62 static void sbr_qmf_pre_shuffle_mips(
float *z)
64 int Temp1, Temp2, Temp3, Temp4, Temp5, Temp6;
72 "lui %[Temp6], 0x8000 \n\t"
74 "lw %[Temp1], 0(%[z2]) \n\t"
75 "lw %[Temp2], 4(%[z2]) \n\t"
76 "lw %[Temp3], 8(%[z2]) \n\t"
77 "lw %[Temp4], 12(%[z2]) \n\t"
78 "lw %[Temp5], 16(%[z2]) \n\t"
79 "xor %[Temp1], %[Temp1], %[Temp6] \n\t"
80 "xor %[Temp2], %[Temp2], %[Temp6] \n\t"
81 "xor %[Temp3], %[Temp3], %[Temp6] \n\t"
82 "xor %[Temp4], %[Temp4], %[Temp6] \n\t"
83 "xor %[Temp5], %[Temp5], %[Temp6] \n\t"
85 "sw %[Temp1], 32(%[z1]) \n\t"
86 "sw %[Temp2], 24(%[z1]) \n\t"
87 "sw %[Temp3], 16(%[z1]) \n\t"
88 "sw %[Temp4], 8(%[z1]) \n\t"
89 "sw %[Temp5], 0(%[z1]) \n\t"
90 "lw %[Temp1], 0(%[z3]) \n\t"
91 "lw %[Temp2], 4(%[z3]) \n\t"
92 "lw %[Temp3], 8(%[z3]) \n\t"
93 "lw %[Temp4], 12(%[z3]) \n\t"
94 "lw %[Temp5], 16(%[z3]) \n\t"
95 "sw %[Temp1], 4(%[z1]) \n\t"
96 "sw %[Temp2], 12(%[z1]) \n\t"
97 "sw %[Temp3], 20(%[z1]) \n\t"
98 "sw %[Temp4], 28(%[z1]) \n\t"
99 "sw %[Temp5], 36(%[z1]) \n\t"
102 "bne %[z1], %[z4], 1b \n\t"
103 "lw %[Temp1], 132(%[z]) \n\t"
104 "lw %[Temp2], 128(%[z]) \n\t"
105 "lw %[Temp3], 0(%[z]) \n\t"
106 "lw %[Temp4], 4(%[z]) \n\t"
107 "xor %[Temp1], %[Temp1], %[Temp6] \n\t"
108 "sw %[Temp1], 504(%[z]) \n\t"
109 "sw %[Temp2], 508(%[z]) \n\t"
110 "sw %[Temp3], 256(%[z]) \n\t"
111 "sw %[Temp4], 260(%[z]) \n\t"
113 : [Temp1]
"=&r"(Temp1), [Temp2]
"=&r"(Temp2),
114 [Temp3]
"=&r"(Temp3), [Temp4]
"=&r"(Temp4),
115 [Temp5]
"=&r"(Temp5), [Temp6]
"=&r"(Temp6),
116 [z1]
"+r"(z1), [z2]
"+r"(z2), [z3]
"+r"(z3)
117 : [z4]
"r"(z4), [z]
"r"(z)
122 static void sbr_qmf_post_shuffle_mips(
float W[32][2],
const float *z)
124 int Temp1, Temp2, Temp3, Temp4, Temp5;
125 float *W_ptr = (
float *)W;
126 float *z1 = (
float *)z;
127 float *z2 = (
float *)&z[60];
128 float *z_end = z1 + 32;
132 "lui %[Temp5], 0x8000 \n\t"
134 "lw %[Temp1], 0(%[z2]) \n\t"
135 "lw %[Temp2], 4(%[z2]) \n\t"
136 "lw %[Temp3], 8(%[z2]) \n\t"
137 "lw %[Temp4], 12(%[z2]) \n\t"
138 "xor %[Temp1], %[Temp1], %[Temp5] \n\t"
139 "xor %[Temp2], %[Temp2], %[Temp5] \n\t"
140 "xor %[Temp3], %[Temp3], %[Temp5] \n\t"
141 "xor %[Temp4], %[Temp4], %[Temp5] \n\t"
143 "sw %[Temp1], 24(%[W_ptr]) \n\t"
144 "sw %[Temp2], 16(%[W_ptr]) \n\t"
145 "sw %[Temp3], 8(%[W_ptr]) \n\t"
146 "sw %[Temp4], 0(%[W_ptr]) \n\t"
147 "lw %[Temp1], 0(%[z1]) \n\t"
148 "lw %[Temp2], 4(%[z1]) \n\t"
149 "lw %[Temp3], 8(%[z1]) \n\t"
150 "lw %[Temp4], 12(%[z1]) \n\t"
151 "sw %[Temp1], 4(%[W_ptr]) \n\t"
152 "sw %[Temp2], 12(%[W_ptr]) \n\t"
153 "sw %[Temp3], 20(%[W_ptr]) \n\t"
154 "sw %[Temp4], 28(%[W_ptr]) \n\t"
157 "bne %[z1], %[z_end], 1b \n\t"
159 : [Temp1]
"=&r"(Temp1), [Temp2]
"=&r"(Temp2),
160 [Temp3]
"=&r"(Temp3), [Temp4]
"=&r"(Temp4),
161 [Temp5]
"=&r"(Temp5), [z1]
"+r"(z1),
162 [z2]
"+r"(z2), [W_ptr]
"+r"(W_ptr)
169 static void sbr_sum64x5_mips(
float *z)
173 float f1, f2, f3, f4, f5, f6, f7, f8;
174 for (k = 0; k < 64; k += 8) {
180 "lwc1 $f0, 0(%[z1]) \n\t"
181 "lwc1 $f1, 256(%[z1]) \n\t"
182 "lwc1 $f2, 4(%[z1]) \n\t"
183 "lwc1 $f3, 260(%[z1]) \n\t"
184 "lwc1 $f4, 8(%[z1]) \n\t"
185 "add.s %[f1], $f0, $f1 \n\t"
186 "lwc1 $f5, 264(%[z1]) \n\t"
187 "add.s %[f2], $f2, $f3 \n\t"
188 "lwc1 $f6, 12(%[z1]) \n\t"
189 "lwc1 $f7, 268(%[z1]) \n\t"
190 "add.s %[f3], $f4, $f5 \n\t"
191 "lwc1 $f8, 16(%[z1]) \n\t"
192 "lwc1 $f9, 272(%[z1]) \n\t"
193 "add.s %[f4], $f6, $f7 \n\t"
194 "lwc1 $f10, 20(%[z1]) \n\t"
195 "lwc1 $f11, 276(%[z1]) \n\t"
196 "add.s %[f5], $f8, $f9 \n\t"
197 "lwc1 $f12, 24(%[z1]) \n\t"
