Mercurial > hg > pmhd

annotate ffmpeg/libavcodec/mips/aacsbr_mips.h @ 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 * Authors: Djordje Pesut (djordje@mips.com)
yading@10 30 * Mirjana Vulin (mvulin@mips.com)
yading@10 31 *
yading@10 32 * This file is part of FFmpeg.
yading@10 33 *
yading@10 34 * FFmpeg is free software; you can redistribute it and/or
yading@10 35 * modify it under the terms of the GNU Lesser General Public
yading@10 36 * License as published by the Free Software Foundation; either
yading@10 37 * version 2.1 of the License, or (at your option) any later version.
yading@10 38 *
yading@10 39 * FFmpeg is distributed in the hope that it will be useful,
yading@10 40 * but WITHOUT ANY WARRANTY; without even the implied warranty of
yading@10 41 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
yading@10 42 * Lesser General Public License for more details.
yading@10 43 *
yading@10 44 * You should have received a copy of the GNU Lesser General Public
yading@10 45 * License along with FFmpeg; if not, write to the Free Software
yading@10 46 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
yading@10 47 */
yading@10 48
yading@10 49 /**
yading@10 50 * @file
yading@10 51 * Reference: libavcodec/aacsbr.c
yading@10 52 */
yading@10 53
yading@10 54 #ifndef AVCODEC_MIPS_AACSBR_FLOAT_H
yading@10 55 #define AVCODEC_MIPS_AACSBR_FLOAT_H
yading@10 56
yading@10 57 #include "libavcodec/aac.h"
yading@10 58 #include "libavcodec/sbr.h"
yading@10 59
yading@10 60 #if HAVE_INLINE_ASM
yading@10 61 static void sbr_qmf_analysis_mips(AVFloatDSPContext *fdsp, FFTContext *mdct,
yading@10 62 SBRDSPContext *sbrdsp, const float *in, float *x,
yading@10 63 float z[320], float W[2][32][32][2], int buf_idx)
yading@10 64 {
yading@10 65 int i;
yading@10 66 float *w0;
yading@10 67 float *w1;
yading@10 68 int temp0, temp1, temp2, temp3, temp4, temp5, temp6, temp7;
yading@10 69
yading@10 70 w0 = x;
yading@10 71 w1 = x + 1024;
yading@10 72 for(i = 0; i < 36; i++)
yading@10 73 {
yading@10 74 /* loop unrolled 8 times */
yading@10 75 __asm__ volatile(
yading@10 76 "lw %[temp0], 0(%[w1]) \n\t"
yading@10 77 "lw %[temp1], 4(%[w1]) \n\t"
yading@10 78 "lw %[temp2], 8(%[w1]) \n\t"
yading@10 79 "lw %[temp3], 12(%[w1]) \n\t"
yading@10 80 "lw %[temp4], 16(%[w1]) \n\t"
yading@10 81 "lw %[temp5], 20(%[w1]) \n\t"
yading@10 82 "lw %[temp6], 24(%[w1]) \n\t"
yading@10 83 "lw %[temp7], 28(%[w1]) \n\t"
yading@10 84 "sw %[temp0], 0(%[w0]) \n\t"
yading@10 85 "sw %[temp1], 4(%[w0]) \n\t"
yading@10 86 "sw %[temp2], 8(%[w0]) \n\t"
yading@10 87 "sw %[temp3], 12(%[w0]) \n\t"
yading@10 88 "sw %[temp4], 16(%[w0]) \n\t"
yading@10 89 "sw %[temp5], 20(%[w0]) \n\t"
yading@10 90 "sw %[temp6], 24(%[w0]) \n\t"
yading@10 91 "sw %[temp7], 28(%[w0]) \n\t"
yading@10 92 "addiu %[w0], %[w0], 32 \n\t"
yading@10 93 "addiu %[w1], %[w1], 32 \n\t"
yading@10 94
yading@10 95 : [w0]"+r"(w0), [w1]"+r"(w1),
