flint-sys 0.9.0

Bindings to the FLINT C library
Documentation 
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

/*
    Copyright 2021 Daniel Schultz

    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 "fmpz.h"
#include "fmpz_mat.h"
#include "profiler.h"

int main(void)
{
    slong m, k, n;
    slong dim, i, reps;
    slong new_total, new_mint, new_maxt;
    slong old_total, old_mint, old_maxt;
    double total_den, den;
    flint_bitcnt_t Abits, Bbits;
    timeit_t timer;
    FLINT_TEST_INIT(state);

    flint_set_num_threads(1);

    flint_printf("(m, k, n) *** timings for mul_small nanoseconds per m*k*n (mul in parentheses) ***\n");

    for (m = 3; m < 10; m++)
    for (k = 3; k < 10; k++)
    for (n = 3; n < 10; n++)
    {
        fmpz_mat_t A, B, C, D;

        fmpz_mat_init(A, m, k);
        fmpz_mat_init(B, k, n);
        fmpz_mat_init(C, m, n);
        fmpz_mat_init(D, m, n);

        reps = 1 + 8000000/m/n/k;

        den = reps*m*n*k;

        total_den = 0;

        new_total = 0;
        new_mint = 10000000000;
        new_maxt = 0;

        old_total = 0;
        old_mint = 10000000000;
        old_maxt = 0;

        for (Abits = 14; Abits <= SMALL_FMPZ_BITCOUNT_MAX; Abits += 16)
        for (Bbits = Abits; Bbits <= SMALL_FMPZ_BITCOUNT_MAX; Bbits += 16)
        {
            fmpz_mat_randtest(A, state, Abits);
            fmpz_mat_randtest(B, state, Bbits);

            total_den += den;

            timeit_start(timer);
            for (i = reps; i > 0; i--)
                _fmpz_mat_mul_small(C, A, B);
            timeit_stop(timer);

            new_total += timer->wall;
            new_mint = FLINT_MIN(new_mint, timer->wall);
            new_maxt = FLINT_MAX(new_maxt, timer->wall);

            timeit_start(timer);
            for (i = reps; i > 0; i--)
                fmpz_mat_mul(D, A, B);
            timeit_stop(timer);

            old_total += timer->wall;
            old_mint = FLINT_MIN(old_mint, timer->wall);
            old_maxt = FLINT_MAX(old_maxt, timer->wall);
        }

        flint_printf("(%2wd,%2wd,%2wd): min %.3f ns (%.3f)  max %.3f ns (%.3f)  aver %.3f ns (%.3f)  ratio %.3f\n",
                     m, k, n,
                     1000000*new_mint/den, 1000000*old_mint/den,
                     1000000*new_maxt/den, 1000000*old_maxt/den,
                     1000000*new_total/total_den, 1000000*old_total/total_den,
                     (double)new_total/old_total);

        fmpz_mat_clear(A);
        fmpz_mat_clear(B);
        fmpz_mat_clear(C);
        fmpz_mat_clear(D);
    }

    flint_printf("*** timings for mul_small nanoseconds per dim^3 (mul in parentheses) ***\n");

    for (dim = 5; dim <= 1100; dim += 2 + dim/4)
    {
        fmpz_mat_t A, B, C, D;

        fmpz_mat_init(A, dim, dim);
        fmpz_mat_init(B, dim, dim);
        fmpz_mat_init(C, dim, dim);
        fmpz_mat_init(D, dim, dim);

        reps = 1 + 5000000/dim/dim/dim;

        den = reps*dim*dim*dim;

        total_den = 0;

        new_total = 0;
        new_mint = 10000000000;
        new_maxt = 0;

        old_total = 0;
        old_mint = 10000000000;
        old_maxt = 0;

        for (Abits = 14; Abits <= SMALL_FMPZ_BITCOUNT_MAX; Abits += 8)
        for (Bbits = Abits; Bbits <= SMALL_FMPZ_BITCOUNT_MAX; Bbits += 8)
        {
            fmpz_mat_randtest(A, state, Abits);
            fmpz_mat_randtest(B, state, Bbits);

            total_den += den;

            timeit_start(timer);
            for (i = reps; i > 0; i--)
                _fmpz_mat_mul_small(C, A, B);
            timeit_stop(timer);

            new_total += timer->wall;
            new_mint = FLINT_MIN(new_mint, timer->wall);
            new_maxt = FLINT_MAX(new_maxt, timer->wall);

            timeit_start(timer);
            for (i = reps; i > 0; i--)
                fmpz_mat_mul(D, A, B);
            timeit_stop(timer);

            old_total += timer->wall;
            old_mint = FLINT_MIN(old_mint, timer->wall);
            old_maxt = FLINT_MAX(old_maxt, timer->wall);

            if (!fmpz_mat_equal(C, D))
            {
                flint_printf("oops C != D %wu %wu\n", Abits, Bbits);
                flint_abort();
            }
        }

        flint_printf("%4wd: min %.3f ns (%.3f)  max %.3f ns (%.3f)  aver %.3f ns (%.3f)  ratio %.3f\n",
                     dim,
                     1000000*new_mint/den, 1000000*old_mint/den,
                     1000000*new_maxt/den, 1000000*old_maxt/den,
                     1000000*new_total/total_den, 1000000*old_total/total_den,
                     (double)new_total/old_total);

        fmpz_mat_clear(A);
        fmpz_mat_clear(B);
        fmpz_mat_clear(C);
        fmpz_mat_clear(D);
    }

    FLINT_TEST_CLEAR(state);
    return 0;
}