Commit | Line | Data |
---|---|---|
2ba45a60 DM |
1 | /* |
2 | * Copyright (c) 2012 | |
3 | * MIPS Technologies, Inc., California. | |
4 | * | |
5 | * Redistribution and use in source and binary forms, with or without | |
6 | * modification, are permitted provided that the following conditions | |
7 | * are met: | |
8 | * 1. Redistributions of source code must retain the above copyright | |
9 | * notice, this list of conditions and the following disclaimer. | |
10 | * 2. Redistributions in binary form must reproduce the above copyright | |
11 | * notice, this list of conditions and the following disclaimer in the | |
12 | * documentation and/or other materials provided with the distribution. | |
13 | * 3. Neither the name of the MIPS Technologies, Inc., nor the names of its | |
14 | * contributors may be used to endorse or promote products derived from | |
15 | * this software without specific prior written permission. | |
16 | * | |
17 | * THIS SOFTWARE IS PROVIDED BY THE MIPS TECHNOLOGIES, INC. ``AS IS'' AND | |
18 | * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE | |
19 | * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE | |
20 | * ARE DISCLAIMED. IN NO EVENT SHALL THE MIPS TECHNOLOGIES, INC. BE LIABLE | |
21 | * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL | |
22 | * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS | |
23 | * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) | |
24 | * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT | |
25 | * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY | |
26 | * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF | |
27 | * SUCH DAMAGE. | |
28 | * | |
29 | * Authors: Djordje Pesut (djordje@mips.com) | |
30 | * Mirjana Vulin (mvulin@mips.com) | |
31 | * | |
32 | * This file is part of FFmpeg. | |
33 | * | |
34 | * FFmpeg is free software; you can redistribute it and/or | |
35 | * modify it under the terms of the GNU Lesser General Public | |
36 | * License as published by the Free Software Foundation; either | |
37 | * version 2.1 of the License, or (at your option) any later version. | |
38 | * | |
39 | * FFmpeg is distributed in the hope that it will be useful, | |
40 | * but WITHOUT ANY WARRANTY; without even the implied warranty of | |
41 | * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU | |
42 | * Lesser General Public License for more details. | |
43 | * | |
44 | * You should have received a copy of the GNU Lesser General Public | |
45 | * License along with FFmpeg; if not, write to the Free Software | |
46 | * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA | |
47 | */ | |
48 | ||
49 | /** | |
50 | * @file | |
51 | * Reference: libavcodec/aacsbr.c | |
52 | */ | |
53 | ||
54 | #ifndef AVCODEC_MIPS_AACSBR_FLOAT_H | |
55 | #define AVCODEC_MIPS_AACSBR_FLOAT_H | |
56 | ||
