| OLD | NEW |
|---|
| (Empty) |
| 1 /* | |
| 2 * Copyright (c) 2014 The WebRTC project authors. All Rights Reserved. | |
| 3 * | |
| 4 * Use of this source code is governed by a BSD-style license | |
| 5 * that can be found in the LICENSE file in the root of the source | |
| 6 * tree. An additional intellectual property rights grant can be found | |
| 7 * in the file PATENTS. All contributing project authors may | |
| 8 * be found in the AUTHORS file in the root of the source tree. | |
| 9 */ | |
| 10 | |
| 11 #include "webrtc/modules/audio_processing/aec/aec_rdft.h" | |
| 12 #include "webrtc/typedefs.h" | |
| 13 | |
| 14 static void bitrv2_128_mips(float* a) { | |
| 15 // n is 128 | |
| 16 float xr, xi, yr, yi; | |
| 17 | |
| 18 xr = a[8]; | |
| 19 xi = a[9]; | |
| 20 yr = a[16]; | |
| 21 yi = a[17]; | |
| 22 a[8] = yr; | |
| 23 a[9] = yi; | |
| 24 a[16] = xr; | |
| 25 a[17] = xi; | |
| 26 | |
| 27 xr = a[64]; | |
| 28 xi = a[65]; | |
| 29 yr = a[2]; | |
| 30 yi = a[3]; | |
| 31 a[64] = yr; | |
| 32 a[65] = yi; | |
| 33 a[2] = xr; | |
| 34 a[3] = xi; | |
| 35 | |
| 36 xr = a[72]; | |
| 37 xi = a[73]; | |
| 38 yr = a[18]; | |
| 39 yi = a[19]; | |
| 40 a[72] = yr; | |
| 41 a[73] = yi; | |
| 42 a[18] = xr; | |
| 43 a[19] = xi; | |
| 44 | |
| 45 xr = a[80]; | |
| 46 xi = a[81]; | |
| 47 yr = a[10]; | |
| 48 yi = a[11]; | |
| 49 a[80] = yr; | |
| 50 a[81] = yi; | |
| 51 a[10] = xr; | |
| 52 a[11] = xi; | |
| 53 | |
| 54 xr = a[88]; | |
| 55 xi = a[89]; | |
| 56 yr = a[26]; | |
| 57 yi = a[27]; | |
| 58 a[88] = yr; | |
| 59 a[89] = yi; | |
| 60 a[26] = xr; | |
| 61 a[27] = xi; | |
| 62 | |
| 63 xr = a[74]; | |
| 64 xi = a[75]; | |
| 65 yr = a[82]; | |
| 66 yi = a[83]; | |
| 67 a[74] = yr; | |
| 68 a[75] = yi; | |
| 69 a[82] = xr; | |
| 70 a[83] = xi; | |
| 71 | |
| 72 xr = a[32]; | |
| 73 xi = a[33]; | |
| 74 yr = a[4]; | |
| 75 yi = a[5]; | |
| 76 a[32] = yr; | |
| 77 a[33] = yi; | |
| 78 a[4] = xr; | |
| 79 a[5] = xi; | |
| 80 | |
| 81 xr = a[40]; | |
| 82 xi = a[41]; | |
| 83 yr = a[20]; | |
| 84 yi = a[21]; | |
| 85 a[40] = yr; | |
| 86 a[41] = yi; | |
| 87 a[20] = xr; | |
| 88 a[21] = xi; | |
| 89 | |
| 90 xr = a[48]; | |
| 91 xi = a[49]; | |
| 92 yr = a[12]; | |
| 93 yi = a[13]; | |
| 94 a[48] = yr; | |
| 95 a[49] = yi; | |
| 96 a[12] = xr; | |
| 97 a[13] = xi; | |
| 98 | |
| 99 xr = a[56]; | |
| 100 xi = a[57]; | |
| 101 yr = a[28]; | |
| 102 yi = a[29]; | |
| 103 a[56] = yr; | |
| 104 a[57] = yi; | |
| 105 a[28] = xr; | |
| 106 a[29] = xi; | |
| 107 | |
| 108 xr = a[34]; | |
| 109 xi = a[35]; | |
| 110 yr = a[68]; | |
| 111 yi = a[69]; | |
| 112 a[34] = yr; | |
| 113 a[35] = yi; | |
| 114 a[68] = xr; | |
| 115 a[69] = xi; | |
| 116 | |
| 117 xr = a[42]; | |
| 118 xi = a[43]; | |
| 119 yr = a[84]; | |
| 120 yi = a[85]; | |
| 121 a[42] = yr; | |
| 122 a[43] = yi; | |
| 123 a[84] = xr; | |
| 124 a[85] = xi; | |
| 125 | |
| 126 xr = a[50]; | |
| 127 xi = a[51]; | |
| 128 yr = a[76]; | |
| 129 yi = a[77]; | |
| 130 a[50] = yr; | |
| 131 a[51] = yi; | |
| 132 a[76] = xr; | |
| 133 a[77] = xi; | |
| 134 | |
| 135 xr = a[58]; | |
| 136 xi = a[59]; | |
| 137 yr = a[92]; | |
| 138 yi = a[93]; | |
| 139 a[58] = yr; | |
| 140 a[59] = yi; | |
| 141 a[92] = xr; | |
| 142 a[93] = xi; | |
| 143 | |
| 144 xr = a[44]; | |
| 145 xi = a[45]; | |
| 146 yr = a[52]; | |
| 147 yi = a[53]; | |
| 148 a[44] = yr; | |
| 149 a[45] = yi; | |
| 150 a[52] = xr; | |
| 151 a[53] = xi; | |
| 152 | |
| 153 xr = a[96]; | |
| 154 xi = a[97]; | |
| 155 yr = a[6]; | |
| 156 yi = a[7]; | |
| 157 a[96] = yr; | |
| 158 a[97] = yi; | |
| 159 a[6] = xr; | |
| 160 a[7] = xi; | |
| 161 | |
| 162 xr = a[104]; | |
| 163 xi = a[105]; | |
| 164 yr = a[22]; | |
| 165 yi = a[23]; | |
| 166 a[104] = yr; | |
| 167 a[105] = yi; | |
| 168 a[22] = xr; | |
| 169 a[23] = xi; | |
| 170 | |
| 171 xr = a[112]; | |
| 172 xi = a[113]; | |
| 173 yr = a[14]; | |
| 174 yi = a[15]; | |
| 175 a[112] = yr; | |
| 176 a[113] = yi; | |
| 177 a[14] = xr; | |
| 178 a[15] = xi; | |
| 179 | |
| 180 xr = a[120]; | |
| 181 xi = a[121]; | |
| 182 yr = a[30]; | |
| 183 yi = a[31]; | |
| 184 a[120] = yr; | |
| 185 a[121] = yi; | |
| 186 a[30] = xr; | |
| 187 a[31] = xi; | |
| 188 | |
| 189 xr = a[98]; | |
| 190 xi = a[99]; | |
| 191 yr = a[70]; | |
| 192 yi = a[71]; | |
| 193 a[98] = yr; | |
| 194 a[99] = yi; | |
| 195 a[70] = xr; | |
| 196 a[71] = xi; | |
| 197 | |
| 198 xr = a[106]; | |
| 199 xi = a[107]; | |
| 200 yr = a[86]; | |
| 201 yi = a[87]; | |
| 202 a[106] = yr; | |
| 203 a[107] = yi; | |
| 204 a[86] = xr; | |
| 205 a[87] = xi; | |
| 206 | |
| 207 xr = a[114]; | |
| 208 xi = a[115]; | |
| 209 yr = a[78]; | |
| 210 yi = a[79]; | |
| 211 a[114] = yr; | |
| 212 a[115] = yi; | |
| 213 a[78] = xr; | |
| 214 a[79] = xi; | |
| 215 | |
| 216 xr = a[122]; | |
| 217 xi = a[123]; | |
| 218 yr = a[94]; | |
| 219 yi = a[95]; | |
| 220 a[122] = yr; | |
| 221 a[123] = yi; | |
| 222 a[94] = xr; | |
| 223 a[95] = xi; | |
| 224 | |
| 225 xr = a[100]; | |
| 226 xi = a[101]; | |
| 227 yr = a[38]; | |
| 228 yi = a[39]; | |
| 229 a[100] = yr; | |
| 230 a[101] = yi; | |
| 231 a[38] = xr; | |
| 232 a[39] = xi; | |
| 233 | |
| 234 xr = a[108]; | |
| 235 xi = a[109]; | |
| 236 yr = a[54]; | |
| 237 yi = a[55]; | |
| 238 a[108] = yr; | |
| 239 a[109] = yi; | |
| 240 a[54] = xr; | |
| 241 a[55] = xi; | |
| 242 | |
| 243 xr = a[116]; | |
| 244 xi = a[117]; | |
| 245 yr = a[46]; | |
| 246 yi = a[47]; | |
| 247 a[116] = yr; | |
| 248 a[117] = yi; | |
| 249 a[46] = xr; | |