198 "lwc1 $f13, 280(%[z1]) \n\t"
199 "add.s %[f6], $f10, $f11 \n\t"
200 "lwc1 $f14, 28(%[z1]) \n\t"
201 "lwc1 $f15, 284(%[z1]) \n\t"
202 "add.s %[f7], $f12, $f13 \n\t"
203 "lwc1 $f0, 512(%[z1]) \n\t"
204 "lwc1 $f1, 516(%[z1]) \n\t"
205 "add.s %[f8], $f14, $f15 \n\t"
206 "lwc1 $f2, 520(%[z1]) \n\t"
207 "add.s %[f1], %[f1], $f0 \n\t"
208 "add.s %[f2], %[f2], $f1 \n\t"
209 "lwc1 $f3, 524(%[z1]) \n\t"
210 "add.s %[f3], %[f3], $f2 \n\t"
211 "lwc1 $f4, 528(%[z1]) \n\t"
212 "lwc1 $f5, 532(%[z1]) \n\t"
213 "add.s %[f4], %[f4], $f3 \n\t"
214 "lwc1 $f6, 536(%[z1]) \n\t"
215 "add.s %[f5], %[f5], $f4 \n\t"
216 "add.s %[f6], %[f6], $f5 \n\t"
217 "lwc1 $f7, 540(%[z1]) \n\t"
218 "add.s %[f7], %[f7], $f6 \n\t"
219 "lwc1 $f0, 768(%[z1]) \n\t"
220 "lwc1 $f1, 772(%[z1]) \n\t"
221 "add.s %[f8], %[f8], $f7 \n\t"
222 "lwc1 $f2, 776(%[z1]) \n\t"
223 "add.s %[f1], %[f1], $f0 \n\t"
224 "add.s %[f2], %[f2], $f1 \n\t"
225 "lwc1 $f3, 780(%[z1]) \n\t"
226 "add.s %[f3], %[f3], $f2 \n\t"
227 "lwc1 $f4, 784(%[z1]) \n\t"
228 "lwc1 $f5, 788(%[z1]) \n\t"
229 "add.s %[f4], %[f4], $f3 \n\t"
230 "lwc1 $f6, 792(%[z1]) \n\t"
231 "add.s %[f5], %[f5], $f4 \n\t"
232 "add.s %[f6], %[f6], $f5 \n\t"
233 "lwc1 $f7, 796(%[z1]) \n\t"
234 "add.s %[f7], %[f7], $f6 \n\t"
235 "lwc1 $f0, 1024(%[z1]) \n\t"
236 "lwc1 $f1, 1028(%[z1]) \n\t"
237 "add.s %[f8], %[f8], $f7 \n\t"
238 "lwc1 $f2, 1032(%[z1]) \n\t"
239 "add.s %[f1], %[f1], $f0 \n\t"
240 "add.s %[f2], %[f2], $f1 \n\t"
241 "lwc1 $f3, 1036(%[z1]) \n\t"
242 "add.s %[f3], %[f3], $f2 \n\t"
243 "lwc1 $f4, 1040(%[z1]) \n\t"
244 "lwc1 $f5, 1044(%[z1]) \n\t"
245 "add.s %[f4], %[f4], $f3 \n\t"
246 "lwc1 $f6, 1048(%[z1]) \n\t"
247 "add.s %[f5], %[f5], $f4 \n\t"
248 "add.s %[f6], %[f6], $f5 \n\t"
249 "lwc1 $f7, 1052(%[z1]) \n\t"
250 "add.s %[f7], %[f7], $f6 \n\t"
251 "swc1 %[f1], 0(%[z1]) \n\t"
252 "swc1 %[f2], 4(%[z1]) \n\t"
253 "add.s %[f8], %[f8], $f7 \n\t"
254 "swc1 %[f3], 8(%[z1]) \n\t"
255 "swc1 %[f4], 12(%[z1]) \n\t"
256 "swc1 %[f5], 16(%[z1]) \n\t"
257 "swc1 %[f6], 20(%[z1]) \n\t"
258 "swc1 %[f7], 24(%[z1]) \n\t"
259 "swc1 %[f8], 28(%[z1]) \n\t"
261 : [f1]
"=&f"(f1), [f2]
"=&f"(f2), [f3]
"=&f"(f3),
262 [f4]
"=&f"(f4), [f5]
"=&f"(f5), [f6]
"=&f"(f6),
263 [f7]
"=&f"(f7), [f8]
"=&f"(f8)
265 :
"$f0",
"$f1",
"$f2",
"$f3",
"$f4",
"$f5",
266 "$f6",
"$f7",
"$f8",
"$f9",
"$f10",
"$f11",
267 "$f12",
"$f13",
"$f14",
"$f15",
273 static float sbr_sum_square_mips(
float (*x)[2],
int n)
275 float sum0 = 0.0f, sum1 = 0.0f;
277 float temp0, temp1, temp2, temp3;
280 loop_end = p_x + (n >> 1)*4 - 4;
284 ".set noreorder \n\t"
285 "lwc1 %[temp0], 0(%[p_x]) \n\t"
286 "lwc1 %[temp1], 4(%[p_x]) \n\t"
287 "lwc1 %[temp2], 8(%[p_x]) \n\t"
288 "lwc1 %[temp3], 12(%[p_x]) \n\t"
291 "madd.s %[sum0], %[sum0], %[temp0], %[temp0] \n\t"
292 "lwc1 %[temp0], 0(%[p_x]) \n\t"
293 "madd.s %[sum1], %[sum1], %[temp1], %[temp1] \n\t"
294 "lwc1 %[temp1], 4(%[p_x]) \n\t"
295 "madd.s %[sum0], %[sum0], %[temp2], %[temp2] \n\t"
296 "lwc1 %[temp2], 8(%[p_x]) \n\t"
297 "madd.s %[sum1], %[sum1], %[temp3], %[temp3] \n\t"
298 "bne %[p_x], %[loop_end], 1b \n\t"
299 " lwc1 %[temp3], 12(%[p_x]) \n\t"
300 "madd.s %[sum0], %[sum0], %[temp0], %[temp0] \n\t"
301 "madd.s %[sum1], %[sum1], %[temp1], %[temp1] \n\t"
302 "madd.s %[sum0], %[sum0], %[temp2], %[temp2] \n\t"
303 "madd.s %[sum1], %[sum1], %[temp3], %[temp3] \n\t"
306 : [temp0]
"=&f"(temp0), [temp1]
"=&f"(temp1), [temp2]
"=&f"(temp2),
307 [temp3]
"=&f"(temp3), [sum0]
"+f"(sum0), [sum1]
"+f"(sum1),
309 : [loop_end]
"r"(loop_end)