yading@10 96 [temp0]"=&r"(temp0), [temp1]"=&r"(temp1),
yading@10 97 [temp2]"=&r"(temp2), [temp3]"=&r"(temp3),
yading@10 98 [temp4]"=&r"(temp4), [temp5]"=&r"(temp5),
yading@10 99 [temp6]"=&r"(temp6), [temp7]"=&r"(temp7)
yading@10 100 :
yading@10 101 : "memory"
yading@10 102 );
yading@10 103 }
yading@10 104
yading@10 105 w0 = x + 288;
yading@10 106 w1 = (float*)in;
yading@10 107 for(i = 0; i < 128; i++)
yading@10 108 {
yading@10 109 /* loop unrolled 8 times */
yading@10 110 __asm__ volatile(
yading@10 111 "lw %[temp0], 0(%[w1]) \n\t"
yading@10 112 "lw %[temp1], 4(%[w1]) \n\t"
yading@10 113 "lw %[temp2], 8(%[w1]) \n\t"
yading@10 114 "lw %[temp3], 12(%[w1]) \n\t"
yading@10 115 "lw %[temp4], 16(%[w1]) \n\t"
yading@10 116 "lw %[temp5], 20(%[w1]) \n\t"
yading@10 117 "lw %[temp6], 24(%[w1]) \n\t"
yading@10 118 "lw %[temp7], 28(%[w1]) \n\t"
yading@10 119 "sw %[temp0], 0(%[w0]) \n\t"
yading@10 120 "sw %[temp1], 4(%[w0]) \n\t"
yading@10 121 "sw %[temp2], 8(%[w0]) \n\t"
yading@10 122 "sw %[temp3], 12(%[w0]) \n\t"
yading@10 123 "sw %[temp4], 16(%[w0]) \n\t"
yading@10 124 "sw %[temp5], 20(%[w0]) \n\t"
yading@10 125 "sw %[temp6], 24(%[w0]) \n\t"
yading@10 126 "sw %[temp7], 28(%[w0]) \n\t"
yading@10 127 "addiu %[w0], %[w0], 32 \n\t"
yading@10 128 "addiu %[w1], %[w1], 32 \n\t"
yading@10 129
yading@10 130 : [w0]"+r"(w0), [w1]"+r"(w1),
yading@10 131 [temp0]"=&r"(temp0), [temp1]"=&r"(temp1),
yading@10 132 [temp2]"=&r"(temp2), [temp3]"=&r"(temp3),
yading@10 133 [temp4]"=&r"(temp4), [temp5]"=&r"(temp5),
yading@10 134 [temp6]"=&r"(temp6), [temp7]"=&r"(temp7)
yading@10 135 :
yading@10 136 : "memory"
yading@10 137 );
yading@10 138 }
yading@10 139
yading@10 140 for (i = 0; i < 32; i++) { // numTimeSlots*RATE = 16*2 as 960 sample frames
yading@10 141 // are not supported
yading@10 142 fdsp->vector_fmul_reverse(z, sbr_qmf_window_ds, x, 320);
yading@10 143 sbrdsp->sum64x5(z);
yading@10 144 sbrdsp->qmf_pre_shuffle(z);
yading@10 145 mdct->imdct_half(mdct, z, z+64);
yading@10 146 sbrdsp->qmf_post_shuffle(W[buf_idx][i], z);
yading@10 147 x += 32;
yading@10 148 }
yading@10 149 }
yading@10 150
yading@10 151 #if HAVE_MIPSFPU
yading@10 152 static void sbr_qmf_synthesis_mips(FFTContext *mdct,
yading@10 153 SBRDSPContext *sbrdsp, AVFloatDSPContext *fdsp,
yading@10 154 float *out, float X[2][38][64],
yading@10 155 float mdct_buf[2][64],
yading@10 156 float *v0, int *v_off, const unsigned int div)
yading@10 157 {
yading@10 158 int i, n;
yading@10 159 const float *sbr_qmf_window = div ? sbr_qmf_window_ds : sbr_qmf_window_us;
yading@10 160 const int step = 128 >> div;
yading@10 161 float *v;
yading@10 162 float temp0, temp1, temp2, temp3, temp4, temp5, temp6, temp7, temp8, temp9, temp10, temp11, temp12, temp13;
yading@10 163 float temp14, temp15, temp16, temp17, temp18, temp19;
yading@10 164 float *vv0, *s0, *dst;
yading@10 165 dst = out;
yading@10 166
