Mercurial > hg > pmhd

annotate ffmpeg/libavcodec/mips/mpegaudiodsp_mips_float.c @ 13:844d341cf643 tip

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
Back up before ISMIR
author Yading Song <yading.song@eecs.qmul.ac.uk>
date Thu, 31 Oct 2013 13:17:06 +0000
parents 6840f77b83aa
children
rev   line source
yading@10 1 /*
yading@10 2 * Copyright (c) 2012
yading@10 3 * MIPS Technologies, Inc., California.
yading@10 4 *
yading@10 5 * Redistribution and use in source and binary forms, with or without
yading@10 6 * modification, are permitted provided that the following conditions
yading@10 7 * are met:
yading@10 8 * 1. Redistributions of source code must retain the above copyright
yading@10 9 * notice, this list of conditions and the following disclaimer.
yading@10 10 * 2. Redistributions in binary form must reproduce the above copyright
yading@10 11 * notice, this list of conditions and the following disclaimer in the
yading@10 12 * documentation and/or other materials provided with the distribution.
yading@10 13 * 3. Neither the name of the MIPS Technologies, Inc., nor the names of its
yading@10 14 * contributors may be used to endorse or promote products derived from
yading@10 15 * this software without specific prior written permission.
yading@10 16 *
yading@10 17 * THIS SOFTWARE IS PROVIDED BY THE MIPS TECHNOLOGIES, INC. ``AS IS'' AND
yading@10 18 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
yading@10 19 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
yading@10 20 * ARE DISCLAIMED. IN NO EVENT SHALL THE MIPS TECHNOLOGIES, INC. BE LIABLE
yading@10 21 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
yading@10 22 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
yading@10 23 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
yading@10 24 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
yading@10 25 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
yading@10 26 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
yading@10 27 * SUCH DAMAGE.
yading@10 28 *
yading@10 29 * Author: Bojan Zivkovic (bojan@mips.com)
yading@10 30 *
yading@10 31 * MPEG Audio decoder optimized for MIPS floating-point architecture
yading@10 32 *
yading@10 33 * This file is part of FFmpeg.
yading@10 34 *
yading@10 35 * FFmpeg is free software; you can redistribute it and/or
yading@10 36 * modify it under the terms of the GNU Lesser General Public
yading@10 37 * License as published by the Free Software Foundation; either
yading@10 38 * version 2.1 of the License, or (at your option) any later version.
yading@10 39 *
yading@10 40 * FFmpeg is distributed in the hope that it will be useful,
yading@10 41 * but WITHOUT ANY WARRANTY; without even the implied warranty of
yading@10 42 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
yading@10 43 * Lesser General Public License for more details.
yading@10 44 *
yading@10 45 * You should have received a copy of the GNU Lesser General Public
yading@10 46 * License along with FFmpeg; if not, write to the Free Software
yading@10 47 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
yading@10 48 */
yading@10 49
yading@10 50 /**
yading@10 51 * @file
yading@10 52 * Reference: libavcodec/mpegaudiodsp_template.c
yading@10 53 * libavcodec/dct32.c
yading@10 54 */
yading@10 55
yading@10 56 #include <string.h>
yading@10 57
yading@10 58 #include "libavcodec/mpegaudiodsp.h"
yading@10 59
yading@10 60 static void ff_mpadsp_apply_window_mips_float(float *synth_buf, float *window,
yading@10 61 int *dither_state, float *samples, int incr)
yading@10 62 {
yading@10 63 register const float *w, *w2, *p;
yading@10 64 int j;
yading@10 65 float *samples2;
yading@10 66 float sum, sum2;
yading@10 67 /* temporary variables */
yading@10 68 int incr1 = incr << 2;
yading@10 69 int t_sample;
yading@10 70 float in1, in2, in3, in4, in5, in6, in7, in8;
yading@10 71 float *p2;
yading@10 72
yading@10 73 /* copy to avoid wrap */
yading@10 74 memcpy(synth_buf + 512, synth_buf, 32 * sizeof(*synth_buf));
yading@10 75
yading@10 76 /**
yading@10 77 * instructions are scheduled to minimize pipeline stall.
yading@10 78 * use of round_sample function from the original code is
yading@10 79 * changed with appropriate assembly instructions.
yading@10 80 */
yading@10 81
yading@10 82 __asm__ volatile (
yading@10 83 "lwc1 %[sum], 0(%[dither_state]) \t\n"
yading@10 84 "sll %[t_sample], %[incr1], 5 \t\n"
yading@10 85 "sub %[t_sample], %[t_sample], %[incr1] \n\t"
yading@10 86 "li %[j], 4 \t\n"
yading@10 87 "lwc1 %[in1], 0(%[window]) \t\n"
yading@10 88 "lwc1 %[in2], 16*4(%[synth_buf]) \t\n"
yading@10 89 "sw $zero, 0(%[dither_state]) \t\n"
yading@10 90 "lwc1 %[in3], 64*4(%[window]) \t\n"
yading@10 91 "lwc1 %[in4], 80*4(%[synth_buf]) \t\n"
yading@10 92 "addu %[samples2], %[samples], %[t_sample] \t\n"
yading@10 93 "madd.s %[sum], %[sum], %[in1], %[in2] \t\n"
yading@10 94 "lwc1 %[in5], 128*4(%[window]) \t\n"
yading@10 95 "lwc1 %[in6], 144*4(%[synth_buf]) \t\n"
yading@10 96 "lwc1 %[in7], 192*4(%[window]) \t\n"
yading@10 97 "madd.s %[sum], %[sum], %[in3], %[in4] \t\n"
yading@10 98 "lwc1 %[in8], 208*4(%[synth_buf]) \t\n"
yading@10 99 "lwc1 %[in1], 256*4(%[window]) \t\n"
yading@10 100 "lwc1 %[in2], 272*4(%[synth_buf]) \t\n"
yading@10 101 "madd.s %[sum], %[sum], %[in5], %[in6] \t\n"
yading@10 102 "lwc1 %[in3], 320*4(%[window]) \t\n"
yading@10 103 "lwc1 %[in4], 336*4(%[synth_buf]) \t\n"
yading@10 104 "lwc1 %[in5], 384*4(%[window]) \t\n"
yading@10 105 "madd.s %[sum], %[sum], %[in7], %[in8] \t\n"
yading@10 106 "lwc1 %[in6], 400*4(%[synth_buf]) \t\n"
yading@10 107 "lwc1 %[in7], 448*4(%[window]) \t\n"
yading@10 108 "lwc1 %[in8], 464*4(%[synth_buf]) \t\n"
yading@10 109 "madd.s %[sum], %[sum], %[in1], %[in2] \t\n"
yading@10 110 "lwc1 %[in1], 32*4(%[window]) \t\n"
yading@10 111 "lwc1 %[in2], 48*4(%[synth_buf]) \t\n"
yading@10 112 "madd.s %[sum], %[sum], %[in3], %[in4] \t\n"
yading@10 113 "lwc1 %[in3], 96*4(%[window]) \t\n"
yading@10 114 "lwc1 %[in4], 112*4(%[synth_buf]) \t\n"
yading@10 115 "madd.s %[sum], %[sum], %[in5], %[in6] \t\n"
yading@10 116 "lwc1 %[in5], 160*4(%[window]) \t\n"
yading@10 117 "lwc1 %[in6], 176*4(%[synth_buf]) \t\n"
yading@10 118 "madd.s %[sum], %[sum], %[in7], %[in8] \t\n"
yading@10 119 "lwc1 %[in7], 224*4(%[window]) \t\n"
yading@10 120 "lwc1 %[in8], 240*4(%[synth_buf]) \t\n"
yading@10 121 "nmsub.s %[sum], %[sum], %[in1], %[in2] \t\n"
yading@10 122 "lwc1 %[in1], 288*4(%[window]) \t\n"
yading@10 123 "lwc1 %[in2], 304*4(%[synth_buf]) \t\n"
yading@10 124 "nmsub.s %[sum], %[sum], %[in3], %[in4] \t\n"
yading@10 125 "lwc1 %[in3], 352*4(%[window]) \t\n"
yading@10 126 "lwc1 %[in4], 368*4(%[synth_buf]) \t\n"
yading@10 127 "nmsub.s %[sum], %[sum], %[in5], %[in6] \t\n"
yading@10 128 "lwc1 %[in5], 416*4(%[window]) \t\n"
yading@10 129 "lwc1 %[in6], 432*4(%[synth_buf]) \t\n"
yading@10 130 "nmsub.s %[sum], %[sum], %[in7], %[in8] \t\n"
yading@10 131 "lwc1 %[in7], 480*4(%[window]) \t\n"
yading@10 132 "lwc1 %[in8], 496*4(%[synth_buf]) \t\n"
yading@10 133 "nmsub.s %[sum], %[sum], %[in1], %[in2] \t\n"
yading@10 134 "addu %[w], %[window], 4 \t\n"
yading@10 135 "nmsub.s %[sum], %[sum], %[in3], %[in4] \t\n"
yading@10 136 "addu %[w2], %[window], 124 \t\n"
yading@10 137 "addiu %[p], %[synth_buf], 68 \t\n"
yading@10 138 "addiu %[p2], %[synth_buf], 188 \t\n"
yading@10 139 "nmsub.s %[sum], %[sum], %[in5], %[in6] \t\n"
yading@10 140 "nmsub.s %[sum], %[sum], %[in7], %[in8] \t\n"
yading@10 141 "swc1 %[sum], 0(%[samples]) \t\n"
yading@10 142 "addu %[samples], %[samples], %[incr1] \t\n"