315 static void sbr_qmf_deint_bfly_mips(
float *
v,
const float *src0,
const float *src1)
318 float temp0, temp1, temp2, temp3, temp4, temp5;
319 float temp6, temp7, temp8, temp9, temp10, temp11;
322 float *psrc0 = (
float*)src0;
323 float *psrc1 = (
float*)&src1[63];
325 for (i = 0; i < 4; i++) {
329 "lwc1 %[temp0], 0(%[src0]) \n\t"
330 "lwc1 %[temp1], 0(%[src1]) \n\t"
331 "lwc1 %[temp3], 4(%[src0]) \n\t"
332 "lwc1 %[temp4], -4(%[src1]) \n\t"
333 "lwc1 %[temp6], 8(%[src0]) \n\t"
334 "lwc1 %[temp7], -8(%[src1]) \n\t"
335 "lwc1 %[temp9], 12(%[src0]) \n\t"
336 "lwc1 %[temp10], -12(%[src1]) \n\t"
337 "add.s %[temp2], %[temp0], %[temp1] \n\t"
338 "add.s %[temp5], %[temp3], %[temp4] \n\t"
339 "add.s %[temp8], %[temp6], %[temp7] \n\t"
340 "add.s %[temp11], %[temp9], %[temp10] \n\t"
341 "sub.s %[temp0], %[temp0], %[temp1] \n\t"
342 "sub.s %[temp3], %[temp3], %[temp4] \n\t"
343 "sub.s %[temp6], %[temp6], %[temp7] \n\t"
344 "sub.s %[temp9], %[temp9], %[temp10] \n\t"
345 "swc1 %[temp2], 0(%[v1]) \n\t"
346 "swc1 %[temp0], 0(%[v0]) \n\t"
347 "swc1 %[temp5], -4(%[v1]) \n\t"
348 "swc1 %[temp3], 4(%[v0]) \n\t"
349 "swc1 %[temp8], -8(%[v1]) \n\t"
350 "swc1 %[temp6], 8(%[v0]) \n\t"
351 "swc1 %[temp11], -12(%[v1]) \n\t"
352 "swc1 %[temp9], 12(%[v0]) \n\t"
353 "lwc1 %[temp0], 16(%[src0]) \n\t"
354 "lwc1 %[temp1], -16(%[src1]) \n\t"
355 "lwc1 %[temp3], 20(%[src0]) \n\t"
356 "lwc1 %[temp4], -20(%[src1]) \n\t"
357 "lwc1 %[temp6], 24(%[src0]) \n\t"
358 "lwc1 %[temp7], -24(%[src1]) \n\t"
359 "lwc1 %[temp9], 28(%[src0]) \n\t"
360 "lwc1 %[temp10], -28(%[src1]) \n\t"
361 "add.s %[temp2], %[temp0], %[temp1] \n\t"
362 "add.s %[temp5], %[temp3], %[temp4] \n\t"
363 "add.s %[temp8], %[temp6], %[temp7] \n\t"
364 "add.s %[temp11], %[temp9], %[temp10] \n\t"
365 "sub.s %[temp0], %[temp0], %[temp1] \n\t"
366 "sub.s %[temp3], %[temp3], %[temp4] \n\t"
367 "sub.s %[temp6], %[temp6], %[temp7] \n\t"
368 "sub.s %[temp9], %[temp9], %[temp10] \n\t"
369 "swc1 %[temp2], -16(%[v1]) \n\t"
370 "swc1 %[temp0], 16(%[v0]) \n\t"
371 "swc1 %[temp5], -20(%[v1]) \n\t"
372 "swc1 %[temp3], 20(%[v0]) \n\t"
373 "swc1 %[temp8], -24(%[v1]) \n\t"
374 "swc1 %[temp6], 24(%[v0]) \n\t"
375 "swc1 %[temp11], -28(%[v1]) \n\t"
376 "swc1 %[temp9], 28(%[v0]) \n\t"
377 "lwc1 %[temp0], 32(%[src0]) \n\t"
378 "lwc1 %[temp1], -32(%[src1]) \n\t"
379 "lwc1 %[temp3], 36(%[src0]) \n\t"
380 "lwc1 %[temp4], -36(%[src1]) \n\t"
381 "lwc1 %[temp6], 40(%[src0]) \n\t"
382 "lwc1 %[temp7], -40(%[src1]) \n\t"
383 "lwc1 %[temp9], 44(%[src0]) \n\t"
384 "lwc1 %[temp10], -44(%[src1]) \n\t"
385 "add.s %[temp2], %[temp0], %[temp1] \n\t"
386 "add.s %[temp5], %[temp3], %[temp4] \n\t"
387 "add.s %[temp8], %[temp6], %[temp7] \n\t"
388 "add.s %[temp11], %[temp9], %[temp10] \n\t"
389 "sub.s %[temp0], %[temp0], %[temp1] \n\t"
390 "sub.s %[temp3], %[temp3], %[temp4] \n\t"
391 "sub.s %[temp6], %[temp6], %[temp7] \n\t"
392 "sub.s %[temp9], %[temp9], %[temp10] \n\t"
393 "swc1 %[temp2], -32(%[v1]) \n\t"
394 "swc1 %[temp0], 32(%[v0]) \n\t"
395 "swc1 %[temp5], -36(%[v1]) \n\t"
396 "swc1 %[temp3], 36(%[v0]) \n\t"
397 "swc1 %[temp8], -40(%[v1]) \n\t"
398 "swc1 %[temp6], 40(%[v0]) \n\t"
399 "swc1 %[temp11], -44(%[v1]) \n\t"
400 "swc1 %[temp9], 44(%[v0]) \n\t"
401 "lwc1 %[temp0], 48(%[src0]) \n\t"
402 "lwc1 %[temp1], -48(%[src1]) \n\t"
403 "lwc1 %[temp3], 52(%[src0]) \n\t"
404 "lwc1 %[temp4], -52(%[src1]) \n\t"
405 "lwc1 %[temp6], 56(%[src0]) \n\t"
406 "lwc1 %[temp7], -56(%[src1]) \n\t"
407 "lwc1 %[temp9], 60(%[src0]) \n\t"
408 "lwc1 %[temp10], -60(%[src1]) \n\t"
