File: mullow_SS.c

package info (click to toggle)
flint 2.5.2-19
  • links: PTS, VCS
  • area: main
  • in suites: buster
  • size: 30,308 kB
  • sloc: ansic: 289,367; cpp: 11,210; python: 1,280; sh: 649; makefile: 283
file content (148 lines) | stat: -rw-r--r-- 4,809 bytes parent folder | download | duplicates (3)
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
/*=============================================================================

    This file is part of FLINT.

    FLINT 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 2 of the License, or
    (at your option) any later version.

    FLINT 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 FLINT; if not, write to the Free Software
    Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301 USA

=============================================================================*/
/******************************************************************************

    Copyright (C) 2008-2011 William Hart
    
******************************************************************************/

#include <stdlib.h>
#include "fmpz_poly.h"
#include "fft.h"
#include "fft_tuning.h"

void _fmpz_poly_mullow_SS(fmpz * output, const fmpz * input1, slong len1, 
               const fmpz * input2, slong len2, slong trunc)
{
    slong len_out, loglen, loglen2, n;
    slong output_bits, limbs, size, i;
    mp_limb_t * ptr, * t1, * t2, * tt, * s1, ** ii, ** jj;
    slong bits1, bits2;
    ulong size1, size2;
    int sign = 0;

    len1 = FLINT_MIN(len1, trunc);
    len2 = FLINT_MIN(len2, trunc);

    len_out = len1 + len2 - 1;
    loglen  = FLINT_CLOG2(len_out);
    loglen2 = FLINT_CLOG2(len2);
    n = (WORD(1) << (loglen - 2));

    size1 = _fmpz_vec_max_limbs(input1, len1); 
    size2 = _fmpz_vec_max_limbs(input2, len2);

    /* Start with an upper bound on the number of bits needed */
    output_bits = FLINT_BITS * (size1 + size2) + loglen2 + 1; 
    
    /* round up for sqrt2 trick */
    output_bits = (((output_bits - 1) >> (loglen - 2)) + 1) << (loglen - 2);

    limbs = (output_bits - 1) / FLINT_BITS + 1; /* initial size of FFT coeffs */
    if (limbs > FFT_MULMOD_2EXPP1_CUTOFF) /* can't be worse than next power of 2 limbs */
        limbs = (WORD(1) << FLINT_CLOG2(limbs));
    size = limbs + 1;

    /* allocate space for ffts */
    ii = flint_malloc((4*(n + n*size) + 5*size)*sizeof(mp_limb_t));
    for (i = 0, ptr = (mp_limb_t *) ii + 4*n; i < 4*n; i++, ptr += size) 
        ii[i] = ptr;
    t1 = ptr;
    t2 = t1 + size;
    s1 = t2 + size;
    tt = s1 + size;

    if (input1 != input2)
    {
        jj = flint_malloc(4*(n + n*size)*sizeof(mp_limb_t));
        for (i = 0, ptr = (mp_limb_t *) jj + 4*n; i < 4*n; i++, ptr += size) 
            jj[i] = ptr;
    } else jj = ii;

    /* put coefficients into FFT vecs */
    bits1 = _fmpz_vec_get_fft(ii, input1, limbs, len1);
    for (i = len1; i < 4*n; i++)
        flint_mpn_zero(ii[i], limbs + 1);

    if (input1 != input2) 
    {
        bits2 = _fmpz_vec_get_fft(jj, input2, limbs, len2);
        for (i = len2; i < 4*n; i++)
            flint_mpn_zero(jj[i], limbs + 1);
    }
    else bits2 = bits1;

    if (bits1 < WORD(0) || bits2 < WORD(0)) 
    {
        sign = 1;  
        bits1 = FLINT_ABS(bits1);
        bits2 = FLINT_ABS(bits2);
    }

    /* Recompute the number of bits/limbs now that we know how large everything is */
    output_bits = bits1 + bits2 + loglen2 + sign;

    /* round up output bits for sqrt2 */
    output_bits = (((output_bits - 1) >> (loglen - 2)) + 1) << (loglen - 2);

    limbs = (output_bits - 1) / FLINT_BITS + 1;
    limbs = fft_adjust_limbs(limbs); /* round up limbs for Nussbaumer */
    
    fft_convolution(ii, jj, loglen - 2, limbs, len_out, &t1, &t2, &s1, tt); 

    _fmpz_vec_set_fft(output, trunc, ii, limbs, sign); /* write output */

    flint_free(ii); 
    if (input1 != input2) 
        flint_free(jj);
}

void
fmpz_poly_mullow_SS(fmpz_poly_t res,
                    const fmpz_poly_t poly1, const fmpz_poly_t poly2, slong n)
{
    const slong len1 = poly1->length;
    const slong len2 = poly2->length;

    if (len1 == 0 || len2 == 0 || n == 0)
    {
        fmpz_poly_zero(res);
        return;
    }

    if (len1 <= 2 || len2 <= 2 || n <= 2)
    {
        fmpz_poly_mullow_classical(res, poly1, poly2, n);
        return;
    }

    n = FLINT_MIN(n, len1 + len2 - 1);
    fmpz_poly_fit_length(res, n);

    if (len1 >= len2)
        _fmpz_poly_mullow_SS(res->coeffs, poly1->coeffs, len1,
                                          poly2->coeffs, len2, n);
    else
        _fmpz_poly_mullow_SS(res->coeffs, poly2->coeffs, len2,
                                          poly1->coeffs, len1, n);

    _fmpz_poly_set_length(res, n);
    _fmpz_poly_normalise(res);
}