yading@10 167 for (i = 0; i < 32; i++) {
yading@10 168 if (*v_off < step) {
yading@10 169 int saved_samples = (1280 - 128) >> div;
yading@10 170 memcpy(&v0[SBR_SYNTHESIS_BUF_SIZE - saved_samples], v0, saved_samples * sizeof(float));
yading@10 171 *v_off = SBR_SYNTHESIS_BUF_SIZE - saved_samples - step;
yading@10 172 } else {
yading@10 173 *v_off -= step;
yading@10 174 }
yading@10 175 v = v0 + *v_off;
yading@10 176 if (div) {
yading@10 177 for (n = 0; n < 32; n++) {
yading@10 178 X[0][i][ n] = -X[0][i][n];
yading@10 179 X[0][i][32+n] = X[1][i][31-n];
yading@10 180 }
yading@10 181 mdct->imdct_half(mdct, mdct_buf[0], X[0][i]);
yading@10 182 sbrdsp->qmf_deint_neg(v, mdct_buf[0]);
yading@10 183 } else {
yading@10 184 sbrdsp->neg_odd_64(X[1][i]);
yading@10 185 mdct->imdct_half(mdct, mdct_buf[0], X[0][i]);
yading@10 186 mdct->imdct_half(mdct, mdct_buf[1], X[1][i]);
yading@10 187 sbrdsp->qmf_deint_bfly(v, mdct_buf[1], mdct_buf[0]);
yading@10 188 }
yading@10 189
yading@10 190 if(div == 0)
yading@10 191 {
yading@10 192 float *v0_end;
yading@10 193 vv0 = v;
yading@10 194 v0_end = v + 60;
yading@10 195 s0 = (float*)sbr_qmf_window;
yading@10 196
yading@10 197 /* 10 calls of function vector_fmul_add merged into one loop
yading@10 198 and loop unrolled 4 times */
yading@10 199 __asm__ volatile(
yading@10 200 ".set push \n\t"
yading@10 201 ".set noreorder \n\t"
yading@10 202 "lwc1 %[temp4], 0(%[v0]) \n\t"
yading@10 203 "lwc1 %[temp5], 0(%[s0]) \n\t"
yading@10 204 "lwc1 %[temp6], 4(%[v0]) \n\t"
yading@10 205 "lwc1 %[temp7], 4(%[s0]) \n\t"
yading@10 206 "lwc1 %[temp8], 8(%[v0]) \n\t"
yading@10 207 "lwc1 %[temp9], 8(%[s0]) \n\t"
yading@10 208 "lwc1 %[temp10], 12(%[v0]) \n\t"
yading@10 209 "lwc1 %[temp11], 12(%[s0]) \n\t"
yading@10 210 "lwc1 %[temp12], 768(%[v0]) \n\t"
yading@10 211 "lwc1 %[temp13], 256(%[s0]) \n\t"
yading@10 212 "lwc1 %[temp14], 772(%[v0]) \n\t"
yading@10 213 "lwc1 %[temp15], 260(%[s0]) \n\t"
yading@10 214 "lwc1 %[temp16], 776(%[v0]) \n\t"
yading@10 215 "lwc1 %[temp17], 264(%[s0]) \n\t"
yading@10 216 "lwc1 %[temp18], 780(%[v0]) \n\t"
yading@10 217 "lwc1 %[temp19], 268(%[s0]) \n\t"
yading@10 218 "1: \n\t"
yading@10 219 "mul.s %[temp0], %[temp4], %[temp5] \n\t"
yading@10 220 "lwc1 %[temp4], 1024(%[v0]) \n\t"
yading@10 221 "mul.s %[temp1], %[temp6], %[temp7] \n\t"
yading@10 222 "lwc1 %[temp5], 512(%[s0]) \n\t"
yading@10 223 "mul.s %[temp2], %[temp8], %[temp9] \n\t"
yading@10 224 "lwc1 %[temp6], 1028(%[v0]) \n\t"
yading@10 225 "mul.s %[temp3], %[temp10], %[temp11] \n\t"
yading@10 226 "lwc1 %[temp7], 516(%[s0]) \n\t"
yading@10 227 "madd.s %[temp0], %[temp0], %[temp12], %[temp13] \n\t"
yading@10 228 "lwc1 %[temp8], 1032(%[v0]) \n\t"
yading@10 229 "madd.s %[temp1], %[temp1], %[temp14], %[temp15] \n\t"
yading@10 230 "lwc1 %[temp9], 520(%[s0]) \n\t"
yading@10 231 "madd.s %[temp2], %[temp2], %[temp16], %[temp17] \n\t"
yading@10 232 "lwc1 %[temp10], 1036(%[v0]) \n\t"