409 "add.s %[temp2], %[temp0], %[temp1] \n\t"
410 "add.s %[temp5], %[temp3], %[temp4] \n\t"
411 "add.s %[temp8], %[temp6], %[temp7] \n\t"
412 "add.s %[temp11], %[temp9], %[temp10] \n\t"
413 "sub.s %[temp0], %[temp0], %[temp1] \n\t"
414 "sub.s %[temp3], %[temp3], %[temp4] \n\t"
415 "sub.s %[temp6], %[temp6], %[temp7] \n\t"
416 "sub.s %[temp9], %[temp9], %[temp10] \n\t"
417 "swc1 %[temp2], -48(%[v1]) \n\t"
418 "swc1 %[temp0], 48(%[v0]) \n\t"
419 "swc1 %[temp5], -52(%[v1]) \n\t"
420 "swc1 %[temp3], 52(%[v0]) \n\t"
421 "swc1 %[temp8], -56(%[v1]) \n\t"
422 "swc1 %[temp6], 56(%[v0]) \n\t"
423 "swc1 %[temp11], -60(%[v1]) \n\t"
424 "swc1 %[temp9], 60(%[v0]) \n\t"
430 : [
v0]
"+r"(
v0), [v1]
"+r"(v1), [src0]
"+r"(psrc0), [src1]
"+r"(psrc1),
431 [temp0]
"=&f"(temp0), [temp1]
"=&f"(temp1), [temp2]
"=&f"(temp2),
432 [temp3]
"=&f"(temp3), [temp4]
"=&f"(temp4), [temp5]
"=&f"(temp5),
433 [temp6]
"=&f"(temp6), [temp7]
"=&f"(temp7), [temp8]
"=&f"(temp8),
434 [temp9]
"=&f"(temp9), [temp10]
"=&f"(temp10), [temp11]
"=&f"(temp11)
442 static void sbr_autocorrelate_mips(
const float x[40][2],
float phi[3][2][2])
445 float real_sum_0 = 0.0f;
446 float real_sum_1 = 0.0f;
447 float real_sum_2 = 0.0f;
448 float imag_sum_1 = 0.0f;
449 float imag_sum_2 = 0.0f;
451 float temp0, temp1, temp2, temp3, temp4, temp5, temp6;
452 float temp7, temp_r, temp_r1, temp_r2, temp_r3, temp_r4;
453 p_x = (
float*)&x[0][0];
454 p_phi = &phi[0][0][0];
457 "lwc1 %[temp0], 8(%[p_x]) \n\t"
458 "lwc1 %[temp1], 12(%[p_x]) \n\t"
459 "lwc1 %[temp2], 16(%[p_x]) \n\t"
460 "lwc1 %[temp3], 20(%[p_x]) \n\t"
461 "lwc1 %[temp4], 24(%[p_x]) \n\t"
462 "lwc1 %[temp5], 28(%[p_x]) \n\t"
463 "mul.s %[temp_r], %[temp1], %[temp1] \n\t"
464 "mul.s %[temp_r1], %[temp1], %[temp3] \n\t"
465 "mul.s %[temp_r2], %[temp1], %[temp2] \n\t"
466 "mul.s %[temp_r3], %[temp1], %[temp5] \n\t"
467 "mul.s %[temp_r4], %[temp1], %[temp4] \n\t"
468 "madd.s %[temp_r], %[temp_r], %[temp0], %[temp0] \n\t"
469 "madd.s %[temp_r1], %[temp_r1], %[temp0], %[temp2] \n\t"
470 "msub.s %[temp_r2], %[temp_r2], %[temp0], %[temp3] \n\t"
471 "madd.s %[temp_r3], %[temp_r3], %[temp0], %[temp4] \n\t"
472 "msub.s %[temp_r4], %[temp_r4], %[temp0], %[temp5] \n\t"
473 "add.s %[real_sum_0], %[real_sum_0], %[temp_r] \n\t"
474 "add.s %[real_sum_1], %[real_sum_1], %[temp_r1] \n\t"
475 "add.s %[imag_sum_1], %[imag_sum_1], %[temp_r2] \n\t"
476 "add.s %[real_sum_2], %[real_sum_2], %[temp_r3] \n\t"
477 "add.s %[imag_sum_2], %[imag_sum_2], %[temp_r4] \n\t"
480 : [temp0]
"=&f"(temp0), [temp1]
"=&f"(temp1), [temp2]
"=&f"(temp2),
481 [temp3]
"=&f"(temp3), [temp4]
"=&f"(temp4), [temp5]
"=&f"(temp5),
482 [real_sum_0]
"+f"(real_sum_0), [real_sum_1]
"+f"(real_sum_1),
483 [imag_sum_1]
"+f"(imag_sum_1), [real_sum_2]
"+f"(real_sum_2),
484 [temp_r]
"=&f"(temp_r), [temp_r1]
"=&f"(temp_r1), [temp_r2]
"=&f"(temp_r2),
485 [temp_r3]
"=&f"(temp_r3), [temp_r4]
"=&f"(temp_r4),
486 [p_x]
"+r"(p_x), [imag_sum_2]
"+f"(imag_sum_2)
491 for (i = 0; i < 12; i++) {
493 "lwc1 %[temp0], 8(%[p_x]) \n\t"
494 "lwc1 %[temp1], 12(%[p_x]) \n\t"
495 "lwc1 %[temp2], 16(%[p_x]) \n\t"
496 "lwc1 %[temp3], 20(%[p_x]) \n\t"
497 "lwc1 %[temp4], 24(%[p_x]) \n\t"
498 "lwc1 %[temp5], 28(%[p_x]) \n\t"
499 "mul.s %[temp_r], %[temp1], %[temp1] \n\t"
500 "mul.s %[temp_r1], %[temp1], %[temp3] \n\t"
501 "mul.s %[temp_r2], %[temp1], %[temp2] \n\t"
502 "mul.s %[temp_r3], %[temp1], %[temp5] \n\t"
503 "mul.s %[temp_r4], %[temp1], %[temp4] \n\t"
504 "madd.s %[temp_r], %[temp_r], %[temp0], %[temp0] \n\t"