| 250 a[47] = xi; | |
| 251 | |
| 252 xr = a[124]; | |
| 253 xi = a[125]; | |
| 254 yr = a[62]; | |
| 255 yi = a[63]; | |
| 256 a[124] = yr; | |
| 257 a[125] = yi; | |
| 258 a[62] = xr; | |
| 259 a[63] = xi; | |
| 260 | |
| 261 xr = a[110]; | |
| 262 xi = a[111]; | |
| 263 yr = a[118]; | |
| 264 yi = a[119]; | |
| 265 a[110] = yr; | |
| 266 a[111] = yi; | |
| 267 a[118] = xr; | |
| 268 a[119] = xi; | |
| 269 } | |
| 270 | |
| 271 static void cft1st_128_mips(float* a) { | |
| 272 float f0, f1, f2, f3, f4, f5, f6, f7, f8, f9, f10, f11, f12, f13, f14; | |
| 273 int a_ptr, p1_rdft, p2_rdft, count; | |
| 274 const float* first = rdft_wk3ri_first; | |
| 275 const float* second = rdft_wk3ri_second; | |
| 276 | |
| 277 __asm __volatile ( | |
| 278 ".set push \n\t" | |
| 279 ".set noreorder \n\t" | |
| 280 // first 8 | |
| 281 "lwc1 %[f0], 0(%[a]) \n\t" | |
| 282 "lwc1 %[f1], 4(%[a]) \n\t" | |
| 283 "lwc1 %[f2], 8(%[a]) \n\t" | |
| 284 "lwc1 %[f3], 12(%[a]) \n\t" | |
| 285 "lwc1 %[f4], 16(%[a]) \n\t" | |
| 286 "lwc1 %[f5], 20(%[a]) \n\t" | |
| 287 "lwc1 %[f6], 24(%[a]) \n\t" | |
| 288 "lwc1 %[f7], 28(%[a]) \n\t" | |
| 289 "add.s %[f8], %[f0], %[f2] \n\t" | |
| 290 "sub.s %[f0], %[f0], %[f2] \n\t" | |
| 291 "add.s %[f2], %[f4], %[f6] \n\t" | |
| 292 "sub.s %[f4], %[f4], %[f6] \n\t" | |
| 293 "add.s %[f6], %[f1], %[f3] \n\t" | |
| 294 "sub.s %[f1], %[f1], %[f3] \n\t" | |
| 295 "add.s %[f3], %[f5], %[f7] \n\t" | |
| 296 "sub.s %[f5], %[f5], %[f7] \n\t" | |
| 297 "add.s %[f7], %[f8], %[f2] \n\t" | |
| 298 "sub.s %[f8], %[f8], %[f2] \n\t" | |
| 299 "sub.s %[f2], %[f1], %[f4] \n\t" | |
| 300 "add.s %[f1], %[f1], %[f4] \n\t" | |
| 301 "add.s %[f4], %[f6], %[f3] \n\t" | |
| 302 "sub.s %[f6], %[f6], %[f3] \n\t" | |
| 303 "sub.s %[f3], %[f0], %[f5] \n\t" | |
| 304 "add.s %[f0], %[f0], %[f5] \n\t" | |
| 305 "swc1 %[f7], 0(%[a]) \n\t" | |
| 306 "swc1 %[f8], 16(%[a]) \n\t" | |
| 307 "swc1 %[f2], 28(%[a]) \n\t" | |
| 308 "swc1 %[f1], 12(%[a]) \n\t" | |
| 309 "swc1 %[f4], 4(%[a]) \n\t" | |
| 310 "swc1 %[f6], 20(%[a]) \n\t" | |
| 311 "swc1 %[f3], 8(%[a]) \n\t" | |
| 312 "swc1 %[f0], 24(%[a]) \n\t" | |
| 313 // second 8 | |
| 314 "lwc1 %[f0], 32(%[a]) \n\t" | |
| 315 "lwc1 %[f1], 36(%[a]) \n\t" | |
| 316 "lwc1 %[f2], 40(%[a]) \n\t" | |
| 317 "lwc1 %[f3], 44(%[a]) \n\t" | |
| 318 "lwc1 %[f4], 48(%[a]) \n\t" | |
| 319 "lwc1 %[f5], 52(%[a]) \n\t" | |
| 320 "lwc1 %[f6], 56(%[a]) \n\t" | |
| 321 "lwc1 %[f7], 60(%[a]) \n\t" | |
| 322 "add.s %[f8], %[f4], %[f6] \n\t" | |
| 323 "sub.s %[f4], %[f4], %[f6] \n\t" | |
| 324 "add.s %[f6], %[f1], %[f3] \n\t" | |
| 325 "sub.s %[f1], %[f1], %[f3] \n\t" | |
| 326 "add.s %[f3], %[f0], %[f2] \n\t" | |
| 327 "sub.s %[f0], %[f0], %[f2] \n\t" | |
| 328 "add.s %[f2], %[f5], %[f7] \n\t" | |
| 329 "sub.s %[f5], %[f5], %[f7] \n\t" | |
| 330 "add.s %[f7], %[f4], %[f1] \n\t" | |
| 331 "sub.s %[f4], %[f4], %[f1] \n\t" | |
| 332 "add.s %[f1], %[f3], %[f8] \n\t" | |
| 333 "sub.s %[f3], %[f3], %[f8] \n\t" | |
| 334 "sub.s %[f8], %[f0], %[f5] \n\t" | |
| 335 "add.s %[f0], %[f0], %[f5] \n\t" | |
| 336 "add.s %[f5], %[f6], %[f2] \n\t" | |
| 337 "sub.s %[f6], %[f2], %[f6] \n\t" | |
| 338 "lwc1 %[f9], 8(%[rdft_w]) \n\t" | |
| 339 "sub.s %[f2], %[f8], %[f7] \n\t" | |
| 340 "add.s %[f8], %[f8], %[f7] \n\t" | |
| 341 "sub.s %[f7], %[f4], %[f0] \n\t" | |
| 342 "add.s %[f4], %[f4], %[f0] \n\t" | |
| 343 // prepare for loop | |
| 344 "addiu %[a_ptr], %[a], 64 \n\t" | |
| 345 "addiu %[p1_rdft], %[rdft_w], 8 \n\t" | |
| 346 "addiu %[p2_rdft], %[rdft_w], 16 \n\t" | |
| 347 "addiu %[count], $zero, 7 \n\t" | |
| 348 // finish second 8 | |
| 349 "mul.s %[f2], %[f9], %[f2] \n\t" | |
| 350 "mul.s %[f8], %[f9], %[f8] \n\t" | |
| 351 "mul.s %[f7], %[f9], %[f7] \n\t" | |
| 352 "mul.s %[f4], %[f9], %[f4] \n\t" | |
| 353 "swc1 %[f1], 32(%[a]) \n\t" | |
| 354 "swc1 %[f3], 52(%[a]) \n\t" | |
| 355 "swc1 %[f5], 36(%[a]) \n\t" | |
| 356 "swc1 %[f6], 48(%[a]) \n\t" | |
| 357 "swc1 %[f2], 40(%[a]) \n\t" | |
| 358 "swc1 %[f8], 44(%[a]) \n\t" | |
| 359 "swc1 %[f7], 56(%[a]) \n\t" | |
| 360 "swc1 %[f4], 60(%[a]) \n\t" | |
| 361 // loop | |
| 362 "1: \n\t" | |
| 363 "lwc1 %[f0], 0(%[a_ptr]) \n\t" | |
| 364 "lwc1 %[f1], 4(%[a_ptr]) \n\t" | |
| 365 "lwc1 %[f2], 8(%[a_ptr]) \n\t" | |
| 366 "lwc1 %[f3], 12(%[a_ptr]) \n\t" | |
| 367 "lwc1 %[f4], 16(%[a_ptr]) \n\t" | |
| 368 "lwc1 %[f5], 20(%[a_ptr]) \n\t" | |
| 369 "lwc1 %[f6], 24(%[a_ptr]) \n\t" | |
| 370 "lwc1 %[f7], 28(%[a_ptr]) \n\t" | |
| 371 "add.s %[f8], %[f0], %[f2] \n\t" | |
| 372 "sub.s %[f0], %[f0], %[f2] \n\t" | |
| 373 "add.s %[f2], %[f4], %[f6] \n\t" | |
| 374 "sub.s %[f4], %[f4], %[f6] \n\t" | |
| 375 "add.s %[f6], %[f1], %[f3] \n\t" | |
| 376 "sub.s %[f1], %[f1], %[f3] \n\t" | |
| 377 "add.s %[f3], %[f5], %[f7] \n\t" | |
| 378 "sub.s %[f5], %[f5], %[f7] \n\t" | |
| 379 "lwc1 %[f10], 4(%[p1_rdft]) \n\t" | |
| 380 "lwc1 %[f11], 0(%[p2_rdft]) \n\t" | |
| 381 "lwc1 %[f12], 4(%[p2_rdft]) \n\t" | |
| 382 "lwc1 %[f13], 8(%[first]) \n\t" | |
| 383 "lwc1 %[f14], 12(%[first]) \n\t" | |
| 384 "add.s %[f7], %[f8], %[f2] \n\t" | |
| 385 "sub.s %[f8], %[f8], %[f2] \n\t" | |
| 386 "add.s %[f2], %[f6], %[f3] \n\t" | |
| 387 "sub.s %[f6], %[f6], %[f3] \n\t" | |
| 388 "add.s %[f3], %[f0], %[f5] \n\t" | |
| 389 "sub.s %[f0], %[f0], %[f5] \n\t" | |
| 390 "add.s %[f5], %[f1], %[f4] \n\t" | |
| 391 "sub.s %[f1], %[f1], %[f4] \n\t" | |
| 392 "swc1 %[f7], 0(%[a_ptr]) \n\t" | |
| 393 "swc1 %[f2], 4(%[a_ptr]) \n\t" | |
| 394 "mul.s %[f4], %[f9], %[f8] \n\t" | |
