• Home
  • Line#
  • Scopes#
  • Navigate#
  • Raw
  • Download
1 /* origin: FreeBSD /usr/src/lib/msun/src/s_fmal.c */
2 /*-
3  * Copyright (c) 2005-2011 David Schultz <das@FreeBSD.ORG>
4  * All rights reserved.
5  *
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  *
15  * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
16  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
17  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
18  * ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
19  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
20  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
21  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
22  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
23  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
24  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
25  * SUCH DAMAGE.
26  */
27 
28 
29 #include "libm.h"
30 #if LDBL_MANT_DIG == 53 && LDBL_MAX_EXP == 1024
fmal(long double x,long double y,long double z)31 long double fmal(long double x, long double y, long double z)
32 {
33 	return fma(x, y, z);
34 }
35 #elif (LDBL_MANT_DIG == 64 || LDBL_MANT_DIG == 113) && LDBL_MAX_EXP == 16384
36 #include <fenv.h>
37 #if LDBL_MANT_DIG == 64
38 #define LASTBIT(u) (u.i.m & 1)
39 #define SPLIT (0x1p32L + 1)
40 #elif LDBL_MANT_DIG == 113
41 #define LASTBIT(u) (u.i.lo & 1)
42 #define SPLIT (0x1p57L + 1)
43 #endif
44 
45 /*
46  * A struct dd represents a floating-point number with twice the precision
47  * of a long double.  We maintain the invariant that "hi" stores the high-order
48  * bits of the result.
49  */
50 struct dd {
51 	long double hi;
52 	long double lo;
53 };
54 
55 /*
56  * Compute a+b exactly, returning the exact result in a struct dd.  We assume
57  * that both a and b are finite, but make no assumptions about their relative
58  * magnitudes.
59  */
dd_add(long double a,long double b)60 static inline struct dd dd_add(long double a, long double b)
61 {
62 	struct dd ret;
63 	long double s;
64 
65 	ret.hi = a + b;
66 	s = ret.hi - a;
67 	ret.lo = (a - (ret.hi - s)) + (b - s);
68 	return (ret);
69 }
70 
71 /*
72  * Compute a+b, with a small tweak:  The least significant bit of the
73  * result is adjusted into a sticky bit summarizing all the bits that
74  * were lost to rounding.  This adjustment negates the effects of double
75  * rounding when the result is added to another number with a higher
76  * exponent.  For an explanation of round and sticky bits, see any reference
77  * on FPU design, e.g.,
78  *
79  *     J. Coonen.  An Implementation Guide to a Proposed Standard for
80  *     Floating-Point Arithmetic.  Computer, vol. 13, no. 1, Jan 1980.
81  */
add_adjusted(long double a,long double b)82 static inline long double add_adjusted(long double a, long double b)
83 {
84 	struct dd sum;
85 	union ldshape u;
86 
87 	sum = dd_add(a, b);
88 	if (sum.lo != 0) {
89 		u.f = sum.hi;
90 		if (!LASTBIT(u))
91 			sum.hi = nextafterl(sum.hi, INFINITY * sum.lo);
92 	}
93 	return (sum.hi);
94 }
95 
96 /*
97  * Compute ldexp(a+b, scale) with a single rounding error. It is assumed
98  * that the result will be subnormal, and care is taken to ensure that
99  * double rounding does not occur.
100  */
add_and_denormalize(long double a,long double b,int scale)101 static inline long double add_and_denormalize(long double a, long double b, int scale)
102 {
103 	struct dd sum;
104 	int bits_lost;
105 	union ldshape u;
106 
107 	sum = dd_add(a, b);
108 
109 	/*
110 	 * If we are losing at least two bits of accuracy to denormalization,
111 	 * then the first lost bit becomes a round bit, and we adjust the
112 	 * lowest bit of sum.hi to make it a sticky bit summarizing all the
113 	 * bits in sum.lo. With the sticky bit adjusted, the hardware will
114 	 * break any ties in the correct direction.
115 	 *
116 	 * If we are losing only one bit to denormalization, however, we must
117 	 * break the ties manually.
118 	 */
119 	if (sum.lo != 0) {
120 		u.f = sum.hi;
121 		bits_lost = -u.i.se - scale + 1;
122 		if ((bits_lost != 1) ^ LASTBIT(u))
123 			sum.hi = nextafterl(sum.hi, INFINITY * sum.lo);
124 	}
125 	return scalbnl(sum.hi, scale);
126 }
127 
128 /*
129  * Compute a*b exactly, returning the exact result in a struct dd.  We assume
130  * that both a and b are normalized, so no underflow or overflow will occur.
131  * The current rounding mode must be round-to-nearest.
132  */
dd_mul(long double a,long double b)133 static inline struct dd dd_mul(long double a, long double b)
134 {
135 	struct dd ret;
136 	long double ha, hb, la, lb, p, q;
137 
138 	p = a * SPLIT;
139 	ha = a - p;
140 	ha += p;
141 	la = a - ha;
142 
143 	p = b * SPLIT;
144 	hb = b - p;
145 	hb += p;
146 	lb = b - hb;
147 
148 	p = ha * hb;
149 	q = ha * lb + la * hb;
150 
151 	ret.hi = p + q;
152 	ret.lo = p - ret.hi + q + la * lb;
153 	return (ret);
154 }
155 
156 /*
157  * Fused multiply-add: Compute x * y + z with a single rounding error.
