yading@10: /* yading@10: * Copyright (c) 2012 yading@10: * MIPS Technologies, Inc., California. yading@10: * yading@10: * Redistribution and use in source and binary forms, with or without yading@10: * modification, are permitted provided that the following conditions yading@10: * are met: yading@10: * 1. Redistributions of source code must retain the above copyright yading@10: * notice, this list of conditions and the following disclaimer. yading@10: * 2. Redistributions in binary form must reproduce the above copyright yading@10: * notice, this list of conditions and the following disclaimer in the yading@10: * documentation and/or other materials provided with the distribution. yading@10: * 3. Neither the name of the MIPS Technologies, Inc., nor the names of its yading@10: * contributors may be used to endorse or promote products derived from yading@10: * this software without specific prior written permission. yading@10: * yading@10: * THIS SOFTWARE IS PROVIDED BY THE MIPS TECHNOLOGIES, INC. ``AS IS'' AND yading@10: * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE yading@10: * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE yading@10: * ARE DISCLAIMED. IN NO EVENT SHALL THE MIPS TECHNOLOGIES, INC. BE LIABLE yading@10: * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL yading@10: * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS yading@10: * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) yading@10: * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT yading@10: * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY yading@10: * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF yading@10: * SUCH DAMAGE. yading@10: * yading@10: * Authors: Djordje Pesut (djordje@mips.com) yading@10: * Mirjana Vulin (mvulin@mips.com) yading@10: * yading@10: * This file is part of FFmpeg. yading@10: * yading@10: * FFmpeg is free software; you can redistribute it and/or yading@10: * modify it under the terms of the GNU Lesser General Public yading@10: * License as published by the Free Software Foundation; either yading@10: * version 2.1 of the License, or (at your option) any later version. yading@10: * yading@10: * FFmpeg is distributed in the hope that it will be useful, yading@10: * but WITHOUT ANY WARRANTY; without even the implied warranty of yading@10: * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU yading@10: * Lesser General Public License for more details. yading@10: * yading@10: * You should have received a copy of the GNU Lesser General Public yading@10: * License along with FFmpeg; if not, write to the Free Software yading@10: * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA yading@10: */ yading@10: yading@10: /** yading@10: * @file yading@10: * Reference: libavcodec/aacsbr.c yading@10: */ yading@10: yading@10: #ifndef AVCODEC_MIPS_AACSBR_FLOAT_H yading@10: #define AVCODEC_MIPS_AACSBR_FLOAT_H yading@10: yading@10: #include "libavcodec/aac.h" yading@10: #include "libavcodec/sbr.h" yading@10: yading@10: #if HAVE_INLINE_ASM yading@10: static void sbr_qmf_analysis_mips(AVFloatDSPContext *fdsp, FFTContext *mdct, yading@10: SBRDSPContext *sbrdsp, const float *in, float *x, yading@10: float z[320], float W[2][32][32][2], int buf_idx) yading@10: { yading@10: int i; yading@10: float *w0; yading@10: float *w1; yading@10: int temp0, temp1, temp2, temp3, temp4, temp5, temp6, temp7; yading@10: yading@10: w0 = x; yading@10: w1 = x + 1024; yading@10: for(i = 0; i < 36; i++) yading@10: { yading@10: /* loop unrolled 8 times */ yading@10: __asm__ volatile( yading@10: "lw %[temp0], 0(%[w1]) \n\t" yading@10: "lw %[temp1], 4(%[w1]) \n\t" yading@10: "lw %[temp2], 8(%[w1]) \n\t" yading@10: "lw %[temp3], 12(%[w1]) \n\t" yading@10: "lw %[temp4], 16(%[w1]) \n\t" yading@10: "lw %[temp5], 20(%[w1]) \n\t" yading@10: "lw %[temp6], 24(%[w1]) \n\t" yading@10: "lw %[temp7], 28(%[w1]) \n\t" yading@10: "sw %[temp0], 0(%[w0]) \n\t" yading@10: "sw %[temp1], 4(%[w0]) \n\t" yading@10: "sw %[temp2], 8(%[w0]) \n\t" yading@10: "sw %[temp3], 12(%[w0]) \n\t" yading@10: "sw %[temp4], 16(%[w0]) \n\t" yading@10: "sw %[temp5], 20(%[w0]) \n\t" yading@10: "sw %[temp6], 24(%[w0]) \n\t" yading@10: "sw %[temp7], 28(%[w0]) \n\t" yading@10: "addiu %[w0], %[w0], 32 \n\t" yading@10: "addiu %[w1], %[w1], 32 \n\t" yading@10: yading@10: : [w0]"+r"(w0), [w1]"+r"(w1), yading@10: [temp0]"=&r"(temp0), [temp1]"=&r"(temp1), yading@10: [temp2]"=&r"(temp2), [temp3]"=&r"(temp3), yading@10: [temp4]"=&r"(temp4), [temp5]"=&r"(temp5), yading@10: [temp6]"=&r"(temp6), [temp7]"=&r"(temp7) yading@10: : yading@10: : "memory" yading@10: ); yading@10: } yading@10: yading@10: w0 = x + 288; yading@10: w1 = (float*)in; yading@10: for(i = 0; i < 128; i++) yading@10: { yading@10: /* loop unrolled 8 times */ yading@10: __asm__ volatile( yading@10: "lw %[temp0], 0(%[w1]) \n\t" yading@10: "lw %[temp1], 4(%[w1]) \n\t" yading@10: "lw %[temp2], 8(%[w1]) \n\t" yading@10: "lw %[temp3], 12(%[w1]) \n\t" yading@10: "lw %[temp4], 16(%[w1]) \n\t" yading@10: "lw %[temp5], 20(%[w1]) \n\t" yading@10: "lw %[temp6], 24(%[w1]) \n\t" yading@10: "lw %[temp7], 28(%[w1]) \n\t" yading@10: "sw %[temp0], 0(%[w0]) \n\t" yading@10: "sw %[temp1], 4(%[w0]) \n\t" yading@10: "sw %[temp2], 8(%[w0]) \n\t" yading@10: "sw %[temp3], 12(%[w0]) \n\t" yading@10: "sw %[temp4], 16(%[w0]) \n\t" yading@10: "sw %[temp5], 20(%[w0]) \n\t" yading@10: "sw %[temp6], 24(%[w0]) \n\t" yading@10: "sw %[temp7], 28(%[w0]) \n\t" yading@10: "addiu %[w0], %[w0], 32 \n\t" yading@10: "addiu %[w1], %[w1], 32 \n\t" yading@10: yading@10: : [w0]"+r"(w0), [w1]"+r"(w1), yading@10: [temp0]"=&r"(temp0), [temp1]"=&r"(temp1), yading@10: [temp2]"=&r"(temp2), [temp3]"=&r"(temp3), yading@10: [temp4]"=&r"(temp4), [temp5]"=&r"(temp5), yading@10: [temp6]"=&r"(temp6), [temp7]"=&r"(temp7) yading@10: : yading@10: : "memory" yading@10: ); yading@10: } yading@10: yading@10: for (i = 0; i < 32; i++) { // numTimeSlots*RATE = 16*2 as 960 sample