| 1 | n/a | /* | 
|---|
| 2 | n/a | * Copyright (c) 2008-2016 Stefan Krah. All rights reserved. | 
|---|
| 3 | n/a | * | 
|---|
| 4 | n/a | * Redistribution and use in source and binary forms, with or without | 
|---|
| 5 | n/a | * modification, are permitted provided that the following conditions | 
|---|
| 6 | n/a | * are met: | 
|---|
| 7 | n/a | * | 
|---|
| 8 | n/a | * 1. Redistributions of source code must retain the above copyright | 
|---|
| 9 | n/a | *    notice, this list of conditions and the following disclaimer. | 
|---|
| 10 | n/a | * | 
|---|
| 11 | n/a | * 2. Redistributions in binary form must reproduce the above copyright | 
|---|
| 12 | n/a | *    notice, this list of conditions and the following disclaimer in the | 
|---|
| 13 | n/a | *    documentation and/or other materials provided with the distribution. | 
|---|
| 14 | n/a | * | 
|---|
| 15 | n/a | * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS "AS IS" AND | 
|---|
| 16 | n/a | * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE | 
|---|
| 17 | n/a | * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE | 
|---|
| 18 | n/a | * ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE | 
|---|
| 19 | n/a | * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL | 
|---|
| 20 | n/a | * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS | 
|---|
| 21 | n/a | * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) | 
|---|
| 22 | n/a | * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT | 
|---|
| 23 | n/a | * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY | 
|---|
| 24 | n/a | * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF | 
|---|
| 25 | n/a | * SUCH DAMAGE. | 
|---|
| 26 | n/a | */ | 
|---|
| 27 | n/a |  | 
|---|
| 28 | n/a |  | 
|---|
| 29 | n/a | #include "mpdecimal.h" | 
|---|
| 30 | n/a | #include <stdio.h> | 
|---|
| 31 | n/a | #include "bits.h" | 
|---|
| 32 | n/a | #include "constants.h" | 
|---|
| 33 | n/a | #include "fnt.h" | 
|---|
| 34 | n/a | #include "fourstep.h" | 
|---|
| 35 | n/a | #include "numbertheory.h" | 
|---|
| 36 | n/a | #include "sixstep.h" | 
|---|
| 37 | n/a | #include "umodarith.h" | 
|---|
| 38 | n/a | #include "convolute.h" | 
|---|
| 39 | n/a |  | 
|---|
| 40 | n/a |  | 
|---|
| 41 | n/a | /* Bignum: Fast convolution using the Number Theoretic Transform. Used for | 
|---|
| 42 | n/a | the multiplication of very large coefficients. */ | 
|---|
| 43 | n/a |  | 
|---|
| 44 | n/a |  | 
|---|
| 45 | n/a | /* Convolute the data in c1 and c2. Result is in c1. */ | 
|---|
| 46 | n/a | int | 
|---|
| 47 | n/a | fnt_convolute(mpd_uint_t *c1, mpd_uint_t *c2, mpd_size_t n, int modnum) | 
|---|
| 48 | n/a | { | 
|---|
| 49 | n/a | int (*fnt)(mpd_uint_t *, mpd_size_t, int); | 
|---|
| 50 | n/a | int (*inv_fnt)(mpd_uint_t *, mpd_size_t, int); | 
|---|
| 51 | n/a | #ifdef PPRO | 
|---|
| 52 | n/a | double dmod; | 
|---|
| 53 | n/a | uint32_t dinvmod[3]; | 
|---|
| 54 | n/a | #endif | 
|---|
| 55 | n/a | mpd_uint_t n_inv, umod; | 
|---|
| 56 | n/a | mpd_size_t i; | 
|---|
| 57 | n/a |  | 
|---|
| 58 | n/a |  | 
|---|
| 59 | n/a | SETMODULUS(modnum); | 
|---|
| 60 | n/a | n_inv = POWMOD(n, (umod-2)); | 
|---|
| 61 | n/a |  | 
|---|
| 62 | n/a | if (ispower2(n)) { | 
|---|
| 63 | n/a | if (n > SIX_STEP_THRESHOLD) { | 
|---|
| 64 | n/a | fnt = six_step_fnt; | 
|---|
| 65 | n/a | inv_fnt = inv_six_step_fnt; | 
|---|
| 66 | n/a | } | 
|---|
| 67 | n/a | else { | 
|---|
| 68 | n/a | fnt = std_fnt; | 
|---|
| 69 | n/a | inv_fnt = std_inv_fnt; | 
|---|
| 70 | n/a | } | 
|---|
| 71 | n/a | } | 
|---|
| 72 | n/a | else { | 
|---|
| 73 | n/a | fnt = four_step_fnt; | 
|---|
| 74 | n/a | inv_fnt = inv_four_step_fnt; | 
|---|
| 75 | n/a | } | 
|---|
| 76 | n/a |  | 
|---|
| 77 | n/a | if (!fnt(c1, n, modnum)) { | 