505 "madd.s %[temp_r1], %[temp_r1], %[temp0], %[temp2] \n\t"
506 "msub.s %[temp_r2], %[temp_r2], %[temp0], %[temp3] \n\t"
507 "madd.s %[temp_r3], %[temp_r3], %[temp0], %[temp4] \n\t"
508 "msub.s %[temp_r4], %[temp_r4], %[temp0], %[temp5] \n\t"
509 "add.s %[real_sum_0], %[real_sum_0], %[temp_r] \n\t"
510 "add.s %[real_sum_1], %[real_sum_1], %[temp_r1] \n\t"
511 "add.s %[imag_sum_1], %[imag_sum_1], %[temp_r2] \n\t"
512 "add.s %[real_sum_2], %[real_sum_2], %[temp_r3] \n\t"
513 "add.s %[imag_sum_2], %[imag_sum_2], %[temp_r4] \n\t"
514 "lwc1 %[temp0], 32(%[p_x]) \n\t"
515 "lwc1 %[temp1], 36(%[p_x]) \n\t"
516 "mul.s %[temp_r], %[temp3], %[temp3] \n\t"
517 "mul.s %[temp_r1], %[temp3], %[temp5] \n\t"
518 "mul.s %[temp_r2], %[temp3], %[temp4] \n\t"
519 "mul.s %[temp_r3], %[temp3], %[temp1] \n\t"
520 "mul.s %[temp_r4], %[temp3], %[temp0] \n\t"
521 "madd.s %[temp_r], %[temp_r], %[temp2], %[temp2] \n\t"
522 "madd.s %[temp_r1], %[temp_r1], %[temp2], %[temp4] \n\t"
523 "msub.s %[temp_r2], %[temp_r2], %[temp2], %[temp5] \n\t"
524 "madd.s %[temp_r3], %[temp_r3], %[temp2], %[temp0] \n\t"
525 "msub.s %[temp_r4], %[temp_r4], %[temp2], %[temp1] \n\t"
526 "add.s %[real_sum_0], %[real_sum_0], %[temp_r] \n\t"
527 "add.s %[real_sum_1], %[real_sum_1], %[temp_r1] \n\t"
528 "add.s %[imag_sum_1], %[imag_sum_1], %[temp_r2] \n\t"
529 "add.s %[real_sum_2], %[real_sum_2], %[temp_r3] \n\t"
530 "add.s %[imag_sum_2], %[imag_sum_2], %[temp_r4] \n\t"
531 "lwc1 %[temp2], 40(%[p_x]) \n\t"
532 "lwc1 %[temp3], 44(%[p_x]) \n\t"
533 "mul.s %[temp_r], %[temp5], %[temp5] \n\t"
534 "mul.s %[temp_r1], %[temp5], %[temp1] \n\t"
535 "mul.s %[temp_r2], %[temp5], %[temp0] \n\t"
536 "mul.s %[temp_r3], %[temp5], %[temp3] \n\t"
537 "mul.s %[temp_r4], %[temp5], %[temp2] \n\t"
538 "madd.s %[temp_r], %[temp_r], %[temp4], %[temp4] \n\t"
539 "madd.s %[temp_r1], %[temp_r1], %[temp4], %[temp0] \n\t"
540 "msub.s %[temp_r2], %[temp_r2], %[temp4], %[temp1] \n\t"
541 "madd.s %[temp_r3], %[temp_r3], %[temp4], %[temp2] \n\t"
542 "msub.s %[temp_r4], %[temp_r4], %[temp4], %[temp3] \n\t"
543 "add.s %[real_sum_0], %[real_sum_0], %[temp_r] \n\t"
544 "add.s %[real_sum_1], %[real_sum_1], %[temp_r1] \n\t"
545 "add.s %[imag_sum_1], %[imag_sum_1], %[temp_r2] \n\t"
546 "add.s %[real_sum_2], %[real_sum_2], %[temp_r3] \n\t"
547 "add.s %[imag_sum_2], %[imag_sum_2], %[temp_r4] \n\t"
550 : [temp0]
"=&f"(temp0), [temp1]
"=&f"(temp1), [temp2]
"=&f"(temp2),
551 [temp3]
"=&f"(temp3), [temp4]
"=&f"(temp4), [temp5]
"=&f"(temp5),
552 [real_sum_0]
"+f"(real_sum_0), [real_sum_1]
"+f"(real_sum_1),
553 [imag_sum_1]
"+f"(imag_sum_1), [real_sum_2]
"+f"(real_sum_2),
554 [temp_r]
"=&f"(temp_r), [temp_r1]
"=&f"(temp_r1),
555 [temp_r2]
"=&f"(temp_r2), [temp_r3]
"=&f"(temp_r3),
556 [temp_r4]
"=&f"(temp_r4), [p_x]
"+r"(p_x),
557 [imag_sum_2]
"+f"(imag_sum_2)
563 "lwc1 %[temp0], -296(%[p_x]) \n\t"
564 "lwc1 %[temp1], -292(%[p_x]) \n\t"
565 "lwc1 %[temp2], 8(%[p_x]) \n\t"
566 "lwc1 %[temp3], 12(%[p_x]) \n\t"
567 "lwc1 %[temp4], -288(%[p_x]) \n\t"
568 "lwc1 %[temp5], -284(%[p_x]) \n\t"
569 "lwc1 %[temp6], -280(%[p_x]) \n\t"
570 "lwc1 %[temp7], -276(%[p_x]) \n\t"
571 "madd.s %[temp_r], %[real_sum_0], %[temp0], %[temp0] \n\t"
572 "madd.s %[temp_r1], %[real_sum_0], %[temp2], %[temp2] \n\t"
573 "madd.s %[temp_r2], %[real_sum_1], %[temp0], %[temp4] \n\t"
574 "madd.s %[temp_r3], %[imag_sum_1], %[temp0], %[temp5] \n\t"
575 "madd.s %[temp_r], %[temp_r], %[temp1], %[temp1] \n\t"
576 "madd.s %[temp_r1], %[temp_r1], %[temp3], %[temp3] \n\t"