yading@10 143
yading@10 144 /* calculate two samples at the same time to avoid one memory
yading@10 145 access per two sample */
yading@10 146
yading@10 147 "ff_mpadsp_apply_window_loop%=: \t\n"
yading@10 148 "lwc1 %[in1], 0(%[w]) \t\n"
yading@10 149 "lwc1 %[in2], 0(%[p]) \t\n"
yading@10 150 "lwc1 %[in3], 0(%[w2]) \t\n"
yading@10 151 "lwc1 %[in4], 64*4(%[w]) \t\n"
yading@10 152 "lwc1 %[in5], 64*4(%[p]) \t\n"
yading@10 153 "lwc1 %[in6], 64*4(%[w2]) \t\n"
yading@10 154 "mul.s %[sum], %[in1], %[in2] \t\n"
yading@10 155 "mul.s %[sum2], %[in2], %[in3] \t\n"
yading@10 156 "lwc1 %[in1], 128*4(%[w]) \t\n"
yading@10 157 "lwc1 %[in2], 128*4(%[p]) \t\n"
yading@10 158 "madd.s %[sum], %[sum], %[in4], %[in5] \t\n"
yading@10 159 "nmadd.s %[sum2], %[sum2], %[in5], %[in6] \t\n"
yading@10 160 "lwc1 %[in3], 128*4(%[w2]) \t\n"
yading@10 161 "lwc1 %[in4], 192*4(%[w]) \t\n"
yading@10 162 "madd.s %[sum], %[sum], %[in1], %[in2] \t\n"
yading@10 163 "lwc1 %[in5], 192*4(%[p]) \t\n"
yading@10 164 "lwc1 %[in6], 192*4(%[w2]) \t\n"
yading@10 165 "nmsub.s %[sum2], %[sum2], %[in2], %[in3] \t\n"
yading@10 166 "lwc1 %[in1], 256*4(%[w]) \t\n"
yading@10 167 "lwc1 %[in2], 256*4(%[p]) \t\n"
yading@10 168 "madd.s %[sum], %[sum], %[in4], %[in5] \t\n"
yading@10 169 "nmsub.s %[sum2], %[sum2], %[in5], %[in6] \t\n"
yading@10 170 "lwc1 %[in3], 256*4(%[w2]) \t\n"
yading@10 171 "lwc1 %[in4], 320*4(%[w]) \t\n"
yading@10 172 "madd.s %[sum], %[sum], %[in1], %[in2] \t\n"
yading@10 173 "lwc1 %[in5], 320*4(%[p]) \t\n"
yading@10 174 "lwc1 %[in6], 320*4(%[w2]) \t\n"
yading@10 175 "nmsub.s %[sum2], %[sum2], %[in2], %[in3] \t\n"
yading@10 176 "lwc1 %[in1], 384*4(%[w]) \t\n"
yading@10 177 "lwc1 %[in2], 384*4(%[p]) \t\n"
yading@10 178 "madd.s %[sum], %[sum], %[in4], %[in5] \t\n"
yading@10 179 "nmsub.s %[sum2], %[sum2], %[in5], %[in6] \t\n"
yading@10 180 "lwc1 %[in3], 384*4(%[w2]) \t\n"
yading@10 181 "lwc1 %[in4], 448*4(%[w]) \t\n"
yading@10 182 "madd.s %[sum], %[sum], %[in1], %[in2] \t\n"
yading@10 183 "lwc1 %[in5], 448*4(%[p]) \t\n"
yading@10 184 "lwc1 %[in6], 448*4(%[w2]) \t\n"
yading@10 185 "nmsub.s %[sum2], %[sum2], %[in2], %[in3] \t\n"
yading@10 186 "madd.s %[sum], %[sum], %[in4], %[in5] \t\n"
yading@10 187 "lwc1 %[in1], 32*4(%[w]) \t\n"
yading@10 188 "lwc1 %[in2], 0(%[p2]) \t\n"
yading@10 189 "nmsub.s %[sum2], %[sum2], %[in5], %[in6] \t\n"
yading@10 190 "lwc1 %[in3], 32*4(%[w2]) \t\n"
yading@10 191 "lwc1 %[in4], 96*4(%[w]) \t\n"
yading@10 192 "lwc1 %[in5], 64*4(%[p2]) \t\n"
yading@10 193 "nmsub.s %[sum], %[sum], %[in1], %[in2] \t\n"
yading@10 194 "lwc1 %[in6], 96*4(%[w2]) \t\n"
yading@10 195 "nmsub.s %[sum2], %[sum2], %[in2], %[in3] \t\n"
yading@10 196 "lwc1 %[in1], 160*4(%[w]) \t\n"
yading@10 197 "nmsub.s %[sum], %[sum], %[in4], %[in5] \t\n"
yading@10 198 "lwc1 %[in2], 128*4(%[p2]) \t\n"
yading@10 199 "nmsub.s %[sum2], %[sum2], %[in5], %[in6] \t\n"
yading@10 200 "lwc1 %[in3], 160*4(%[w2]) \t\n"
yading@10 201 "lwc1 %[in4], 224*4(%[w]) \t\n"
yading@10 202 "lwc1 %[in5], 192*4(%[p2]) \t\n"
yading@10 203 "nmsub.s %[sum], %[sum], %[in1], %[in2] \t\n"
yading@10 204 "lwc1 %[in6], 224*4(%[w2]) \t\n"
yading@10 205 "nmsub.s %[sum2], %[sum2], %[in2], %[in3] \t\n"
yading@10 206 "lwc1 %[in1], 288*4(%[w]) \t\n"
yading@10 207 "nmsub.s %[sum], %[sum], %[in4], %[in5] \t\n"
yading@10 208 "lwc1 %[in2], 256*4(%[p2]) \t\n"
yading@10 209 "nmsub.s %[sum2], %[sum2], %[in5], %[in6] \t\n"
yading@10 210 "lwc1 %[in3], 288*4(%[w2]) \t\n"
yading@10 211 "lwc1 %[in4], 352*4(%[w]) \t\n"
yading@10 212 "lwc1 %[in5], 320*4(%[p2]) \t\n"
yading@10 213 "nmsub.s %[sum], %[sum], %[in1], %[in2] \t\n"
yading@10 214 "lwc1 %[in6], 352*4(%[w2]) \t\n"
yading@10 215 "nmsub.s %[sum2], %[sum2], %[in2], %[in3] \t\n"
yading@10 216 "lwc1 %[in1], 416*4(%[w]) \t\n"
yading@10 217 "nmsub.s %[sum], %[sum], %[in4], %[in5] \t\n"
yading@10 218 "lwc1 %[in2], 384*4(%[p2]) \t\n"
yading@10 219 "nmsub.s %[sum2], %[sum2], %[in5], %[in6] \t\n"
yading@10 220 "lwc1 %[in3], 416*4(%[w2]) \t\n"
yading@10 221 "lwc1 %[in4], 480*4(%[w]) \t\n"
yading@10 222 "lwc1 %[in5], 448*4(%[p2]) \t\n"
yading@10 223 "nmsub.s %[sum], %[sum], %[in1], %[in2] \t\n"
yading@10 224 "lwc1 %[in6], 480*4(%[w2]) \t\n"
yading@10 225 "nmsub.s %[sum2], %[sum2], %[in2], %[in3] \t\n"
yading@10 226 "addiu %[w], %[w], 4 \t\n"
yading@10 227 "nmsub.s %[sum], %[sum], %[in4], %[in5] \t\n"
yading@10 228 "addiu %[w2], %[w2], -4 \t\n"
yading@10 229 "nmsub.s %[sum2], %[sum2], %[in5], %[in6] \t\n"
yading@10 230 "addu %[j], %[j], 4 \t\n"
yading@10 231 "addiu %[p], 4 \t\n"
yading@10 232 "swc1 %[sum], 0(%[samples]) \t\n"
yading@10 233 "addiu %[p2], -4 \t\n"
yading@10 234 "swc1 %[sum2], 0(%[samples2]) \t\n"
yading@10 235 "addu %[samples], %[samples], %[incr1] \t\n"
yading@10 236 "subu %[samples2], %[samples2], %[incr1] \t\n"
yading@10 237 "bne %[j], 64, ff_mpadsp_apply_window_loop%= \t\n"
yading@10 238
yading@10 239 "lwc1 %[in1], 48*4(%[window]) \t\n"
yading@10 240 "lwc1 %[in2], 32*4(%[synth_buf]) \t\n"
yading@10 241 "lwc1 %[in3], 112*4(%[window]) \t\n"
yading@10 242 "lwc1 %[in4], 96*4(%[synth_buf]) \t\n"
yading@10 243 "lwc1 %[in5], 176*4(%[window]) \t\n"
yading@10 244 "lwc1 %[in6], 160*4(%[synth_buf]) \t\n"
yading@10 245 "mul.s %[sum], %[in1], %[in2] \t\n"
yading@10 246 "lwc1 %[in7], 240*4(%[window]) \t\n"
yading@10 247 "lwc1 %[in8], 224*4(%[synth_buf]) \t\n"
yading@10 248 "lwc1 %[in1], 304*4(%[window]) \t\n"
yading@10 249 "nmadd.s %[sum], %[sum], %[in3], %[in4] \t\n"
yading@10 250 "lwc1 %[in2], 288*4(%[synth_buf]) \t\n"
yading@10 251 "lwc1 %[in3], 368*4(%[window]) \t\n"
yading@10 252 "lwc1 %[in4], 352*4(%[synth_buf]) \t\n"
yading@10 253 "nmsub.s %[sum], %[sum], %[in5], %[in6] \t\n"
yading@10 254 "nmsub.s %[sum], %[sum], %[in7], %[in8] \t\n"
yading@10 255 "lwc1 %[in5], 432*4(%[window]) \t\n"
yading@10 256 "lwc1 %[in6], 416*4(%[synth_buf]) \t\n"
yading@10 257 "nmsub.s %[sum], %[sum], %[in1], %[in2] \t\n"
yading@10 258 "lwc1 %[in7], 496*4(%[window]) \t\n"
yading@10 259 "lwc1 %[in8], 480*4(%[synth_buf]) \t\n"
yading@10 260 "nmsub.s %[sum], %[sum], %[in3], %[in4] \t\n"
yading@10 261 "nmsub.s %[sum], %[sum], %[in5], %[in6] \t\n"
yading@10 262 "nmsub.s %[sum], %[sum], %[in7], %[in8] \t\n"
yading@10 263 "swc1 %[sum], 0(%[samples]) \t\n"
yading@10 264
yading@10 265 : [sum] "=&f" (sum), [sum2] "=&f" (sum2),
yading@10 266 [w2] "=&r" (w2), [w] "=&r" (w),
yading@10 267 [p] "=&r" (p), [p2] "=&r" (p2), [j] "=&r" (j),
yading@10 268 [samples] "+r" (samples), [samples2] "=&r" (samples2),
yading@10 269 [in1] "=&f" (in1), [in2] "=&f" (in2),
yading@10 270 [in3] "=&f" (in3), [in4] "=&f" (in4),
yading@10 271 [in5] "=&f" (in5), [in6] "=&f" (in6),
yading@10 272 [in7] "=&f" (in7), [in8] "=&f" (in8),
yading@10 273 [t_sample] "=&r" (t_sample)
yading@10 274 : [synth_buf] "r" (synth_buf), [window] "r" (window),
yading@10 275 [dither_state] "r" (dither_state), [incr1] "r" (incr1)
yading@10 276 : "memory"
yading@10 277 );
yading@10 278 }
yading@10 279
yading@10 280 static void ff_dct32_mips_float(float *out, const float *tab)
yading@10 281 {
yading@10 282 float val0 , val1 , val2 , val3 , val4 , val5 , val6 , val7,
yading@10 283 val8 , val9 , val10, val11, val12, val13, val14, val15,
yading@10 284 val16, val17, val18, val19, val20, val21, val22, val23,
yading@10 285 val24, val25, val26, val27, val28, val29, val30, val31;
yading@10 286 float fTmp1, fTmp2, fTmp3, fTmp4, fTmp5, fTmp6, fTmp7, fTmp8,
yading@10 287 fTmp9, fTmp10, fTmp11;
yading@10 288
yading@10 289 /**
yading@10 290 * instructions are scheduled to minimize pipeline stall.