57 | #include "libavcodec/aac.h" | |
58 | #include "libavcodec/sbr.h" | |
59 | ||
60 | #if HAVE_INLINE_ASM | |
61 | static void sbr_qmf_analysis_mips(AVFloatDSPContext *fdsp, FFTContext *mdct, | |
62 | SBRDSPContext *sbrdsp, const float *in, float *x, | |
63 | float z[320], float W[2][32][32][2], int buf_idx) | |
64 | { | |
65 | int i; | |
66 | float *w0; | |
67 | float *w1; | |
68 | int temp0, temp1, temp2, temp3, temp4, temp5, temp6, temp7; | |
69 | ||
70 | w0 = x; | |
71 | w1 = x + 1024; | |
72 | for(i = 0; i < 36; i++) | |
73 | { | |
74 | /* loop unrolled 8 times */ | |
75 | __asm__ volatile( | |
76 | "lw %[temp0], 0(%[w1]) \n\t" | |
77 | "lw %[temp1], 4(%[w1]) \n\t" | |
78 | "lw %[temp2], 8(%[w1]) \n\t" | |
79 | "lw %[temp3], 12(%[w1]) \n\t" | |
80 | "lw %[temp4], 16(%[w1]) \n\t" | |
81 | "lw %[temp5], 20(%[w1]) \n\t" | |
82 | "lw %[temp6], 24(%[w1]) \n\t" | |
83 | "lw %[temp7], 28(%[w1]) \n\t" | |
84 | "sw %[temp0], 0(%[w0]) \n\t" | |
85 | "sw %[temp1], 4(%[w0]) \n\t" | |
86 | "sw %[temp2], 8(%[w0]) \n\t" | |
87 | "sw %[temp3], 12(%[w0]) \n\t" | |
88 | "sw %[temp4], 16(%[w0]) \n\t" | |
89 | "sw %[temp5], 20(%[w0]) \n\t" | |
90 | "sw %[temp6], 24(%[w0]) \n\t" | |
91 | "sw %[temp7], 28(%[w0]) \n\t" | |
92 | "addiu %[w0], %[w0], 32 \n\t" | |
93 | "addiu %[w1], %[w1], 32 \n\t" | |
94 | ||
95 | : [w0]"+r"(w0), [w1]"+r"(w1), | |
96 | [temp0]"=&r"(temp0), [temp1]"=&r"(temp1), | |
97 | [temp2]"=&r"(temp2), [temp3]"=&r"(temp3), | |
98 | [temp4]"=&r"(temp4), [temp5]"=&r"(temp5), | |
99 | [temp6]"=&r"(temp6), [temp7]"=&r"(temp7) | |
100 | : | |
101 | : "memory" | |
102 | ); | |
103 | } | |
104 | ||
105 | w0 = x + 288; | |
106 | w1 = (float*)in; | |
107 | for(i = 0; i < 128; i++) | |
108 | { | |
109 | /* loop unrolled 8 times */ | |
110 | __asm__ volatile( | |
111 | "lw %[temp0], 0(%[w1]) \n\t" | |
112 | "lw %[temp1], 4(%[w1]) \n\t" | |
113 | "lw %[temp2], 8(%[w1]) \n\t" | |
114 | "lw %[temp3], 12(%[w1]) \n\t" | |
115 | "lw %[temp4], 16(%[w1]) \n\t" | |
116 | "lw %[temp5], 20(%[w1]) \n\t" | |
117 | "lw %[temp6], 24(%[w1]) \n\t" | |
118 | "lw %[temp7], 28(%[w1]) \n\t" | |
119 | "sw %[temp0], 0(%[w0]) \n\t" | |
120 | "sw %[temp1], 4(%[w0]) \n\t" | |
121 | "sw %[temp2], 8(%[w0]) \n\t" | |
122 | "sw %[temp3], 12(%[w0]) \n\t" | |
123 | "sw %[temp4], 16(%[w0]) \n\t" | |
124 | "sw %[temp5], 20(%[w0]) \n\t" | |
125 | "sw %[temp6], 24(%[w0]) \n\t" | |
126 | "sw %[temp7], 28(%[w0]) \n\t" | |
127 | "addiu %[w0], %[w0], 32 \n\t" | |
128 | "addiu %[w1], %[w1], 32 \n\t" | |
129 | ||
130 | : [w0]"+r"(w0), [w1]"+r"(w1), | |
131 | [temp0]"=&r"(temp0), [temp1]"=&r"(temp1), | |
132 | [temp2]"=&r"(temp2), [temp3]"=&r"(temp3), | |
133 | [temp4]"=&r"(temp4), [temp5]"=&r"(temp5), | |
134 | [temp6]"=&r"(temp6), [temp7]"=&r"(temp7) | |
135 | : | |
136 | : "memory" | |
137 | ); | |
138 | } | |
139 | ||
