02110-1301, USA. */
22 
23 #define MPFR_NEED_LONGLONG_H
24 #include "mpfr-impl.h"
25 
26 static int
27 mpfr_cos_fast (mpfr_ptr y, mpfr_srcptr x, mpfr_rnd_t rnd_mode)
28 {
29   int inex;
30 
31   inex = mpfr_sincos_fast (NULL, y, x, rnd_mode);
32   inex = inex >> 2; /* 0: exact, 1: rounded up, 2: rounded down */
33   return (inex == 2) ? -1 : inex;
34 }
35 
36 /* f <- 1 - r/2! + r^2/4! + ... + (-1)^l r^l/(2l)! + ...
37    Assumes |r| < 1/2, and f, r have the same precision.
38    Returns e such that the error on f is bounded by 2^e ulps.
39 */
40 static int
41 mpfr_cos2_aux (mpfr_ptr f, mpfr_srcptr r)
42 {
43   mpz_t x, t, s;
44   mpfr_exp_t ex, l, m;
45   mpfr_prec_t p, q;
46   unsigned long i, maxi, imax;
47 
48   MPFR_ASSERTD(mpfr_get_exp (r) <= -1);
49 
50   /* compute minimal i such that i*(i+1) does not fit in an unsigned long,
51      assuming that there are no padding bits. */
52   maxi = 1UL << (sizeof(unsigned long) * CHAR_BIT / 2);
53   if (maxi * (maxi / 2) == 0) /* test checked at compile time */
54     {
55       /* can occur only when there are padding bits. */
56       /* maxi * (maxi-1) is representable iff maxi * (maxi / 2) != 0 */
57       do
58         maxi /= 2;
59       while (maxi * (maxi / 2) == 0);
60     }
61 
62   mpz_init (x);
63   mpz_init (s);
64   mpz_init (t);
65   ex = mpfr_get_z_2exp (x, r); /* r = x*2^ex */
66 
67   /* Remove trailing zeroes.
68      Since x comes from a regular MPFR number, due to the constraints on the
69      exponent and the precision, there can be no integer overflow below. */
70   l = mpz_scan1 (x, 0);
71   ex += l;
72   mpz_fdiv_q_2exp (x, x, l);
73 
74   /* since |r| < 1, r = x*2^ex, and x is an integer, necessarily ex < 0 */
75 
76   p = mpfr_get_prec (f); /* same than r */
77   /* bound for number of iterations */
78   imax = p / (-mpfr_get_exp (r));
79   imax += (imax == 0);
80   q = 2 * MPFR_INT_CEIL_LOG2(imax) + 4; /* bound for (3l)^2 */
81 
82   mpz_set_ui (s, 1); /* initialize sum with 1 */
83   mpz_mul_2exp (s, s, p + q); /* scale all values by 2^(p+q) */
84   mpz_set (t, s); /* invariant: t is previous term */
85   for (i = 1; (m = mpz_sizeinbase (t, 2)) >= q; i += 2)
86     {
87       /* adjust precision of x to that of t */
88       l = mpz_sizeinbase (x, 2);
89       if (l > m)
90         {
91           l -= m;
92           mpz_fdiv_q_2exp (x, x, l);
93           ex += l;
94         }
95       /* multiply t by r */
96       mpz_mul (t, t, x);
97       mpz_fdiv_q_2exp (t, t, -ex);
98       /* divide t by i*(i+1) */
99       if (i < maxi)
100         mpz_fdiv_q_ui (t, t, i * (i + 1));
101       else
102         {
103           mpz_fdiv_q_ui (t, t, i);
104           mpz_fdiv_q_ui (t, t, i + 1);
105         }
106       /* if m is the (current) number of bits of t, we can consider that
107          all operations on t so far had precision >= m, so we can prove
108          by induction that the relative error on t is of the form
109          (1+u)^(3l)-1, where |u| <= 2^(-m), and l=(i+1)/2 is the # of loops.
110          Since |(1+x^2)^(1/x) - 1| <= 4x/3 for |x| <= 1/2,
111          for |u| <= 1/(3l)^2, the absolute error is bounded by
112          4/3*(3l)*2^(-m)*t <= 4*l since |t| < 2^m.