yading@10 291 */
yading@10 292 __asm__ volatile (
yading@10 293 "lwc1 %[fTmp1], 0*4(%[tab]) \n\t"
yading@10 294 "lwc1 %[fTmp2], 31*4(%[tab]) \n\t"
yading@10 295 "lwc1 %[fTmp3], 15*4(%[tab]) \n\t"
yading@10 296 "lwc1 %[fTmp4], 16*4(%[tab]) \n\t"
yading@10 297 "li.s %[fTmp7], 0.50241928618815570551 \n\t"
yading@10 298 "add.s %[fTmp5], %[fTmp1], %[fTmp2] \n\t"
yading@10 299 "sub.s %[fTmp8], %[fTmp1], %[fTmp2] \n\t"
yading@10 300 "add.s %[fTmp6], %[fTmp3], %[fTmp4] \n\t"
yading@10 301 "sub.s %[fTmp9], %[fTmp3], %[fTmp4] \n\t"
yading@10 302 "li.s %[fTmp10], 0.50060299823519630134 \n\t"
yading@10 303 "li.s %[fTmp11], 10.19000812354805681150 \n\t"
yading@10 304 "mul.s %[fTmp8], %[fTmp8], %[fTmp10] \n\t"
yading@10 305 "add.s %[val0], %[fTmp5], %[fTmp6] \n\t"
yading@10 306 "sub.s %[val15], %[fTmp5], %[fTmp6] \n\t"
yading@10 307 "lwc1 %[fTmp1], 7*4(%[tab]) \n\t"
yading@10 308 "lwc1 %[fTmp2], 24*4(%[tab]) \n\t"
yading@10 309 "madd.s %[val16], %[fTmp8], %[fTmp9], %[fTmp11] \n\t"
yading@10 310 "nmsub.s %[val31], %[fTmp8], %[fTmp9], %[fTmp11] \n\t"
yading@10 311 "mul.s %[val15], %[val15], %[fTmp7] \n\t"
yading@10 312 "lwc1 %[fTmp3], 8*4(%[tab]) \n\t"
yading@10 313 "lwc1 %[fTmp4], 23*4(%[tab]) \n\t"
yading@10 314 "add.s %[fTmp5], %[fTmp1], %[fTmp2] \n\t"
yading@10 315 "mul.s %[val31], %[val31], %[fTmp7] \n\t"
yading@10 316 "sub.s %[fTmp8], %[fTmp1], %[fTmp2] \n\t"
yading@10 317 "add.s %[fTmp6], %[fTmp3], %[fTmp4] \n\t"
yading@10 318 "sub.s %[fTmp9], %[fTmp3], %[fTmp4] \n\t"
yading@10 319 "li.s %[fTmp7], 5.10114861868916385802 \n\t"
yading@10 320 "li.s %[fTmp10], 0.67480834145500574602 \n\t"
yading@10 321 "li.s %[fTmp11], 0.74453627100229844977 \n\t"
yading@10 322 "add.s %[val7], %[fTmp5], %[fTmp6] \n\t"
yading@10 323 "sub.s %[val8], %[fTmp5], %[fTmp6] \n\t"
yading@10 324 "mul.s %[fTmp8], %[fTmp8], %[fTmp10] \n\t"
yading@10 325 "li.s %[fTmp1], 0.50979557910415916894 \n\t"
yading@10 326 "sub.s %[fTmp2], %[val0], %[val7] \n\t"
yading@10 327 "mul.s %[val8], %[val8], %[fTmp7] \n\t"
yading@10 328 "madd.s %[val23], %[fTmp8], %[fTmp9], %[fTmp11] \n\t"
yading@10 329 "nmsub.s %[val24], %[fTmp8], %[fTmp9], %[fTmp11] \n\t"
yading@10 330 "add.s %[val0], %[val0], %[val7] \n\t"
yading@10 331 "mul.s %[val7], %[fTmp1], %[fTmp2] \n\t"
yading@10 332 "sub.s %[fTmp2], %[val15], %[val8] \n\t"
yading@10 333 "add.s %[val8], %[val15], %[val8] \n\t"
yading@10 334 "mul.s %[val24], %[val24], %[fTmp7] \n\t"
yading@10 335 "sub.s %[fTmp3], %[val16], %[val23] \n\t"
yading@10 336 "add.s %[val16], %[val16], %[val23] \n\t"
yading@10 337 "mul.s %[val15], %[fTmp1], %[fTmp2] \n\t"
yading@10 338 "sub.s %[fTmp4], %[val31], %[val24] \n\t"
yading@10 339 "mul.s %[val23], %[fTmp1], %[fTmp3] \n\t"
yading@10 340 "add.s %[val24], %[val31], %[val24] \n\t"
yading@10 341 "mul.s %[val31], %[fTmp1], %[fTmp4] \n\t"
yading@10 342
yading@10 343 : [fTmp1] "=&f" (fTmp1), [fTmp2] "=&f" (fTmp2), [fTmp3] "=&f" (fTmp3),
yading@10 344 [fTmp4] "=&f" (fTmp4), [fTmp5] "=&f" (fTmp5), [fTmp6] "=&f" (fTmp6),
yading@10 345 [fTmp7] "=&f" (fTmp7), [fTmp8] "=&f" (fTmp8), [fTmp9] "=&f" (fTmp9),
yading@10 346 [fTmp10] "=&f" (fTmp10), [fTmp11] "=&f" (fTmp11),
yading@10 347 [val0] "=f" (val0), [val7] "=f" (val7),
yading@10 348 [val8] "=f" (val8), [val15] "=f" (val15),
yading@10 349 [val16] "=f" (val16), [val23] "=f" (val23),
yading@10 350 [val24] "=f" (val24), [val31] "=f" (val31)
yading@10 351 : [tab] "r" (tab)
yading@10 352 : "memory"
yading@10 353 );
yading@10 354
yading@10 355 __asm__ volatile (
yading@10 356 "lwc1 %[fTmp1], 3*4(%[tab]) \n\t"
yading@10 357 "lwc1 %[fTmp2], 28*4(%[tab]) \n\t"
yading@10 358 "lwc1 %[fTmp3], 12*4(%[tab]) \n\t"
yading@10 359 "lwc1 %[fTmp4], 19*4(%[tab]) \n\t"
yading@10 360 "li.s %[fTmp7], 0.64682178335999012954 \n\t"
yading@10 361 "add.s %[fTmp5], %[fTmp1], %[fTmp2] \n\t"
yading@10 362 "sub.s %[fTmp8], %[fTmp1], %[fTmp2] \n\t"
yading@10 363 "add.s %[fTmp6], %[fTmp3], %[fTmp4] \n\t"
yading@10 364 "sub.s %[fTmp9], %[fTmp3], %[fTmp4] \n\t"
yading@10 365 "li.s %[fTmp10], 0.53104259108978417447 \n\t"
yading@10 366 "li.s %[fTmp11], 1.48416461631416627724 \n\t"
yading@10 367 "mul.s %[fTmp8], %[fTmp8], %[fTmp10] \n\t"
yading@10 368 "add.s %[val3], %[fTmp5], %[fTmp6] \n\t"
yading@10 369 "sub.s %[val12], %[fTmp5], %[fTmp6] \n\t"
yading@10 370 "lwc1 %[fTmp1], 4*4(%[tab]) \n\t"
yading@10 371 "lwc1 %[fTmp2], 27*4(%[tab]) \n\t"
yading@10 372 "madd.s %[val19], %[fTmp8], %[fTmp9], %[fTmp11] \n\t"
yading@10 373 "nmsub.s %[val28], %[fTmp8], %[fTmp9], %[fTmp11] \n\t"
yading@10 374 "mul.s %[val12], %[val12], %[fTmp7] \n\t"
yading@10 375 "lwc1 %[fTmp3], 11*4(%[tab]) \n\t"
yading@10 376 "lwc1 %[fTmp4], 20*4(%[tab]) \n\t"
yading@10 377 "add.s %[fTmp5], %[fTmp1], %[fTmp2] \n\t"
yading@10 378 "mul.s %[val28], %[val28], %[fTmp7] \n\t"
yading@10 379 "sub.s %[fTmp8], %[fTmp1], %[fTmp2] \n\t"
yading@10 380 "li.s %[fTmp7], 0.78815462345125022473 \n\t"
yading@10 381 "add.s %[fTmp6], %[fTmp3], %[fTmp4] \n\t"
yading@10 382 "sub.s %[fTmp9], %[fTmp3], %[fTmp4] \n\t"
yading@10 383 "li.s %[fTmp10], 0.55310389603444452782 \n\t"
yading@10 384 "li.s %[fTmp11], 1.16943993343288495515 \n\t"
yading@10 385 "mul.s %[fTmp8], %[fTmp8], %[fTmp10] \n\t"
yading@10 386 "add.s %[val4], %[fTmp5], %[fTmp6] \n\t"
yading@10 387 "sub.s %[val11], %[fTmp5], %[fTmp6] \n\t"
yading@10 388 "li.s %[fTmp1], 2.56291544774150617881 \n\t"
yading@10 389 "madd.s %[val20], %[fTmp8], %[fTmp9], %[fTmp11] \n\t"
yading@10 390 "nmsub.s %[val27], %[fTmp8], %[fTmp9], %[fTmp11] \n\t"
yading@10 391 "mul.s %[val11], %[val11], %[fTmp7] \n\t"
yading@10 392 "sub.s %[fTmp2], %[val3], %[val4] \n\t"
yading@10 393 "add.s %[val3], %[val3], %[val4] \n\t"
yading@10 394 "sub.s %[fTmp4], %[val19], %[val20] \n\t"
yading@10 395 "mul.s %[val27], %[val27], %[fTmp7] \n\t"
yading@10 396 "sub.s %[fTmp3], %[val12], %[val11] \n\t"
yading@10 397 "mul.s %[val4], %[fTmp1], %[fTmp2] \n\t"
yading@10 398 "add.s %[val11], %[val12], %[val11] \n\t"
yading@10 399 "add.s %[val19], %[val19], %[val20] \n\t"
yading@10 400 "mul.s %[val20], %[fTmp1], %[fTmp4] \n\t"
yading@10 401 "mul.s %[val12], %[fTmp1], %[fTmp3] \n\t"
yading@10 402 "sub.s %[fTmp2], %[val28], %[val27] \n\t"
yading@10 403 "add.s %[val27], %[val28], %[val27] \n\t"
yading@10 404 "mul.s %[val28], %[fTmp1], %[fTmp2] \n\t"
yading@10 405
yading@10 406 : [fTmp1] "=&f" (fTmp1), [fTmp2] "=&f" (fTmp2), [fTmp3] "=&f" (fTmp3),
yading@10 407 [fTmp4] "=&f" (fTmp4), [fTmp5] "=&f" (fTmp5), [fTmp6] "=&f" (fTmp6),
yading@10 408 [fTmp7] "=&f" (fTmp7), [fTmp8] "=&f" (fTmp8), [fTmp9] "=&f" (fTmp9),
yading@10 409 [fTmp10] "=&f" (fTmp10), [fTmp11] "=&f" (fTmp11),
yading@10 410 [val3] "=f" (val3), [val4] "=f" (val4),
yading@10 411 [val11] "=f" (val11), [val12] "=f" (val12),
yading@10 412 [val19] "=f" (val19), [val20] "=f" (val20),
yading@10 413 [val27] "=f" (val27), [val28] "=f" (val28)
yading@10 414 : [tab] "r" (tab)
yading@10 415 : "memory"
yading@10 416 );
yading@10 417
yading@10 418 __asm__ volatile (
yading@10 419 "li.s %[fTmp1], 0.54119610014619698439 \n\t"
yading@10 420 "sub.s %[fTmp2], %[val0], %[val3] \n\t"
yading@10 421 "add.s %[val0], %[val0], %[val3] \n\t"
yading@10 422 "sub.s %[fTmp3], %[val7], %[val4] \n\t"
yading@10 423 "add.s %[val4], %[val7], %[val4] \n\t"
yading@10 424 "sub.s %[fTmp4], %[val8], %[val11] \n\t"
yading@10 425 "mul.s %[val3], %[fTmp1], %[fTmp2] \n\t"
yading@10 426 "add.s %[val8], %[val8], %[val11] \n\t"
yading@10 427 "mul.s %[val7], %[fTmp1], %[fTmp3] \n\t"
yading@10 428 "sub.s %[fTmp2], %[val15], %[val12] \n\t"
yading@10 429 "mul.s %[val11], %[fTmp1], %[fTmp4] \n\t"
yading@10 430 "add.s %[val12], %[val15], %[val12] \n\t"
yading@10 431 "mul.s %[val15], %[fTmp1], %[fTmp2] \n\t"
yading@10 432
yading@10 433 : [val0] "+f" (val0), [val3] "+f" (val3),
yading@10 434 [val4] "+f" (val4), [val7] "+f" (val7),
yading@10 435 [val8] "+f" (val8), [val11] "+f" (val11),
yading@10 436 [val12] "+f" (val12), [val15] "+f" (val15),
yading@10 437 [fTmp1] "=f" (fTmp1), [fTmp2] "=&f" (fTmp2),
yading@10 438 [fTmp3] "=&f" (fTmp3), [fTmp4] "=&f" (fTmp4)
yading@10 439 :
yading@10 440 );