140 | for (i = 0; i < 32; i++) { // numTimeSlots*RATE = 16*2 as 960 sample frames | |
141 | // are not supported | |
142 | fdsp->vector_fmul_reverse(z, sbr_qmf_window_ds, x, 320); | |
143 | sbrdsp->sum64x5(z); | |
144 | sbrdsp->qmf_pre_shuffle(z); | |
145 | mdct->imdct_half(mdct, z, z+64); | |
146 | sbrdsp->qmf_post_shuffle(W[buf_idx][i], z); | |
147 | x += 32; | |
148 | } | |
149 | } | |
150 | ||
151 | #if HAVE_MIPSFPU | |
152 | static void sbr_qmf_synthesis_mips(FFTContext *mdct, | |
153 | SBRDSPContext *sbrdsp, AVFloatDSPContext *fdsp, | |
154 | float *out, float X[2][38][64], | |
155 | float mdct_buf[2][64], | |
156 | float *v0, int *v_off, const unsigned int div) | |
157 | { | |
158 | int i, n; | |
159 | const float *sbr_qmf_window = div ? sbr_qmf_window_ds : sbr_qmf_window_us; | |
160 | const int step = 128 >> div; | |
161 | float *v; | |
162 | float temp0, temp1, temp2, temp3, temp4, temp5, temp6, temp7, temp8, temp9, temp10, temp11, temp12, temp13; | |
163 | float temp14, temp15, temp16, temp17, temp18, temp19; | |
164 | float *vv0, *s0, *dst; | |
165 | dst = out; | |
166 | ||
167 | for (i = 0; i < 32; i++) { | |
168 | if (*v_off < step) { | |
169 | int saved_samples = (1280 - 128) >> div; | |
170 | memcpy(&v0[SBR_SYNTHESIS_BUF_SIZE - saved_samples], v0, saved_samples * sizeof(float)); | |
171 | *v_off = SBR_SYNTHESIS_BUF_SIZE - saved_samples - step; | |
172 | } else { | |
173 | *v_off -= step; | |
174 | } | |
175 | v = v0 + *v_off; | |
176 | if (div) { | |
177 | for (n = 0; n < 32; n++) { | |
178 | X[0][i][ n] = -X[0][i][n]; | |
179 | X[0][i][32+n] = X[1][i][31-n]; | |
180 | } | |
181 | mdct->imdct_half(mdct, mdct_buf[0], X[0][i]); | |
182 | sbrdsp->qmf_deint_neg(v, mdct_buf[0]); | |
183 | } else { | |
184 | sbrdsp->neg_odd_64(X[1][i]); | |
185 | mdct->imdct_half(mdct, mdct_buf[0], X[0][i]); | |
186 | mdct->imdct_half(mdct, mdct_buf[1], X[1][i]); | |
187 | sbrdsp->qmf_deint_bfly(v, mdct_buf[1], mdct_buf[0]); | |
188 | } | |
189 | ||
190 | if(div == 0) | |
191 | { | |
192 | float *v0_end; | |
193 | vv0 = v; | |
194 | v0_end = v + 60; | |
195 | s0 = (float*)sbr_qmf_window; | |
196 | ||
197 | /* 10 calls of function vector_fmul_add merged into one loop | |
198 | and loop unrolled 4 times */ | |
199 | __asm__ volatile( | |
200 | ".set push \n\t" | |
201 | ".set noreorder \n\t" | |
202 | "lwc1 %[temp4], 0(%[v0]) \n\t" | |
203 | "lwc1 %[temp5], 0(%[s0]) \n\t" | |
204 | "lwc1 %[temp6], 4(%[v0]) \n\t" | |
205 | "lwc1 %[temp7], 4(%[s0]) \n\t" | |
206 | "lwc1 %[temp8], 8(%[v0]) \n\t" | |
207 | "lwc1 %[temp9], 8(%[s0]) \n\t" | |
208 | "lwc1 %[temp10], 12(%[v0]) \n\t" | |
209 | "lwc1 %[temp11], 12(%[s0]) \n\t" | |
210 | "lwc1 %[temp12], 768(%[v0]) \n\t" | |
211 | "lwc1 %[temp13], 256(%[s0]) \n\t" | |
212 | "lwc1 %[temp14], 772(%[v0]) \n\t" | |
213 | "lwc1 %[temp15], 260(%[s0]) \n\t" | |
214 | "lwc1 %[temp16], 776(%[v0]) \n\t" | |