frames yading@10: // are not supported yading@10: fdsp->vector_fmul_reverse(z, sbr_qmf_window_ds, x, 320); yading@10: sbrdsp->sum64x5(z); yading@10: sbrdsp->qmf_pre_shuffle(z); yading@10: mdct->imdct_half(mdct, z, z+64); yading@10: sbrdsp->qmf_post_shuffle(W[buf_idx][i], z); yading@10: x += 32; yading@10: } yading@10: } yading@10: yading@10: #if HAVE_MIPSFPU yading@10: static void sbr_qmf_synthesis_mips(FFTContext *mdct, yading@10: SBRDSPContext *sbrdsp, AVFloatDSPContext *fdsp, yading@10: float *out, float X[2][38][64], yading@10: float mdct_buf[2][64], yading@10: float *v0, int *v_off, const unsigned int div) yading@10: { yading@10: int i, n; yading@10: const float *sbr_qmf_window = div ? sbr_qmf_window_ds : sbr_qmf_window_us; yading@10: const int step = 128 >> div; yading@10: float *v; yading@10: float temp0, temp1, temp2, temp3, temp4, temp5, temp6, temp7, temp8, temp9, temp10, temp11, temp12, temp13; yading@10: float temp14, temp15, temp16, temp17, temp18, temp19; yading@10: float *vv0, *s0, *dst; yading@10: dst = out; yading@10: yading@10: for (i = 0; i < 32; i++) { yading@10: if (*v_off < step) { yading@10: int saved_samples = (1280 - 128) >> div; yading@10: memcpy(&v0[SBR_SYNTHESIS_BUF_SIZE - saved_samples], v0, saved_samples * sizeof(float)); yading@10: *v_off = SBR_SYNTHESIS_BUF_SIZE - saved_samples - step; yading@10: } else { yading@10: *v_off -= step; yading@10: } yading@10: v = v0 + *v_off; yading@10: if (div) { yading@10: for (n = 0; n < 32; n++) { yading@10: X[0][i][ n] = -X[0][i][n]; yading@10: X[0][i][32+n] = X[1][i][31-n]; yading@10: } yading@10: mdct->imdct_half(mdct, mdct_buf[0], X[0][i]); yading@10: sbrdsp->qmf_deint_neg(v, mdct_buf[0]); yading@10: } else { yading@10: sbrdsp->neg_odd_64(X[1][i]); yading@10: mdct->imdct_half(mdct, mdct_buf[0], X[0][i]); yading@10: mdct->imdct_half(mdct, mdct_buf[1], X[1][i]); yading@10: sbrdsp->qmf_deint_bfly(v, mdct_buf[1], mdct_buf[0]); yading@10: } yading@10: yading@10: if(div == 0) yading@10: { yading@10: float *v0_end; yading@10: vv0 = v; yading@10: v0_end = v + 60; yading@10: s0 = (float*)sbr_qmf_window; yading@10: yading@10: /* 10 calls of function vector_fmul_add merged into one loop yading@10: and loop unrolled 4 times */ yading@10: __asm__ volatile( yading@10: ".set push \n\t" yading@10: ".set noreorder \n\t" yading@10: "lwc1 %[temp4], 0(%[v0]) \n\t" yading@10: "lwc1 %[temp5], 0(%[s0]) \n\t" yading@10: "lwc1 %[temp6], 4(%[v0]) \n\t" yading@10: "lwc1 %[temp7], 4(%[s0]) \n\t" yading@10: "lwc1 %[temp8], 8(%[v0]) \n\t" yading@10: "lwc1 %[temp9], 8(%[s0]) \n\t" yading@10: "lwc1 %[temp10], 12(%[v0]) \n\t" yading@10: "lwc1 %[temp11], 12(%[s0]) \n\t" yading@10: "lwc1 %[temp12], 768(%[v0]) \n\t" yading@10: "lwc1 %[temp13], 256(%[s0]) \n\t" yading@10: "lwc1 %[temp14], 772(%[v0]) \n\t" yading@10: "lwc1 %[temp15], 260(%[s0]) \n\t" yading@10: "lwc1 %[temp16], 776(%[v0]) \n\t" yading@10: "lwc1 %[temp17], 264(%[s0]) \n\t" yading@10: "lwc1 %[temp18], 