| 395 #if defined(MIPS32_R2_LE) | |
| 396 "mul.s %[f8], %[f10], %[f8] \n\t" | |
| 397 "mul.s %[f7], %[f11], %[f0] \n\t" | |
| 398 "mul.s %[f0], %[f12], %[f0] \n\t" | |
| 399 "mul.s %[f2], %[f13], %[f3] \n\t" | |
| 400 "mul.s %[f3], %[f14], %[f3] \n\t" | |
| 401 "nmsub.s %[f4], %[f4], %[f10], %[f6] \n\t" | |
| 402 "madd.s %[f8], %[f8], %[f9], %[f6] \n\t" | |
| 403 "nmsub.s %[f7], %[f7], %[f12], %[f5] \n\t" | |
| 404 "madd.s %[f0], %[f0], %[f11], %[f5] \n\t" | |
| 405 "nmsub.s %[f2], %[f2], %[f14], %[f1] \n\t" | |
| 406 "madd.s %[f3], %[f3], %[f13], %[f1] \n\t" | |
| 407 #else | |
| 408 "mul.s %[f7], %[f10], %[f6] \n\t" | |
| 409 "mul.s %[f6], %[f9], %[f6] \n\t" | |
| 410 "mul.s %[f8], %[f10], %[f8] \n\t" | |
| 411 "mul.s %[f2], %[f11], %[f0] \n\t" | |
| 412 "mul.s %[f11], %[f11], %[f5] \n\t" | |
| 413 "mul.s %[f5], %[f12], %[f5] \n\t" | |
| 414 "mul.s %[f0], %[f12], %[f0] \n\t" | |
| 415 "mul.s %[f12], %[f13], %[f3] \n\t" | |
| 416 "mul.s %[f13], %[f13], %[f1] \n\t" | |
| 417 "mul.s %[f1], %[f14], %[f1] \n\t" | |
| 418 "mul.s %[f3], %[f14], %[f3] \n\t" | |
| 419 "sub.s %[f4], %[f4], %[f7] \n\t" | |
| 420 "add.s %[f8], %[f6], %[f8] \n\t" | |
| 421 "sub.s %[f7], %[f2], %[f5] \n\t" | |
| 422 "add.s %[f0], %[f11], %[f0] \n\t" | |
| 423 "sub.s %[f2], %[f12], %[f1] \n\t" | |
| 424 "add.s %[f3], %[f13], %[f3] \n\t" | |
| 425 #endif | |
| 426 "swc1 %[f4], 16(%[a_ptr]) \n\t" | |
| 427 "swc1 %[f8], 20(%[a_ptr]) \n\t" | |
| 428 "swc1 %[f7], 8(%[a_ptr]) \n\t" | |
| 429 "swc1 %[f0], 12(%[a_ptr]) \n\t" | |
| 430 "swc1 %[f2], 24(%[a_ptr]) \n\t" | |
| 431 "swc1 %[f3], 28(%[a_ptr]) \n\t" | |
| 432 "lwc1 %[f0], 32(%[a_ptr]) \n\t" | |
| 433 "lwc1 %[f1], 36(%[a_ptr]) \n\t" | |
| 434 "lwc1 %[f2], 40(%[a_ptr]) \n\t" | |
| 435 "lwc1 %[f3], 44(%[a_ptr]) \n\t" | |
| 436 "lwc1 %[f4], 48(%[a_ptr]) \n\t" | |
| 437 "lwc1 %[f5], 52(%[a_ptr]) \n\t" | |
| 438 "lwc1 %[f6], 56(%[a_ptr]) \n\t" | |
| 439 "lwc1 %[f7], 60(%[a_ptr]) \n\t" | |
| 440 "add.s %[f8], %[f0], %[f2] \n\t" | |
| 441 "sub.s %[f0], %[f0], %[f2] \n\t" | |
| 442 "add.s %[f2], %[f4], %[f6] \n\t" | |
| 443 "sub.s %[f4], %[f4], %[f6] \n\t" | |
| 444 "add.s %[f6], %[f1], %[f3] \n\t" | |
| 445 "sub.s %[f1], %[f1], %[f3] \n\t" | |
| 446 "add.s %[f3], %[f5], %[f7] \n\t" | |
| 447 "sub.s %[f5], %[f5], %[f7] \n\t" | |
| 448 "lwc1 %[f11], 8(%[p2_rdft]) \n\t" | |
| 449 "lwc1 %[f12], 12(%[p2_rdft]) \n\t" | |
| 450 "lwc1 %[f13], 8(%[second]) \n\t" | |
| 451 "lwc1 %[f14], 12(%[second]) \n\t" | |
| 452 "add.s %[f7], %[f8], %[f2] \n\t" | |
| 453 "sub.s %[f8], %[f2], %[f8] \n\t" | |
| 454 "add.s %[f2], %[f6], %[f3] \n\t" | |
| 455 "sub.s %[f6], %[f3], %[f6] \n\t" | |
| 456 "add.s %[f3], %[f0], %[f5] \n\t" | |
| 457 "sub.s %[f0], %[f0], %[f5] \n\t" | |
| 458 "add.s %[f5], %[f1], %[f4] \n\t" | |
| 459 "sub.s %[f1], %[f1], %[f4] \n\t" | |
| 460 "swc1 %[f7], 32(%[a_ptr]) \n\t" | |
| 461 "swc1 %[f2], 36(%[a_ptr]) \n\t" | |
| 462 "mul.s %[f4], %[f10], %[f8] \n\t" | |
| 463 #if defined(MIPS32_R2_LE) | |
| 464 "mul.s %[f10], %[f10], %[f6] \n\t" | |
| 465 "mul.s %[f7], %[f11], %[f0] \n\t" | |
| 466 "mul.s %[f11], %[f11], %[f5] \n\t" | |
| 467 "mul.s %[f2], %[f13], %[f3] \n\t" | |
| 468 "mul.s %[f13], %[f13], %[f1] \n\t" | |
| 469 "madd.s %[f4], %[f4], %[f9], %[f6] \n\t" | |
| 470 "nmsub.s %[f10], %[f10], %[f9], %[f8] \n\t" | |
| 471 "nmsub.s %[f7], %[f7], %[f12], %[f5] \n\t" | |
| 472 "madd.s %[f11], %[f11], %[f12], %[f0] \n\t" | |
| 473 "nmsub.s %[f2], %[f2], %[f14], %[f1] \n\t" | |
| 474 "madd.s %[f13], %[f13], %[f14], %[f3] \n\t" | |
| 475 #else | |
| 476 "mul.s %[f2], %[f9], %[f6] \n\t" | |
| 477 "mul.s %[f10], %[f10], %[f6] \n\t" | |
| 478 "mul.s %[f9], %[f9], %[f8] \n\t" | |
| 479 "mul.s %[f7], %[f11], %[f0] \n\t" | |
| 480 "mul.s %[f8], %[f12], %[f5] \n\t" | |
| 481 "mul.s %[f11], %[f11], %[f5] \n\t" | |
| 482 "mul.s %[f12], %[f12], %[f0] \n\t" | |
| 483 "mul.s %[f5], %[f13], %[f3] \n\t" | |
| 484 "mul.s %[f0], %[f14], %[f1] \n\t" | |
| 485 "mul.s %[f13], %[f13], %[f1] \n\t" | |
| 486 "mul.s %[f14], %[f14], %[f3] \n\t" | |
| 487 "add.s %[f4], %[f4], %[f2] \n\t" | |
| 488 "sub.s %[f10], %[f10], %[f9] \n\t" | |
| 489 "sub.s %[f7], %[f7], %[f8] \n\t" | |
| 490 "add.s %[f11], %[f11], %[f12] \n\t" | |
| 491 "sub.s %[f2], %[f5], %[f0] \n\t" | |
| 492 "add.s %[f13], %[f13], %[f14] \n\t" | |
| 493 #endif | |
| 494 "swc1 %[f4], 48(%[a_ptr]) \n\t" | |
| 495 "swc1 %[f10], 52(%[a_ptr]) \n\t" | |
| 496 "swc1 %[f7], 40(%[a_ptr]) \n\t" | |
| 497 "swc1 %[f11], 44(%[a_ptr]) \n\t" | |
| 498 "swc1 %[f2], 56(%[a_ptr]) \n\t" | |
| 499 "swc1 %[f13], 60(%[a_ptr]) \n\t" | |
| 500 "addiu %[count], %[count], -1 \n\t" | |
| 501 "lwc1 %[f9], 8(%[p1_rdft]) \n\t" | |
| 502 "addiu %[a_ptr], %[a_ptr], 64 \n\t" | |
| 503 "addiu %[p1_rdft], %[p1_rdft], 8 \n\t" | |
| 504 "addiu %[p2_rdft], %[p2_rdft], 16 \n\t" | |
| 505 "addiu %[first], %[first], 8 \n\t" | |
| 506 "bgtz %[count], 1b \n\t" | |
| 507 " addiu %[second], %[second], 8 \n\t" | |
| 508 ".set pop \n\t" | |
| 509 : [f0] "=&f" (f0), [f1] "=&f" (f1), [f2] "=&f" (f2), [f3] "=&f" (f3), | |
| 510 [f4] "=&f" (f4), [f5] "=&f" (f5), [f6] "=&f" (f6), [f7] "=&f" (f7), | |
| 511 [f8] "=&f" (f8), [f9] "=&f" (f9), [f10] "=&f" (f10), [f11] "=&f" (f11), | |
| 512 [f12] "=&f" (f12), [f13] "=&f" (f13), [f14] "=&f" (f14), | |
| 513 [a_ptr] "=&r" (a_ptr), [p1_rdft] "=&r" (p1_rdft), [first] "+r" (first), | |
| 514 [p2_rdft] "=&r" (p2_rdft), [count] "=&r" (count), [second] "+r" (second) | |
| 515 : [a] "r" (a), [rdft_w] "r" (rdft_w) | |
| 516 : "memory" | |
| 517 ); | |