yading@10 441
yading@10 442 __asm__ volatile (
yading@10 443 "sub.s %[fTmp2], %[val16], %[val19] \n\t"
yading@10 444 "add.s %[val16], %[val16], %[val19] \n\t"
yading@10 445 "sub.s %[fTmp3], %[val23], %[val20] \n\t"
yading@10 446 "add.s %[val20], %[val23], %[val20] \n\t"
yading@10 447 "sub.s %[fTmp4], %[val24], %[val27] \n\t"
yading@10 448 "mul.s %[val19], %[fTmp1], %[fTmp2] \n\t"
yading@10 449 "add.s %[val24], %[val24], %[val27] \n\t"
yading@10 450 "mul.s %[val23], %[fTmp1], %[fTmp3] \n\t"
yading@10 451 "sub.s %[fTmp2], %[val31], %[val28] \n\t"
yading@10 452 "mul.s %[val27], %[fTmp1], %[fTmp4] \n\t"
yading@10 453 "add.s %[val28], %[val31], %[val28] \n\t"
yading@10 454 "mul.s %[val31], %[fTmp1], %[fTmp2] \n\t"
yading@10 455
yading@10 456 : [fTmp2] "=&f" (fTmp2), [fTmp3] "=&f" (fTmp3), [fTmp4] "=&f" (fTmp4),
yading@10 457 [val16] "+f" (val16), [val19] "+f" (val19), [val20] "+f" (val20),
yading@10 458 [val23] "+f" (val23), [val24] "+f" (val24), [val27] "+f" (val27),
yading@10 459 [val28] "+f" (val28), [val31] "+f" (val31)
yading@10 460 : [fTmp1] "f" (fTmp1)
yading@10 461 );
yading@10 462
yading@10 463 __asm__ volatile (
yading@10 464 "lwc1 %[fTmp1], 1*4(%[tab]) \n\t"
yading@10 465 "lwc1 %[fTmp2], 30*4(%[tab]) \n\t"
yading@10 466 "lwc1 %[fTmp3], 14*4(%[tab]) \n\t"
yading@10 467 "lwc1 %[fTmp4], 17*4(%[tab]) \n\t"
yading@10 468 "li.s %[fTmp7], 0.52249861493968888062 \n\t"
yading@10 469 "add.s %[fTmp5], %[fTmp1], %[fTmp2] \n\t"
yading@10 470 "sub.s %[fTmp8], %[fTmp1], %[fTmp2] \n\t"
yading@10 471 "add.s %[fTmp6], %[fTmp3], %[fTmp4] \n\t"
yading@10 472 "sub.s %[fTmp9], %[fTmp3], %[fTmp4] \n\t"
yading@10 473 "li.s %[fTmp10], 0.50547095989754365998 \n\t"
yading@10 474 "li.s %[fTmp11], 3.40760841846871878570 \n\t"
yading@10 475 "mul.s %[fTmp8], %[fTmp8], %[fTmp10] \n\t"
yading@10 476 "add.s %[val1], %[fTmp5], %[fTmp6] \n\t"
yading@10 477 "sub.s %[val14], %[fTmp5], %[fTmp6] \n\t"
yading@10 478 "lwc1 %[fTmp1], 6*4(%[tab]) \n\t"
yading@10 479 "lwc1 %[fTmp2], 25*4(%[tab]) \n\t"
yading@10 480 "madd.s %[val17], %[fTmp8], %[fTmp9], %[fTmp11] \n\t"
yading@10 481 "nmsub.s %[val30], %[fTmp8], %[fTmp9], %[fTmp11] \n\t"
yading@10 482 "mul.s %[val14], %[val14], %[fTmp7] \n\t"
yading@10 483 "lwc1 %[fTmp3], 9*4(%[tab]) \n\t"
yading@10 484 "lwc1 %[fTmp4], 22*4(%[tab]) \n\t"
yading@10 485 "add.s %[fTmp5], %[fTmp1], %[fTmp2] \n\t"
yading@10 486 "mul.s %[val30], %[val30], %[fTmp7] \n\t"
yading@10 487 "sub.s %[fTmp8], %[fTmp1], %[fTmp2] \n\t"
yading@10 488 "add.s %[fTmp6], %[fTmp3], %[fTmp4] \n\t"
yading@10 489 "sub.s %[fTmp9], %[fTmp3], %[fTmp4] \n\t"
yading@10 490 "li.s %[fTmp7], 1.72244709823833392782 \n\t"
yading@10 491 "li.s %[fTmp10], 0.62250412303566481615 \n\t"
yading@10 492 "li.s %[fTmp11], 0.83934964541552703873 \n\t"
yading@10 493 "add.s %[val6], %[fTmp5], %[fTmp6] \n\t"
yading@10 494 "sub.s %[val9], %[fTmp5], %[fTmp6] \n\t"
yading@10 495 "mul.s %[fTmp8], %[fTmp8], %[fTmp10] \n\t"
yading@10 496 "li.s %[fTmp1], 0.60134488693504528054 \n\t"
yading@10 497 "sub.s %[fTmp2], %[val1], %[val6] \n\t"
yading@10 498 "add.s %[val1], %[val1], %[val6] \n\t"
yading@10 499 "mul.s %[val9], %[val9], %[fTmp7] \n\t"
yading@10 500 "madd.s %[val22], %[fTmp8], %[fTmp9], %[fTmp11] \n\t"
yading@10 501 "nmsub.s %[val25], %[fTmp8], %[fTmp9], %[fTmp11] \n\t"
yading@10 502 "mul.s %[val6], %[fTmp1], %[fTmp2] \n\t"
yading@10 503 "sub.s %[fTmp2], %[val14], %[val9] \n\t"
yading@10 504 "add.s %[val9], %[val14], %[val9] \n\t"
yading@10 505 "mul.s %[val25], %[val25], %[fTmp7] \n\t"
yading@10 506 "sub.s %[fTmp3], %[val17], %[val22] \n\t"
yading@10 507 "add.s %[val17], %[val17], %[val22] \n\t"
yading@10 508 "mul.s %[val14], %[fTmp1], %[fTmp2] \n\t"
yading@10 509 "sub.s %[fTmp2], %[val30], %[val25] \n\t"
yading@10 510 "mul.s %[val22], %[fTmp1], %[fTmp3] \n\t"
yading@10 511 "add.s %[val25], %[val30], %[val25] \n\t"
yading@10 512 "mul.s %[val30], %[fTmp1], %[fTmp2] \n\t"
yading@10 513
yading@10 514 : [fTmp1] "=&f" (fTmp1), [fTmp2] "=&f" (fTmp2), [fTmp3] "=&f" (fTmp3),
yading@10 515 [fTmp4] "=&f" (fTmp4), [fTmp5] "=&f" (fTmp5), [fTmp6] "=&f" (fTmp6),
yading@10 516 [fTmp7] "=&f" (fTmp7), [fTmp8] "=&f" (fTmp8), [fTmp9] "=&f" (fTmp9),
yading@10 517 [fTmp10] "=&f" (fTmp10), [fTmp11] "=&f" (fTmp11),
yading@10 518 [val1] "=f" (val1), [val6] "=f" (val6),
yading@10 519 [val9] "=f" (val9), [val14] "=f" (val14),
yading@10 520 [val17] "=f" (val17), [val22] "=f" (val22),
yading@10 521 [val25] "=f" (val25), [val30] "=f" (val30)
yading@10 522 : [tab] "r" (tab)
yading@10 523 : "memory"
yading@10 524 );
yading@10 525
yading@10 526 __asm__ volatile (
yading@10 527 "lwc1 %[fTmp1], 2*4(%[tab]) \n\t"
yading@10 528 "lwc1 %[fTmp2], 29*4(%[tab]) \n\t"
yading@10 529 "lwc1 %[fTmp3], 13*4(%[tab]) \n\t"
yading@10 530 "lwc1 %[fTmp4], 18*4(%[tab]) \n\t"
yading@10 531 "li.s %[fTmp7], 0.56694403481635770368 \n\t"
yading@10 532 "add.s %[fTmp5], %[fTmp1], %[fTmp2] \n\t"
yading@10 533 "sub.s %[fTmp8], %[fTmp1], %[fTmp2] \n\t"
yading@10 534 "add.s %[fTmp6], %[fTmp3], %[fTmp4] \n\t"
yading@10 535 "sub.s %[fTmp9], %[fTmp3], %[fTmp4] \n\t"
yading@10 536 "li.s %[fTmp10], 0.51544730992262454697 \n\t"
yading@10 537 "li.s %[fTmp11], 2.05778100995341155085 \n\t"
yading@10 538 "mul.s %[fTmp8], %[fTmp8], %[fTmp10] \n\t"
yading@10 539 "add.s %[val2], %[fTmp5], %[fTmp6] \n\t"
yading@10 540 "sub.s %[val13], %[fTmp5], %[fTmp6] \n\t"
yading@10 541 "lwc1 %[fTmp1], 5*4(%[tab]) \n\t"
yading@10 542 "lwc1 %[fTmp2], 26*4(%[tab]) \n\t"
yading@10 543 "madd.s %[val18], %[fTmp8], %[fTmp9], %[fTmp11] \n\t"
yading@10 544 "nmsub.s %[val29], %[fTmp8], %[fTmp9], %[fTmp11] \n\t"
yading@10 545 "mul.s %[val13], %[val13], %[fTmp7] \n\t"
yading@10 546 "lwc1 %[fTmp3], 10*4(%[tab]) \n\t"
yading@10 547 "lwc1 %[fTmp4], 21*4(%[tab]) \n\t"
yading@10 548 "mul.s %[val29], %[val29], %[fTmp7] \n\t"
yading@10 549 "add.s %[fTmp5], %[fTmp1], %[fTmp2] \n\t"
yading@10 550 "sub.s %[fTmp8], %[fTmp1], %[fTmp2] \n\t"
yading@10 551 "add.s %[fTmp6], %[fTmp3], %[fTmp4] \n\t"
yading@10 552 "sub.s %[fTmp9], %[fTmp3], %[fTmp4] \n\t"
yading@10 553 "li.s %[fTmp7], 1.06067768599034747134 \n\t"
yading@10 554 "li.s %[fTmp10], 0.58293496820613387367 \n\t"
yading@10 555 "li.s %[fTmp11], 0.97256823786196069369 \n\t"
yading@10 556 "add.s %[val5], %[fTmp5], %[fTmp6] \n\t"
yading@10 557 "sub.s %[val10], %[fTmp5], %[fTmp6] \n\t"
yading@10 558 "mul.s %[fTmp8], %[fTmp8], %[fTmp10] \n\t"
yading@10 559 "li.s %[fTmp1], 0.89997622313641570463 \n\t"
yading@10 560 "sub.s %[fTmp2], %[val2], %[val5] \n\t"
yading@10 561 "mul.s %[val10], %[val10], %[fTmp7] \n\t"
yading@10 562 "madd.s %[val21], %[fTmp8], %[fTmp9], %[fTmp11] \n\t"
yading@10 563 "nmsub.s %[val26], %[fTmp8], %[fTmp9], %[fTmp11] \n\t"
yading@10 564 "add.s %[val2], %[val2], %[val5] \n\t"
yading@10 565 "mul.s %[val5], %[fTmp1], %[fTmp2] \n\t"
yading@10 566 "sub.s %[fTmp3], %[val13], %[val10] \n\t"
yading@10 567 "add.s %[val10], %[val13], %[val10] \n\t"
yading@10 568 "mul.s %[val26], %[val26], %[fTmp7] \n\t"
yading@10 569 "sub.s %[fTmp4], %[val18], %[val21] \n\t"
yading@10 570 "add.s %[val18], %[val18], %[val21] \n\t"
yading@10 571 "mul.s %[val13], %[fTmp1], %[fTmp3] \n\t"
yading@10 572 "sub.s %[fTmp2], %[val29], %[val26] \n\t"
yading@10 573 "add.s %[val26], %[val29], %[val26] \n\t"
yading@10 574 "mul.s %[val21], %[fTmp1], %[fTmp4] \n\t"
yading@10 575 "mul.s %[val29], %[fTmp1], %[fTmp2] \n\t"
yading@10 576
yading@10 577 : [fTmp1] "=&f" (fTmp1), [fTmp2] "=&f" (fTmp2), [fTmp3] "=&f" (fTmp3),
yading@10 578 [fTmp4] "=&f" (fTmp4), [fTmp5] "=&f" (fTmp5), [fTmp6] "=&f" (fTmp6),
yading@10 579 [fTmp7] "=&f" (fTmp7), [fTmp8] "=&f" (fTmp8), [fTmp9] "=&f" (fTmp9),
yading@10 580 [fTmp10] "=&f" (fTmp10), [fTmp11] "=&f" (fTmp11),
yading@10 581 [val2] "=f" (val2), [val5] "=f" (val5),
yading@10 582 [val10] "=f" (val10), [val13] "=f" (val13),
yading@10 583 [val18] "=f" (val18), [val21] "=f" (val21),
yading@10 584 [val26] "=f" (val26), [val29] "=f" (val29)
yading@10 585 : [tab] "r" (tab)
yading@10 586 : "memory"
yading@10 587 );
yading@10 588
yading@10 589 __asm__ volatile (