yading@10 233 "madd.s %[temp3], %[temp3], %[temp18], %[temp19] \n\t"
yading@10 234 "lwc1 %[temp11], 524(%[s0]) \n\t"
yading@10 235 "lwc1 %[temp12], 1792(%[v0]) \n\t"
yading@10 236 "lwc1 %[temp13], 768(%[s0]) \n\t"
yading@10 237 "lwc1 %[temp14], 1796(%[v0]) \n\t"
yading@10 238 "lwc1 %[temp15], 772(%[s0]) \n\t"
yading@10 239 "lwc1 %[temp16], 1800(%[v0]) \n\t"
yading@10 240 "lwc1 %[temp17], 776(%[s0]) \n\t"
yading@10 241 "lwc1 %[temp18], 1804(%[v0]) \n\t"
yading@10 242 "lwc1 %[temp19], 780(%[s0]) \n\t"
yading@10 243 "madd.s %[temp0], %[temp0], %[temp4], %[temp5] \n\t"
yading@10 244 "lwc1 %[temp4], 2048(%[v0]) \n\t"
yading@10 245 "madd.s %[temp1], %[temp1], %[temp6], %[temp7] \n\t"
yading@10 246 "lwc1 %[temp5], 1024(%[s0]) \n\t"
yading@10 247 "madd.s %[temp2], %[temp2], %[temp8], %[temp9] \n\t"
yading@10 248 "lwc1 %[temp6], 2052(%[v0]) \n\t"
yading@10 249 "madd.s %[temp3], %[temp3], %[temp10], %[temp11] \n\t"
yading@10 250 "lwc1 %[temp7], 1028(%[s0]) \n\t"
yading@10 251 "madd.s %[temp0], %[temp0], %[temp12], %[temp13] \n\t"
yading@10 252 "lwc1 %[temp8], 2056(%[v0]) \n\t"
yading@10 253 "madd.s %[temp1], %[temp1], %[temp14], %[temp15] \n\t"
yading@10 254 "lwc1 %[temp9], 1032(%[s0]) \n\t"
yading@10 255 "madd.s %[temp2], %[temp2], %[temp16], %[temp17] \n\t"
yading@10 256 "lwc1 %[temp10], 2060(%[v0]) \n\t"
yading@10 257 "madd.s %[temp3], %[temp3], %[temp18], %[temp19] \n\t"
yading@10 258 "lwc1 %[temp11], 1036(%[s0]) \n\t"
yading@10 259 "lwc1 %[temp12], 2816(%[v0]) \n\t"
yading@10 260 "lwc1 %[temp13], 1280(%[s0]) \n\t"
yading@10 261 "lwc1 %[temp14], 2820(%[v0]) \n\t"
yading@10 262 "lwc1 %[temp15], 1284(%[s0]) \n\t"
yading@10 263 "lwc1 %[temp16], 2824(%[v0]) \n\t"
yading@10 264 "lwc1 %[temp17], 1288(%[s0]) \n\t"
yading@10 265 "lwc1 %[temp18], 2828(%[v0]) \n\t"
yading@10 266 "lwc1 %[temp19], 1292(%[s0]) \n\t"
yading@10 267 "madd.s %[temp0], %[temp0], %[temp4], %[temp5] \n\t"
yading@10 268 "lwc1 %[temp4], 3072(%[v0]) \n\t"
yading@10 269 "madd.s %[temp1], %[temp1], %[temp6], %[temp7] \n\t"
yading@10 270 "lwc1 %[temp5], 1536(%[s0]) \n\t"
yading@10 271 "madd.s %[temp2], %[temp2], %[temp8], %[temp9] \n\t"
yading@10 272 "lwc1 %[temp6], 3076(%[v0]) \n\t"
yading@10 273 "madd.s %[temp3], %[temp3], %[temp10], %[temp11] \n\t"
yading@10 274 "lwc1 %[temp7], 1540(%[s0]) \n\t"
yading@10 275 "madd.s %[temp0], %[temp0], %[temp12], %[temp13] \n\t"
yading@10 276 "lwc1 %[temp8], 3080(%[v0]) \n\t"
yading@10 277 "madd.s %[temp1], %[temp1], %[temp14], %[temp15] \n\t"
yading@10 278 "lwc1 %[temp9], 1544(%[s0]) \n\t"
yading@10 279 "madd.s %[temp2], %[temp2], %[temp16], %[temp17] \n\t"
yading@10 280 "lwc1 %[temp10], 3084(%[v0]) \n\t"
yading@10 281 "madd.s %[temp3], %[temp3], %[temp18], %[temp19] \n\t"
yading@10 282 "lwc1 %[temp11], 1548(%[s0]) \n\t"
yading@10 283 "lwc1 %[temp12], 3840(%[v0]) \n\t"
yading@10 284 "lwc1 %[temp13], 1792(%[s0]) \n\t"