215 | "lwc1 %[temp17], 264(%[s0]) \n\t" | |
216 | "lwc1 %[temp18], 780(%[v0]) \n\t" | |
217 | "lwc1 %[temp19], 268(%[s0]) \n\t" | |
218 | "1: \n\t" | |
219 | "mul.s %[temp0], %[temp4], %[temp5] \n\t" | |
220 | "lwc1 %[temp4], 1024(%[v0]) \n\t" | |
221 | "mul.s %[temp1], %[temp6], %[temp7] \n\t" | |
222 | "lwc1 %[temp5], 512(%[s0]) \n\t" | |
223 | "mul.s %[temp2], %[temp8], %[temp9] \n\t" | |
224 | "lwc1 %[temp6], 1028(%[v0]) \n\t" | |
225 | "mul.s %[temp3], %[temp10], %[temp11] \n\t" | |
226 | "lwc1 %[temp7], 516(%[s0]) \n\t" | |
227 | "madd.s %[temp0], %[temp0], %[temp12], %[temp13] \n\t" | |
228 | "lwc1 %[temp8], 1032(%[v0]) \n\t" | |
229 | "madd.s %[temp1], %[temp1], %[temp14], %[temp15] \n\t" | |
230 | "lwc1 %[temp9], 520(%[s0]) \n\t" | |
231 | "madd.s %[temp2], %[temp2], %[temp16], %[temp17] \n\t" | |
232 | "lwc1 %[temp10], 1036(%[v0]) \n\t" | |
233 | "madd.s %[temp3], %[temp3], %[temp18], %[temp19] \n\t" | |
234 | "lwc1 %[temp11], 524(%[s0]) \n\t" | |
235 | "lwc1 %[temp12], 1792(%[v0]) \n\t" | |
236 | "lwc1 %[temp13], 768(%[s0]) \n\t" | |
237 | "lwc1 %[temp14], 1796(%[v0]) \n\t" | |
238 | "lwc1 %[temp15], 772(%[s0]) \n\t" | |
239 | "lwc1 %[temp16], 1800(%[v0]) \n\t" | |
240 | "lwc1 %[temp17], 776(%[s0]) \n\t" | |
241 | "lwc1 %[temp18], 1804(%[v0]) \n\t" | |
242 | "lwc1 %[temp19], 780(%[s0]) \n\t" | |
243 | "madd.s %[temp0], %[temp0], %[temp4], %[temp5] \n\t" | |
244 | "lwc1 %[temp4], 2048(%[v0]) \n\t" | |
245 | "madd.s %[temp1], %[temp1], %[temp6], %[temp7] \n\t" | |
246 | "lwc1 %[temp5], 1024(%[s0]) \n\t" | |
247 | "madd.s %[temp2], %[temp2], %[temp8], %[temp9] \n\t" | |
248 | "lwc1 %[temp6], 2052(%[v0]) \n\t" | |
249 | "madd.s %[temp3], %[temp3], %[temp10], %[temp11] \n\t" | |
250 | "lwc1 %[temp7], 1028(%[s0]) \n\t" | |
251 | "madd.s %[temp0], %[temp0], %[temp12], %[temp13] \n\t" | |
252 | "lwc1 %[temp8], 2056(%[v0]) \n\t" | |
253 | "madd.s %[temp1], %[temp1], %[temp14], %[temp15] \n\t" | |
254 | "lwc1 %[temp9], 1032(%[s0]) \n\t" | |
255 | "madd.s %[temp2], %[temp2], %[temp16], %[temp17] \n\t" | |
256 | "lwc1 %[temp10], 2060(%[v0]) \n\t" | |
257 | "madd.s %[temp3], %[temp3], %[temp18], %[temp19] \n\t" | |
258 | "lwc1 %[temp11], 1036(%[s0]) \n\t" | |
259 | "lwc1 %[temp12], 2816(%[v0]) \n\t" | |
260 | "lwc1 %[temp13], 1280(%[s0]) \n\t" | |
261 | "lwc1 %[temp14], 2820(%[v0]) \n\t" | |
262 | "lwc1 %[temp15], 1284(%[s0]) \n\t" | |
263 | "lwc1 %[temp16], 2824(%[v0]) \n\t" | |
264 | "lwc1 %[temp17], 1288(%[s0]) \n\t" | |
265 | "lwc1 %[temp18], 2828(%[v0]) \n\t" | |
266 | "lwc1 %[temp19], 1292(%[s0]) \n\t" | |
267 | "madd.s %[temp0], %[temp0], %[temp4], %[temp5] \n\t" | |
268 | "lwc1 %[temp4], 3072(%[v0]) \n\t" | |
269 | "madd.s %[temp1], %[temp1], %[temp6], %[temp7] \n\t" | |
270 | "lwc1 %[temp5], 1536(%[s0]) \n\t" | |
271 | "madd.s %[temp2], %[temp2], %[temp8], %[temp9] \n\t" | |