yading@10 590 "li.s %[fTmp1], 1.30656296487637652785 \n\t"
yading@10 591 "sub.s %[fTmp2], %[val1], %[val2] \n\t"
yading@10 592 "add.s %[val1], %[val1], %[val2] \n\t"
yading@10 593 "sub.s %[fTmp3], %[val6], %[val5] \n\t"
yading@10 594 "add.s %[val5], %[val6], %[val5] \n\t"
yading@10 595 "sub.s %[fTmp4], %[val9], %[val10] \n\t"
yading@10 596 "mul.s %[val2], %[fTmp1], %[fTmp2] \n\t"
yading@10 597 "add.s %[val9], %[val9], %[val10] \n\t"
yading@10 598 "mul.s %[val6], %[fTmp1], %[fTmp3] \n\t"
yading@10 599 "sub.s %[fTmp2], %[val14], %[val13] \n\t"
yading@10 600 "mul.s %[val10], %[fTmp1], %[fTmp4] \n\t"
yading@10 601 "add.s %[val13], %[val14], %[val13] \n\t"
yading@10 602 "mul.s %[val14], %[fTmp1], %[fTmp2] \n\t"
yading@10 603
yading@10 604 : [fTmp1] "=f" (fTmp1), [fTmp2] "=&f" (fTmp2),
yading@10 605 [fTmp3] "=&f" (fTmp3), [fTmp4] "=&f" (fTmp4),
yading@10 606 [val1] "+f" (val1), [val2] "+f" (val2),
yading@10 607 [val5] "+f" (val5), [val6] "+f" (val6),
yading@10 608 [val9] "+f" (val9), [val10] "+f" (val10),
yading@10 609 [val13] "+f" (val13), [val14] "+f" (val14)
yading@10 610 :
yading@10 611 );
yading@10 612
yading@10 613 __asm__ volatile (
yading@10 614 "sub.s %[fTmp2], %[val17], %[val18] \n\t"
yading@10 615 "add.s %[val17], %[val17], %[val18] \n\t"
yading@10 616 "sub.s %[fTmp3], %[val22], %[val21] \n\t"
yading@10 617 "add.s %[val21], %[val22], %[val21] \n\t"
yading@10 618 "sub.s %[fTmp4], %[val25], %[val26] \n\t"
yading@10 619 "mul.s %[val18], %[fTmp1], %[fTmp2] \n\t"
yading@10 620 "add.s %[val25], %[val25], %[val26] \n\t"
yading@10 621 "mul.s %[val22], %[fTmp1], %[fTmp3] \n\t"
yading@10 622 "sub.s %[fTmp2], %[val30], %[val29] \n\t"
yading@10 623 "mul.s %[val26], %[fTmp1], %[fTmp4] \n\t"
yading@10 624 "add.s %[val29], %[val30], %[val29] \n\t"
yading@10 625 "mul.s %[val30], %[fTmp1], %[fTmp2] \n\t"
yading@10 626
yading@10 627 : [fTmp2] "=&f" (fTmp2), [fTmp3] "=&f" (fTmp3), [fTmp4] "=&f" (fTmp4),
yading@10 628 [val17] "+f" (val17), [val18] "+f" (val18), [val21] "+f" (val21),
yading@10 629 [val22] "+f" (val22), [val25] "+f" (val25), [val26] "+f" (val26),
yading@10 630 [val29] "+f" (val29), [val30] "+f" (val30)
yading@10 631 : [fTmp1] "f" (fTmp1)
yading@10 632 );
yading@10 633
yading@10 634 __asm__ volatile (
yading@10 635 "li.s %[fTmp1], 0.70710678118654752439 \n\t"
yading@10 636 "sub.s %[fTmp2], %[val0], %[val1] \n\t"
yading@10 637 "add.s %[val0], %[val0], %[val1] \n\t"
yading@10 638 "sub.s %[fTmp3], %[val3], %[val2] \n\t"
yading@10 639 "add.s %[val2], %[val3], %[val2] \n\t"
yading@10 640 "sub.s %[fTmp4], %[val4], %[val5] \n\t"
yading@10 641 "mul.s %[val1], %[fTmp1], %[fTmp2] \n\t"
yading@10 642 "swc1 %[val0], 0(%[out]) \n\t"
yading@10 643 "mul.s %[val3], %[fTmp3], %[fTmp1] \n\t"
yading@10 644 "add.s %[val4], %[val4], %[val5] \n\t"
yading@10 645 "mul.s %[val5], %[fTmp1], %[fTmp4] \n\t"
yading@10 646 "swc1 %[val1], 16*4(%[out]) \n\t"
yading@10 647 "sub.s %[fTmp2], %[val7], %[val6] \n\t"
yading@10 648 "add.s %[val2], %[val2], %[val3] \n\t"
yading@10 649 "swc1 %[val3], 24*4(%[out]) \n\t"
yading@10 650 "add.s %[val6], %[val7], %[val6] \n\t"
yading@10 651 "mul.s %[val7], %[fTmp1], %[fTmp2] \n\t"
yading@10 652 "swc1 %[val2], 8*4(%[out]) \n\t"
yading@10 653 "add.s %[val6], %[val6], %[val7] \n\t"
yading@10 654 "swc1 %[val7], 28*4(%[out]) \n\t"
yading@10 655 "add.s %[val4], %[val4], %[val6] \n\t"
yading@10 656 "add.s %[val6], %[val6], %[val5] \n\t"
yading@10 657 "add.s %[val5], %[val5], %[val7] \n\t"
yading@10 658 "swc1 %[val4], 4*4(%[out]) \n\t"
yading@10 659 "swc1 %[val5], 20*4(%[out]) \n\t"
yading@10 660 "swc1 %[val6], 12*4(%[out]) \n\t"
yading@10 661
yading@10 662 : [fTmp1] "=f" (fTmp1), [fTmp2] "=&f" (fTmp2),
yading@10 663 [fTmp3] "=&f" (fTmp3), [fTmp4] "=&f" (fTmp4),
yading@10 664 [val0] "+f" (val0), [val1] "+f" (val1),
yading@10 665 [val2] "+f" (val2), [val3] "+f" (val3),
yading@10 666 [val4] "+f" (val4), [val5] "+f" (val5),
yading@10 667 [val6] "+f" (val6), [val7] "+f" (val7)
yading@10 668 : [out] "r" (out)
yading@10 669 );
yading@10 670
yading@10 671 __asm__ volatile (
yading@10 672 "sub.s %[fTmp2], %[val8], %[val9] \n\t"
yading@10 673 "add.s %[val8], %[val8], %[val9] \n\t"
yading@10 674 "sub.s %[fTmp3], %[val11], %[val10] \n\t"
yading@10 675 "add.s %[val10], %[val11], %[val10] \n\t"
yading@10 676 "sub.s %[fTmp4], %[val12], %[val13] \n\t"
yading@10 677 "mul.s %[val9], %[fTmp1], %[fTmp2] \n\t"
yading@10 678 "add.s %[val12], %[val12], %[val13] \n\t"
yading@10 679 "mul.s %[val11], %[fTmp1], %[fTmp3] \n\t"
yading@10 680 "sub.s %[fTmp2], %[val15], %[val14] \n\t"
yading@10 681 "mul.s %[val13], %[fTmp1], %[fTmp4] \n\t"
yading@10 682 "add.s %[val14], %[val15], %[val14] \n\t"
yading@10 683 "add.s %[val10], %[val10], %[val11] \n\t"
yading@10 684 "mul.s %[val15], %[fTmp1], %[fTmp2] \n\t"
yading@10 685 "add.s %[val14], %[val14], %[val15] \n\t"
yading@10 686 "add.s %[val12], %[val12], %[val14] \n\t"
yading@10 687 "add.s %[val14], %[val14], %[val13] \n\t"
yading@10 688 "add.s %[val13], %[val13], %[val15] \n\t"
yading@10 689 "add.s %[val8], %[val8], %[val12] \n\t"
yading@10 690 "add.s %[val12], %[val12], %[val10] \n\t"
yading@10 691 "add.s %[val10], %[val10], %[val14] \n\t"
yading@10 692 "add.s %[val14], %[val14], %[val9] \n\t"
yading@10 693 "add.s %[val9], %[val9], %[val13] \n\t"
yading@10 694 "add.s %[val13], %[val13], %[val11] \n\t"
yading@10 695 "add.s %[val11], %[val11], %[val15] \n\t"
yading@10 696 "swc1 %[val8], 2*4(%[out]) \n\t"
yading@10 697 "swc1 %[val9], 18*4(%[out]) \n\t"
yading@10 698 "swc1 %[val10], 10*4(%[out]) \n\t"
yading@10 699 "swc1 %[val11], 26*4(%[out]) \n\t"
yading@10 700 "swc1 %[val12], 6*4(%[out]) \n\t"
yading@10 701 "swc1 %[val13], 22*4(%[out]) \n\t"
yading@10 702 "swc1 %[val14], 14*4(%[out]) \n\t"
yading@10 703 "swc1 %[val15], 30*4(%[out]) \n\t"
yading@10 704
yading@10 705 : [fTmp2] "=&f" (fTmp2), [fTmp3] "=&f" (fTmp3), [fTmp4] "=&f" (fTmp4),
yading@10 706 [val8] "+f" (val8), [val9] "+f" (val9), [val10] "+f" (val10),
yading@10 707 [val11] "+f" (val11), [val12] "+f" (val12), [val13] "+f" (val13),
yading@10 708 [val14] "+f" (val14), [val15] "+f" (val15)
yading@10 709 : [fTmp1] "f" (fTmp1), [out] "r" (out)
yading@10 710 );
yading@10 711
yading@10 712 __asm__ volatile (
yading@10 713 "sub.s %[fTmp2], %[val16], %[val17] \n\t"
yading@10 714 "add.s %[val16], %[val16], %[val17] \n\t"
yading@10 715 "sub.s %[fTmp3], %[val19], %[val18] \n\t"
yading@10 716 "add.s %[val18], %[val19], %[val18] \n\t"
yading@10 717 "sub.s %[fTmp4], %[val20], %[val21] \n\t"
yading@10 718 "mul.s %[val17], %[fTmp1], %[fTmp2] \n\t"
yading@10 719 "add.s %[val20], %[val20], %[val21] \n\t"
yading@10 720 "mul.s %[val19], %[fTmp1], %[fTmp3] \n\t"
yading@10 721 "sub.s %[fTmp2], %[val23], %[val22] \n\t"
yading@10 722 "mul.s %[val21], %[fTmp1], %[fTmp4] \n\t"
yading@10 723 "add.s %[val22], %[val23], %[val22] \n\t"
yading@10 724 "add.s %[val18], %[val18], %[val19] \n\t"
yading@10 725 "mul.s %[val23], %[fTmp1], %[fTmp2] \n\t"
yading@10 726 "add.s %[val22], %[val22], %[val23] \n\t"
yading@10 727 "add.s %[val20], %[val20], %[val22] \n\t"
yading@10 728 "add.s %[val22], %[val22], %[val21] \n\t"
yading@10 729 "add.s %[val21], %[val21], %[val23] \n\t"
yading@10 730
yading@10 731 : [fTmp2] "=&f" (fTmp2), [fTmp3] "=&f" (fTmp3), [fTmp4] "=&f" (fTmp4),
yading@10 732 [val16] "+f" (val16), [val17] "+f" (val17), [val18] "+f" (val18),
yading@10 733 [val19] "+f" (val19), [val20] "+f" (val20), [val21] "+f" (val21),
yading@10 734 [val22] "+f" (val22), [val23] "+f" (val23)
yading@10 735 : [fTmp1] "f" (fTmp1)
yading@10 736 );
yading@10 737
yading@10 738 __asm__ volatile (
yading@10 739 "sub.s %[fTmp2], %[val24], %[val25] \n\t"
yading@10 740 "add.s %[val24], %[val24], %[val25] \n\t"
yading@10 741 "sub.s %[fTmp3], %[val27], %[val26] \n\t"
yading@10 742 "add.s %[val26], %[val27], %[val26] \n\t"
yading@10 743 "sub.s %[fTmp4], %[val28], %[val29] \n\t"
yading@10 744 "mul.s %[val25], %[fTmp1], %[fTmp2] \n\t"
yading@10 745 "add.s %[val28], %[val28], %[val29] \n\t"
yading@10 746 "mul.s %[val27], %[fTmp1], %[fTmp3] \n\t"