113          Therefore the error on s is bounded by 2*l*(l+1). */
114       /* add or subtract to s */
115       if (i % 4 == 1)
116         mpz_sub (s, s, t);
117       else
118         mpz_add (s, s, t);
119     }
120 
121   mpfr_set_z (f, s, MPFR_RNDN);
122   mpfr_div_2ui (f, f, p + q, MPFR_RNDN);
123 
124   mpz_clear (x);
125   mpz_clear (s);
126   mpz_clear (t);
127 
128   l = (i - 1) / 2; /* number of iterations */
129   return 2 * MPFR_INT_CEIL_LOG2 (l + 1) + 1; /* bound is 2l(l+1) */
130 }
131 
132 int
133 mpfr_cos (mpfr_ptr y, mpfr_srcptr x, mpfr_rnd_t rnd_mode)
134 {
135   mpfr_prec_t K0, K, precy, m, k, l;
136   int inexact, reduce = 0;
137   mpfr_t r, s, xr, c;
138   mpfr_exp_t exps, cancel = 0, expx;
139   MPFR_ZIV_DECL (loop);
140   MPFR_SAVE_EXPO_DECL (expo);
141   MPFR_GROUP_DECL (group);
142 
143   MPFR_LOG_FUNC (
144     ("x[%Pu]=%*.Rg rnd=%d", mpfr_get_prec (x), mpfr_log_prec, x, rnd_mode),
145     ("y[%Pu]=%*.Rg inexact=%d", mpfr_get_prec (y), mpfr_log_prec, y,
146      inexact));
147 
148   if (MPFR_UNLIKELY (MPFR_IS_SINGULAR (x)))
149     {
150       if (MPFR_IS_NAN (x) || MPFR_IS_INF (x))
151         {
152           MPFR_SET_NAN (y);
153           MPFR_RET_NAN;
154         }
155       else
156         {
157           MPFR_ASSERTD (MPFR_IS_ZERO (x));
158           return mpfr_set_ui (y, 1, rnd_mode);
159         }
160     }
161 
162   MPFR_SAVE_EXPO_MARK (expo);
163 
164   /* cos(x) = 1-x^2/2 + ..., so error < 2^(2*EXP(x)-1) */
165   expx = MPFR_GET_EXP (x);
166   MPFR_SMALL_INPUT_AFTER_SAVE_EXPO (y, __gmpfr_one, -2 * expx,
167                                     1, 0, rnd_mode, expo, {});
168 
169   /* Compute initial precision */
170   precy = MPFR_PREC (y);
171 
172   if (precy >= MPFR_SINCOS_THRESHOLD)
173     {
174       inexact = mpfr_cos_fast (y, x, rnd_mode);
175       goto end;
176     }
177 
178   K0 = __gmpfr_isqrt (precy / 3);
179   m = precy + 2 * MPFR_INT_CEIL_LOG2 (precy) + 2 * K0 + 4;
180 
181   if (expx >= 3)
182     {
183       reduce = 1;
184       /* As expx + m - 1 will silently be converted into mpfr_prec_t
185          in the mpfr_init2 call, the assert below may be useful to
186          avoid undefined behavior. */
187       MPFR_ASSERTN (expx + m - 1 <= MPFR_PREC_MAX);
188       mpfr_init2 (c, expx + m - 1);
189       mpfr_init2 (xr, m);
190     }
191 
192   MPFR_GROUP_INIT_2 (group, m, r, s);
193   MPFR_ZIV_INIT (loop, m);
194   for (;;)
195     {
196       /* If |x| >= 4, first reduce x cmod (2*Pi) into xr, using mpfr_remainder:
197          let e = EXP(x) >= 3, and m the target precision:
198          (1) c <- 2*Pi              [precision e+m-1, nearest]
199          (2) xr <- remainder (x, c) [precision m, nearest]
200          We have |c - 2*Pi| <= 1/2ulp(c) = 2^(3-e-m)
201                  |xr - x - k c| <= 1/2ulp(xr) <= 2^(1-m)
202                  |k| <= |x|/(2*Pi) <= 2^(e-2)