272 | "lwc1 %[temp6], 3076(%[v0]) \n\t" | |
273 | "madd.s %[temp3], %[temp3], %[temp10], %[temp11] \n\t" | |
274 | "lwc1 %[temp7], 1540(%[s0]) \n\t" | |
275 | "madd.s %[temp0], %[temp0], %[temp12], %[temp13] \n\t" | |
276 | "lwc1 %[temp8], 3080(%[v0]) \n\t" | |
277 | "madd.s %[temp1], %[temp1], %[temp14], %[temp15] \n\t" | |
278 | "lwc1 %[temp9], 1544(%[s0]) \n\t" | |
279 | "madd.s %[temp2], %[temp2], %[temp16], %[temp17] \n\t" | |
280 | "lwc1 %[temp10], 3084(%[v0]) \n\t" | |
281 | "madd.s %[temp3], %[temp3], %[temp18], %[temp19] \n\t" | |
282 | "lwc1 %[temp11], 1548(%[s0]) \n\t" | |
283 | "lwc1 %[temp12], 3840(%[v0]) \n\t" | |
284 | "lwc1 %[temp13], 1792(%[s0]) \n\t" | |
285 | "lwc1 %[temp14], 3844(%[v0]) \n\t" | |
286 | "lwc1 %[temp15], 1796(%[s0]) \n\t" | |
287 | "lwc1 %[temp16], 3848(%[v0]) \n\t" | |
288 | "lwc1 %[temp17], 1800(%[s0]) \n\t" | |
289 | "lwc1 %[temp18], 3852(%[v0]) \n\t" | |
290 | "lwc1 %[temp19], 1804(%[s0]) \n\t" | |
291 | "madd.s %[temp0], %[temp0], %[temp4], %[temp5] \n\t" | |
292 | "lwc1 %[temp4], 4096(%[v0]) \n\t" | |
293 | "madd.s %[temp1], %[temp1], %[temp6], %[temp7] \n\t" | |
294 | "lwc1 %[temp5], 2048(%[s0]) \n\t" | |
295 | "madd.s %[temp2], %[temp2], %[temp8], %[temp9] \n\t" | |
296 | "lwc1 %[temp6], 4100(%[v0]) \n\t" | |
297 | "madd.s %[temp3], %[temp3], %[temp10], %[temp11] \n\t" | |
298 | "lwc1 %[temp7], 2052(%[s0]) \n\t" | |
299 | "madd.s %[temp0], %[temp0], %[temp12], %[temp13] \n\t" | |
300 | "lwc1 %[temp8], 4104(%[v0]) \n\t" | |
301 | "addiu %[dst], %[dst], 16 \n\t" | |
302 | "madd.s %[temp1], %[temp1], %[temp14], %[temp15] \n\t" | |
303 | "lwc1 %[temp9], 2056(%[s0]) \n\t" | |
304 | "addiu %[s0], %[s0], 16 \n\t" | |
305 | "madd.s %[temp2], %[temp2], %[temp16], %[temp17] \n\t" | |
306 | "lwc1 %[temp10], 4108(%[v0]) \n\t" | |
307 | "addiu %[v0], %[v0], 16 \n\t" | |
308 | "madd.s %[temp3], %[temp3], %[temp18], %[temp19] \n\t" | |
309 | "lwc1 %[temp11], 2044(%[s0]) \n\t" | |
310 | "lwc1 %[temp12], 4848(%[v0]) \n\t" | |
311 | "lwc1 %[temp13], 2288(%[s0]) \n\t" | |
312 | "lwc1 %[temp14], 4852(%[v0]) \n\t" | |
313 | "lwc1 %[temp15], 2292(%[s0]) \n\t" | |
314 | "lwc1 %[temp16], 4856(%[v0]) \n\t" | |
315 | "lwc1 %[temp17], 2296(%[s0]) \n\t" | |
316 | "lwc1 %[temp18], 4860(%[v0]) \n\t" | |
317 | "lwc1 %[temp19], 2300(%[s0]) \n\t" | |
318 | "madd.s %[temp0], %[temp0], %[temp4], %[temp5] \n\t" | |
319 | "lwc1 %[temp4], 0(%[v0]) \n\t" | |
320 | "madd.s %[temp1], %[temp1], %[temp6], %[temp7] \n\t" | |
321 | "lwc1 %[temp5], 0(%[s0]) \n\t" | |
322 | "madd.s %[temp2], %[temp2], %[temp8], %[temp9] \n\t" | |
323 | "lwc1 %[temp6], 4(%[v0]) \n\t" | |
324 | "madd.s %[temp3], %[temp3], %[temp10], %[temp11] \n\t" | |
325 | "lwc1 %[temp7], 4(%[s0]) \n\t" | |
326 | "madd.s %[temp0], %[temp0], %[temp12], %[temp13] \n\t" | |
327 | "lwc1 %[temp8], 8(%[v0]) \n\t" | |