| 518 } | |
| 519 | |
| 520 static void cftmdl_128_mips(float* a) { | |
| 521 float f0, f1, f2, f3, f4, f5, f6, f7, f8, f9, f10, f11, f12, f13, f14; | |
| 522 int tmp_a, count; | |
| 523 __asm __volatile ( | |
| 524 ".set push \n\t" | |
| 525 ".set noreorder \n\t" | |
| 526 "addiu %[tmp_a], %[a], 0 \n\t" | |
| 527 "addiu %[count], $zero, 4 \n\t" | |
| 528 "1: \n\t" | |
| 529 "addiu %[count], %[count], -1 \n\t" | |
| 530 "lwc1 %[f0], 0(%[tmp_a]) \n\t" | |
| 531 "lwc1 %[f2], 32(%[tmp_a]) \n\t" | |
| 532 "lwc1 %[f4], 64(%[tmp_a]) \n\t" | |
| 533 "lwc1 %[f6], 96(%[tmp_a]) \n\t" | |
| 534 "lwc1 %[f1], 4(%[tmp_a]) \n\t" | |
| 535 "lwc1 %[f3], 36(%[tmp_a]) \n\t" | |
| 536 "lwc1 %[f5], 68(%[tmp_a]) \n\t" | |
| 537 "lwc1 %[f7], 100(%[tmp_a]) \n\t" | |
| 538 "add.s %[f8], %[f0], %[f2] \n\t" | |
| 539 "sub.s %[f0], %[f0], %[f2] \n\t" | |
| 540 "add.s %[f2], %[f4], %[f6] \n\t" | |
| 541 "sub.s %[f4], %[f4], %[f6] \n\t" | |
| 542 "add.s %[f6], %[f1], %[f3] \n\t" | |
| 543 "sub.s %[f1], %[f1], %[f3] \n\t" | |
| 544 "add.s %[f3], %[f5], %[f7] \n\t" | |
| 545 "sub.s %[f5], %[f5], %[f7] \n\t" | |
| 546 "add.s %[f7], %[f8], %[f2] \n\t" | |
| 547 "sub.s %[f8], %[f8], %[f2] \n\t" | |
| 548 "add.s %[f2], %[f1], %[f4] \n\t" | |
| 549 "sub.s %[f1], %[f1], %[f4] \n\t" | |
| 550 "add.s %[f4], %[f6], %[f3] \n\t" | |
| 551 "sub.s %[f6], %[f6], %[f3] \n\t" | |
| 552 "sub.s %[f3], %[f0], %[f5] \n\t" | |
| 553 "add.s %[f0], %[f0], %[f5] \n\t" | |
| 554 "swc1 %[f7], 0(%[tmp_a]) \n\t" | |
| 555 "swc1 %[f8], 64(%[tmp_a]) \n\t" | |
| 556 "swc1 %[f2], 36(%[tmp_a]) \n\t" | |
| 557 "swc1 %[f1], 100(%[tmp_a]) \n\t" | |
| 558 "swc1 %[f4], 4(%[tmp_a]) \n\t" | |
| 559 "swc1 %[f6], 68(%[tmp_a]) \n\t" | |
| 560 "swc1 %[f3], 32(%[tmp_a]) \n\t" | |
| 561 "swc1 %[f0], 96(%[tmp_a]) \n\t" | |
| 562 "bgtz %[count], 1b \n\t" | |
| 563 " addiu %[tmp_a], %[tmp_a], 8 \n\t" | |
| 564 ".set pop \n\t" | |
| 565 : [f0] "=&f" (f0), [f1] "=&f" (f1), [f2] "=&f" (f2), [f3] "=&f" (f3), | |
| 566 [f4] "=&f" (f4), [f5] "=&f" (f5), [f6] "=&f" (f6), [f7] "=&f" (f7), | |
| 567 [f8] "=&f" (f8), [tmp_a] "=&r" (tmp_a), [count] "=&r" (count) | |
| 568 : [a] "r" (a) | |
| 569 : "memory" | |
| 570 ); | |
| 571 f9 = rdft_w[2]; | |
| 572 __asm __volatile ( | |
| 573 ".set push \n\t" | |
| 574 ".set noreorder \n\t" | |
| 575 "addiu %[tmp_a], %[a], 128 \n\t" | |
| 576 "addiu %[count], $zero, 4 \n\t" | |
| 577 "1: \n\t" | |
| 578 "addiu %[count], %[count], -1 \n\t" | |
| 579 "lwc1 %[f0], 0(%[tmp_a]) \n\t" | |
| 580 "lwc1 %[f2], 32(%[tmp_a]) \n\t" | |
| 581 "lwc1 %[f5], 68(%[tmp_a]) \n\t" | |
| 582 "lwc1 %[f7], 100(%[tmp_a]) \n\t" | |
| 583 "lwc1 %[f1], 4(%[tmp_a]) \n\t" | |
| 584 "lwc1 %[f3], 36(%[tmp_a]) \n\t" | |
| 585 "lwc1 %[f4], 64(%[tmp_a]) \n\t" | |
| 586 "lwc1 %[f6], 96(%[tmp_a]) \n\t" | |
| 587 "sub.s %[f8], %[f0], %[f2] \n\t" | |
| 588 "add.s %[f0], %[f0], %[f2] \n\t" | |
| 589 "sub.s %[f2], %[f5], %[f7] \n\t" | |
| 590 "add.s %[f5], %[f5], %[f7] \n\t" | |
| 591 "sub.s %[f7], %[f1], %[f3] \n\t" | |
| 592 "add.s %[f1], %[f1], %[f3] \n\t" | |
| 593 "sub.s %[f3], %[f4], %[f6] \n\t" | |
| 594 "add.s %[f4], %[f4], %[f6] \n\t" | |
| 595 "sub.s %[f6], %[f8], %[f2] \n\t" | |
| 596 "add.s %[f8], %[f8], %[f2] \n\t" | |
| 597 "add.s %[f2], %[f5], %[f1] \n\t" | |
| 598 "sub.s %[f5], %[f5], %[f1] \n\t" | |
| 599 "add.s %[f1], %[f3], %[f7] \n\t" | |
| 600 "sub.s %[f3], %[f3], %[f7] \n\t" | |
| 601 "add.s %[f7], %[f0], %[f4] \n\t" | |
| 602 "sub.s %[f0], %[f0], %[f4] \n\t" | |
| 603 "sub.s %[f4], %[f6], %[f1] \n\t" | |
| 604 "add.s %[f6], %[f6], %[f1] \n\t" | |
| 605 "sub.s %[f1], %[f3], %[f8] \n\t" | |
| 606 "add.s %[f3], %[f3], %[f8] \n\t" | |
| 607 "mul.s %[f4], %[f4], %[f9] \n\t" | |
| 608 "mul.s %[f6], %[f6], %[f9] \n\t" | |
| 609 "mul.s %[f1], %[f1], %[f9] \n\t" | |
| 610 "mul.s %[f3], %[f3], %[f9] \n\t" | |
| 611 "swc1 %[f7], 0(%[tmp_a]) \n\t" | |
| 612 "swc1 %[f2], 4(%[tmp_a]) \n\t" | |
| 613 "swc1 %[f5], 64(%[tmp_a]) \n\t" | |
| 614 "swc1 %[f0], 68(%[tmp_a]) \n\t" | |
| 615 "swc1 %[f4], 32(%[tmp_a]) \n\t" | |
| 616 "swc1 %[f6], 36(%[tmp_a]) \n\t" | |
| 617 "swc1 %[f1], 96(%[tmp_a]) \n\t" | |
| 618 "swc1 %[f3], 100(%[tmp_a]) \n\t" | |
| 619 "bgtz %[count], 1b \n\t" | |
| 620 " addiu %[tmp_a], %[tmp_a], 8 \n\t" | |
| 621 ".set pop \n\t" | |
| 622 : [f0] "=&f" (f0), [f1] "=&f" (f1), [f2] "=&f" (f2), [f3] "=&f" (f3), | |
| 623 [f4] "=&f" (f4), [f5] "=&f" (f5), [f6] "=&f" (f6), [f7] "=&f" (f7), | |
| 624 [f8] "=&f" (f8), [tmp_a] "=&r" (tmp_a), [count] "=&r" (count) | |
| 625 : [a] "r" (a), [f9] "f" (f9) | |
| 626 : "memory" | |
| 627 ); | |
| 628 f10 = rdft_w[3]; | |
| 629 f11 = rdft_w[4]; | |
| 630 f12 = rdft_w[5]; | |
| 631 f13 = rdft_wk3ri_first[2]; | |
| 632 f14 = rdft_wk3ri_first[3]; | |
| 633 | |
| 634 __asm __volatile ( | |
| 635 ".set push \n\t" | |
| 636 ".set noreorder \n\t" | |
| 637 "addiu %[tmp_a], %[a], 256 \n\t" | |
| 638 "addiu %[count], $zero, 4 \n\t" | |
| 639 "1: \n\t" | |
| 640 "addiu %[count], %[count], -1 \n\t" | |
| 641 "lwc1 %[f0], 0(%[tmp_a]) \n\t" | |
| 642 "lwc1 %[f2], 32(%[tmp_a]) \n\t" | |
| 643 "lwc1 %[f4], 64(%[tmp_a]) \n\t" | |
| 644 "lwc1 %[f6], 96(%[tmp_a]) \n\t" | |
| 645 "lwc1 %[f1], 4(%[tmp_a]) \n\t" | |
| 646 "lwc1 %[f3], 36(%[tmp_a]) \n\t" | |
| 647 "lwc1 %[f5], 68(%[tmp_a]) \n\t" | |
| 648 "lwc1 %[f7], 100(%[tmp_a]) \n\t" | |
| 649 "add.s %[f8], %[f0], %[f2] \n\t" | |
| 650 "sub.s %[f0], %[f0], %[f2] \n\t" | |
| 651 "add.s %[f2], %[f4], %[f6] \n\t" | |
| 652 "sub.s %[f4], %[f4], %[f6] \n\t" | |