|---|
| 78 | n/a | return 0; | 
|---|
| 79 | n/a | } | 
|---|
| 80 | n/a | if (!fnt(c2, n, modnum)) { | 
|---|
| 81 | n/a | return 0; | 
|---|
| 82 | n/a | } | 
|---|
| 83 | n/a | for (i = 0; i < n-1; i += 2) { | 
|---|
| 84 | n/a | mpd_uint_t x0 = c1[i]; | 
|---|
| 85 | n/a | mpd_uint_t y0 = c2[i]; | 
|---|
| 86 | n/a | mpd_uint_t x1 = c1[i+1]; | 
|---|
| 87 | n/a | mpd_uint_t y1 = c2[i+1]; | 
|---|
| 88 | n/a | MULMOD2(&x0, y0, &x1, y1); | 
|---|
| 89 | n/a | c1[i] = x0; | 
|---|
| 90 | n/a | c1[i+1] = x1; | 
|---|
| 91 | n/a | } | 
|---|
| 92 | n/a |  | 
|---|
| 93 | n/a | if (!inv_fnt(c1, n, modnum)) { | 
|---|
| 94 | n/a | return 0; | 
|---|
| 95 | n/a | } | 
|---|
| 96 | n/a | for (i = 0; i < n-3; i += 4) { | 
|---|
| 97 | n/a | mpd_uint_t x0 = c1[i]; | 
|---|
| 98 | n/a | mpd_uint_t x1 = c1[i+1]; | 
|---|
| 99 | n/a | mpd_uint_t x2 = c1[i+2]; | 
|---|
| 100 | n/a | mpd_uint_t x3 = c1[i+3]; | 
|---|
| 101 | n/a | MULMOD2C(&x0, &x1, n_inv); | 
|---|
| 102 | n/a | MULMOD2C(&x2, &x3, n_inv); | 
|---|
| 103 | n/a | c1[i] = x0; | 
|---|
| 104 | n/a | c1[i+1] = x1; | 
|---|
| 105 | n/a | c1[i+2] = x2; | 
|---|
| 106 | n/a | c1[i+3] = x3; | 
|---|
| 107 | n/a | } | 
|---|
| 108 | n/a |  | 
|---|
| 109 | n/a | return 1; | 
|---|
| 110 | n/a | } | 
|---|
| 111 | n/a |  | 
|---|
| 112 | n/a | /* Autoconvolute the data in c1. Result is in c1. */ | 
|---|
| 113 | n/a | int | 
|---|
| 114 | n/a | fnt_autoconvolute(mpd_uint_t *c1, mpd_size_t n, int modnum) | 
|---|
| 115 | n/a | { | 
|---|
| 116 | n/a | int (*fnt)(mpd_uint_t *, mpd_size_t, int); | 
|---|
| 117 | n/a | int (*inv_fnt)(mpd_uint_t *, mpd_size_t, int); | 
|---|
| 118 | n/a | #ifdef PPRO | 
|---|
| 119 | n/a | double dmod; | 
|---|
| 120 | n/a | uint32_t dinvmod[3]; | 
|---|
| 121 | n/a | #endif | 
|---|
| 122 | n/a | mpd_uint_t n_inv, umod; | 
|---|
| 123 | n/a | mpd_size_t i; | 
|---|
| 124 | n/a |  | 
|---|
| 125 | n/a |  | 
|---|
| 126 | n/a | SETMODULUS(modnum); | 
|---|
| 127 | n/a | n_inv = POWMOD(n, (umod-2)); | 
|---|
| 128 | n/a |  | 
|---|
| 129 | n/a | if (ispower2(n)) { | 
|---|
| 130 | n/a | if (n > SIX_STEP_THRESHOLD) { | 
|---|
| 131 | n/a | fnt = six_step_fnt; | 
|---|
| 132 | n/a | inv_fnt = inv_six_step_fnt; | 
|---|
| 133 | n/a | } | 
|---|
| 134 | n/a | else { | 
|---|
| 135 | n/a | fnt = std_fnt; | 
|---|
| 136 | n/a | inv_fnt = std_inv_fnt; | 
|---|
| 137 | n/a | } | 
|---|
| 138 | n/a | } | 
|---|
| 139 | n/a | else { | 
|---|
| 140 | n/a | fnt = four_step_fnt; | 
|---|
| 141 | n/a | inv_fnt = inv_four_step_fnt; | 
|---|
| 142 | n/a | } | 
|---|
| 143 | n/a |  | 
|---|
| 144 | n/a | if (!fnt(c1, n, modnum)) { | 
|---|
| 145 | n/a | return 0; | 
|---|
| 146 | n/a | } | 
|---|
| 147 | n/a | for (i = 0; i < n-1; i += 2) { | 
|---|
| 148 | n/a | mpd_uint_t x0 = c1[i]; | 
|---|
| 149 | n/a | mpd_uint_t x1 = c1[i+1]; | 
|---|
| 150 | n/a | MULMOD2(&x0, x0, &x1, x1); | 
|---|
| 151 | n/a | c1[i] = x0; | 
|---|
| 152 | n/a | c1[i+1] = x1; | 
|---|
| 153 | n/a | } | 
|---|
| 154 | n/a |  | 
|---|
| 155 | n/a | if (!inv_fnt(c1, n, modnum)) { | 
|---|
| 156 | n/a | return 0; | 
|---|
| 157 | n/a | } | 
|---|
| 158 | n/a | for (i = 0; i < n-3; i += 4) { | 
|---|
| 159 | n/a | mpd_uint_t x0 = c1[i]; | 
|---|
| 160 | n/a | mpd_uint_t x1 = c1[i+1]; | 
|---|
| 161 | n/a | mpd_uint_t x2 = c1[i+2]; | 
|---|
| 162 | n/a | mpd_uint_t x3 = c1[i+3]; | 
|---|
| 163 | n/a | MULMOD2C(&x0, &x1, n_inv); | 
|---|
| 164 | n/a | MULMOD2C(&x2, &x3, n_inv); | 
|---|
| 165 | n/a | c1[i] = x0; | 
|---|
| 166 | n/a | c1[i+1] = x1; | 
|---|
| 167 | n/a | c1[i+2] = x2; | 
|---|
| 168 | n/a | c1[i+3] = x3; | 
|---|
| 169 | n/a | } | 
|---|
| 170 | n/a |  | 
|---|
| 171 | n/a | return 1; | 
|---|
| 172 | n/a | } | 
|---|
| 173 | n/a |  | 
|---|
| 174 | n/a |  | 
|---|