yading@10 747 "sub.s %[fTmp2], %[val31], %[val30] \n\t"
yading@10 748 "mul.s %[val29], %[fTmp1], %[fTmp4] \n\t"
yading@10 749 "add.s %[val30], %[val31], %[val30] \n\t"
yading@10 750 "add.s %[val26], %[val26], %[val27] \n\t"
yading@10 751 "mul.s %[val31], %[fTmp1], %[fTmp2] \n\t"
yading@10 752 "add.s %[val30], %[val30], %[val31] \n\t"
yading@10 753 "add.s %[val28], %[val28], %[val30] \n\t"
yading@10 754 "add.s %[val30], %[val30], %[val29] \n\t"
yading@10 755 "add.s %[val29], %[val29], %[val31] \n\t"
yading@10 756 "add.s %[val24], %[val24], %[val28] \n\t"
yading@10 757 "add.s %[val28], %[val28], %[val26] \n\t"
yading@10 758 "add.s %[val26], %[val26], %[val30] \n\t"
yading@10 759 "add.s %[val30], %[val30], %[val25] \n\t"
yading@10 760 "add.s %[val25], %[val25], %[val29] \n\t"
yading@10 761 "add.s %[val29], %[val29], %[val27] \n\t"
yading@10 762 "add.s %[val27], %[val27], %[val31] \n\t"
yading@10 763
yading@10 764 : [fTmp2] "=&f" (fTmp2), [fTmp3] "=&f" (fTmp3), [fTmp4] "=&f" (fTmp4),
yading@10 765 [val24] "+f" (val24), [val25] "+f" (val25), [val26] "+f" (val26),
yading@10 766 [val27] "+f" (val27), [val28] "+f" (val28), [val29] "+f" (val29),
yading@10 767 [val30] "+f" (val30), [val31] "+f" (val31)
yading@10 768 : [fTmp1] "f" (fTmp1)
yading@10 769 );
yading@10 770
yading@10 771 out[ 1] = val16 + val24;
yading@10 772 out[17] = val17 + val25;
yading@10 773 out[ 9] = val18 + val26;
yading@10 774 out[25] = val19 + val27;
yading@10 775 out[ 5] = val20 + val28;
yading@10 776 out[21] = val21 + val29;
yading@10 777 out[13] = val22 + val30;
yading@10 778 out[29] = val23 + val31;
yading@10 779 out[ 3] = val24 + val20;
yading@10 780 out[19] = val25 + val21;
yading@10 781 out[11] = val26 + val22;
yading@10 782 out[27] = val27 + val23;
yading@10 783 out[ 7] = val28 + val18;
yading@10 784 out[23] = val29 + val19;
yading@10 785 out[15] = val30 + val17;
yading@10 786 out[31] = val31;
yading@10 787 }
yading@10 788
yading@10 789 static void imdct36_mips_float(float *out, float *buf, float *in, float *win)
yading@10 790 {
yading@10 791 float t0, t1, t2, t3, s0, s1, s2, s3;
yading@10 792 float tmp[18];
yading@10 793 /* temporary variables */
yading@10 794 float in1, in2, in3, in4, in5, in6;
yading@10 795 float out1, out2, out3, out4, out5;
yading@10 796 float c1, c2, c3, c4, c5, c6, c7, c8, c9;
yading@10 797
yading@10 798 /**
yading@10 799 * all loops are unrolled totally, and instructions are scheduled to
yading@10 800 * minimize pipeline stall. instructions of the first two loops are
yading@10 801 * reorganized, in order to eliminate unnecessary readings and
yading@10 802 * writings into array. values defined in macros and tables are
yading@10 803 * eliminated - they are directly loaded in appropriate variables
yading@10 804 */
yading@10 805
yading@10 806 /* loop 1 and 2 */
yading@10 807 __asm__ volatile (
yading@10 808 "lwc1 %[in1], 17*4(%[in]) \t\n"
yading@10 809 "lwc1 %[in2], 16*4(%[in]) \t\n"
yading@10 810 "lwc1 %[in3], 15*4(%[in]) \t\n"
yading@10 811 "lwc1 %[in4], 14*4(%[in]) \t\n"
yading@10 812 "lwc1 %[in5], 13*4(%[in]) \t\n"
yading@10 813 "lwc1 %[in6], 12*4(%[in]) \t\n"
yading@10 814 "add.s %[out1], %[in1], %[in2] \t\n"
yading@10 815 "add.s %[out2], %[in2], %[in3] \t\n"
yading@10 816 "add.s %[out3], %[in3], %[in4] \t\n"
yading@10 817 "add.s %[out4], %[in4], %[in5] \t\n"
yading@10 818 "add.s %[out5], %[in5], %[in6] \t\n"
yading@10 819 "lwc1 %[in1], 11*4(%[in]) \t\n"
yading@10 820 "swc1 %[out2], 16*4(%[in]) \t\n"
yading@10 821 "add.s %[out1], %[out1], %[out3] \t\n"
yading@10 822 "swc1 %[out4], 14*4(%[in]) \t\n"
yading@10 823 "add.s %[out3], %[out3], %[out5] \t\n"
yading@10 824 "lwc1 %[in2], 10*4(%[in]) \t\n"
yading@10 825 "lwc1 %[in3], 9*4(%[in]) \t\n"
yading@10 826 "swc1 %[out1], 17*4(%[in]) \t\n"
yading@10 827 "lwc1 %[in4], 8*4(%[in]) \t\n"
yading@10 828 "swc1 %[out3], 15*4(%[in]) \t\n"
yading@10 829 "add.s %[out1], %[in6], %[in1] \t\n"
yading@10 830 "add.s %[out2], %[in1], %[in2] \t\n"
yading@10 831 "add.s %[out3], %[in2], %[in3] \t\n"
yading@10 832 "add.s %[out4], %[in3], %[in4] \t\n"
yading@10 833 "lwc1 %[in5], 7*4(%[in]) \t\n"
yading@10 834 "swc1 %[out1], 12*4(%[in]) \t\n"
yading@10 835 "add.s %[out5], %[out5], %[out2] \t\n"
yading@10 836 "swc1 %[out3], 10*4(%[in]) \t\n"
yading@10 837 "add.s %[out2], %[out2], %[out4] \t\n"
yading@10 838 "lwc1 %[in6], 6*4(%[in]) \t\n"
yading@10 839 "lwc1 %[in1], 5*4(%[in]) \t\n"
yading@10 840 "swc1 %[out5], 13*4(%[in]) \t\n"
yading@10 841 "lwc1 %[in2], 4*4(%[in]) \t\n"
yading@10 842 "swc1 %[out2], 11*4(%[in]) \t\n"
yading@10 843 "add.s %[out5], %[in4], %[in5] \t\n"
yading@10 844 "add.s %[out1], %[in5], %[in6] \t\n"
yading@10 845 "add.s %[out2], %[in6], %[in1] \t\n"
yading@10 846 "add.s %[out3], %[in1], %[in2] \t\n"
yading@10 847 "lwc1 %[in3], 3*4(%[in]) \t\n"
yading@10 848 "swc1 %[out5], 8*4(%[in]) \t\n"
yading@10 849 "add.s %[out4], %[out4], %[out1] \t\n"
yading@10 850 "swc1 %[out2], 6*4(%[in]) \t\n"
yading@10 851 "add.s %[out1], %[out1], %[out3] \t\n"
yading@10 852 "lwc1 %[in4], 2*4(%[in]) \t\n"
yading@10 853 "lwc1 %[in5], 1*4(%[in]) \t\n"
yading@10 854 "swc1 %[out4], 9*4(%[in]) \t\n"
yading@10 855 "lwc1 %[in6], 0(%[in]) \t\n"
yading@10 856 "swc1 %[out1], 7*4(%[in]) \t\n"
yading@10 857 "add.s %[out4], %[in2], %[in3] \t\n"
yading@10 858 "add.s %[out5], %[in3], %[in4] \t\n"
yading@10 859 "add.s %[out1], %[in4], %[in5] \t\n"
yading@10 860 "add.s %[out2], %[in5], %[in6] \t\n"
yading@10 861 "swc1 %[out4], 4*4(%[in]) \t\n"
yading@10 862 "add.s %[out3], %[out3], %[out5] \t\n"
yading@10 863 "swc1 %[out1], 2*4(%[in]) \t\n"
yading@10 864 "add.s %[out5], %[out5], %[out2] \t\n"
yading@10 865 "swc1 %[out2], 1*4(%[in]) \t\n"
yading@10 866 "swc1 %[out3], 5*4(%[in]) \t\n"
yading@10 867 "swc1 %[out5], 3*4(%[in]) \t\n"
yading@10 868
yading@10 869 : [in1] "=&f" (in1), [in2] "=&f" (in2),
yading@10 870 [in3] "=&f" (in3), [in4] "=&f" (in4),
yading@10 871 [in5] "=&f" (in5), [in6] "=&f" (in6),
yading@10 872 [out1] "=&f" (out1), [out2] "=&f" (out2),
yading@10 873 [out3] "=&f" (out3), [out4] "=&f" (out4),
yading@10 874 [out5] "=&f" (out5)
yading@10 875 : [in] "r" (in)
yading@10 876 : "memory"
yading@10 877 );
yading@10 878
yading@10 879 /* loop 3 */
yading@10 880 __asm__ volatile (
yading@10 881 "li.s %[c1], 0.5 \t\n"
yading@10 882 "lwc1 %[in1], 8*4(%[in]) \t\n"
yading@10 883 "lwc1 %[in2], 16*4(%[in]) \t\n"
yading@10 884 "lwc1 %[in3], 4*4(%[in]) \t\n"
yading@10 885 "lwc1 %[in4], 0(%[in]) \t\n"
yading@10 886 "lwc1 %[in5], 12*4(%[in]) \t\n"
yading@10 887 "li.s %[c2], 0.93969262078590838405 \t\n"
yading@10 888 "add.s %[t2], %[in1], %[in2] \t\n"
yading@10 889 "add.s %[t0], %[in1], %[in3] \t\n"
yading@10 890 "li.s %[c3], -0.76604444311897803520 \t\n"
yading@10 891 "madd.s %[t3], %[in4], %[in5], %[c1] \t\n"
yading@10 892 "sub.s %[t1], %[in4], %[in5] \t\n"
yading@10 893 "sub.s %[t2], %[t2], %[in3] \t\n"
yading@10 894 "mul.s %[t0], %[t0], %[c2] \t\n"
yading@10 895 "li.s %[c4], -0.17364817766693034885 \t\n"
yading@10 896 "li.s %[c5], -0.86602540378443864676 \t\n"
yading@10 897 "li.s %[c6], 0.98480775301220805936 \t\n"
yading@10 898 "nmsub.s %[out1], %[t1], %[t2], %[c1] \t\n"
yading@10 899 "add.s %[out2], %[t1], %[t2] \t\n"
yading@10 900 "add.s %[t2], %[in2], %[in3] \t\n"
yading@10 901 "sub.s %[t1], %[in1], %[in2] \t\n"
yading@10 902 "sub.s %[out3], %[t3], %[t0] \t\n"
yading@10 903 "swc1 %[out1], 6*4(%[tmp]) \t\n"
yading@10 904 "swc1 %[out2], 16*4(%[tmp]) \t\n"
yading@10 905 "mul.s %[t2], %[t2], %[c3] \t\n"
yading@10 906 "mul.s %[t1], %[t1], %[c4] \t\n"
yading@10 907 "add.s %[out1], %[t3], %[t0] \t\n"
yading@10 908 "lwc1 %[in1], 10*4(%[in]) \t\n"
yading@10 909 "lwc1 %[in2], 14*4(%[in]) \t\n"
yading@10 910 "sub.s %[out3], %[out3], %[t2] \t\n"
yading@10 911 "add.s %[out2], %[t3], %[t2] \t\n"
yading@10 912 "add.s %[out1], %[out1], %[t1] \t\n"
yading@10 913 "lwc1 %[in3], 2*4(%[in]) \t\n"
yading@10 914 "lwc1 %[in4], 6*4(%[in]) \t\n"
yading@10 915 "swc1 %[out3], 10*4(%[tmp]) \t\n"
yading@10 916 "sub.s %[out2], %[out2], %[t1] \t\n"
yading@10 917 "swc1 %[out1], 2*4(%[tmp]) \t\n"