yading@10 285 "lwc1 %[temp14], 3844(%[v0]) \n\t"
yading@10 286 "lwc1 %[temp15], 1796(%[s0]) \n\t"
yading@10 287 "lwc1 %[temp16], 3848(%[v0]) \n\t"
yading@10 288 "lwc1 %[temp17], 1800(%[s0]) \n\t"
yading@10 289 "lwc1 %[temp18], 3852(%[v0]) \n\t"
yading@10 290 "lwc1 %[temp19], 1804(%[s0]) \n\t"
yading@10 291 "madd.s %[temp0], %[temp0], %[temp4], %[temp5] \n\t"
yading@10 292 "lwc1 %[temp4], 4096(%[v0]) \n\t"
yading@10 293 "madd.s %[temp1], %[temp1], %[temp6], %[temp7] \n\t"
yading@10 294 "lwc1 %[temp5], 2048(%[s0]) \n\t"
yading@10 295 "madd.s %[temp2], %[temp2], %[temp8], %[temp9] \n\t"
yading@10 296 "lwc1 %[temp6], 4100(%[v0]) \n\t"
yading@10 297 "madd.s %[temp3], %[temp3], %[temp10], %[temp11] \n\t"
yading@10 298 "lwc1 %[temp7], 2052(%[s0]) \n\t"
yading@10 299 "madd.s %[temp0], %[temp0], %[temp12], %[temp13] \n\t"
yading@10 300 "lwc1 %[temp8], 4104(%[v0]) \n\t"
yading@10 301 "addiu %[dst], %[dst], 16 \n\t"
yading@10 302 "madd.s %[temp1], %[temp1], %[temp14], %[temp15] \n\t"
yading@10 303 "lwc1 %[temp9], 2056(%[s0]) \n\t"
yading@10 304 "addiu %[s0], %[s0], 16 \n\t"
yading@10 305 "madd.s %[temp2], %[temp2], %[temp16], %[temp17] \n\t"
yading@10 306 "lwc1 %[temp10], 4108(%[v0]) \n\t"
yading@10 307 "addiu %[v0], %[v0], 16 \n\t"
yading@10 308 "madd.s %[temp3], %[temp3], %[temp18], %[temp19] \n\t"
yading@10 309 "lwc1 %[temp11], 2044(%[s0]) \n\t"
yading@10 310 "lwc1 %[temp12], 4848(%[v0]) \n\t"
yading@10 311 "lwc1 %[temp13], 2288(%[s0]) \n\t"
yading@10 312 "lwc1 %[temp14], 4852(%[v0]) \n\t"
yading@10 313 "lwc1 %[temp15], 2292(%[s0]) \n\t"
yading@10 314 "lwc1 %[temp16], 4856(%[v0]) \n\t"
yading@10 315 "lwc1 %[temp17], 2296(%[s0]) \n\t"
yading@10 316 "lwc1 %[temp18], 4860(%[v0]) \n\t"
yading@10 317 "lwc1 %[temp19], 2300(%[s0]) \n\t"
yading@10 318 "madd.s %[temp0], %[temp0], %[temp4], %[temp5] \n\t"
yading@10 319 "lwc1 %[temp4], 0(%[v0]) \n\t"
yading@10 320 "madd.s %[temp1], %[temp1], %[temp6], %[temp7] \n\t"
yading@10 321 "lwc1 %[temp5], 0(%[s0]) \n\t"
yading@10 322 "madd.s %[temp2], %[temp2], %[temp8], %[temp9] \n\t"
yading@10 323 "lwc1 %[temp6], 4(%[v0]) \n\t"
yading@10 324 "madd.s %[temp3], %[temp3], %[temp10], %[temp11] \n\t"
yading@10 325 "lwc1 %[temp7], 4(%[s0]) \n\t"
yading@10 326 "madd.s %[temp0], %[temp0], %[temp12], %[temp13] \n\t"
yading@10 327 "lwc1 %[temp8], 8(%[v0]) \n\t"
yading@10 328 "madd.s %[temp1], %[temp1], %[temp14], %[temp15] \n\t"
yading@10 329 "lwc1 %[temp9], 8(%[s0]) \n\t"
yading@10 330 "madd.s %[temp2], %[temp2], %[temp16], %[temp17] \n\t"
yading@10 331 "lwc1 %[temp10], 12(%[v0]) \n\t"
yading@10 332 "madd.s %[temp3], %[temp3], %[temp18], %[temp19] \n\t"
yading@10 333 "lwc1 %[temp11], 12(%[s0]) \n\t"
yading@10 334 "lwc1 %[temp12], 768(%[v0]) \n\t"
yading@10 335 "lwc1 %[temp13], 256(%[s0]) \n\t"
yading@10 336 "lwc1 %[temp14], 772(%[v0]) \n\t"
yading@10 337 "lwc1 %[temp15], 260(%[s0]) \n\t"