328 | "madd.s %[temp1], %[temp1], %[temp14], %[temp15] \n\t" | |
329 | "lwc1 %[temp9], 8(%[s0]) \n\t" | |
330 | "madd.s %[temp2], %[temp2], %[temp16], %[temp17] \n\t" | |
331 | "lwc1 %[temp10], 12(%[v0]) \n\t" | |
332 | "madd.s %[temp3], %[temp3], %[temp18], %[temp19] \n\t" | |
333 | "lwc1 %[temp11], 12(%[s0]) \n\t" | |
334 | "lwc1 %[temp12], 768(%[v0]) \n\t" | |
335 | "lwc1 %[temp13], 256(%[s0]) \n\t" | |
336 | "lwc1 %[temp14], 772(%[v0]) \n\t" | |
337 | "lwc1 %[temp15], 260(%[s0]) \n\t" | |
338 | "lwc1 %[temp16], 776(%[v0]) \n\t" | |
339 | "lwc1 %[temp17], 264(%[s0]) \n\t" | |
340 | "lwc1 %[temp18], 780(%[v0]) \n\t" | |
341 | "lwc1 %[temp19], 268(%[s0]) \n\t" | |
342 | "swc1 %[temp0], -16(%[dst]) \n\t" | |
343 | "swc1 %[temp1], -12(%[dst]) \n\t" | |
344 | "swc1 %[temp2], -8(%[dst]) \n\t" | |
345 | "bne %[v0], %[v0_end], 1b \n\t" | |
346 | " swc1 %[temp3], -4(%[dst]) \n\t" | |
347 | "mul.s %[temp0], %[temp4], %[temp5] \n\t" | |
348 | "lwc1 %[temp4], 1024(%[v0]) \n\t" | |
349 | "mul.s %[temp1], %[temp6], %[temp7] \n\t" | |
350 | "lwc1 %[temp5], 512(%[s0]) \n\t" | |
351 | "mul.s %[temp2], %[temp8], %[temp9] \n\t" | |
352 | "lwc1 %[temp6], 1028(%[v0]) \n\t" | |
353 | "mul.s %[temp3], %[temp10], %[temp11] \n\t" | |
354 | "lwc1 %[temp7], 516(%[s0]) \n\t" | |
355 | "madd.s %[temp0], %[temp0], %[temp12], %[temp13] \n\t" | |
356 | "lwc1 %[temp8], 1032(%[v0]) \n\t" | |
357 | "madd.s %[temp1], %[temp1], %[temp14], %[temp15] \n\t" | |
358 | "lwc1 %[temp9], 520(%[s0]) \n\t" | |
359 | "madd.s %[temp2], %[temp2], %[temp16], %[temp17] \n\t" | |
360 | "lwc1 %[temp10], 1036(%[v0]) \n\t" | |
361 | "madd.s %[temp3], %[temp3], %[temp18], %[temp19] \n\t" | |
362 | "lwc1 %[temp11], 524(%[s0]) \n\t" | |
363 | "lwc1 %[temp12], 1792(%[v0]) \n\t" | |
364 | "lwc1 %[temp13], 768(%[s0]) \n\t" | |
365 | "lwc1 %[temp14], 1796(%[v0]) \n\t" | |
366 | "lwc1 %[temp15], 772(%[s0]) \n\t" | |
367 | "lwc1 %[temp16], 1800(%[v0]) \n\t" | |
368 | "lwc1 %[temp17], 776(%[s0]) \n\t" | |
369 | "lwc1 %[temp18], 1804(%[v0]) \n\t" | |
370 | "lwc1 %[temp19], 780(%[s0]) \n\t" | |
371 | "madd.s %[temp0], %[temp0], %[temp4], %[temp5] \n\t" | |
372 | "lwc1 %[temp4], 2048(%[v0]) \n\t" | |
373 | "madd.s %[temp1], %[temp1], %[temp6], %[temp7] \n\t" | |
374 | "lwc1 %[temp5], 1024(%[s0]) \n\t" | |
375 | "madd.s %[temp2], %[temp2], %[temp8], %[temp9] \n\t" | |
376 | "lwc1 %[temp6], 2052(%[v0]) \n\t" | |
377 | "madd.s %[temp3], %[temp3], %[temp10], %[temp11] \n\t" | |
378 | "lwc1 %[temp7], 1028(%[s0]) \n\t" | |
379 | "madd.s %[temp0], %[temp0], %[temp12], %[temp13] \n\t" | |
380 | "lwc1 %[temp8], 2056(%[v0]) \n\t" | |
381 | "madd.s %[temp1], %[temp1], %[temp14], %[temp15] \n\t" | |
382 | "lwc1 %[temp9], 1032(%[s0]) \n\t" | |
383 | "madd.s %[temp2], %[temp2], %[temp16], %[temp17] \n\t" | |
384 | "lwc1 %[temp10], 2060(%[v0]) \n\t" | |
