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
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
|
/*
Copyright (C) 2012, 2013 Fredrik Johansson
This file is part of Arb.
Arb is free software: you can redistribute it and/or modify it under
the terms of the GNU Lesser General Public License (LGPL) as published
by the Free Software Foundation; either version 2.1 of the License, or
(at your option) any later version. See <http://www.gnu.org/licenses/>.
*/
#include "fmpr.h"
#define ADD_STACK_ALLOC 40
#define ADD_TLS_ALLOC 1000
TLS_PREFIX mp_ptr __add_tmp = NULL;
TLS_PREFIX slong __add_alloc = 0;
void _add_tmp_cleanup(void)
{
flint_free(__add_tmp);
__add_tmp = NULL;
__add_alloc = 0;
}
#define ADD_TMP_ALLOC \
if (alloc <= ADD_STACK_ALLOC) \
{ \
tmp = tmp_stack; \
} \
else if (alloc <= ADD_TLS_ALLOC) \
{ \
if (__add_alloc < alloc) \
{ \
if (__add_alloc == 0) \
{ \
flint_register_cleanup_function(_add_tmp_cleanup); \
} \
__add_tmp = flint_realloc(__add_tmp, sizeof(mp_limb_t) * alloc); \
__add_alloc = alloc; \
} \
tmp = __add_tmp; \
} \
else \
{ \
tmp = flint_malloc(sizeof(mp_limb_t) * alloc); \
}
#define ADD_TMP_FREE \
if (alloc > ADD_TLS_ALLOC) \
flint_free(tmp);
/* computes x + y * 2^shift (optionally negated) */
slong
_fmpr_add_mpn(fmpr_t z,
mp_srcptr xman, mp_size_t xn, int xsign, const fmpz_t xexp,
mp_srcptr yman, mp_size_t yn, int ysign, const fmpz_t yexp,
slong shift, slong prec, fmpr_rnd_t rnd)
{
slong tn, zn, alloc, ret, shift_bits, shift_limbs;
int negative;
mp_limb_t tmp_stack[ADD_STACK_ALLOC];
mp_limb_t cy;
mp_ptr tmp, tmp2;
shift_limbs = shift / FLINT_BITS;
shift_bits = shift % FLINT_BITS;
/* x does not overlap with y or the result -- outcome is
equivalent to adding/subtracting a small number to/from y
and rounding */
if (shift > xn * FLINT_BITS &&
prec != FMPR_PREC_EXACT &&
xn * FLINT_BITS + prec - (FLINT_BITS * (yn - 1)) < shift)
{
zn = (prec + FLINT_BITS - 1) / FLINT_BITS;
zn = FLINT_MAX(zn, yn) + 2;
shift_limbs = zn - yn;
alloc = zn;
ADD_TMP_ALLOC
flint_mpn_zero(tmp, shift_limbs);
flint_mpn_copyi(tmp + shift_limbs, yman, yn);
if (xsign == ysign)
{
tmp[0] = 1;
}
else
{
mpn_sub_1(tmp, tmp, zn, 1);
while (tmp[zn-1] == 0)
zn--;
}
ret = _fmpr_set_round_mpn(&shift, fmpr_manref(z), tmp, zn, ysign, prec, rnd);
shift -= shift_limbs * FLINT_BITS;
fmpz_add_si_inline(fmpr_expref(z), yexp, shift);
ADD_TMP_FREE
return ret;
}
if (xsign == ysign)
{
negative = xsign;
/* certainly no overlap between x and y * 2^shift
XXXXX
YYYYY
|-------|
shift_limbs
todo: handle huge case efficiently!
*/
if (shift_limbs >= xn)
{
alloc = shift_limbs + yn + 1;
ADD_TMP_ALLOC
flint_mpn_copyi(tmp, xman, xn);
flint_mpn_zero(tmp + xn, shift_limbs - xn);
if (shift_bits == 0)
{
flint_mpn_copyi(tmp + shift_limbs, yman, yn);
zn = shift_limbs + yn;
}
else
{
cy = mpn_lshift(tmp + shift_limbs, yman, yn, shift_bits);
tmp[shift_limbs + yn] = cy;
zn = shift_limbs + yn + (cy != 0);
}
}
else /* may have overlap between x and y * 2^shift */
{
slong alloc1, alloc2;
alloc2 = yn + 1; /* shifted value of y */
alloc1 = FLINT_MAX(alloc2 + shift_limbs, xn) + 1; /* space for sum */
alloc = alloc1 + alloc2;
ADD_TMP_ALLOC
tmp2 = tmp + alloc1;
flint_mpn_copyi(tmp, xman, shift_limbs);
/* {t, tn} = shift of y (ignoring shift_limbs zero limbs) */
if (shift_bits == 0)
{
flint_mpn_copyi(tmp2, yman, yn);
tn = yn;
}
else
{
cy = mpn_lshift(tmp2, yman, yn, shift_bits);
tmp2[yn] = cy;
tn = yn + (cy != 0);
}
if (tn + shift_limbs == xn)
{
/* XXXXXXXX
TTTTT */
cy = mpn_add_n(tmp + shift_limbs, xman + shift_limbs, tmp2, tn);
tmp[xn] = cy;
zn = xn + (cy != 0);
}
else if (tn + shift_limbs < xn)
{
/* XXXXXXXX
TTTTT */
cy = mpn_add_n(tmp + shift_limbs, xman + shift_limbs, tmp2, tn);
cy = mpn_add_1(tmp + shift_limbs + tn, xman + shift_limbs + tn, xn - tn - shift_limbs, cy);
tmp[xn] = cy;
zn = xn + (cy != 0);
}
else
{
/* XXXXXXXX
TTTTTTT */
cy = mpn_add_n(tmp + shift_limbs, xman + shift_limbs, tmp2, xn - shift_limbs);
cy = mpn_add_1(tmp + xn, tmp2 + xn - shift_limbs, tn + shift_limbs - xn, cy);
tmp[tn + shift_limbs] = cy;
zn = tn + shift_limbs + (cy != 0);
}
}
}
else
{
/* certainly no overlap between x and y * 2^shift
XXXXX
YYYYY
|-------|
shift_limbs
todo: handle huge case efficiently!