yading@10 338 "lwc1 %[temp16], 776(%[v0]) \n\t"
yading@10 339 "lwc1 %[temp17], 264(%[s0]) \n\t"
yading@10 340 "lwc1 %[temp18], 780(%[v0]) \n\t"
yading@10 341 "lwc1 %[temp19], 268(%[s0]) \n\t"
yading@10 342 "swc1 %[temp0], -16(%[dst]) \n\t"
yading@10 343 "swc1 %[temp1], -12(%[dst]) \n\t"
yading@10 344 "swc1 %[temp2], -8(%[dst]) \n\t"
yading@10 345 "bne %[v0], %[v0_end], 1b \n\t"
yading@10 346 " swc1 %[temp3], -4(%[dst]) \n\t"
yading@10 347 "mul.s %[temp0], %[temp4], %[temp5] \n\t"
yading@10 348 "lwc1 %[temp4], 1024(%[v0]) \n\t"
yading@10 349 "mul.s %[temp1], %[temp6], %[temp7] \n\t"
yading@10 350 "lwc1 %[temp5], 512(%[s0]) \n\t"
yading@10 351 "mul.s %[temp2], %[temp8], %[temp9] \n\t"
yading@10 352 "lwc1 %[temp6], 1028(%[v0]) \n\t"
yading@10 353 "mul.s %[temp3], %[temp10], %[temp11] \n\t"
yading@10 354 "lwc1 %[temp7], 516(%[s0]) \n\t"
yading@10 355 "madd.s %[temp0], %[temp0], %[temp12], %[temp13] \n\t"
yading@10 356 "lwc1 %[temp8], 1032(%[v0]) \n\t"
yading@10 357 "madd.s %[temp1], %[temp1], %[temp14], %[temp15] \n\t"
yading@10 358 "lwc1 %[temp9], 520(%[s0]) \n\t"
yading@10 359 "madd.s %[temp2], %[temp2], %[temp16], %[temp17] \n\t"
yading@10 360 "lwc1 %[temp10], 1036(%[v0]) \n\t"
yading@10 361 "madd.s %[temp3], %[temp3], %[temp18], %[temp19] \n\t"
yading@10 362 "lwc1 %[temp11], 524(%[s0]) \n\t"
yading@10 363 "lwc1 %[temp12], 1792(%[v0]) \n\t"
yading@10 364 "lwc1 %[temp13], 768(%[s0]) \n\t"
yading@10 365 "lwc1 %[temp14], 1796(%[v0]) \n\t"
yading@10 366 "lwc1 %[temp15], 772(%[s0]) \n\t"
yading@10 367 "lwc1 %[temp16], 1800(%[v0]) \n\t"
yading@10 368 "lwc1 %[temp17], 776(%[s0]) \n\t"
yading@10 369 "lwc1 %[temp18], 1804(%[v0]) \n\t"
yading@10 370 "lwc1 %[temp19], 780(%[s0]) \n\t"
yading@10 371 "madd.s %[temp0], %[temp0], %[temp4], %[temp5] \n\t"
yading@10 372 "lwc1 %[temp4], 2048(%[v0]) \n\t"
yading@10 373 "madd.s %[temp1], %[temp1], %[temp6], %[temp7] \n\t"
yading@10 374 "lwc1 %[temp5], 1024(%[s0]) \n\t"
yading@10 375 "madd.s %[temp2], %[temp2], %[temp8], %[temp9] \n\t"
yading@10 376 "lwc1 %[temp6], 2052(%[v0]) \n\t"
yading@10 377 "madd.s %[temp3], %[temp3], %[temp10], %[temp11] \n\t"
yading@10 378 "lwc1 %[temp7], 1028(%[s0]) \n\t"
yading@10 379 "madd.s %[temp0], %[temp0], %[temp12], %[temp13] \n\t"
yading@10 380 "lwc1 %[temp8], 2056(%[v0]) \n\t"
yading@10 381 "madd.s %[temp1], %[temp1], %[temp14], %[temp15] \n\t"
yading@10 382 "lwc1 %[temp9], 1032(%[s0]) \n\t"
yading@10 383 "madd.s %[temp2], %[temp2], %[temp16], %[temp17] \n\t"
yading@10 384 "lwc1 %[temp10], 2060(%[v0]) \n\t"
yading@10 385 "madd.s %[temp3], %[temp3], %[temp18], %[temp19] \n\t"
yading@10 386 "lwc1 %[temp11], 1036(%[s0]) \n\t"
yading@10 387 "lwc1 %[temp12], 2816(%[v0]) \n\t"
yading@10 388 "lwc1 %[temp13], 1280(%[s0]) \n\t"
yading@10 389 "lwc1 %[temp14], 2820(%[v0]) \n\t"
yading@10 390 "lwc1 %[temp15], 1284(%[s0]) \n\t"
yading@10 391 "lwc1 %[temp16], 2824(%[v0]) \n\t"
yading@10 392 "lwc1 %[temp17], 1288(%[s0]) \n\t"