yading@10 918 "add.s %[out1], %[in1], %[in2] \t\n"
yading@10 919 "add.s %[t2], %[in1], %[in3] \t\n"
yading@10 920 "sub.s %[t3], %[in1], %[in2] \t\n"
yading@10 921 "swc1 %[out2], 14*4(%[tmp]) \t\n"
yading@10 922 "li.s %[c7], -0.34202014332566873304 \t\n"
yading@10 923 "sub.s %[out1], %[out1], %[in3] \t\n"
yading@10 924 "mul.s %[t2], %[t2], %[c6] \t\n"
yading@10 925 "mul.s %[t3], %[t3], %[c7] \t\n"
yading@10 926 "li.s %[c8], 0.86602540378443864676 \t\n"
yading@10 927 "mul.s %[t0], %[in4], %[c8] \t\n"
yading@10 928 "mul.s %[out1], %[out1], %[c5] \t\n"
yading@10 929 "add.s %[t1], %[in2], %[in3] \t\n"
yading@10 930 "li.s %[c9], -0.64278760968653932632 \t\n"
yading@10 931 "add.s %[out2], %[t2], %[t3] \t\n"
yading@10 932 "lwc1 %[in1], 9*4(%[in]) \t\n"
yading@10 933 "swc1 %[out1], 4*4(%[tmp]) \t\n"
yading@10 934 "mul.s %[t1], %[t1], %[c9] \t\n"
yading@10 935 "lwc1 %[in2], 17*4(%[in]) \t\n"
yading@10 936 "add.s %[out2], %[out2], %[t0] \t\n"
yading@10 937 "lwc1 %[in3], 5*4(%[in]) \t\n"
yading@10 938 "lwc1 %[in4], 1*4(%[in]) \t\n"
yading@10 939 "add.s %[out3], %[t2], %[t1] \t\n"
yading@10 940 "sub.s %[out1], %[t3], %[t1] \t\n"
yading@10 941 "swc1 %[out2], 0(%[tmp]) \t\n"
yading@10 942 "lwc1 %[in5], 13*4(%[in]) \t\n"
yading@10 943 "add.s %[t2], %[in1], %[in2] \t\n"
yading@10 944 "sub.s %[out3], %[out3], %[t0] \t\n"
yading@10 945 "sub.s %[out1], %[out1], %[t0] \t\n"
yading@10 946 "add.s %[t0], %[in1], %[in3] \t\n"
yading@10 947 "madd.s %[t3], %[in4], %[in5], %[c1] \t\n"
yading@10 948 "sub.s %[t2], %[t2], %[in3] \t\n"
yading@10 949 "swc1 %[out3], 12*4(%[tmp]) \t\n"
yading@10 950 "swc1 %[out1], 8*4(%[tmp]) \t\n"
yading@10 951 "sub.s %[t1], %[in4], %[in5] \t\n"
yading@10 952 "mul.s %[t0], %[t0], %[c2] \t\n"
yading@10 953 "nmsub.s %[out1], %[t1], %[t2], %[c1] \t\n"
yading@10 954 "add.s %[out2], %[t1], %[t2] \t\n"
yading@10 955 "add.s %[t2], %[in2], %[in3] \t\n"
yading@10 956 "sub.s %[t1], %[in1], %[in2] \t\n"
yading@10 957 "sub.s %[out3], %[t3], %[t0] \t\n"
yading@10 958 "swc1 %[out1], 7*4(%[tmp]) \t\n"
yading@10 959 "swc1 %[out2], 17*4(%[tmp]) \t\n"
yading@10 960 "mul.s %[t2], %[t2], %[c3] \t\n"
yading@10 961 "mul.s %[t1], %[t1], %[c4] \t\n"
yading@10 962 "add.s %[out1], %[t3], %[t0] \t\n"
yading@10 963 "lwc1 %[in1], 11*4(%[in]) \t\n"
yading@10 964 "lwc1 %[in2], 15*4(%[in]) \t\n"
yading@10 965 "sub.s %[out3], %[out3], %[t2] \t\n"
yading@10 966 "add.s %[out2], %[t3], %[t2] \t\n"
yading@10 967 "add.s %[out1], %[out1], %[t1] \t\n"
yading@10 968 "lwc1 %[in3], 3*4(%[in]) \t\n"
yading@10 969 "lwc1 %[in4], 7*4(%[in]) \t\n"
yading@10 970 "swc1 %[out3], 11*4(%[tmp]) \t\n"
yading@10 971 "sub.s %[out2], %[out2], %[t1] \t\n"
yading@10 972 "swc1 %[out1], 3*4(%[tmp]) \t\n"
yading@10 973 "add.s %[out3], %[in1], %[in2] \t\n"
yading@10 974 "add.s %[t2], %[in1], %[in3] \t\n"
yading@10 975 "sub.s %[t3], %[in1], %[in2] \t\n"
yading@10 976 "swc1 %[out2], 15*4(%[tmp]) \t\n"
yading@10 977 "mul.s %[t0], %[in4], %[c8] \t\n"
yading@10 978 "sub.s %[out3], %[out3], %[in3] \t\n"
yading@10 979 "mul.s %[t2], %[t2], %[c6] \t\n"
yading@10 980 "mul.s %[t3], %[t3], %[c7] \t\n"
yading@10 981 "add.s %[t1], %[in2], %[in3] \t\n"
yading@10 982 "mul.s %[out3], %[out3], %[c5] \t\n"
yading@10 983 "add.s %[out1], %[t2], %[t3] \t\n"
yading@10 984 "mul.s %[t1], %[t1], %[c9] \t\n"
yading@10 985 "swc1 %[out3], 5*4(%[tmp]) \t\n"
yading@10 986 "add.s %[out1], %[out1], %[t0] \t\n"
yading@10 987 "add.s %[out2], %[t2], %[t1] \t\n"
yading@10 988 "sub.s %[out3], %[t3], %[t1] \t\n"
yading@10 989 "swc1 %[out1], 1*4(%[tmp]) \t\n"
yading@10 990 "sub.s %[out2], %[out2], %[t0] \t\n"
yading@10 991 "sub.s %[out3], %[out3], %[t0] \t\n"
yading@10 992 "swc1 %[out2], 13*4(%[tmp]) \t\n"
yading@10 993 "swc1 %[out3], 9*4(%[tmp]) \t\n"
yading@10 994
yading@10 995 : [t0] "=&f" (t0), [t1] "=&f" (t1),
yading@10 996 [t2] "=&f" (t2), [t3] "=&f" (t3),
yading@10 997 [in1] "=&f" (in1), [in2] "=&f" (in2),
yading@10 998 [in3] "=&f" (in3), [in4] "=&f" (in4),
yading@10 999 [in5] "=&f" (in5),
yading@10 1000 [out1] "=&f" (out1), [out2] "=&f" (out2),
yading@10 1001 [out3] "=&f" (out3),
yading@10 1002 [c1] "=&f" (c1), [c2] "=&f" (c2),
yading@10 1003 [c3] "=&f" (c3), [c4] "=&f" (c4),
yading@10 1004 [c5] "=&f" (c5), [c6] "=&f" (c6),
yading@10 1005 [c7] "=&f" (c7), [c8] "=&f" (c8),
yading@10 1006 [c9] "=&f" (c9)
yading@10 1007 : [in] "r" (in), [tmp] "r" (tmp)
yading@10 1008 : "memory"
yading@10 1009 );
yading@10 1010
yading@10 1011 /* loop 4 */
yading@10 1012 __asm__ volatile (
yading@10 1013 "lwc1 %[in1], 2*4(%[tmp]) \t\n"
yading@10 1014 "lwc1 %[in2], 0(%[tmp]) \t\n"
yading@10 1015 "lwc1 %[in3], 3*4(%[tmp]) \t\n"
yading@10 1016 "lwc1 %[in4], 1*4(%[tmp]) \t\n"
yading@10 1017 "li.s %[c1], 0.50190991877167369479 \t\n"
yading@10 1018 "li.s %[c2], 5.73685662283492756461 \t\n"
yading@10 1019 "add.s %[s0], %[in1], %[in2] \t\n"
yading@10 1020 "sub.s %[s2], %[in1], %[in2] \t\n"
yading@10 1021 "add.s %[s1], %[in3], %[in4] \t\n"
yading@10 1022 "sub.s %[s3], %[in3], %[in4] \t\n"
yading@10 1023 "lwc1 %[in1], 9*4(%[win]) \t\n"
yading@10 1024 "lwc1 %[in2], 4*9*4(%[buf]) \t\n"
yading@10 1025 "lwc1 %[in3], 8*4(%[win]) \t\n"
yading@10 1026 "mul.s %[s1], %[s1], %[c1] \t\n"
yading@10 1027 "mul.s %[s3], %[s3], %[c2] \t\n"
yading@10 1028 "lwc1 %[in4], 4*8*4(%[buf]) \t\n"
yading@10 1029 "lwc1 %[in5], 29*4(%[win]) \t\n"
yading@10 1030 "lwc1 %[in6], 28*4(%[win]) \t\n"
yading@10 1031 "add.s %[t0], %[s0], %[s1] \t\n"
yading@10 1032 "sub.s %[t1], %[s0], %[s1] \t\n"
yading@10 1033 "li.s %[c1], 0.51763809020504152469 \t\n"
yading@10 1034 "li.s %[c2], 1.93185165257813657349 \t\n"
yading@10 1035 "mul.s %[out3], %[in5], %[t0] \t\n"
yading@10 1036 "madd.s %[out1], %[in2], %[in1], %[t1] \t\n"
yading@10 1037 "madd.s %[out2], %[in4], %[in3], %[t1] \t\n"
yading@10 1038 "mul.s %[out4], %[in6], %[t0] \t\n"
yading@10 1039 "add.s %[t0], %[s2], %[s3] \t\n"
yading@10 1040 "swc1 %[out3], 4*9*4(%[buf]) \t\n"
yading@10 1041 "swc1 %[out1], 288*4(%[out]) \t\n"
yading@10 1042 "swc1 %[out2], 256*4(%[out]) \t\n"
yading@10 1043 "swc1 %[out4], 4*8*4(%[buf]) \t\n"
yading@10 1044 "sub.s %[t1], %[s2], %[s3] \t\n"
yading@10 1045 "lwc1 %[in1], 17*4(%[win]) \t\n"
yading@10 1046 "lwc1 %[in2], 4*17*4(%[buf]) \t\n"
yading@10 1047 "lwc1 %[in3], 0(%[win]) \t\n"
yading@10 1048 "lwc1 %[in4], 0(%[buf]) \t\n"
yading@10 1049 "lwc1 %[in5], 37*4(%[win]) \t\n"
yading@10 1050 "lwc1 %[in6], 20*4(%[win]) \t\n"
yading@10 1051 "madd.s %[out1], %[in2], %[in1], %[t1] \t\n"
yading@10 1052 "lwc1 %[in1], 6*4(%[tmp]) \t\n"
yading@10 1053 "madd.s %[out2], %[in4], %[in3], %[t1] \t\n"
yading@10 1054 "mul.s %[out3], %[t0], %[in5] \t\n"
yading@10 1055 "mul.s %[out4], %[t0], %[in6] \t\n"
yading@10 1056 "swc1 %[out1], 544*4(%[out]) \t\n"
yading@10 1057 "lwc1 %[in2], 4*4(%[tmp]) \t\n"
yading@10 1058 "swc1 %[out2], 0(%[out]) \t\n"
yading@10 1059 "swc1 %[out3], 4*17*4(%[buf]) \t\n"
yading@10 1060 "swc1 %[out4], 0(%[buf]) \t\n"
yading@10 1061 "lwc1 %[in3], 7*4(%[tmp]) \t\n"
yading@10 1062 "add.s %[s0], %[in1], %[in2] \t\n"
yading@10 1063 "sub.s %[s2], %[in1], %[in2] \t\n"
yading@10 1064 "lwc1 %[in4], 5*4(%[tmp]) \t\n"
yading@10 1065 "add.s %[s1], %[in3], %[in4] \t\n"
yading@10 1066 "sub.s %[s3], %[in3], %[in4] \t\n"
yading@10 1067 "lwc1 %[in1], 10*4(%[win]) \t\n"
yading@10 1068 "lwc1 %[in2], 4*10*4(%[buf]) \t\n"
yading@10 1069 "lwc1 %[in3], 7*4(%[win]) \t\n"
yading@10 1070 "mul.s %[s1], %[s1], %[c1] \t\n"
yading@10 1071 "mul.s %[s3], %[s3], %[c2] \t\n"
yading@10 1072 "add.s %[t0], %[s0], %[s1] \t\n"
yading@10 1073 "sub.s %[t1], %[s0], %[s1] \t\n"
yading@10 1074 "lwc1 %[in4], 4*7*4(%[buf]) \t\n"
yading@10 1075 "lwc1 %[in5], 30*4(%[win]) \t\n"
yading@10 1076 "lwc1 %[in6], 27*4(%[win]) \t\n"
yading@10 1077 "li.s %[c1], 0.55168895948124587824 \t\n"
yading@10 1078 "madd.s %[out1], %[in2], %[in1], %[t1] \t\n"
yading@10 1079 "madd.s %[out2], %[in4], %[in3], %[t1] \t\n"
yading@10 1080 "mul.s %[out3], %[t0], %[in5] \t\n"
yading@10 1081 "mul.s %[out4], %[t0], %[in6] \t\n"
yading@10 1082 "add.s %[t0], %[s2], %[s3] \t\n"
yading@10 1083 "swc1 %[out1], 320*4(%[out]) \t\n"