385 | "madd.s %[temp3], %[temp3], %[temp18], %[temp19] \n\t" | |
386 | "lwc1 %[temp11], 1036(%[s0]) \n\t" | |
387 | "lwc1 %[temp12], 2816(%[v0]) \n\t" | |
388 | "lwc1 %[temp13], 1280(%[s0]) \n\t" | |
389 | "lwc1 %[temp14], 2820(%[v0]) \n\t" | |
390 | "lwc1 %[temp15], 1284(%[s0]) \n\t" | |
391 | "lwc1 %[temp16], 2824(%[v0]) \n\t" | |
392 | "lwc1 %[temp17], 1288(%[s0]) \n\t" | |
393 | "lwc1 %[temp18], 2828(%[v0]) \n\t" | |
394 | "lwc1 %[temp19], 1292(%[s0]) \n\t" | |
395 | "madd.s %[temp0], %[temp0], %[temp4], %[temp5] \n\t" | |
396 | "lwc1 %[temp4], 3072(%[v0]) \n\t" | |
397 | "madd.s %[temp1], %[temp1], %[temp6], %[temp7] \n\t" | |
398 | "lwc1 %[temp5], 1536(%[s0]) \n\t" | |
399 | "madd.s %[temp2], %[temp2], %[temp8], %[temp9] \n\t" | |
400 | "lwc1 %[temp6], 3076(%[v0]) \n\t" | |
401 | "madd.s %[temp3], %[temp3], %[temp10], %[temp11] \n\t" | |
402 | "lwc1 %[temp7], 1540(%[s0]) \n\t" | |
403 | "madd.s %[temp0], %[temp0], %[temp12], %[temp13] \n\t" | |
404 | "lwc1 %[temp8], 3080(%[v0]) \n\t" | |
405 | "madd.s %[temp1], %[temp1], %[temp14], %[temp15] \n\t" | |
406 | "lwc1 %[temp9], 1544(%[s0]) \n\t" | |
407 | "madd.s %[temp2], %[temp2], %[temp16], %[temp17] \n\t" | |
408 | "lwc1 %[temp10], 3084(%[v0]) \n\t" | |
409 | "madd.s %[temp3], %[temp3], %[temp18], %[temp19] \n\t" | |
410 | "lwc1 %[temp11], 1548(%[s0]) \n\t" | |
411 | "lwc1 %[temp12], 3840(%[v0]) \n\t" | |
412 | "lwc1 %[temp13], 1792(%[s0]) \n\t" | |
413 | "lwc1 %[temp14], 3844(%[v0]) \n\t" | |
414 | "lwc1 %[temp15], 1796(%[s0]) \n\t" | |
415 | "lwc1 %[temp16], 3848(%[v0]) \n\t" | |
416 | "lwc1 %[temp17], 1800(%[s0]) \n\t" | |
417 | "lwc1 %[temp18], 3852(%[v0]) \n\t" | |
418 | "lwc1 %[temp19], 1804(%[s0]) \n\t" | |
419 | "madd.s %[temp0], %[temp0], %[temp4], %[temp5] \n\t" | |
420 | "lwc1 %[temp4], 4096(%[v0]) \n\t" | |
421 | "madd.s %[temp1], %[temp1], %[temp6], %[temp7] \n\t" | |
422 | "lwc1 %[temp5], 2048(%[s0]) \n\t" | |
423 | "madd.s %[temp2], %[temp2], %[temp8], %[temp9] \n\t" | |
424 | "lwc1 %[temp6], 4100(%[v0]) \n\t" | |
425 | "madd.s %[temp3], %[temp3], %[temp10], %[temp11] \n\t" | |
426 | "lwc1 %[temp7], 2052(%[s0]) \n\t" | |
427 | "madd.s %[temp0], %[temp0], %[temp12], %[temp13] \n\t" | |
428 | "lwc1 %[temp8], 4104(%[v0]) \n\t" | |
429 | "madd.s %[temp1], %[temp1], %[temp14], %[temp15] \n\t" | |
430 | "lwc1 %[temp9], 2056(%[s0]) \n\t" | |
431 | "madd.s %[temp2], %[temp2], %[temp16], %[temp17] \n\t" | |
432 | "lwc1 %[temp10], 4108(%[v0]) \n\t" | |
433 | "madd.s %[temp3], %[temp3], %[temp18], %[temp19] \n\t" | |
434 | "lwc1 %[temp11], 2060(%[s0]) \n\t" | |
435 | "lwc1 %[temp12], 4864(%[v0]) \n\t" | |
436 | "lwc1 %[temp13], 2304(%[s0]) \n\t" | |
437 | "lwc1 %[temp14], 4868(%[v0]) \n\t" | |
438 | "lwc1 %[temp15], 2308(%[s0]) \n\t" | |
439 | "madd.s %[temp0], %[temp0], %[temp4], %[temp5] \n\t" | |
440 | "lwc1 %[temp16], 4872(%[v0]) \n\t" | |