| 653 "add.s %[f6], %[f1], %[f3] \n\t" | |
| 654 "sub.s %[f1], %[f1], %[f3] \n\t" | |
| 655 "add.s %[f3], %[f5], %[f7] \n\t" | |
| 656 "sub.s %[f5], %[f5], %[f7] \n\t" | |
| 657 "sub.s %[f7], %[f8], %[f2] \n\t" | |
| 658 "add.s %[f8], %[f8], %[f2] \n\t" | |
| 659 "add.s %[f2], %[f1], %[f4] \n\t" | |
| 660 "sub.s %[f1], %[f1], %[f4] \n\t" | |
| 661 "sub.s %[f4], %[f6], %[f3] \n\t" | |
| 662 "add.s %[f6], %[f6], %[f3] \n\t" | |
| 663 "sub.s %[f3], %[f0], %[f5] \n\t" | |
| 664 "add.s %[f0], %[f0], %[f5] \n\t" | |
| 665 "swc1 %[f8], 0(%[tmp_a]) \n\t" | |
| 666 "swc1 %[f6], 4(%[tmp_a]) \n\t" | |
| 667 "mul.s %[f5], %[f9], %[f7] \n\t" | |
| 668 #if defined(MIPS32_R2_LE) | |
| 669 "mul.s %[f7], %[f10], %[f7] \n\t" | |
| 670 "mul.s %[f8], %[f11], %[f3] \n\t" | |
| 671 "mul.s %[f3], %[f12], %[f3] \n\t" | |
| 672 "mul.s %[f6], %[f13], %[f0] \n\t" | |
| 673 "mul.s %[f0], %[f14], %[f0] \n\t" | |
| 674 "nmsub.s %[f5], %[f5], %[f10], %[f4] \n\t" | |
| 675 "madd.s %[f7], %[f7], %[f9], %[f4] \n\t" | |
| 676 "nmsub.s %[f8], %[f8], %[f12], %[f2] \n\t" | |
| 677 "madd.s %[f3], %[f3], %[f11], %[f2] \n\t" | |
| 678 "nmsub.s %[f6], %[f6], %[f14], %[f1] \n\t" | |
| 679 "madd.s %[f0], %[f0], %[f13], %[f1] \n\t" | |
| 680 "swc1 %[f5], 64(%[tmp_a]) \n\t" | |
| 681 "swc1 %[f7], 68(%[tmp_a]) \n\t" | |
| 682 #else | |
| 683 "mul.s %[f8], %[f10], %[f4] \n\t" | |
| 684 "mul.s %[f4], %[f9], %[f4] \n\t" | |
| 685 "mul.s %[f7], %[f10], %[f7] \n\t" | |
| 686 "mul.s %[f6], %[f11], %[f3] \n\t" | |
| 687 "mul.s %[f3], %[f12], %[f3] \n\t" | |
| 688 "sub.s %[f5], %[f5], %[f8] \n\t" | |
| 689 "mul.s %[f8], %[f12], %[f2] \n\t" | |
| 690 "mul.s %[f2], %[f11], %[f2] \n\t" | |
| 691 "add.s %[f7], %[f4], %[f7] \n\t" | |
| 692 "mul.s %[f4], %[f13], %[f0] \n\t" | |
| 693 "mul.s %[f0], %[f14], %[f0] \n\t" | |
| 694 "sub.s %[f8], %[f6], %[f8] \n\t" | |
| 695 "mul.s %[f6], %[f14], %[f1] \n\t" | |
| 696 "mul.s %[f1], %[f13], %[f1] \n\t" | |
| 697 "add.s %[f3], %[f2], %[f3] \n\t" | |
| 698 "swc1 %[f5], 64(%[tmp_a]) \n\t" | |
| 699 "swc1 %[f7], 68(%[tmp_a]) \n\t" | |
| 700 "sub.s %[f6], %[f4], %[f6] \n\t" | |
| 701 "add.s %[f0], %[f1], %[f0] \n\t" | |
| 702 #endif | |
| 703 "swc1 %[f8], 32(%[tmp_a]) \n\t" | |
| 704 "swc1 %[f3], 36(%[tmp_a]) \n\t" | |
| 705 "swc1 %[f6], 96(%[tmp_a]) \n\t" | |
| 706 "swc1 %[f0], 100(%[tmp_a]) \n\t" | |
| 707 "bgtz %[count], 1b \n\t" | |
| 708 " addiu %[tmp_a], %[tmp_a], 8 \n\t" | |
| 709 ".set pop \n\t" | |
| 710 : [f0] "=&f" (f0), [f1] "=&f" (f1), [f2] "=&f" (f2), [f3] "=&f" (f3), | |
| 711 [f4] "=&f" (f4), [f5] "=&f" (f5), [f6] "=&f" (f6), [f7] "=&f" (f7), | |
| 712 [f8] "=&f" (f8), [tmp_a] "=&r" (tmp_a), [count] "=&r" (count) | |
| 713 : [a] "r" (a), [f9] "f" (f9), [f10] "f" (f10), [f11] "f" (f11), | |
| 714 [f12] "f" (f12), [f13] "f" (f13), [f14] "f" (f14) | |
| 715 : "memory" | |
| 716 ); | |
| 717 f11 = rdft_w[6]; | |
| 718 f12 = rdft_w[7]; | |
| 719 f13 = rdft_wk3ri_second[2]; | |
| 720 f14 = rdft_wk3ri_second[3]; | |
| 721 __asm __volatile ( | |
| 722 ".set push \n\t" | |
| 723 ".set noreorder \n\t" | |
| 724 "addiu %[tmp_a], %[a], 384 \n\t" | |
| 725 "addiu %[count], $zero, 4 \n\t" | |
| 726 "1: \n\t" | |
| 727 "addiu %[count], %[count], -1 \n\t" | |
| 728 "lwc1 %[f0], 0(%[tmp_a]) \n\t" | |
| 729 "lwc1 %[f1], 4(%[tmp_a]) \n\t" | |
| 730 "lwc1 %[f2], 32(%[tmp_a]) \n\t" | |
| 731 "lwc1 %[f3], 36(%[tmp_a]) \n\t" | |
| 732 "lwc1 %[f4], 64(%[tmp_a]) \n\t" | |
| 733 "lwc1 %[f5], 68(%[tmp_a]) \n\t" | |
| 734 "lwc1 %[f6], 96(%[tmp_a]) \n\t" | |
| 735 "lwc1 %[f7], 100(%[tmp_a]) \n\t" | |
| 736 "add.s %[f8], %[f0], %[f2] \n\t" | |
| 737 "sub.s %[f0], %[f0], %[f2] \n\t" | |
| 738 "add.s %[f2], %[f4], %[f6] \n\t" | |
| 739 "sub.s %[f4], %[f4], %[f6] \n\t" | |
| 740 "add.s %[f6], %[f1], %[f3] \n\t" | |
| 741 "sub.s %[f1], %[f1], %[f3] \n\t" | |
| 742 "add.s %[f3], %[f5], %[f7] \n\t" | |
| 743 "sub.s %[f5], %[f5], %[f7] \n\t" | |
| 744 "sub.s %[f7], %[f2], %[f8] \n\t" | |
| 745 "add.s %[f2], %[f2], %[f8] \n\t" | |
| 746 "add.s %[f8], %[f1], %[f4] \n\t" | |
| 747 "sub.s %[f1], %[f1], %[f4] \n\t" | |
| 748 "sub.s %[f4], %[f3], %[f6] \n\t" | |
| 749 "add.s %[f3], %[f3], %[f6] \n\t" | |
| 750 "sub.s %[f6], %[f0], %[f5] \n\t" | |
| 751 "add.s %[f0], %[f0], %[f5] \n\t" | |
| 752 "swc1 %[f2], 0(%[tmp_a]) \n\t" | |
| 753 "swc1 %[f3], 4(%[tmp_a]) \n\t" | |
| 754 "mul.s %[f5], %[f10], %[f7] \n\t" | |
| 755 #if defined(MIPS32_R2_LE) | |
| 756 "mul.s %[f7], %[f9], %[f7] \n\t" | |
| 757 "mul.s %[f2], %[f12], %[f8] \n\t" | |
| 758 "mul.s %[f8], %[f11], %[f8] \n\t" | |
| 759 "mul.s %[f3], %[f14], %[f1] \n\t" | |
| 760 "mul.s %[f1], %[f13], %[f1] \n\t" | |
| 761 "madd.s %[f5], %[f5], %[f9], %[f4] \n\t" | |
| 762 "msub.s %[f7], %[f7], %[f10], %[f4] \n\t" | |
| 763 "msub.s %[f2], %[f2], %[f11], %[f6] \n\t" | |
| 764 "madd.s %[f8], %[f8], %[f12], %[f6] \n\t" | |
| 765 "msub.s %[f3], %[f3], %[f13], %[f0] \n\t" | |
| 766 "madd.s %[f1], %[f1], %[f14], %[f0] \n\t" | |
| 767 "swc1 %[f5], 64(%[tmp_a]) \n\t" | |
| 768 "swc1 %[f7], 68(%[tmp_a]) \n\t" | |
| 769 #else | |
| 770 "mul.s %[f2], %[f9], %[f4] \n\t" | |
| 771 "mul.s %[f4], %[f10], %[f4] \n\t" | |
| 772 "mul.s %[f7], %[f9], %[f7] \n\t" | |
| 773 "mul.s %[f3], %[f11], %[f6] \n\t" | |
| 774 "mul.s %[f6], %[f12], %[f6] \n\t" | |
| 775 "add.s %[f5], %[f5], %[f2] \n\t" | |
| 776 "sub.s %[f7], %[f4], %[f7] \n\t" | |
| 777 "mul.s %[f2], %[f12], %[f8] \n\t" | |
| 778 "mul.s %[f8], %[f11], %[f8] \n\t" | |
| 779 "mul.s %[f4], %[f14], %[f1] \n\t" | |
| 780 "mul.s %[f1], %[f13], %[f1] \n\t" | |