yading@10 393 "lwc1 %[temp18], 2828(%[v0]) \n\t"
yading@10 394 "lwc1 %[temp19], 1292(%[s0]) \n\t"
yading@10 395 "madd.s %[temp0], %[temp0], %[temp4], %[temp5] \n\t"
yading@10 396 "lwc1 %[temp4], 3072(%[v0]) \n\t"
yading@10 397 "madd.s %[temp1], %[temp1], %[temp6], %[temp7] \n\t"
yading@10 398 "lwc1 %[temp5], 1536(%[s0]) \n\t"
yading@10 399 "madd.s %[temp2], %[temp2], %[temp8], %[temp9] \n\t"
yading@10 400 "lwc1 %[temp6], 3076(%[v0]) \n\t"
yading@10 401 "madd.s %[temp3], %[temp3], %[temp10], %[temp11] \n\t"
yading@10 402 "lwc1 %[temp7], 1540(%[s0]) \n\t"
yading@10 403 "madd.s %[temp0], %[temp0], %[temp12], %[temp13] \n\t"
yading@10 404 "lwc1 %[temp8], 3080(%[v0]) \n\t"
yading@10 405 "madd.s %[temp1], %[temp1], %[temp14], %[temp15] \n\t"
yading@10 406 "lwc1 %[temp9], 1544(%[s0]) \n\t"
yading@10 407 "madd.s %[temp2], %[temp2], %[temp16], %[temp17] \n\t"
yading@10 408 "lwc1 %[temp10], 3084(%[v0]) \n\t"
yading@10 409 "madd.s %[temp3], %[temp3], %[temp18], %[temp19] \n\t"
yading@10 410 "lwc1 %[temp11], 1548(%[s0]) \n\t"
yading@10 411 "lwc1 %[temp12], 3840(%[v0]) \n\t"
yading@10 412 "lwc1 %[temp13], 1792(%[s0]) \n\t"
yading@10 413 "lwc1 %[temp14], 3844(%[v0]) \n\t"
yading@10 414 "lwc1 %[temp15], 1796(%[s0]) \n\t"
yading@10 415 "lwc1 %[temp16], 3848(%[v0]) \n\t"
yading@10 416 "lwc1 %[temp17], 1800(%[s0]) \n\t"
yading@10 417 "lwc1 %[temp18], 3852(%[v0]) \n\t"
yading@10 418 "lwc1 %[temp19], 1804(%[s0]) \n\t"
yading@10 419 "madd.s %[temp0], %[temp0], %[temp4], %[temp5] \n\t"
yading@10 420 "lwc1 %[temp4], 4096(%[v0]) \n\t"
yading@10 421 "madd.s %[temp1], %[temp1], %[temp6], %[temp7] \n\t"
yading@10 422 "lwc1 %[temp5], 2048(%[s0]) \n\t"
yading@10 423 "madd.s %[temp2], %[temp2], %[temp8], %[temp9] \n\t"
yading@10 424 "lwc1 %[temp6], 4100(%[v0]) \n\t"
yading@10 425 "madd.s %[temp3], %[temp3], %[temp10], %[temp11] \n\t"
yading@10 426 "lwc1 %[temp7], 2052(%[s0]) \n\t"
yading@10 427 "madd.s %[temp0], %[temp0], %[temp12], %[temp13] \n\t"
yading@10 428 "lwc1 %[temp8], 4104(%[v0]) \n\t"
yading@10 429 "madd.s %[temp1], %[temp1], %[temp14], %[temp15] \n\t"
yading@10 430 "lwc1 %[temp9], 2056(%[s0]) \n\t"
yading@10 431 "madd.s %[temp2], %[temp2], %[temp16], %[temp17] \n\t"
yading@10 432 "lwc1 %[temp10], 4108(%[v0]) \n\t"
yading@10 433 "madd.s %[temp3], %[temp3], %[temp18], %[temp19] \n\t"
yading@10 434 "lwc1 %[temp11], 2060(%[s0]) \n\t"
yading@10 435 "lwc1 %[temp12], 4864(%[v0]) \n\t"
yading@10 436 "lwc1 %[temp13], 2304(%[s0]) \n\t"
yading@10 437 "lwc1 %[temp14], 4868(%[v0]) \n\t"
yading@10 438 "lwc1 %[temp15], 2308(%[s0]) \n\t"
yading@10 439 "madd.s %[temp0], %[temp0], %[temp4], %[temp5] \n\t"
yading@10 440 "lwc1 %[temp16], 4872(%[v0]) \n\t"
yading@10 441 "madd.s %[temp1], %[temp1], %[temp6], %[temp7] \n\t"
yading@10 442 "lwc1 %[temp17], 2312(%[s0]) \n\t"
yading@10 443 "madd.s %[temp2], %[temp2], %[temp8], %[temp9] \n\t"