158  *
159  * We use scaling to avoid overflow/underflow, along with the
160  * canonical precision-doubling technique adapted from:
161  *
162  *      Dekker, T.  A Floating-Point Technique for Extending the
163  *      Available Precision.  Numer. Math. 18, 224-242 (1971).
164  */
fmal(long double x,long double y,long double z)165 long double fmal(long double x, long double y, long double z)
166 {
167 	#pragma STDC FENV_ACCESS ON
168 	long double xs, ys, zs, adj;
169 	struct dd xy, r;
170 	int oround;
171 	int ex, ey, ez;
172 	int spread;
173 
174 	/*
175 	 * Handle special cases. The order of operations and the particular
176 	 * return values here are crucial in handling special cases involving
177 	 * infinities, NaNs, overflows, and signed zeroes correctly.
178 	 */
179 	if (!isfinite(x) || !isfinite(y))
180 		return (x * y + z);
181 	if (!isfinite(z))
182 		return (z);
183 	if (x == 0.0 || y == 0.0)
184 		return (x * y + z);
185 	if (z == 0.0)
186 		return (x * y);
187 
188 	xs = frexpl(x, &ex);
189 	ys = frexpl(y, &ey);
190 	zs = frexpl(z, &ez);
191 	oround = fegetround();
192 	spread = ex + ey - ez;
193 
194 	/*
195 	 * If x * y and z are many orders of magnitude apart, the scaling
196 	 * will overflow, so we handle these cases specially.  Rounding
197 	 * modes other than FE_TONEAREST are painful.
198 	 */
199 	if (spread < -LDBL_MANT_DIG) {
200 #ifdef FE_INEXACT
201 		feraiseexcept(FE_INEXACT);
202 #endif
203 #ifdef FE_UNDERFLOW
204 		if (!isnormal(z))
205 			feraiseexcept(FE_UNDERFLOW);
206 #endif
207 		switch (oround) {
208 		default: /* FE_TONEAREST */
209 			return (z);
210 #ifdef FE_TOWARDZERO
211 		case FE_TOWARDZERO:
212 			if (x > 0.0 ^ y < 0.0 ^ z < 0.0)
213 				return (z);
214 			else
215 				return (nextafterl(z, 0));
216 #endif
217 #ifdef FE_DOWNWARD
218 		case FE_DOWNWARD:
219 			if (x > 0.0 ^ y < 0.0)
220 				return (z);
221 			else
222 				return (nextafterl(z, -INFINITY));
223 #endif
224 #ifdef FE_UPWARD
225 		case FE_UPWARD:
226 			if (x > 0.0 ^ y < 0.0)
227 				return (nextafterl(z, INFINITY));
228 			else
229 				return (z);
230 #endif
231 		}
232 	}
233 	if (spread <= LDBL_MANT_DIG * 2)
234 		zs = scalbnl(zs, -spread);
235 	else
236 		zs = copysignl(LDBL_MIN, zs);
237 
238 	fesetround(FE_TONEAREST);
239 
240 	/*
241 	 * Basic approach for round-to-nearest:
242 	 *
243 	 *     (xy.hi, xy.lo) = x * y           (exact)
244 	 *     (r.hi, r.lo)   = xy.hi + z       (exact)
245 	 *     adj = xy.lo + r.lo               (inexact; low bit is sticky)
246 	 *     result = r.hi + adj              (correctly rounded)
247 	 */
248 	xy = dd_mul(xs, ys);
249 	r = dd_add(xy.hi, zs);
250 
251 	spread = ex + ey;
252 
253 	if (r.hi == 0.0) {
254 		/*
255 		 * When the addends cancel to 0, ensure that the result has
256 		 * the correct sign.
257 		 */
258 		fesetround(oround);
259 		volatile long double vzs = zs; /* XXX gcc CSE bug workaround */
260 		return xy.hi + vzs + scalbnl(xy.lo, spread);
261 	}
262 
263 	if (oround != FE_TONEAREST) {
264 		/*
265 		 * There is no need to worry about double rounding in directed
266 		 * rounding modes.
267 		 * But underflow may not be raised correctly, example in downward rounding:
268 		 * fmal(0x1.0000000001p-16000L, 0x1.0000000001p-400L, -0x1p-16440L)
269 		 */
270 		long double ret;
271 #if defined(FE_INEXACT) && defined(FE_UNDERFLOW)
272 		int e = fetestexcept(FE_INEXACT);
273 		feclearexcept(FE_INEXACT);
274 #endif
275 		fesetround(oround);
276 		adj = r.lo + xy.lo;
277 		ret = scalbnl(r.hi + adj, spread);
278 #if defined(FE_INEXACT) && defined(FE_UNDERFLOW)
279 		if (ilogbl(ret) < -16382 && fetestexcept(FE_INEXACT))
280 			feraiseexcept(FE_UNDERFLOW);
281 		else if (e)
282 			feraiseexcept(FE_INEXACT);
283 #endif
284 		return ret;
285 	}
286 
287 	adj = add_adjusted(r.lo, xy.lo);
288 	if (spread + ilogbl(r.hi) > -16383)
289 		return scalbnl(r.hi + adj, spread);
290 	else
291 		return add_and_denormalize(r.hi, adj, spread);
292 }
293 #endif
294