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
/*
Copyright (C) 2009, 2020 William Hart
Copyright (C) 2021 Albin Ahlbäck
Copyright (C) 2022 Fredrik Johansson
This file is part of FLINT.
FLINT 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 3 of the License, or
(at your option) any later version. See <https://www.gnu.org/licenses/>.
*/
#include "gmpcompat.h"
#include "mpn_extras.h"
#include "fmpz.h"
/* This can only be called from fmpz_mul, and assumes
x and y are not small. */
static void
flint_mpz_mul(mpz_ptr z, mpz_srcptr x, mpz_srcptr y)
{
slong xn, yn, zn, sgn;
nn_srcptr xd, yd;
nn_ptr zd;
ulong top;
TMP_INIT;
xn = x->_mp_size;
yn = y->_mp_size;
sgn = xn ^ yn;
xn = FLINT_ABS(xn);
yn = FLINT_ABS(yn);
if (xn < yn)
{
mpz_srcptr t;
slong tn;
t = x;
x = y;
y = t;
tn = xn;
xn = yn;
yn = tn;
}
zn = xn + yn;
zd = FLINT_MPZ_REALLOC(z, zn);
/* Important: read after possibly resizing z, so that the
pointers are valid in case of aliasing. */
xd = x->_mp_d;
yd = y->_mp_d;
if (xn == yn)
{
if (xn == 2)
{
ulong r3, r2, r1, r0;
FLINT_MPN_MUL_2X2(r3, r2, r1, r0, xd[1], xd[0], yd[1], yd[0]);
zd[0] = r0;
zd[1] = r1;
zd[2] = r2;
zd[3] = r3;
zn -= (r3 == 0);
z->_mp_size = (sgn >= 0) ? zn : -zn;
return;
}
if (xn == 1)
{
ulong hi, lo;
umul_ppmm(hi, lo, xd[0], yd[0]);
zd[0] = lo;
zd[1] = hi;
/* The result cannot be 1 limb, because that would
require a coefficient smaller than COEFF_MAX. */
FLINT_ASSERT(hi != 0);
z->_mp_size = (sgn >= 0) ? 2 : -2;
return;
}
}
/* Unlikely case since operands up to FLINT_BITS-2 bits get
caught in the fmpz fast path, but we should still optimize
for this, especially since we don't need to handle aliasing. */
if (yn == 1)
{
if (xn == 2)
{
ulong r2, r1, r0;
FLINT_MPN_MUL_2X1(r2, r1, r0, xd[1], xd[0], yd[0]);
zd[0] = r0;
zd[1] = r1;
zd[2] = top = r2;
}
else
{
top = zd[xn] = mpn_mul_1(zd, xd, xn, yd[0]);
}
zn -= (top == 0);
z->_mp_size = (sgn >= 0) ? zn : -zn;
return;
}
TMP_START;
/* In case of aliasing, we need to copy the input so that
we do not overwrite it during the multiplication. */
if (zd == xd)
{
nn_ptr tmp = TMP_ALLOC(xn * sizeof(ulong));
flint_mpn_copyi(tmp, xd, xn);
xd = tmp;
}
else if (zd == yd)
{
nn_ptr tmp = TMP_ALLOC(yn * sizeof(ulong));
flint_mpn_copyi(tmp, yd, yn);
yd = tmp;
}
if (x == y)
{
flint_mpn_sqr(zd, xd, xn);
top = zd[zn - 1];
}
else
{
top = flint_mpn_mul(zd, xd, xn, yd, yn);
}
zn -= (top == 0);
z->_mp_size = (sgn >= 0) ? zn : -zn;
TMP_END;
}
void
fmpz_mul(fmpz_t f, const fmpz_t g, const fmpz_t h)
{
mpz_ptr mf;
fmpz c1 = *g;
fmpz c2 = *h;
if (!COEFF_IS_MPZ(c1))
{
if (!COEFF_IS_MPZ(c2))
{
ulong th, tl;
smul_ppmm(th, tl, c1, c2);
fmpz_set_signed_uiui(f, th, tl);
return;
}
else if (c1 != 0)
{
mf = _fmpz_promote(f);
flint_mpz_mul_si(mf, COEFF_TO_PTR(c2), c1);
return;
}
}
if (!COEFF_IS_MPZ(*f))
{
if (c1 == 0 || c2 == 0)
{
*f = 0;
return;
}
mf = _fmpz_new_mpz();
(*f) = PTR_TO_COEFF(mf);
}
else
{
if (c1 == 0 || c2 == 0)
{
_fmpz_clear_mpz(*f);
*f = 0;
return;
}
mf = COEFF_TO_PTR(*f);
}
if (!COEFF_IS_MPZ(c2))
flint_mpz_mul_si(mf, COEFF_TO_PTR(c1), c2);
else
flint_mpz_mul(mf, COEFF_TO_PTR(c1), COEFF_TO_PTR(c2));
}
void
fmpz_mul_si(fmpz_t f, const fmpz_t g, slong x)
{
fmpz c2 = *g;
if (!COEFF_IS_MPZ(c2)) /* c2 is small */
{
ulong th, tl;
/* limb by limb multiply (assembly for most CPU's) */
smul_ppmm(th, tl, c2, x);
fmpz_set_signed_uiui(f, th, tl);
}
else /* c2 is large */
{
mpz_ptr mf;
if (!COEFF_IS_MPZ(*f))
{
if (x == 0)
{
*f = 0;
return;
}
mf = _fmpz_new_mpz();
*f = PTR_TO_COEFF(mf);
}
else
{
if (x == 0)
{
_fmpz_clear_mpz(*f);
*f = 0;
return;
}
mf = COEFF_TO_PTR(*f);
}
flint_mpz_mul_si(mf, COEFF_TO_PTR(c2), x);
}
}
void
fmpz_mul_ui(fmpz_t f, const fmpz_t g, ulong x)
{
fmpz c2 = *g;
if (!COEFF_IS_MPZ(c2)) /* c2 is small */
{
ulong th, tl;
ulong uc2 = FLINT_ABS(c2);
/* unsigned limb by limb multiply (assembly for most CPU's) */
umul_ppmm(th, tl, uc2, x);
if (c2 >= 0)
fmpz_set_uiui(f, th, tl);
else
fmpz_neg_uiui(f, th, tl);
}
else /* c2 is large */
{
mpz_ptr mf;
if (!COEFF_IS_MPZ(*f))
{
if (x == 0)
{
*f = 0;
return;
}
mf = _fmpz_new_mpz();
*f = PTR_TO_COEFF(mf);
}
else
{
if (x == 0)
{
_fmpz_clear_mpz(*f);
*f = 0;
return;
}
mf = COEFF_TO_PTR(*f);
}
flint_mpz_mul_ui(mf, COEFF_TO_PTR(c2), x);
}
}