yading@10 1084 "swc1 %[out2], 224*4(%[out]) \t\n"
yading@10 1085 "swc1 %[out3], 4*10*4(%[buf]) \t\n"
yading@10 1086 "swc1 %[out4], 4*7*4(%[buf]) \t\n"
yading@10 1087 "sub.s %[t1], %[s2], %[s3] \t\n"
yading@10 1088 "lwc1 %[in1], 16*4(%[win]) \t\n"
yading@10 1089 "lwc1 %[in2], 4*16*4(%[buf]) \t\n"
yading@10 1090 "lwc1 %[in3], 1*4(%[win]) \t\n"
yading@10 1091 "lwc1 %[in4], 4*1*4(%[buf]) \t\n"
yading@10 1092 "lwc1 %[in5], 36*4(%[win]) \t\n"
yading@10 1093 "lwc1 %[in6], 21*4(%[win]) \t\n"
yading@10 1094 "madd.s %[out1], %[in2], %[in1], %[t1] \t\n"
yading@10 1095 "lwc1 %[in1], 10*4(%[tmp]) \t\n"
yading@10 1096 "madd.s %[out2], %[in4], %[in3], %[t1] \t\n"
yading@10 1097 "mul.s %[out3], %[in5], %[t0] \t\n"
yading@10 1098 "mul.s %[out4], %[in6], %[t0] \t\n"
yading@10 1099 "swc1 %[out1], 512*4(%[out]) \t\n"
yading@10 1100 "lwc1 %[in2], 8*4(%[tmp]) \t\n"
yading@10 1101 "swc1 %[out2], 32*4(%[out]) \t\n"
yading@10 1102 "swc1 %[out3], 4*16*4(%[buf]) \t\n"
yading@10 1103 "swc1 %[out4], 4*1*4(%[buf]) \t\n"
yading@10 1104 "li.s %[c2], 1.18310079157624925896 \t\n"
yading@10 1105 "add.s %[s0], %[in1], %[in2] \t\n"
yading@10 1106 "sub.s %[s2], %[in1], %[in2] \t\n"
yading@10 1107 "lwc1 %[in3], 11*4(%[tmp]) \t\n"
yading@10 1108 "lwc1 %[in4], 9*4(%[tmp]) \t\n"
yading@10 1109 "add.s %[s1], %[in3], %[in4] \t\n"
yading@10 1110 "sub.s %[s3], %[in3], %[in4] \t\n"
yading@10 1111 "lwc1 %[in1], 11*4(%[win]) \t\n"
yading@10 1112 "lwc1 %[in2], 4*11*4(%[buf]) \t\n"
yading@10 1113 "lwc1 %[in3], 6*4(%[win]) \t\n"
yading@10 1114 "mul.s %[s1], %[s1], %[c1] \t\n"
yading@10 1115 "mul.s %[s3], %[s3], %[c2] \t\n"
yading@10 1116 "lwc1 %[in4], 4*6*4(%[buf]) \t\n"
yading@10 1117 "lwc1 %[in5], 31*4(%[win]) \t\n"
yading@10 1118 "lwc1 %[in6], 26*4(%[win]) \t\n"
yading@10 1119 "add.s %[t0], %[s0], %[s1] \t\n"
yading@10 1120 "sub.s %[t1], %[s0], %[s1] \t\n"
yading@10 1121 "mul.s %[out3], %[t0], %[in5] \t\n"
yading@10 1122 "mul.s %[out4], %[t0], %[in6] \t\n"
yading@10 1123 "add.s %[t0], %[s2], %[s3] \t\n"
yading@10 1124 "madd.s %[out1], %[in2], %[in1], %[t1] \t\n"
yading@10 1125 "madd.s %[out2], %[in4], %[in3], %[t1] \t\n"
yading@10 1126 "swc1 %[out3], 4*11*4(%[buf]) \t\n"
yading@10 1127 "swc1 %[out4], 4*6*4(%[buf]) \t\n"
yading@10 1128 "sub.s %[t1], %[s2], %[s3] \t\n"
yading@10 1129 "swc1 %[out1], 352*4(%[out]) \t\n"
yading@10 1130 "swc1 %[out2], 192*4(%[out]) \t\n"
yading@10 1131 "lwc1 %[in1], 15*4(%[win]) \t\n"
yading@10 1132 "lwc1 %[in2], 4*15*4(%[buf]) \t\n"
yading@10 1133 "lwc1 %[in3], 2*4(%[win]) \t\n"
yading@10 1134 "lwc1 %[in4], 4*2*4(%[buf]) \t\n"
yading@10 1135 "lwc1 %[in5], 35*4(%[win]) \t\n"
yading@10 1136 "lwc1 %[in6], 22*4(%[win]) \t\n"
yading@10 1137 "madd.s %[out1], %[in2], %[in1], %[t1] \t\n"
yading@10 1138 "lwc1 %[in1], 14*4(%[tmp]) \t\n"
yading@10 1139 "madd.s %[out2], %[in4], %[in3], %[t1] \t\n"
yading@10 1140 "mul.s %[out3], %[t0], %[in5] \t\n"
yading@10 1141 "mul.s %[out4], %[t0], %[in6] \t\n"
yading@10 1142 "swc1 %[out1], 480*4(%[out]) \t\n"
yading@10 1143 "lwc1 %[in2], 12*4(%[tmp]) \t\n"
yading@10 1144 "swc1 %[out2], 64*4(%[out]) \t\n"
yading@10 1145 "swc1 %[out3], 4*15*4(%[buf]) \t\n"
yading@10 1146 "swc1 %[out4], 4*2*4(%[buf]) \t\n"
yading@10 1147 "lwc1 %[in3], 15*4(%[tmp]) \t\n"
yading@10 1148 "add.s %[s0], %[in1], %[in2] \t\n"
yading@10 1149 "sub.s %[s2], %[in1], %[in2] \t\n"
yading@10 1150 "lwc1 %[in4], 13*4(%[tmp]) \t\n"
yading@10 1151 "li.s %[c1], 0.61038729438072803416 \t\n"
yading@10 1152 "li.s %[c2], 0.87172339781054900991 \t\n"
yading@10 1153 "add.s %[s1], %[in3], %[in4] \t\n"
yading@10 1154 "sub.s %[s3], %[in3], %[in4] \t\n"
yading@10 1155 "lwc1 %[in1], 12*4(%[win]) \t\n"
yading@10 1156 "lwc1 %[in2], 4*12*4(%[buf]) \t\n"
yading@10 1157 "lwc1 %[in3], 5*4(%[win]) \t\n"
yading@10 1158 "mul.s %[s1], %[s1], %[c1] \t\n"
yading@10 1159 "mul.s %[s3], %[s3], %[c2] \t\n"
yading@10 1160 "lwc1 %[in4], 4*5*4(%[buf]) \t\n"
yading@10 1161 "lwc1 %[in5], 32*4(%[win]) \t\n"
yading@10 1162 "lwc1 %[in6], 25*4(%[win]) \t\n"
yading@10 1163 "add.s %[t0], %[s0], %[s1] \t\n"
yading@10 1164 "sub.s %[t1], %[s0], %[s1] \t\n"
yading@10 1165 "lwc1 %[s0], 16*4(%[tmp]) \t\n"
yading@10 1166 "lwc1 %[s1], 17*4(%[tmp]) \t\n"
yading@10 1167 "li.s %[c1], 0.70710678118654752439 \t\n"
yading@10 1168 "mul.s %[out3], %[t0], %[in5] \t\n"
yading@10 1169 "madd.s %[out1], %[in2], %[in1], %[t1] \t\n"
yading@10 1170 "madd.s %[out2], %[in4], %[in3], %[t1] \t\n"
yading@10 1171 "mul.s %[out4], %[t0], %[in6] \t\n"
yading@10 1172 "add.s %[t0], %[s2], %[s3] \t\n"
yading@10 1173 "swc1 %[out3], 4*12*4(%[buf]) \t\n"
yading@10 1174 "swc1 %[out1], 384*4(%[out]) \t\n"
yading@10 1175 "swc1 %[out2], 160*4(%[out]) \t\n"
yading@10 1176 "swc1 %[out4], 4*5*4(%[buf]) \t\n"
yading@10 1177 "sub.s %[t1], %[s2], %[s3] \t\n"
yading@10 1178 "lwc1 %[in1], 14*4(%[win]) \t\n"
yading@10 1179 "lwc1 %[in2], 4*14*4(%[buf]) \t\n"
yading@10 1180 "lwc1 %[in3], 3*4(%[win]) \t\n"
yading@10 1181 "lwc1 %[in4], 4*3*4(%[buf]) \t\n"
yading@10 1182 "lwc1 %[in5], 34*4(%[win]) \t\n"
yading@10 1183 "lwc1 %[in6], 23*4(%[win]) \t\n"
yading@10 1184 "madd.s %[out1], %[in2], %[in1], %[t1] \t\n"
yading@10 1185 "mul.s %[s1], %[s1], %[c1] \t\n"
yading@10 1186 "madd.s %[out2], %[in4], %[in3], %[t1] \t\n"
yading@10 1187 "mul.s %[out3], %[in5], %[t0] \t\n"
yading@10 1188 "mul.s %[out4], %[in6], %[t0] \t\n"
yading@10 1189 "swc1 %[out1], 448*4(%[out]) \t\n"
yading@10 1190 "add.s %[t0], %[s0], %[s1] \t\n"
yading@10 1191 "swc1 %[out2], 96*4(%[out]) \t\n"
yading@10 1192 "swc1 %[out3], 4*14*4(%[buf]) \t\n"
yading@10 1193 "swc1 %[out4], 4*3*4(%[buf]) \t\n"
yading@10 1194 "sub.s %[t1], %[s0], %[s1] \t\n"
yading@10 1195 "lwc1 %[in1], 13*4(%[win]) \t\n"
yading@10 1196 "lwc1 %[in2], 4*13*4(%[buf]) \t\n"
yading@10 1197 "lwc1 %[in3], 4*4(%[win]) \t\n"
yading@10 1198 "lwc1 %[in4], 4*4*4(%[buf]) \t\n"
yading@10 1199 "lwc1 %[in5], 33*4(%[win]) \t\n"
yading@10 1200 "lwc1 %[in6], 24*4(%[win]) \t\n"
yading@10 1201 "madd.s %[out1], %[in2], %[in1], %[t1] \t\n"
yading@10 1202 "madd.s %[out2], %[in4], %[in3], %[t1] \t\n"
yading@10 1203 "mul.s %[out3], %[t0], %[in5] \t\n"
yading@10 1204 "mul.s %[out4], %[t0], %[in6] \t\n"
yading@10 1205 "swc1 %[out1], 416*4(%[out]) \t\n"
yading@10 1206 "swc1 %[out2], 128*4(%[out]) \t\n"
yading@10 1207 "swc1 %[out3], 4*13*4(%[buf]) \t\n"
yading@10 1208 "swc1 %[out4], 4*4*4(%[buf]) \t\n"
yading@10 1209
yading@10 1210 : [c1] "=&f" (c1), [c2] "=&f" (c2),
yading@10 1211 [in1] "=&f" (in1), [in2] "=&f" (in2),
yading@10 1212 [in3] "=&f" (in3), [in4] "=&f" (in4),
yading@10 1213 [in5] "=&f" (in5), [in6] "=&f" (in6),
yading@10 1214 [out1] "=&f" (out1), [out2] "=&f" (out2),
yading@10 1215 [out3] "=&f" (out3), [out4] "=&f" (out4),
yading@10 1216 [t0] "=&f" (t0), [t1] "=&f" (t1),
yading@10 1217 [t2] "=&f" (t2), [t3] "=&f" (t3),
yading@10 1218 [s0] "=&f" (s0), [s1] "=&f" (s1),
yading@10 1219 [s2] "=&f" (s2), [s3] "=&f" (s3)
yading@10 1220 : [tmp] "r" (tmp), [win] "r" (win),
yading@10 1221 [buf] "r" (buf), [out] "r" (out)
yading@10 1222 : "memory"
yading@10 1223 );
yading@10 1224 }
yading@10 1225
yading@10 1226 static void ff_imdct36_blocks_mips_float(float *out, float *buf, float *in,
yading@10 1227 int count, int switch_point, int block_type)
yading@10 1228 {
yading@10 1229 int j;
yading@10 1230 for (j=0 ; j < count; j++) {
yading@10 1231 /* apply window & overlap with previous buffer */
yading@10 1232
yading@10 1233 /* select window */
yading@10 1234 int win_idx = (switch_point && j < 2) ? 0 : block_type;
yading@10 1235 float *win = ff_mdct_win_float[win_idx + (4 & -(j & 1))];
yading@10 1236
yading@10 1237 imdct36_mips_float(out, buf, in, win);
yading@10 1238
yading@10 1239 in += 18;
yading@10 1240 buf += ((j&3) != 3 ? 1 : (72-3));
yading@10 1241 out++;
yading@10 1242 }
yading@10 1243 }
yading@10 1244
yading@10 1245 void ff_mpadsp_init_mipsfpu(MPADSPContext *s)
yading@10 1246 {
yading@10 1247 s->apply_window_float = ff_mpadsp_apply_window_mips_float;
yading@10 1248 s->imdct36_blocks_float = ff_imdct36_blocks_mips_float;
yading@10 1249 s->dct32_float = ff_dct32_mips_float;
yading@10 1250 }