203          Thus |xr - x - 2kPi| <= |k| |c - 2Pi| + 2^(1-m) <= 2^(2-m).
204          It follows |cos(xr) - cos(x)| <= 2^(2-m). */
205       if (reduce)
206         {
207           mpfr_const_pi (c, MPFR_RNDN);
208           mpfr_mul_2ui (c, c, 1, MPFR_RNDN); /* 2Pi */
209           mpfr_remainder (xr, x, c, MPFR_RNDN);
210           if (MPFR_IS_ZERO(xr))
211             goto ziv_next;
212           /* now |xr| <= 4, thus r <= 16 below */
213           mpfr_sqr (r, xr, MPFR_RNDU); /* err <= 1 ulp */
214         }
215       else
216         mpfr_sqr (r, x, MPFR_RNDU); /* err <= 1 ulp */
217 
218       /* now |x| < 4 (or xr if reduce = 1), thus |r| <= 16 */
219 
220       /* we need |r| < 1/2 for mpfr_cos2_aux, i.e., EXP(r) - 2K <= -1 */
221       K = K0 + 1 + MAX(0, MPFR_GET_EXP(r)) / 2;
222       /* since K0 >= 0, if EXP(r) < 0, then K >= 1, thus EXP(r) - 2K <= -3;
223          otherwise if EXP(r) >= 0, then K >= 1/2 + EXP(r)/2, thus
224          EXP(r) - 2K <= -1 */
225 
226       MPFR_SET_EXP (r, MPFR_GET_EXP (r) - 2 * K); /* Can't overflow! */
227 
228       /* s <- 1 - r/2! + ... + (-1)^l r^l/(2l)! */
229       l = mpfr_cos2_aux (s, r);
230       /* l is the error bound in ulps on s */
231       MPFR_SET_ONE (r);
232       for (k = 0; k < K; k++)
233         {
234           mpfr_sqr (s, s, MPFR_RNDU);            /* err <= 2*olderr */
235           MPFR_SET_EXP (s, MPFR_GET_EXP (s) + 1); /* Can't overflow */
236           mpfr_sub (s, s, r, MPFR_RNDN);         /* err <= 4*olderr */
237           if (MPFR_IS_ZERO(s))
238             goto ziv_next;
239           MPFR_ASSERTD (MPFR_GET_EXP (s) <= 1);
240         }
241 
242       /* The absolute error on s is bounded by (2l+1/3)*2^(2K-m)
243          2l+1/3 <= 2l+1.
244          If |x| >= 4, we need to add 2^(2-m) for the argument reduction
245          by 2Pi: if K = 0, this amounts to add 4 to 2l+1/3, i.e., to add
246          2 to l; if K >= 1, this amounts to add 1 to 2*l+1/3. */
247       l = 2 * l + 1;
248       if (reduce)
249         l += (K == 0) ? 4 : 1;
250       k = MPFR_INT_CEIL_LOG2 (l) + 2 * K;
251       /* now the error is bounded by 2^(k-m) = 2^(EXP(s)-err) */
252 
253       exps = MPFR_GET_EXP (s);
254       if (MPFR_LIKELY (MPFR_CAN_ROUND (s, exps + m - k, precy, rnd_mode)))
255         break;
256 
257       if (MPFR_UNLIKELY (exps == 1))
258         /* s = 1 or -1, and except x=0 which was already checked above,
259            cos(x) cannot be 1 or -1, so we can round if the error is less
260            than 2^(-precy) for directed rounding, or 2^(-precy-1) for rounding
261            to nearest. */
262         {
263           if (m > k && (m - k >= precy + (rnd_mode == MPFR_RNDN)))
264             {
265               /* If round to nearest or away, result is s = 1 or -1,
266                  otherwise it is round(nexttoward (s, 0)). However in order to
267                  have the inexact flag correctly set below, we set |s| to
268                  1 - 2^(-m) in all cases. */
269               mpfr_nexttozero (s);
270               break;
271             }
272         }
273 
274       if (exps < cancel)
275         {
276           m += cancel - exps;
277           cancel = exps;
278         }
279 
280     ziv_next:
281       MPFR_ZIV_NEXT (loop, m);
282       MPFR_GROUP_REPREC_2 (group, m, r, s);
283       if (reduce)
284         {
285           mpfr_set_prec (xr, m);
286           mpfr_set_prec (c, expx + m - 1);
287         }
288     }
289   MPFR_ZIV_FREE (loop);
290   inexact = mpfr_set (y, s, rnd_mode);
291   MPFR_GROUP_CLEAR (group);
292   if (reduce)
293     {
294       mpfr_clear (xr);
295       mpfr_clear (c);
296     }
297 
298  end:
299   MPFR_SAVE_EXPO_FREE (expo);
300   return mpfr_check_range (y, inexact, rnd_mode);
301 }
302