Mercurial > hg > sv-dependency-builds

annotate src/fftw-3.3.5/kernel/transpose.c @ 84:08ae793730bd

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
Add null config files
author Chris Cannam
date Mon, 02 Mar 2020 14:03:47 +0000
parents 2cd0e3b3e1fd
children
rev   line source
Chris@42 1 /*
Chris@42 2 * Copyright (c) 2003, 2007-14 Matteo Frigo
Chris@42 3 * Copyright (c) 2003, 2007-14 Massachusetts Institute of Technology
Chris@42 4 *
Chris@42 5 * This program is free software; you can redistribute it and/or modify
Chris@42 6 * it under the terms of the GNU General Public License as published by
Chris@42 7 * the Free Software Foundation; either version 2 of the License, or
Chris@42 8 * (at your option) any later version.
Chris@42 9 *
Chris@42 10 * This program is distributed in the hope that it will be useful,
Chris@42 11 * but WITHOUT ANY WARRANTY; without even the implied warranty of
Chris@42 12 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
Chris@42 13 * GNU General Public License for more details.
Chris@42 14 *
Chris@42 15 * You should have received a copy of the GNU General Public License
Chris@42 16 * along with this program; if not, write to the Free Software
Chris@42 17 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
Chris@42 18 *
Chris@42 19 */
Chris@42 20
Chris@42 21 #include "ifftw.h"
Chris@42 22
Chris@42 23 /* in place square transposition, iterative */
Chris@42 24 void X(transpose)(R *I, INT n, INT s0, INT s1, INT vl)
Chris@42 25 {
Chris@42 26 INT i0, i1, v;
Chris@42 27
Chris@42 28 switch (vl) {
Chris@42 29 case 1:
Chris@42 30 for (i1 = 1; i1 < n; ++i1) {
Chris@42 31 for (i0 = 0; i0 < i1; ++i0) {
Chris@42 32 R x0 = I[i1 * s0 + i0 * s1];
Chris@42 33 R y0 = I[i1 * s1 + i0 * s0];
Chris@42 34 I[i1 * s1 + i0 * s0] = x0;
Chris@42 35 I[i1 * s0 + i0 * s1] = y0;
Chris@42 36 }
Chris@42 37 }
Chris@42 38 break;
Chris@42 39 case 2:
Chris@42 40 for (i1 = 1; i1 < n; ++i1) {
Chris@42 41 for (i0 = 0; i0 < i1; ++i0) {
Chris@42 42 R x0 = I[i1 * s0 + i0 * s1];
Chris@42 43 R x1 = I[i1 * s0 + i0 * s1 + 1];
Chris@42 44 R y0 = I[i1 * s1 + i0 * s0];
Chris@42 45 R y1 = I[i1 * s1 + i0 * s0 + 1];
Chris@42 46 I[i1 * s1 + i0 * s0] = x0;
Chris@42 47 I[i1 * s1 + i0 * s0 + 1] = x1;
Chris@42 48 I[i1 * s0 + i0 * s1] = y0;
Chris@42 49 I[i1 * s0 + i0 * s1 + 1] = y1;
Chris@42 50 }
Chris@42 51 }
Chris@42 52 break;
Chris@42 53 default:
Chris@42 54 for (i1 = 1; i1 < n; ++i1) {
Chris@42 55 for (i0 = 0; i0 < i1; ++i0) {
Chris@42 56 for (v = 0; v < vl; ++v) {
Chris@42 57 R x0 = I[i1 * s0 + i0 * s1 + v];
Chris@42 58 R y0 = I[i1 * s1 + i0 * s0 + v];
Chris@42 59 I[i1 * s1 + i0 * s0 + v] = x0;
Chris@42 60 I[i1 * s0 + i0 * s1 + v] = y0;
Chris@42 61 }
Chris@42 62 }
Chris@42 63 }
Chris@42 64 break;
Chris@42 65 }
Chris@42 66 }
Chris@42 67
Chris@42 68 struct transpose_closure {
Chris@42 69 R *I;
Chris@42 70 INT s0, s1, vl, tilesz;
Chris@42 71 R *buf0, *buf1;
Chris@42 72 };
Chris@42 73
Chris@42 74 static void dotile(INT n0l, INT n0u, INT n1l, INT n1u, void *args)
Chris@42 75 {
Chris@42 76 struct transpose_closure *k = (struct transpose_closure *)args;
Chris@42 77 R *I = k->I;
Chris@42 78 INT s0 = k->s0, s1 = k->s1, vl = k->vl;
Chris@42 79 INT i0, i1, v;
Chris@42 80
Chris@42 81 switch (vl) {
Chris@42 82 case 1:
Chris@42 83 for (i1 = n1l; i1 < n1u; ++i1) {
Chris@42 84 for (i0 = n0l; i0 < n0u; ++i0) {
Chris@42 85 R x0 = I[i1 * s0 + i0 * s1];
Chris@42 86 R y0 = I[i1 * s1 + i0 * s0];
Chris@42 87 I[i1 * s1 + i0 * s0] = x0;
Chris@42 88 I[i1 * s0 + i0 * s1] = y0;
Chris@42 89 }
Chris@42 90 }
Chris@42 91 break;
Chris@42 92 case 2:
Chris@42 93 for (i1 = n1l; i1 < n1u; ++i1) {