yading@10 444 "lwc1 %[temp18], 4876(%[v0]) \n\t"
yading@10 445 "madd.s %[temp3], %[temp3], %[temp10], %[temp11] \n\t"
yading@10 446 "lwc1 %[temp19], 2316(%[s0]) \n\t"
yading@10 447 "madd.s %[temp0], %[temp0], %[temp12], %[temp13] \n\t"
yading@10 448 "addiu %[dst], %[dst], 16 \n\t"
yading@10 449 "madd.s %[temp1], %[temp1], %[temp14], %[temp15] \n\t"
yading@10 450 "madd.s %[temp2], %[temp2], %[temp16], %[temp17] \n\t"
yading@10 451 "madd.s %[temp3], %[temp3], %[temp18], %[temp19] \n\t"
yading@10 452 "swc1 %[temp0], -16(%[dst]) \n\t"
yading@10 453 "swc1 %[temp1], -12(%[dst]) \n\t"
yading@10 454 "swc1 %[temp2], -8(%[dst]) \n\t"
yading@10 455 "swc1 %[temp3], -4(%[dst]) \n\t"
yading@10 456 ".set pop \n\t"
yading@10 457
yading@10 458 : [dst]"+r"(dst), [v0]"+r"(vv0), [s0]"+r"(s0),
yading@10 459 [temp0]"=&f"(temp0), [temp1]"=&f"(temp1), [temp2]"=&f"(temp2),
yading@10 460 [temp3]"=&f"(temp3), [temp4]"=&f"(temp4), [temp5]"=&f"(temp5),
yading@10 461 [temp6]"=&f"(temp6), [temp7]"=&f"(temp7), [temp8]"=&f"(temp8),
yading@10 462 [temp9]"=&f"(temp9), [temp10]"=&f"(temp10), [temp11]"=&f"(temp11),
yading@10 463 [temp12]"=&f"(temp12), [temp13]"=&f"(temp13), [temp14]"=&f"(temp14),
yading@10 464 [temp15]"=&f"(temp15), [temp16]"=&f"(temp16), [temp17]"=&f"(temp17),
yading@10 465 [temp18]"=&f"(temp18), [temp19]"=&f"(temp19)
yading@10 466 : [v0_end]"r"(v0_end)
yading@10 467 : "memory"
yading@10 468 );
yading@10 469 }
yading@10 470 else
yading@10 471 {
yading@10 472 fdsp->vector_fmul (out, v , sbr_qmf_window , 64 >> div);
yading@10 473 fdsp->vector_fmul_add(out, v + ( 192 >> div), sbr_qmf_window + ( 64 >> div), out , 64 >> div);
yading@10 474 fdsp->vector_fmul_add(out, v + ( 256 >> div), sbr_qmf_window + (128 >> div), out , 64 >> div);
yading@10 475 fdsp->vector_fmul_add(out, v + ( 448 >> div), sbr_qmf_window + (192 >> div), out , 64 >> div);
yading@10 476 fdsp->vector_fmul_add(out, v + ( 512 >> div), sbr_qmf_window + (256 >> div), out , 64 >> div);
yading@10 477 fdsp->vector_fmul_add(out, v + ( 704 >> div), sbr_qmf_window + (320 >> div), out , 64 >> div);
yading@10 478 fdsp->vector_fmul_add(out, v + ( 768 >> div), sbr_qmf_window + (384 >> div), out , 64 >> div);
yading@10 479 fdsp->vector_fmul_add(out, v + ( 960 >> div), sbr_qmf_window + (448 >> div), out , 64 >> div);
yading@10 480 fdsp->vector_fmul_add(out, v + (1024 >> div), sbr_qmf_window + (512 >> div), out , 64 >> div);
yading@10 481 fdsp->vector_fmul_add(out, v + (1216 >> div), sbr_qmf_window + (576 >> div), out , 64 >> div);
yading@10 482 out += 64 >> div;
yading@10 483 }
yading@10 484 }
yading@10 485 }
yading@10 486
yading@10 487 #define sbr_qmf_analysis sbr_qmf_analysis_mips
yading@10 488 #define sbr_qmf_synthesis sbr_qmf_synthesis_mips
yading@10 489
yading@10 490 #endif /* HAVE_MIPSFPU */
yading@10 491 #endif /* HAVE_INLINE_ASM */
yading@10 492
yading@10 493 #endif /* AVCODEC_MIPS_AACSBR_FLOAT_H */