| 781 "sub.s %[f2], %[f3], %[f2] \n\t" | |
| 782 "mul.s %[f3], %[f13], %[f0] \n\t" | |
| 783 "mul.s %[f0], %[f14], %[f0] \n\t" | |
| 784 "add.s %[f8], %[f8], %[f6] \n\t" | |
| 785 "swc1 %[f5], 64(%[tmp_a]) \n\t" | |
| 786 "swc1 %[f7], 68(%[tmp_a]) \n\t" | |
| 787 "sub.s %[f3], %[f3], %[f4] \n\t" | |
| 788 "add.s %[f1], %[f1], %[f0] \n\t" | |
| 789 #endif | |
| 790 "swc1 %[f2], 32(%[tmp_a]) \n\t" | |
| 791 "swc1 %[f8], 36(%[tmp_a]) \n\t" | |
| 792 "swc1 %[f3], 96(%[tmp_a]) \n\t" | |
| 793 "swc1 %[f1], 100(%[tmp_a]) \n\t" | |
| 794 "bgtz %[count], 1b \n\t" | |
| 795 " addiu %[tmp_a], %[tmp_a], 8 \n\t" | |
| 796 ".set pop \n\t" | |
| 797 : [f0] "=&f" (f0), [f1] "=&f" (f1), [f2] "=&f" (f2), [f3] "=&f" (f3), | |
| 798 [f4] "=&f" (f4), [f5] "=&f" (f5), [f6] "=&f" (f6), [f7] "=&f" (f7), | |
| 799 [f8] "=&f" (f8), [tmp_a] "=&r" (tmp_a), [count] "=&r" (count) | |
| 800 : [a] "r" (a), [f9] "f" (f9), [f10] "f" (f10), [f11] "f" (f11), | |
| 801 [f12] "f" (f12), [f13] "f" (f13), [f14] "f" (f14) | |
| 802 : "memory" | |
| 803 ); | |
| 804 } | |
| 805 | |
| 806 static void cftfsub_128_mips(float* a) { | |
| 807 float f0, f1, f2, f3, f4, f5, f6, f7, f8; | |
| 808 int tmp_a, count; | |
| 809 | |
| 810 cft1st_128(a); | |
| 811 cftmdl_128(a); | |
| 812 | |
| 813 __asm __volatile ( | |
| 814 ".set push \n\t" | |
| 815 ".set noreorder \n\t" | |
| 816 "addiu %[tmp_a], %[a], 0 \n\t" | |
| 817 "addiu %[count], $zero, 16 \n\t" | |
| 818 "1: \n\t" | |
| 819 "addiu %[count], %[count], -1 \n\t" | |
| 820 "lwc1 %[f0], 0(%[tmp_a]) \n\t" | |
| 821 "lwc1 %[f2], 128(%[tmp_a]) \n\t" | |
| 822 "lwc1 %[f4], 256(%[tmp_a]) \n\t" | |
| 823 "lwc1 %[f6], 384(%[tmp_a]) \n\t" | |
| 824 "lwc1 %[f1], 4(%[tmp_a]) \n\t" | |
| 825 "lwc1 %[f3], 132(%[tmp_a]) \n\t" | |
| 826 "lwc1 %[f5], 260(%[tmp_a]) \n\t" | |
| 827 "lwc1 %[f7], 388(%[tmp_a]) \n\t" | |
| 828 "add.s %[f8], %[f0], %[f2] \n\t" | |
| 829 "sub.s %[f0], %[f0], %[f2] \n\t" | |
| 830 "add.s %[f2], %[f4], %[f6] \n\t" | |
| 831 "sub.s %[f4], %[f4], %[f6] \n\t" | |
| 832 "add.s %[f6], %[f1], %[f3] \n\t" | |
| 833 "sub.s %[f1], %[f1], %[f3] \n\t" | |
| 834 "add.s %[f3], %[f5], %[f7] \n\t" | |
| 835 "sub.s %[f5], %[f5], %[f7] \n\t" | |
| 836 "add.s %[f7], %[f8], %[f2] \n\t" | |
| 837 "sub.s %[f8], %[f8], %[f2] \n\t" | |
| 838 "add.s %[f2], %[f1], %[f4] \n\t" | |
| 839 "sub.s %[f1], %[f1], %[f4] \n\t" | |
| 840 "add.s %[f4], %[f6], %[f3] \n\t" | |
| 841 "sub.s %[f6], %[f6], %[f3] \n\t" | |
| 842 "sub.s %[f3], %[f0], %[f5] \n\t" | |
| 843 "add.s %[f0], %[f0], %[f5] \n\t" | |
| 844 "swc1 %[f7], 0(%[tmp_a]) \n\t" | |
| 845 "swc1 %[f8], 256(%[tmp_a]) \n\t" | |
| 846 "swc1 %[f2], 132(%[tmp_a]) \n\t" | |
| 847 "swc1 %[f1], 388(%[tmp_a]) \n\t" | |
| 848 "swc1 %[f4], 4(%[tmp_a]) \n\t" | |
| 849 "swc1 %[f6], 260(%[tmp_a]) \n\t" | |
| 850 "swc1 %[f3], 128(%[tmp_a]) \n\t" | |
| 851 "swc1 %[f0], 384(%[tmp_a]) \n\t" | |
| 852 "bgtz %[count], 1b \n\t" | |
| 853 " addiu %[tmp_a], %[tmp_a], 8 \n\t" | |
| 854 ".set pop \n\t" | |
| 855 : [f0] "=&f" (f0), [f1] "=&f" (f1), [f2] "=&f" (f2), [f3] "=&f" (f3), | |
| 856 [f4] "=&f" (f4), [f5] "=&f" (f5), [f6] "=&f" (f6), [f7] "=&f" (f7), | |
| 857 [f8] "=&f" (f8), [tmp_a] "=&r" (tmp_a), | |
| 858 [count] "=&r" (count) | |
| 859 : [a] "r" (a) | |
| 860 : "memory" | |
| 861 ); | |
| 862 } | |
| 863 | |
| 864 static void cftbsub_128_mips(float* a) { | |
| 865 float f0, f1, f2, f3, f4, f5, f6, f7, f8; | |
| 866 int tmp_a, count; | |
| 867 | |
| 868 cft1st_128(a); | |
| 869 cftmdl_128(a); | |
| 870 | |
| 871 __asm __volatile ( | |
| 872 ".set push \n\t" | |
| 873 ".set noreorder \n\t" | |
| 874 "addiu %[tmp_a], %[a], 0 \n\t" | |
| 875 "addiu %[count], $zero, 16 \n\t" | |
| 876 "1: \n\t" | |
| 877 "addiu %[count], %[count], -1 \n\t" | |
| 878 "lwc1 %[f0], 0(%[tmp_a]) \n\t" | |
| 879 "lwc1 %[f2], 128(%[tmp_a]) \n\t" | |
| 880 "lwc1 %[f4], 256(%[tmp_a]) \n\t" | |
| 881 "lwc1 %[f6], 384(%[tmp_a]) \n\t" | |
| 882 "lwc1 %[f1], 4(%[tmp_a]) \n\t" | |
| 883 "lwc1 %[f3], 132(%[tmp_a]) \n\t" | |
| 884 "lwc1 %[f5], 260(%[tmp_a]) \n\t" | |
| 885 "lwc1 %[f7], 388(%[tmp_a]) \n\t" | |
| 886 "add.s %[f8], %[f0], %[f2] \n\t" | |
| 887 "sub.s %[f0], %[f0], %[f2] \n\t" | |
| 888 "add.s %[f2], %[f4], %[f6] \n\t" | |
| 889 "sub.s %[f4], %[f4], %[f6] \n\t" | |
| 890 "add.s %[f6], %[f1], %[f3] \n\t" | |
| 891 "sub.s %[f1], %[f3], %[f1] \n\t" | |
| 892 "add.s %[f3], %[f5], %[f7] \n\t" | |
| 893 "sub.s %[f5], %[f5], %[f7] \n\t" | |
| 894 "add.s %[f7], %[f8], %[f2] \n\t" | |
| 895 "sub.s %[f8], %[f8], %[f2] \n\t" | |
| 896 "sub.s %[f2], %[f1], %[f4] \n\t" | |
| 897 "add.s %[f1], %[f1], %[f4] \n\t" | |
| 898 "add.s %[f4], %[f3], %[f6] \n\t" | |
| 899 "sub.s %[f6], %[f3], %[f6] \n\t" | |
| 900 "sub.s %[f3], %[f0], %[f5] \n\t" | |
| 901 "add.s %[f0], %[f0], %[f5] \n\t" | |
| 902 "neg.s %[f4], %[f4] \n\t" | |
| 903 "swc1 %[f7], 0(%[tmp_a]) \n\t" | |
| 904 "swc1 %[f8], 256(%[tmp_a]) \n\t" | |
| 905 "swc1 %[f2], 132(%[tmp_a]) \n\t" | |
| 906 "swc1 %[f1], 388(%[tmp_a]) \n\t" | |
| 907 "swc1 %[f6], 260(%[tmp_a]) \n\t" | |
| 908 "swc1 %[f3], 128(%[tmp_a]) \n\t" | |
| 909 "swc1 %[f0], 384(%[tmp_a]) \n\t" | |
| 910 "swc1 %[f4], 4(%[tmp_a]) \n\t" | |
| 911 "bgtz %[count], 1b \n\t" | |
| 912 " addiu %[tmp_a], %[tmp_a], 8 \n\t" | |
| 913 ".set pop \n\t" | |
| 914 : [f0] "=&f" (f0), [f1] "=&f" (f1), [f2] "=&f" (f2), [f3] "=&f" (f3), | |
| 915 [f4] "=&f" (f4), [f5] "=&f" (f5), [f6] "=&f" (f6), [f7] "=&f" (f7), | |
| 916 [f8] "=&f" (f8), [tmp_a] "=&r" (tmp_a), [count] "=&r" (count) | |