*/
if (shift_limbs >= xn)
{
alloc = shift_limbs + yn + 1;
ADD_TMP_ALLOC
mpn_neg_n(tmp, xman, xn);
/* since x is nonzero, there is certainly carry (borrow) */
flint_mpn_store(tmp + xn, shift_limbs - xn, (mp_limb_t) -1);
if (shift_bits == 0)
{
flint_mpn_copyi(tmp + shift_limbs, yman, yn);
zn = shift_limbs + yn;
}
else
{
cy = mpn_lshift(tmp + shift_limbs, yman, yn, shift_bits);
tmp[shift_limbs + yn] = cy;
zn = shift_limbs + yn + (cy != 0);
}
mpn_sub_1(tmp + shift_limbs, tmp + shift_limbs, zn - shift_limbs, 1);
/* y has larger absolute value, and determines the sign */
negative = ysign;
}
else /* may have overlap between x and y * 2^shift */
{
slong alloc1, alloc2;
alloc2 = yn + 1; /* shifted value of y */
alloc1 = FLINT_MAX(alloc2 + shift_limbs, xn); /* space for difference */
alloc = alloc1 + alloc2;
ADD_TMP_ALLOC
tmp2 = tmp + alloc1;
/* {t, tn} = shift of y (ignoring shift_limbs zero limbs) */
if (shift_bits == 0)
{
flint_mpn_copyi(tmp2, yman, yn);
tn = yn;
}
else
{
cy = mpn_lshift(tmp2, yman, yn, shift_bits);
tmp2[yn] = cy;
tn = yn + (cy != 0);
}
if (tn + shift_limbs == xn)
{
int comp;
/* XXXXXXXX
TTTTT */
comp = mpn_cmp(xman + shift_limbs, tmp2, tn);
if (comp == 0)
{
slong i;
for (i = 0; i < shift_limbs; i++)
{
if (xman[i] != 0)
{
comp = 1;
break;
}
}
/* the result is exactly zero */
if (comp == 0)
{
fmpr_zero(z);
ret = FMPR_RESULT_EXACT;
goto cleanup;
}
}
/* x is larger */
if (comp > 0)
{
flint_mpn_copyi(tmp, xman, shift_limbs);
mpn_sub_n(tmp + shift_limbs, xman + shift_limbs, tmp2, tn);
negative = xsign;
}
else
{
mpn_sub_n(tmp + shift_limbs, tmp2, xman + shift_limbs, tn);
if (shift_limbs != 0)
{
cy = mpn_neg_n(tmp, xman, shift_limbs);
mpn_sub_1(tmp + shift_limbs, tmp + shift_limbs, tn, cy);
}
negative = ysign;
}
zn = xn;
}
else if (tn + shift_limbs < xn)
{
/* XXXXXXXX
TTTTT */
flint_mpn_copyi(tmp, xman, shift_limbs);
cy = mpn_sub_n(tmp + shift_limbs, xman + shift_limbs, tmp2, tn);
mpn_sub_1(tmp + shift_limbs + tn, xman + shift_limbs + tn, xn - tn - shift_limbs, cy);
zn = xn;
negative = xsign;
}
else
{
/* XXXXXXXX
TTTTTTT */
cy = mpn_sub_n(tmp + shift_limbs, tmp2, xman + shift_limbs, xn - shift_limbs);
mpn_sub_1(tmp + xn, tmp2 + xn - shift_limbs, tn + shift_limbs - xn, cy);
if (shift_limbs != 0)
{
cy = mpn_neg_n(tmp, xman, shift_limbs);
mpn_sub_1(tmp + shift_limbs, tmp + shift_limbs, tn, cy);
}
zn = shift_limbs + tn;
negative = ysign;
}
}
/* there might have been cancellation */
while (tmp[zn-1] == 0)
zn--;
}
ret = _fmpr_set_round_mpn(&shift, fmpr_manref(z), tmp, zn, negative, prec, rnd);
fmpz_add_si_inline(fmpr_expref(z), xexp, shift);
cleanup:
ADD_TMP_FREE
return ret;
}
|