577 "madd.s %[temp_r2], %[temp_r2], %[temp1], %[temp5] \n\t"
578 "nmsub.s %[temp_r3], %[temp_r3], %[temp1], %[temp4] \n\t"
579 "lwc1 %[temp4], 16(%[p_x]) \n\t"
580 "lwc1 %[temp5], 20(%[p_x]) \n\t"
581 "swc1 %[temp_r], 40(%[p_phi]) \n\t"
582 "swc1 %[temp_r1], 16(%[p_phi]) \n\t"
583 "swc1 %[temp_r2], 24(%[p_phi]) \n\t"
584 "swc1 %[temp_r3], 28(%[p_phi]) \n\t"
585 "madd.s %[temp_r], %[real_sum_1], %[temp2], %[temp4] \n\t"
586 "madd.s %[temp_r1], %[imag_sum_1], %[temp2], %[temp5] \n\t"
587 "madd.s %[temp_r2], %[real_sum_2], %[temp0], %[temp6] \n\t"
588 "madd.s %[temp_r3], %[imag_sum_2], %[temp0], %[temp7] \n\t"
589 "madd.s %[temp_r], %[temp_r], %[temp3], %[temp5] \n\t"
590 "nmsub.s %[temp_r1], %[temp_r1], %[temp3], %[temp4] \n\t"
591 "madd.s %[temp_r2], %[temp_r2], %[temp1], %[temp7] \n\t"
592 "nmsub.s %[temp_r3], %[temp_r3], %[temp1], %[temp6] \n\t"
593 "swc1 %[temp_r], 0(%[p_phi]) \n\t"
594 "swc1 %[temp_r1], 4(%[p_phi]) \n\t"
595 "swc1 %[temp_r2], 8(%[p_phi]) \n\t"
596 "swc1 %[temp_r3], 12(%[p_phi]) \n\t"
598 : [temp0]
"=&f"(temp0), [temp1]
"=&f"(temp1), [temp2]
"=&f"(temp2),
599 [temp3]
"=&f"(temp3), [temp4]
"=&f"(temp4), [temp5]
"=&f"(temp5),
600 [temp6]
"=&f"(temp6), [temp7]
"=&f"(temp7), [temp_r]
"=&f"(temp_r),
601 [real_sum_0]
"+f"(real_sum_0), [real_sum_1]
"+f"(real_sum_1),
602 [real_sum_2]
"+f"(real_sum_2), [imag_sum_1]
"+f"(imag_sum_1),
603 [temp_r2]
"=&f"(temp_r2), [temp_r3]
"=&f"(temp_r3),
604 [temp_r1]
"=&f"(temp_r1), [p_phi]
"+r"(p_phi),
605 [imag_sum_2]
"+f"(imag_sum_2)
612 static void sbr_hf_gen_mips(
float (*X_high)[2],
const float (*X_low)[2],
613 const float alpha0[2],
const float alpha1[2],
618 float *p_x_low = (
float*)&X_low[0][0] + 2*start;
619 float *p_x_high = &X_high[0][0] + 2*
start;
620 float temp0, temp1, temp2, temp3, temp4, temp5, temp6;
621 float temp7, temp8, temp9, temp10, temp11, temp12;
623 alpha[0] = alpha1[0] * bw * bw;
624 alpha[1] = alpha1[1] * bw * bw;
625 alpha[2] = alpha0[0] * bw;
626 alpha[3] = alpha0[1] * bw;
628 for (i = start; i <
end; i++) {
630 "lwc1 %[temp0], -16(%[p_x_low]) \n\t"
631 "lwc1 %[temp1], -12(%[p_x_low]) \n\t"
632 "lwc1 %[temp2], -8(%[p_x_low]) \n\t"
633 "lwc1 %[temp3], -4(%[p_x_low]) \n\t"
634 "lwc1 %[temp5], 0(%[p_x_low]) \n\t"
635 "lwc1 %[temp6], 4(%[p_x_low]) \n\t"
636 "lwc1 %[temp7], 0(%[alpha]) \n\t"
637 "lwc1 %[temp8], 4(%[alpha]) \n\t"
638 "lwc1 %[temp9], 8(%[alpha]) \n\t"
639 "lwc1 %[temp10], 12(%[alpha]) \n\t"
640 PTR_ADDIU "%[p_x_high], %[p_x_high], 8 \n\t"
641 PTR_ADDIU "%[p_x_low], %[p_x_low], 8 \n\t"
642 "mul.s %[temp11], %[temp1], %[temp8] \n\t"
643 "msub.s %[temp11], %[temp11], %[temp0], %[temp7] \n\t"
644 "madd.s %[temp11], %[temp11], %[temp2], %[temp9] \n\t"
645 "nmsub.s %[temp11], %[temp11], %[temp3], %[temp10] \n\t"
646 "add.s %[temp11], %[temp11], %[temp5] \n\t"
647 "swc1 %[temp11], -8(%[p_x_high]) \n\t"
648 "mul.s %[temp12], %[temp1], %[temp7] \n\t"
649 "madd.s %[temp12], %[temp12], %[temp0], %[temp8] \n\t"
650 "madd.s %[temp12], %[temp12], %[temp3], %[temp9] \n\t"
651 "madd.s %[temp12], %[temp12], %[temp2], %[temp10] \n\t"
652 "add.s %[temp12], %[temp12], %[temp6] \n\t"
653 "swc1 %[temp12], -4(%[p_x_high]) \n\t"
655 : [temp0]
"=&f"(temp0), [temp1]
"=&f"(temp1), [temp2]
"=&f"(temp2),
656 [temp3]
"=&f"(temp3), [temp4]
"=&f"(temp4), [temp5]
"=&f"(temp5),
657 [temp6]
"=&f"(temp6), [temp7]
"=&f"(temp7), [temp8]
"=&f"(temp8),
658 [temp9]
"=&f"(temp9), [temp10]
"=&f"(temp10), [temp11]
"=&f"(temp11),
659 [temp12]
"=&f"(temp12), [p_x_high]
"+r"(p_x_high),
