1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
|
/* Simple data type for real numbers for the GNU compiler.
Copyright (C) 2002-2014 Free Software Foundation, Inc.
This file is part of GCC.
GCC is free software; you can redistribute it and/or modify it under
the terms of the GNU General Public License as published by the Free
Software Foundation; either version 3, or (at your option) any later
version.
GCC is distributed in the hope that it will be useful, but WITHOUT ANY
WARRANTY; without even the implied warranty of MERCHANTABILITY or
FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
for more details.
You should have received a copy of the GNU General Public License
along with GCC; see the file COPYING3. If not see
<http://www.gnu.org/licenses/>. */
/* This library supports real numbers;
inf and nan are NOT supported.
It is written to be simple and fast.
Value of sreal is
x = sig * 2 ^ exp
where
sig = significant
(for < 64-bit machines sig = sig_lo + sig_hi * 2 ^ SREAL_PART_BITS)
exp = exponent
One uint64_t is used for the significant.
Only a half of significant bits is used (in normalized sreals) so that we do
not have problems with overflow, for example when c->sig = a->sig * b->sig.
So the precision is 32-bit.
Invariant: The numbers are normalized before and after each call of sreal_*.
Normalized sreals:
All numbers (except zero) meet following conditions:
SREAL_MIN_SIG <= sig && sig <= SREAL_MAX_SIG
-SREAL_MAX_EXP <= exp && exp <= SREAL_MAX_EXP
If the number would be too large, it is set to upper bounds of these
conditions.
If the number is zero or would be too small it meets following conditions:
sig == 0 && exp == -SREAL_MAX_EXP
*/
#include <math.h>
#include "config.h"
#include "system.h"
#include "coretypes.h"
#include "sreal.h"
/* Print the content of struct sreal. */
void
sreal::dump (FILE *file) const
{
fprintf (file, "(%" PRIu64 " * 2^%d)", m_sig, m_exp);
}
DEBUG_FUNCTION void
debug (const sreal &ref)
{
ref.dump (stderr);
}
DEBUG_FUNCTION void
debug (const sreal *ptr)
{
if (ptr)
debug (*ptr);
else
fprintf (stderr, "<nil>\n");
}
/* Shift this right by S bits. Needed: 0 < S <= SREAL_BITS.
When the most significant bit shifted out is 1, add 1 to this (rounding).
*/
void
sreal::shift_right (int s)
{
gcc_checking_assert (s > 0);
gcc_checking_assert (s <= SREAL_BITS);
/* Exponent should never be so large because shift_right is used only by
sreal_add and sreal_sub ant thus the number cannot be shifted out from
exponent range. */
gcc_checking_assert (m_exp + s <= SREAL_MAX_EXP);
m_exp += s;
m_sig += (int64_t) 1 << (s - 1);
m_sig >>= s;
}
/* Normalize *this. */
void
sreal::normalize ()
{
int64_t s = m_sig < 0 ? -1 : 1;
unsigned HOST_WIDE_INT sig = absu_hwi (m_sig);
if (sig == 0)
{
m_exp = -SREAL_MAX_EXP;
}
else if (sig < SREAL_MIN_SIG)
{
do
{
sig <<= 1;
m_exp--;
}
while (sig < SREAL_MIN_SIG);
/* Check underflow. */
if (m_exp < -SREAL_MAX_EXP)
{
m_exp = -SREAL_MAX_EXP;
sig = 0;
}
}
else if (sig > SREAL_MAX_SIG)
{
int last_bit;
do
{
last_bit = sig & 1;
sig >>= 1;
m_exp++;
}
while (sig > SREAL_MAX_SIG);
/* Round the number. */
sig += last_bit;
if (sig > SREAL_MAX_SIG)
{
sig >>= 1;
m_exp++;
}
/* Check overflow. */
if (m_exp > SREAL_MAX_EXP)
{
m_exp = SREAL_MAX_EXP;
sig = SREAL_MAX_SIG;
}
}
m_sig = s * sig;
}
/* Return integer value of *this. */
int64_t
sreal::to_int () const
{
int64_t sign = m_sig < 0 ? -1 : 1;
if (m_exp <= -SREAL_BITS)
return 0;
if (m_exp >= SREAL_PART_BITS)
return sign * INTTYPE_MAXIMUM (int64_t);
if (m_exp > 0)
return m_sig << m_exp;
if (m_exp < 0)
return m_sig >> -m_exp;
return m_sig;
}
/* Return value of *this as double.
This should be used for debug output only. */
double
sreal::to_double () const
{
double val = m_sig;
if (m_exp)
val *= exp2 (m_exp);
return val;
}
/* Return *this + other. */
sreal
sreal::operator+ (const sreal &other) const
{
int dexp;
sreal tmp, r;
const sreal *a_p = this, *b_p = &other, *bb;
if (a_p->m_exp < b_p->m_exp)
std::swap (a_p, b_p);
dexp = a_p->m_exp - b_p->m_exp;
r.m_exp = a_p->m_exp;
if (dexp > SREAL_BITS)
{
r.m_sig = a_p->m_sig;
return r;
}
if (dexp == 0)
bb = b_p;
else
{
tmp = *b_p;
tmp.shift_right (dexp);
bb = &tmp;
}
r.m_sig = a_p->m_sig + bb->m_sig;
r.normalize ();
return r;
}
/* Return *this - other. */
sreal
sreal::operator- (const sreal &other) const
{
int dexp;
sreal tmp, r;
const sreal *bb;
const sreal *a_p = this, *b_p = &other;
int64_t sign = 1;
if (a_p->m_exp < b_p->m_exp)
{
sign = -1;
std::swap (a_p, b_p);
}
dexp = a_p->m_exp - b_p->m_exp;
r.m_exp = a_p->m_exp;
if (dexp > SREAL_BITS)
{
r.m_sig = sign * a_p->m_sig;
return r;
}
if (dexp == 0)
bb = b_p;
else
{
tmp = *b_p;
tmp.shift_right (dexp);
bb = &tmp;
}
r.m_sig = sign * (a_p->m_sig - bb->m_sig);
r.normalize ();
return r;
}
/* Return *this * other. */
sreal
sreal::operator* (const sreal &other) const
{
sreal r;
if (absu_hwi (m_sig) < SREAL_MIN_SIG || absu_hwi (other.m_sig) < SREAL_MIN_SIG)
{
r.m_sig = 0;
r.m_exp = -SREAL_MAX_EXP;
}
else
{
r.m_sig = m_sig * other.m_sig;
r.m_exp = m_exp + other.m_exp;
r.normalize ();
}
return r;
}
/* Return *this / other. */
sreal
sreal::operator/ (const sreal &other) const
{
gcc_checking_assert (other.m_sig != 0);
sreal r;
r.m_sig = (m_sig << SREAL_PART_BITS) / other.m_sig;
r.m_exp = m_exp - other.m_exp - SREAL_PART_BITS;
r.normalize ();
return r;
}
|