| 917 : [a] "r" (a) | |
| 918 : "memory" | |
| 919 ); | |
| 920 } | |
| 921 | |
| 922 static void rftfsub_128_mips(float* a) { | |
| 923 const float* c = rdft_w + 32; | |
| 924 const float f0 = 0.5f; | |
| 925 float* a1 = &a[2]; | |
| 926 float* a2 = &a[126]; | |
| 927 const float* c1 = &c[1]; | |
| 928 const float* c2 = &c[31]; | |
| 929 float f1, f2, f3 ,f4, f5, f6, f7, f8, f9, f10, f11, f12, f13, f14, f15; | |
| 930 int count; | |
| 931 | |
| 932 __asm __volatile ( | |
| 933 ".set push \n\t" | |
| 934 ".set noreorder \n\t" | |
| 935 "lwc1 %[f6], 0(%[c2]) \n\t" | |
| 936 "lwc1 %[f1], 0(%[a1]) \n\t" | |
| 937 "lwc1 %[f2], 0(%[a2]) \n\t" | |
| 938 "lwc1 %[f3], 4(%[a1]) \n\t" | |
| 939 "lwc1 %[f4], 4(%[a2]) \n\t" | |
| 940 "lwc1 %[f5], 0(%[c1]) \n\t" | |
| 941 "sub.s %[f6], %[f0], %[f6] \n\t" | |
| 942 "sub.s %[f7], %[f1], %[f2] \n\t" | |
| 943 "add.s %[f8], %[f3], %[f4] \n\t" | |
| 944 "addiu %[count], $zero, 15 \n\t" | |
| 945 "mul.s %[f9], %[f6], %[f7] \n\t" | |
| 946 "mul.s %[f6], %[f6], %[f8] \n\t" | |
| 947 #if !defined(MIPS32_R2_LE) | |
| 948 "mul.s %[f8], %[f5], %[f8] \n\t" | |
| 949 "mul.s %[f5], %[f5], %[f7] \n\t" | |
| 950 "sub.s %[f9], %[f9], %[f8] \n\t" | |
| 951 "add.s %[f6], %[f6], %[f5] \n\t" | |
| 952 #else | |
| 953 "nmsub.s %[f9], %[f9], %[f5], %[f8] \n\t" | |
| 954 "madd.s %[f6], %[f6], %[f5], %[f7] \n\t" | |
| 955 #endif | |
| 956 "sub.s %[f1], %[f1], %[f9] \n\t" | |
| 957 "add.s %[f2], %[f2], %[f9] \n\t" | |
| 958 "sub.s %[f3], %[f3], %[f6] \n\t" | |
| 959 "sub.s %[f4], %[f4], %[f6] \n\t" | |
| 960 "swc1 %[f1], 0(%[a1]) \n\t" | |
| 961 "swc1 %[f2], 0(%[a2]) \n\t" | |
| 962 "swc1 %[f3], 4(%[a1]) \n\t" | |
| 963 "swc1 %[f4], 4(%[a2]) \n\t" | |
| 964 "addiu %[a1], %[a1], 8 \n\t" | |
| 965 "addiu %[a2], %[a2], -8 \n\t" | |
| 966 "addiu %[c1], %[c1], 4 \n\t" | |
| 967 "addiu %[c2], %[c2], -4 \n\t" | |
| 968 "1: \n\t" | |
| 969 "lwc1 %[f6], 0(%[c2]) \n\t" | |
| 970 "lwc1 %[f1], 0(%[a1]) \n\t" | |
| 971 "lwc1 %[f2], 0(%[a2]) \n\t" | |
| 972 "lwc1 %[f3], 4(%[a1]) \n\t" | |
| 973 "lwc1 %[f4], 4(%[a2]) \n\t" | |
| 974 "lwc1 %[f5], 0(%[c1]) \n\t" | |
| 975 "sub.s %[f6], %[f0], %[f6] \n\t" | |
| 976 "sub.s %[f7], %[f1], %[f2] \n\t" | |
| 977 "add.s %[f8], %[f3], %[f4] \n\t" | |
| 978 "lwc1 %[f10], -4(%[c2]) \n\t" | |
| 979 "lwc1 %[f11], 8(%[a1]) \n\t" | |
| 980 "lwc1 %[f12], -8(%[a2]) \n\t" | |
| 981 "mul.s %[f9], %[f6], %[f7] \n\t" | |
| 982 "mul.s %[f6], %[f6], %[f8] \n\t" | |
| 983 #if !defined(MIPS32_R2_LE) | |
| 984 "mul.s %[f8], %[f5], %[f8] \n\t" | |
| 985 "mul.s %[f5], %[f5], %[f7] \n\t" | |
| 986 "lwc1 %[f13], 12(%[a1]) \n\t" | |
| 987 "lwc1 %[f14], -4(%[a2]) \n\t" | |
| 988 "lwc1 %[f15], 4(%[c1]) \n\t" | |
| 989 "sub.s %[f9], %[f9], %[f8] \n\t" | |
| 990 "add.s %[f6], %[f6], %[f5] \n\t" | |
| 991 #else | |
| 992 "lwc1 %[f13], 12(%[a1]) \n\t" | |
| 993 "lwc1 %[f14], -4(%[a2]) \n\t" | |
| 994 "lwc1 %[f15], 4(%[c1]) \n\t" | |
| 995 "nmsub.s %[f9], %[f9], %[f5], %[f8] \n\t" | |
| 996 "madd.s %[f6], %[f6], %[f5], %[f7] \n\t" | |
| 997 #endif | |
| 998 "sub.s %[f10], %[f0], %[f10] \n\t" | |
| 999 "sub.s %[f5], %[f11], %[f12] \n\t" | |
| 1000 "add.s %[f7], %[f13], %[f14] \n\t" | |
| 1001 "sub.s %[f1], %[f1], %[f9] \n\t" | |
| 1002 "add.s %[f2], %[f2], %[f9] \n\t" | |
| 1003 "sub.s %[f3], %[f3], %[f6] \n\t" | |
| 1004 "mul.s %[f8], %[f10], %[f5] \n\t" | |
| 1005 "mul.s %[f10], %[f10], %[f7] \n\t" | |
| 1006 #if !defined(MIPS32_R2_LE) | |
| 1007 "mul.s %[f9], %[f15], %[f7] \n\t" | |
| 1008 "mul.s %[f15], %[f15], %[f5] \n\t" | |
| 1009 "sub.s %[f4], %[f4], %[f6] \n\t" | |
| 1010 "swc1 %[f1], 0(%[a1]) \n\t" | |
| 1011 "swc1 %[f2], 0(%[a2]) \n\t" | |
| 1012 "sub.s %[f8], %[f8], %[f9] \n\t" | |
| 1013 "add.s %[f10], %[f10], %[f15] \n\t" | |
| 1014 #else | |
| 1015 "swc1 %[f1], 0(%[a1]) \n\t" | |
| 1016 "swc1 %[f2], 0(%[a2]) \n\t" | |
| 1017 "sub.s %[f4], %[f4], %[f6] \n\t" | |
| 1018 "nmsub.s %[f8], %[f8], %[f15], %[f7] \n\t" | |
| 1019 "madd.s %[f10], %[f10], %[f15], %[f5] \n\t" | |
| 1020 #endif | |
| 1021 "swc1 %[f3], 4(%[a1]) \n\t" | |
| 1022 "swc1 %[f4], 4(%[a2]) \n\t" | |
| 1023 "sub.s %[f11], %[f11], %[f8] \n\t" | |
| 1024 "add.s %[f12], %[f12], %[f8] \n\t" | |
| 1025 "sub.s %[f13], %[f13], %[f10] \n\t" | |
| 1026 "sub.s %[f14], %[f14], %[f10] \n\t" | |
| 1027 "addiu %[c2], %[c2], -8 \n\t" | |
| 1028 "addiu %[c1], %[c1], 8 \n\t" | |
| 1029 "swc1 %[f11], 8(%[a1]) \n\t" | |
| 1030 "swc1 %[f12], -8(%[a2]) \n\t" | |
| 1031 "swc1 %[f13], 12(%[a1]) \n\t" | |
| 1032 "swc1 %[f14], -4(%[a2]) \n\t" | |
| 1033 "addiu %[a1], %[a1], 16 \n\t" | |
| 1034 "addiu %[count], %[count], -1 \n\t" | |
| 1035 "bgtz %[count], 1b \n\t" | |
| 1036 " addiu %[a2], %[a2], -16 \n\t" | |
| 1037 ".set pop \n\t" | |
| 1038 : [a1] "+r" (a1), [a2] "+r" (a2), [c1] "+r" (c1), [c2] "+r" (c2), | |
| 1039 [f1] "=&f" (f1), [f2] "=&f" (f2), [f3] "=&f" (f3), [f4] "=&f" (f4), | |
| 1040 [f5] "=&f" (f5), [f6] "=&f" (f6), [f7] "=&f" (f7), [f8] "=&f" (f8), | |
| 1041 [f9] "=&f" (f9), [f10] "=&f" (f10), [f11] "=&f" (f11), [f12] "=&f" (f12), | |
| 1042 [f13] "=&f" (f13), [f14] "=&f" (f14), [f15] "=&f" (f15), | |
| 1043 [count] "=&r" (count) | |
| 1044 : [f0] "f" (f0) | |
| 1045 : "memory" | |
| 1046 ); | |
| 1047 } | |
| 1048 | |
| 1049 static void rftbsub_128_mips(float* a) { | |
| 1050 const float *c = rdft_w + 32; | |
| 1051 const float f0 = 0.5f; | |
| 1052 float* a1 = &a[2]; | |
| 1053 float* a2 = &a[126]; | |
| 1054 const float* c1 = &c[1]; | |
| 1055 const float* c2 = &c[31]; | |