Chris@42 94 for (i0 = n0l; i0 < n0u; ++i0) {
Chris@42 95 R x0 = I[i1 * s0 + i0 * s1];
Chris@42 96 R x1 = I[i1 * s0 + i0 * s1 + 1];
Chris@42 97 R y0 = I[i1 * s1 + i0 * s0];
Chris@42 98 R y1 = I[i1 * s1 + i0 * s0 + 1];
Chris@42 99 I[i1 * s1 + i0 * s0] = x0;
Chris@42 100 I[i1 * s1 + i0 * s0 + 1] = x1;
Chris@42 101 I[i1 * s0 + i0 * s1] = y0;
Chris@42 102 I[i1 * s0 + i0 * s1 + 1] = y1;
Chris@42 103 }
Chris@42 104 }
Chris@42 105 break;
Chris@42 106 default:
Chris@42 107 for (i1 = n1l; i1 < n1u; ++i1) {
Chris@42 108 for (i0 = n0l; i0 < n0u; ++i0) {
Chris@42 109 for (v = 0; v < vl; ++v) {
Chris@42 110 R x0 = I[i1 * s0 + i0 * s1 + v];
Chris@42 111 R y0 = I[i1 * s1 + i0 * s0 + v];
Chris@42 112 I[i1 * s1 + i0 * s0 + v] = x0;
Chris@42 113 I[i1 * s0 + i0 * s1 + v] = y0;
Chris@42 114 }
Chris@42 115 }
Chris@42 116 }
Chris@42 117 }
Chris@42 118 }
Chris@42 119
Chris@42 120 static void dotile_buf(INT n0l, INT n0u, INT n1l, INT n1u, void *args)
Chris@42 121 {
Chris@42 122 struct transpose_closure *k = (struct transpose_closure *)args;
Chris@42 123 X(cpy2d_ci)(k->I + n0l * k->s0 + n1l * k->s1,
Chris@42 124 k->buf0,
Chris@42 125 n0u - n0l, k->s0, k->vl,
Chris@42 126 n1u - n1l, k->s1, k->vl * (n0u - n0l),
Chris@42 127 k->vl);
Chris@42 128 X(cpy2d_ci)(k->I + n0l * k->s1 + n1l * k->s0,
Chris@42 129 k->buf1,
Chris@42 130 n0u - n0l, k->s1, k->vl,
Chris@42 131 n1u - n1l, k->s0, k->vl * (n0u - n0l),
Chris@42 132 k->vl);
Chris@42 133 X(cpy2d_co)(k->buf1,
Chris@42 134 k->I + n0l * k->s0 + n1l * k->s1,
Chris@42 135 n0u - n0l, k->vl, k->s0,
Chris@42 136 n1u - n1l, k->vl * (n0u - n0l), k->s1,
Chris@42 137 k->vl);
Chris@42 138 X(cpy2d_co)(k->buf0,
Chris@42 139 k->I + n0l * k->s1 + n1l * k->s0,
Chris@42 140 n0u - n0l, k->vl, k->s1,
Chris@42 141 n1u - n1l, k->vl * (n0u - n0l), k->s0,
Chris@42 142 k->vl);
Chris@42 143 }
Chris@42 144
Chris@42 145 static void transpose_rec(R *I, INT n,
Chris@42 146 void (*f)(INT n0l, INT n0u, INT n1l, INT n1u,
Chris@42 147 void *args),
Chris@42 148 struct transpose_closure *k)
Chris@42 149 {
Chris@42 150 tail:
Chris@42 151 if (n > 1) {
Chris@42 152 INT n2 = n / 2;
Chris@42 153 k->I = I;
Chris@42 154 X(tile2d)(0, n2, n2, n, k->tilesz, f, k);
Chris@42 155 transpose_rec(I, n2, f, k);
Chris@42 156 I += n2 * (k->s0 + k->s1); n -= n2; goto tail;
Chris@42 157 }
Chris@42 158 }
Chris@42 159
Chris@42 160 void X(transpose_tiled)(R *I, INT n, INT s0, INT s1, INT vl)
Chris@42 161 {
Chris@42 162 struct transpose_closure k;
Chris@42 163 k.s0 = s0;
Chris@42 164 k.s1 = s1;
Chris@42 165 k.vl = vl;
Chris@42 166 /* two blocks must be in cache, to be swapped */
Chris@42 167 k.tilesz = X(compute_tilesz)(vl, 2);
Chris@42 168 k.buf0 = k.buf1 = 0; /* unused */
Chris@42 169 transpose_rec(I, n, dotile, &k);
Chris@42 170 }
Chris@42 171
Chris@42 172 void X(transpose_tiledbuf)(R *I, INT n, INT s0, INT s1, INT vl)
Chris@42 173 {
Chris@42 174 struct transpose_closure k;
Chris@42 175 /* Assume that the the rows of I conflict into the same cache
Chris@42 176 lines, and therefore we don't need to reserve cache space for
Chris@42 177 the input. If the rows don't conflict, there is no reason
Chris@42 178 to use tiledbuf at all.*/
Chris@42 179 R buf0[CACHESIZE / (2 * sizeof(R))];
Chris@42 180 R buf1[CACHESIZE / (2 * sizeof(R))];
Chris@42 181 k.s0 = s0;
Chris@42 182 k.s1 = s1;
Chris@42 183 k.vl = vl;
Chris@42 184 k.tilesz = X(compute_tilesz)(vl, 2);
Chris@42 185 k.buf0 = buf0;
Chris@42 186 k.buf1 = buf1;
Chris@42 187 A(k.tilesz * k.tilesz * vl * sizeof(R) <= sizeof(buf0));
Chris@42 188 A(k.tilesz * k.tilesz * vl * sizeof(R) <= sizeof(buf1));
Chris@42 189 transpose_rec(I, n, dotile_buf, &k);
Chris@42 190 }
Chris@42 191