660 [p_x_low]
"+r"(p_x_low)
667 static void sbr_hf_g_filt_mips(
float (*
Y)[2],
const float (*X_high)[40][2],
668 const float *g_filt,
int m_max, intptr_t ixh)
670 const float *p_x, *p_g, *loop_end;
672 float temp0, temp1, temp2;
676 p_x = &X_high[0][ixh][0];
677 loop_end = p_g + m_max;
681 ".set noreorder \n\t"
683 "lwc1 %[temp0], 0(%[p_g]) \n\t"
684 "lwc1 %[temp1], 0(%[p_x]) \n\t"
685 "lwc1 %[temp2], 4(%[p_x]) \n\t"
686 "mul.s %[temp1], %[temp1], %[temp0] \n\t"
687 "mul.s %[temp2], %[temp2], %[temp0] \n\t"
690 "swc1 %[temp1], 0(%[p_y]) \n\t"
691 "swc1 %[temp2], 4(%[p_y]) \n\t"
692 "bne %[p_g], %[loop_end], 1b \n\t"
696 : [temp0]
"=&f"(temp0), [temp1]
"=&f"(temp1),
697 [temp2]
"=&f"(temp2), [p_x]
"+r"(p_x),
698 [p_y]
"+r"(p_y), [p_g]
"+r"(p_g)
699 : [loop_end]
"r"(loop_end)
704 static void sbr_hf_apply_noise_0_mips(
float (*
Y)[2],
const float *s_m,
705 const float *q_filt,
int noise,
710 for (m = 0; m < m_max; m++){
714 float y0,y1, temp1, temp2, temp4, temp5;
716 const float *s_m1=&s_m[
m];
717 const float *q_filt1= &q_filt[
m];
720 "lwc1 %[y0], 0(%[Y1]) \n\t"
721 "lwc1 %[temp1], 0(%[s_m1]) \n\t"
722 "addiu %[noise], %[noise], 1 \n\t"
723 "andi %[noise], %[noise], 0x1ff \n\t"
724 "sll %[temp0], %[noise], 3 \n\t"
725 PTR_ADDU "%[ff_table],%[ff_sbr_noise_table], %[temp0] \n\t"
726 "add.s %[y0], %[y0], %[temp1] \n\t"
727 "mfc1 %[temp3], %[temp1] \n\t"
728 "bne %[temp3], $0, 1f \n\t"
729 "lwc1 %[y1], 4(%[Y1]) \n\t"
730 "lwc1 %[temp2], 0(%[q_filt1]) \n\t"
731 "lwc1 %[temp4], 0(%[ff_table]) \n\t"
732 "lwc1 %[temp5], 4(%[ff_table]) \n\t"
733 "madd.s %[y0], %[y0], %[temp2], %[temp4] \n\t"
734 "madd.s %[y1], %[y1], %[temp2], %[temp5] \n\t"
735 "swc1 %[y1], 4(%[Y1]) \n\t"
737 "swc1 %[y0], 0(%[Y1]) \n\t"
739 : [ff_table]
"=&r"(ff_table), [y0]
"=&f"(y0), [y1]
"=&f"(y1),
740 [temp0]
"=&r"(temp0), [temp1]
"=&f"(temp1), [temp2]
"=&f"(temp2),
741 [temp3]
"=&r"(temp3), [temp4]
"=&f"(temp4), [temp5]
"=&f"(temp5)
743 [Y1]
"r"(Y1), [s_m1]
"r"(s_m1), [q_filt1]
"r"(q_filt1)
749 static void sbr_hf_apply_noise_1_mips(
float (*
Y)[2],
const float *s_m,
750 const float *q_filt,
int noise,
753 float y0,y1,temp1, temp2, temp4, temp5;
755 float phi_sign = 1 - 2 * (kx & 1);
757 for (m = 0; m < m_max; m++) {
761 const float *s_m1=&s_m[
m];
762 const float *q_filt1= &q_filt[
m];
765 "lwc1 %[y1], 4(%[Y1]) \n\t"
766 "lwc1 %[temp1], 0(%[s_m1]) \n\t"
767 "lw %[temp3], 0(%[s_m1]) \n\t"
768 "addiu %[noise], %[noise], 1 \n\t"
769 "andi %[noise], %[noise], 0x1ff \n\t"
770 "sll %[temp0], %[noise], 3 \n\t"
771 PTR_ADDU "%[ff_table],%[ff_sbr_noise_table],%[temp0] \n\t"
772 "madd.s %[y1], %[y1], %[temp1], %[phi_sign] \n\t"
773 "bne %[temp3], $0, 1f \n\t"
774 "lwc1 %[y0], 0(%[Y1]) \n\t"
775 "lwc1 %[temp2], 0(%[q_filt1]) \n\t"
776 "lwc1 %[temp4], 0(%[ff_table]) \n\t"
777 "lwc1 %[temp5], 4(%[ff_table]) \n\t"
778 "madd.s %[y0], %[y0], %[temp2], %[temp4] \n\t"
779 "madd.s %[y1], %[y1], %[temp2], %[temp5] \n\t"
780 "swc1 %[y0], 0(%[Y1]) \n\t"
782 "swc1 %[y1], 4(%[Y1]) \n\t"
784 : [ff_table]
"=&r" (ff_table), [y0]
"=&f" (y0), [y1]
"=&f" (y1),
785 [temp0]
"=&r" (temp0), [temp1]
"=&f" (temp1), [temp2]
"=&f" (temp2),
786 [temp3]
"=&r" (temp3), [temp4]
"=&f" (temp4), [temp5]
"=&f" (temp5)
788 [Y1]
"r" (Y1), [s_m1]
"r" (s_m1), [q_filt1]
"r" (q_filt1),
789 [phi_sign]
"f" (phi_sign)
792 phi_sign = -phi_sign;
796 static void sbr_hf_apply_noise_2_mips(
float (*
Y)[2],
const float *s_m,
797 const float *q_filt,
int noise,
802 float y0,y1, temp0, temp1, temp2, temp3, temp4, temp5;