| 1056 float f1, f2, f3 ,f4, f5, f6, f7, f8, f9, f10, f11, f12, f13, f14, f15; | |
| 1057 int count; | |
| 1058 | |
| 1059 a[1] = -a[1]; | |
| 1060 a[65] = -a[65]; | |
| 1061 | |
| 1062 __asm __volatile ( | |
| 1063 ".set push \n\t" | |
| 1064 ".set noreorder \n\t" | |
| 1065 "lwc1 %[f6], 0(%[c2]) \n\t" | |
| 1066 "lwc1 %[f1], 0(%[a1]) \n\t" | |
| 1067 "lwc1 %[f2], 0(%[a2]) \n\t" | |
| 1068 "lwc1 %[f3], 4(%[a1]) \n\t" | |
| 1069 "lwc1 %[f4], 4(%[a2]) \n\t" | |
| 1070 "lwc1 %[f5], 0(%[c1]) \n\t" | |
| 1071 "sub.s %[f6], %[f0], %[f6] \n\t" | |
| 1072 "sub.s %[f7], %[f1], %[f2] \n\t" | |
| 1073 "add.s %[f8], %[f3], %[f4] \n\t" | |
| 1074 "addiu %[count], $zero, 15 \n\t" | |
| 1075 "mul.s %[f9], %[f6], %[f7] \n\t" | |
| 1076 "mul.s %[f6], %[f6], %[f8] \n\t" | |
| 1077 #if !defined(MIPS32_R2_LE) | |
| 1078 "mul.s %[f8], %[f5], %[f8] \n\t" | |
| 1079 "mul.s %[f5], %[f5], %[f7] \n\t" | |
| 1080 "add.s %[f9], %[f9], %[f8] \n\t" | |
| 1081 "sub.s %[f6], %[f6], %[f5] \n\t" | |
| 1082 #else | |
| 1083 "madd.s %[f9], %[f9], %[f5], %[f8] \n\t" | |
| 1084 "nmsub.s %[f6], %[f6], %[f5], %[f7] \n\t" | |
| 1085 #endif | |
| 1086 "sub.s %[f1], %[f1], %[f9] \n\t" | |
| 1087 "add.s %[f2], %[f2], %[f9] \n\t" | |
| 1088 "sub.s %[f3], %[f6], %[f3] \n\t" | |
| 1089 "sub.s %[f4], %[f6], %[f4] \n\t" | |
| 1090 "swc1 %[f1], 0(%[a1]) \n\t" | |
| 1091 "swc1 %[f2], 0(%[a2]) \n\t" | |
| 1092 "swc1 %[f3], 4(%[a1]) \n\t" | |
| 1093 "swc1 %[f4], 4(%[a2]) \n\t" | |
| 1094 "addiu %[a1], %[a1], 8 \n\t" | |
| 1095 "addiu %[a2], %[a2], -8 \n\t" | |
| 1096 "addiu %[c1], %[c1], 4 \n\t" | |
| 1097 "addiu %[c2], %[c2], -4 \n\t" | |
| 1098 "1: \n\t" | |
| 1099 "lwc1 %[f6], 0(%[c2]) \n\t" | |
| 1100 "lwc1 %[f1], 0(%[a1]) \n\t" | |
| 1101 "lwc1 %[f2], 0(%[a2]) \n\t" | |
| 1102 "lwc1 %[f3], 4(%[a1]) \n\t" | |
| 1103 "lwc1 %[f4], 4(%[a2]) \n\t" | |
| 1104 "lwc1 %[f5], 0(%[c1]) \n\t" | |
| 1105 "sub.s %[f6], %[f0], %[f6] \n\t" | |
| 1106 "sub.s %[f7], %[f1], %[f2] \n\t" | |
| 1107 "add.s %[f8], %[f3], %[f4] \n\t" | |
| 1108 "lwc1 %[f10], -4(%[c2]) \n\t" | |
| 1109 "lwc1 %[f11], 8(%[a1]) \n\t" | |
| 1110 "lwc1 %[f12], -8(%[a2]) \n\t" | |
| 1111 "mul.s %[f9], %[f6], %[f7] \n\t" | |
| 1112 "mul.s %[f6], %[f6], %[f8] \n\t" | |
| 1113 #if !defined(MIPS32_R2_LE) | |
| 1114 "mul.s %[f8], %[f5], %[f8] \n\t" | |
| 1115 "mul.s %[f5], %[f5], %[f7] \n\t" | |
| 1116 "lwc1 %[f13], 12(%[a1]) \n\t" | |
| 1117 "lwc1 %[f14], -4(%[a2]) \n\t" | |
| 1118 "lwc1 %[f15], 4(%[c1]) \n\t" | |
| 1119 "add.s %[f9], %[f9], %[f8] \n\t" | |
| 1120 "sub.s %[f6], %[f6], %[f5] \n\t" | |
| 1121 #else | |
| 1122 "lwc1 %[f13], 12(%[a1]) \n\t" | |
| 1123 "lwc1 %[f14], -4(%[a2]) \n\t" | |
| 1124 "lwc1 %[f15], 4(%[c1]) \n\t" | |
| 1125 "madd.s %[f9], %[f9], %[f5], %[f8] \n\t" | |
| 1126 "nmsub.s %[f6], %[f6], %[f5], %[f7] \n\t" | |
| 1127 #endif | |
| 1128 "sub.s %[f10], %[f0], %[f10] \n\t" | |
| 1129 "sub.s %[f5], %[f11], %[f12] \n\t" | |
| 1130 "add.s %[f7], %[f13], %[f14] \n\t" | |
| 1131 "sub.s %[f1], %[f1], %[f9] \n\t" | |
| 1132 "add.s %[f2], %[f2], %[f9] \n\t" | |
| 1133 "sub.s %[f3], %[f6], %[f3] \n\t" | |
| 1134 "mul.s %[f8], %[f10], %[f5] \n\t" | |
| 1135 "mul.s %[f10], %[f10], %[f7] \n\t" | |
| 1136 #if !defined(MIPS32_R2_LE) | |
| 1137 "mul.s %[f9], %[f15], %[f7] \n\t" | |
| 1138 "mul.s %[f15], %[f15], %[f5] \n\t" | |
| 1139 "sub.s %[f4], %[f6], %[f4] \n\t" | |
| 1140 "swc1 %[f1], 0(%[a1]) \n\t" | |
| 1141 "swc1 %[f2], 0(%[a2]) \n\t" | |
| 1142 "add.s %[f8], %[f8], %[f9] \n\t" | |
| 1143 "sub.s %[f10], %[f10], %[f15] \n\t" | |
| 1144 #else | |
| 1145 "swc1 %[f1], 0(%[a1]) \n\t" | |
| 1146 "swc1 %[f2], 0(%[a2]) \n\t" | |
| 1147 "sub.s %[f4], %[f6], %[f4] \n\t" | |
| 1148 "madd.s %[f8], %[f8], %[f15], %[f7] \n\t" | |
| 1149 "nmsub.s %[f10], %[f10], %[f15], %[f5] \n\t" | |
| 1150 #endif | |
| 1151 "swc1 %[f3], 4(%[a1]) \n\t" | |
| 1152 "swc1 %[f4], 4(%[a2]) \n\t" | |
| 1153 "sub.s %[f11], %[f11], %[f8] \n\t" | |
| 1154 "add.s %[f12], %[f12], %[f8] \n\t" | |
| 1155 "sub.s %[f13], %[f10], %[f13] \n\t" | |
| 1156 "sub.s %[f14], %[f10], %[f14] \n\t" | |
| 1157 "addiu %[c2], %[c2], -8 \n\t" | |
| 1158 "addiu %[c1], %[c1], 8 \n\t" | |
| 1159 "swc1 %[f11], 8(%[a1]) \n\t" | |
| 1160 "swc1 %[f12], -8(%[a2]) \n\t" | |
| 1161 "swc1 %[f13], 12(%[a1]) \n\t" | |
| 1162 "swc1 %[f14], -4(%[a2]) \n\t" | |
| 1163 "addiu %[a1], %[a1], 16 \n\t" | |
| 1164 "addiu %[count], %[count], -1 \n\t" | |
| 1165 "bgtz %[count], 1b \n\t" | |
| 1166 " addiu %[a2], %[a2], -16 \n\t" | |
| 1167 ".set pop \n\t" | |
| 1168 : [a1] "+r" (a1), [a2] "+r" (a2), [c1] "+r" (c1), [c2] "+r" (c2), | |
| 1169 [f1] "=&f" (f1), [f2] "=&f" (f2), [f3] "=&f" (f3), [f4] "=&f" (f4), | |
| 1170 [f5] "=&f" (f5), [f6] "=&f" (f6), [f7] "=&f" (f7), [f8] "=&f" (f8), | |
| 1171 [f9] "=&f" (f9), [f10] "=&f" (f10), [f11] "=&f" (f11), [f12] "=&f" (f12), | |
| 1172 [f13] "=&f" (f13), [f14] "=&f" (f14), [f15] "=&f" (f15), | |
| 1173 [count] "=&r" (count) | |
| 1174 : [f0] "f" (f0) | |
| 1175 : "memory" | |
| 1176 ); | |
| 1177 } | |
| 1178 | |
| 1179 void aec_rdft_init_mips(void) { | |
| 1180 cft1st_128 = cft1st_128_mips; | |
| 1181 cftmdl_128 = cftmdl_128_mips; | |
| 1182 rftfsub_128 = rftfsub_128_mips; | |
| 1183 rftbsub_128 = rftbsub_128_mips; | |
| 1184 cftfsub_128 = cftfsub_128_mips; | |
| 1185 cftbsub_128 = cftbsub_128_mips; | |
| 1186 bitrv2_128 = bitrv2_128_mips; | |
| 1187 } | |
| OLD | NEW |
|---|
Chromium Code Reviews
Issue 2348213002:
Move the aec_rdft* files to a more proper place beneath APM and make them thread-safe. (Closed)