441 | "madd.s %[temp1], %[temp1], %[temp6], %[temp7] \n\t" | |
442 | "lwc1 %[temp17], 2312(%[s0]) \n\t" | |
443 | "madd.s %[temp2], %[temp2], %[temp8], %[temp9] \n\t" | |
444 | "lwc1 %[temp18], 4876(%[v0]) \n\t" | |
445 | "madd.s %[temp3], %[temp3], %[temp10], %[temp11] \n\t" | |
446 | "lwc1 %[temp19], 2316(%[s0]) \n\t" | |
447 | "madd.s %[temp0], %[temp0], %[temp12], %[temp13] \n\t" | |
448 | "addiu %[dst], %[dst], 16 \n\t" | |
449 | "madd.s %[temp1], %[temp1], %[temp14], %[temp15] \n\t" | |
450 | "madd.s %[temp2], %[temp2], %[temp16], %[temp17] \n\t" | |
451 | "madd.s %[temp3], %[temp3], %[temp18], %[temp19] \n\t" | |
452 | "swc1 %[temp0], -16(%[dst]) \n\t" | |
453 | "swc1 %[temp1], -12(%[dst]) \n\t" | |
454 | "swc1 %[temp2], -8(%[dst]) \n\t" | |
455 | "swc1 %[temp3], -4(%[dst]) \n\t" | |
456 | ".set pop \n\t" | |
457 | ||
458 | : [dst]"+r"(dst), [v0]"+r"(vv0), [s0]"+r"(s0), | |
459 | [temp0]"=&f"(temp0), [temp1]"=&f"(temp1), [temp2]"=&f"(temp2), | |
460 | [temp3]"=&f"(temp3), [temp4]"=&f"(temp4), [temp5]"=&f"(temp5), | |
461 | [temp6]"=&f"(temp6), [temp7]"=&f"(temp7), [temp8]"=&f"(temp8), | |
462 | [temp9]"=&f"(temp9), [temp10]"=&f"(temp10), [temp11]"=&f"(temp11), | |
463 | [temp12]"=&f"(temp12), [temp13]"=&f"(temp13), [temp14]"=&f"(temp14), | |
464 | [temp15]"=&f"(temp15), [temp16]"=&f"(temp16), [temp17]"=&f"(temp17), | |
465 | [temp18]"=&f"(temp18), [temp19]"=&f"(temp19) | |
466 | : [v0_end]"r"(v0_end) | |
467 | : "memory" | |
468 | ); | |
469 | } | |
470 | else | |
471 | { | |
472 | fdsp->vector_fmul (out, v , sbr_qmf_window , 64 >> div); | |
473 | fdsp->vector_fmul_add(out, v + ( 192 >> div), sbr_qmf_window + ( 64 >> div), out , 64 >> div); | |
474 | fdsp->vector_fmul_add(out, v + ( 256 >> div), sbr_qmf_window + (128 >> div), out , 64 >> div); | |
475 | fdsp->vector_fmul_add(out, v + ( 448 >> div), sbr_qmf_window + (192 >> div), out , 64 >> div); | |
476 | fdsp->vector_fmul_add(out, v + ( 512 >> div), sbr_qmf_window + (256 >> div), out , 64 >> div); | |
477 | fdsp->vector_fmul_add(out, v + ( 704 >> div), sbr_qmf_window + (320 >> div), out , 64 >> div); | |
478 | fdsp->vector_fmul_add(out, v + ( 768 >> div), sbr_qmf_window + (384 >> div), out , 64 >> div); | |
479 | fdsp->vector_fmul_add(out, v + ( 960 >> div), sbr_qmf_window + (448 >> div), out , 64 >> div); | |
480 | fdsp->vector_fmul_add(out, v + (1024 >> div), sbr_qmf_window + (512 >> div), out , 64 >> div); | |
481 | fdsp->vector_fmul_add(out, v + (1216 >> div), sbr_qmf_window + (576 >> div), out , 64 >> div); | |
482 | out += 64 >> div; | |
483 | } | |
484 | } | |
485 | } | |
486 | ||
487 | #define sbr_qmf_analysis sbr_qmf_analysis_mips | |
488 | #define sbr_qmf_synthesis sbr_qmf_synthesis_mips | |
489 | ||
490 | #endif /* HAVE_MIPSFPU */ | |
491 | #endif /* HAVE_INLINE_ASM */ | |
492 | ||
493 | #endif /* AVCODEC_MIPS_AACSBR_FLOAT_H */ |