804 for (m = 0; m < m_max; m++) {
807 const float *s_m1=&s_m[
m];
808 const float *q_filt1= &q_filt[
m];
811 "lwc1 %[y0], 0(%[Y1]) \n\t"
812 "lwc1 %[temp1], 0(%[s_m1]) \n\t"
813 "addiu %[noise], %[noise], 1 \n\t"
814 "andi %[noise], %[noise], 0x1ff \n\t"
815 "sll %[temp0], %[noise], 3 \n\t"
816 PTR_ADDU "%[ff_table],%[ff_sbr_noise_table],%[temp0] \n\t"
817 "sub.s %[y0], %[y0], %[temp1] \n\t"
818 "mfc1 %[temp3], %[temp1] \n\t"
819 "bne %[temp3], $0, 1f \n\t"
820 "lwc1 %[y1], 4(%[Y1]) \n\t"
821 "lwc1 %[temp2], 0(%[q_filt1]) \n\t"
822 "lwc1 %[temp4], 0(%[ff_table]) \n\t"
823 "lwc1 %[temp5], 4(%[ff_table]) \n\t"
824 "madd.s %[y0], %[y0], %[temp2], %[temp4] \n\t"
825 "madd.s %[y1], %[y1], %[temp2], %[temp5] \n\t"
826 "swc1 %[y1], 4(%[Y1]) \n\t"
828 "swc1 %[y0], 0(%[Y1]) \n\t"
830 : [temp0]
"=&r"(temp0), [ff_table]
"=&r"(ff_table), [y0]
"=&f"(y0),
831 [y1]
"=&f"(y1), [temp1]
"=&f"(temp1), [temp2]
"=&f"(temp2),
832 [temp3]
"=&r"(temp3), [temp4]
"=&f"(temp4), [temp5]
"=&f"(temp5)
834 [Y1]
"r"(Y1), [s_m1]
"r"(s_m1), [q_filt1]
"r"(q_filt1)
840 static void sbr_hf_apply_noise_3_mips(
float (*
Y)[2],
const float *s_m,
841 const float *q_filt,
int noise,
844 float phi_sign = 1 - 2 * (kx & 1);
847 for (m = 0; m < m_max; m++) {
851 float y0,y1, temp1, temp2, temp4, temp5;
853 const float *s_m1=&s_m[
m];
854 const float *q_filt1= &q_filt[
m];
857 "lwc1 %[y1], 4(%[Y1]) \n\t"
858 "lwc1 %[temp1], 0(%[s_m1]) \n\t"
859 "addiu %[noise], %[noise], 1 \n\t"
860 "andi %[noise], %[noise], 0x1ff \n\t"
861 "sll %[temp0], %[noise], 3 \n\t"
862 PTR_ADDU "%[ff_table],%[ff_sbr_noise_table], %[temp0] \n\t"
863 "nmsub.s %[y1], %[y1], %[temp1], %[phi_sign] \n\t"
864 "mfc1 %[temp3], %[temp1] \n\t"
865 "bne %[temp3], $0, 1f \n\t"
866 "lwc1 %[y0], 0(%[Y1]) \n\t"
867 "lwc1 %[temp2], 0(%[q_filt1]) \n\t"
868 "lwc1 %[temp4], 0(%[ff_table]) \n\t"
869 "lwc1 %[temp5], 4(%[ff_table]) \n\t"
870 "madd.s %[y0], %[y0], %[temp2], %[temp4] \n\t"
871 "madd.s %[y1], %[y1], %[temp2], %[temp5] \n\t"
872 "swc1 %[y0], 0(%[Y1]) \n\t"
874 "swc1 %[y1], 4(%[Y1]) \n\t"
876 : [ff_table]
"=&r"(ff_table), [y0]
"=&f"(y0), [y1]
"=&f"(y1),
877 [temp0]
"=&r"(temp0), [temp1]
"=&f"(temp1), [temp2]
"=&f"(temp2),
878 [temp3]
"=&r"(temp3), [temp4]
"=&f"(temp4), [temp5]
"=&f"(temp5)
880 [Y1]
"r"(Y1), [s_m1]
"r"(s_m1), [q_filt1]
"r"(q_filt1),
881 [phi_sign]
"f"(phi_sign)
884 phi_sign = -phi_sign;
902 s->
hf_gen = sbr_hf_gen_mips;
MIPS assembly defines from sys/asm.h but rewritten for use with C inline assembly (rather than from w...
const float ff_sbr_noise_table[][2]
void(* sum64x5)(float *z)
float(* sum_square)(float(*x)[2], int n)
static av_cold int end(AVCodecContext *avctx)
void(* qmf_deint_bfly)(float *v, const float *src0, const float *src1)
static double alpha(void *priv, double x, double y)
void(* hf_apply_noise[4])(float(*Y)[2], const float *s_m, const float *q_filt, int noise, int kx, int m_max)
void ff_sbrdsp_init_mips(SBRDSPContext *s)
static int noise(AVBitStreamFilterContext *bsfc, AVCodecContext *avctx, const char *args, uint8_t **poutbuf, int *poutbuf_size, const uint8_t *buf, int buf_size, int keyframe)
void(* qmf_post_shuffle)(float W[32][2], const float *z)
void(* autocorrelate)(const float x[40][2], float phi[3][2][2])
void(* qmf_pre_shuffle)(float *z)
void(* hf_gen)(float(*X_high)[2], const float(*X_low)[2], const float alpha0[2], const float alpha1[2], float bw, int start, int end)
void(* hf_g_filt)(float(*Y)[2], const float(*X_high)[40][2], const float *g_filt, int m_max, intptr_t ixh)