780(%[v0]) \n\t" yading@10: "lwc1 %[temp19], 268(%[s0]) \n\t" yading@10: "1: \n\t" yading@10: "mul.s %[temp0], %[temp4], %[temp5] \n\t" yading@10: "lwc1 %[temp4], 1024(%[v0]) \n\t" yading@10: "mul.s %[temp1], %[temp6], %[temp7] \n\t" yading@10: "lwc1 %[temp5], 512(%[s0]) \n\t" yading@10: "mul.s %[temp2], %[temp8], %[temp9] \n\t" yading@10: "lwc1 %[temp6], 1028(%[v0]) \n\t" yading@10: "mul.s %[temp3], %[temp10], %[temp11] \n\t" yading@10: "lwc1 %[temp7], 516(%[s0]) \n\t" yading@10: "madd.s %[temp0], %[temp0], %[temp12], %[temp13] \n\t" yading@10: "lwc1 %[temp8], 1032(%[v0]) \n\t" yading@10: "madd.s %[temp1], %[temp1], %[temp14], %[temp15] \n\t" yading@10: "lwc1 %[temp9], 520(%[s0]) \n\t" yading@10: "madd.s %[temp2], %[temp2], %[temp16], %[temp17] \n\t" yading@10: "lwc1 %[temp10], 1036(%[v0]) \n\t" yading@10: "madd.s %[temp3], %[temp3], %[temp18], %[temp19] \n\t" yading@10: "lwc1 %[temp11], 524(%[s0]) \n\t" yading@10: "lwc1 %[temp12], 1792(%[v0]) \n\t" yading@10: "lwc1 %[temp13], 768(%[s0]) \n\t" yading@10: "lwc1 %[temp14], 1796(%[v0]) \n\t" yading@10: "lwc1 %[temp15], 772(%[s0]) \n\t" yading@10: "lwc1 %[temp16], 1800(%[v0]) \n\t" yading@10: "lwc1 %[temp17], 776(%[s0]) \n\t" yading@10: "lwc1 %[temp18], 1804(%[v0]) \n\t" yading@10: "lwc1 %[temp19], 780(%[s0]) \n\t" yading@10: "madd.s %[temp0], %[temp0], %[temp4], %[temp5] \n\t" yading@10: "lwc1 %[temp4], 2048(%[v0]) \n\t" yading@10: "madd.s %[temp1], %[temp1], %[temp6], %[temp7] \n\t" yading@10: "lwc1 %[temp5], 1024(%[s0]) \n\t" yading@10: "madd.s %[temp2], %[temp2], %[temp8], %[temp9] \n\t" yading@10: "lwc1 %[temp6], 2052(%[v0]) \n\t" yading@10: "madd.s %[temp3], %[temp3], %[temp10], %[temp11] \n\t" yading@10: "lwc1 %[temp7], 1028(%[s0]) \n\t" yading@10: "madd.s %[temp0], %[temp0], %[temp12], %[temp13] \n\t" yading@10: "lwc1 %[temp8], 2056(%[v0]) \n\t" yading@10: "madd.s %[temp1], %[temp1], %[temp14], %[temp15] \n\t" yading@10: "lwc1 %[temp9], 1032(%[s0]) \n\t" yading@10: "madd.s %[temp2], %[temp2], %[temp16], %[temp17] \n\t" yading@10: "lwc1 %[temp10], 2060(%[v0]) \n\t" yading@10: "madd.s %[temp3], %[temp3], %[temp18], %[temp19] \n\t" yading@10: "lwc1 %[temp11], 1036(%[s0]) \n\t" yading@10: "lwc1 %[temp12], 2816(%[v0]) \n\t" yading@10: "lwc1 %[temp13], 1280(%[s0]) \n\t" yading@10: "lwc1 %[temp14], 2820(%[v0]) \n\t" yading@10: "lwc1 %[temp15], 1284(%[s0]) \n\t" yading@10: "lwc1 %[temp16], 2824(%[v0]) \n\t" yading@10: "lwc1 %[temp17], 1288(%[s0]) \n\t" yading@10: "lwc1 %[temp18], 2828(%[v0]) \n\t" yading@10: "lwc1 %[temp19], 1292(%[s0]) \n\t" yading@10: "madd.s %[temp0], %[temp0], %[temp4], %[temp5] \n\t" yading@10: "lwc1 %[temp4], 3072(%[v0]) \n\t" yading@10: "madd.s %[temp1], %[temp1], %[temp6], %[temp7] \n\t" yading@10: "lwc1 %[temp5], 1536(%[s0]) \n\t" yading@10: "madd.s %[temp2], %[temp2], %[temp8], %[temp9] \n\t" yading@10: "lwc1 %[temp6], 3076(%[v0]) \n\t" yading@10: "madd.s %[temp3], %[temp3], %[temp10], %[temp11] \n\t" yading@10: "lwc1 %[temp7], 1540(%[s0]) \n\t" yading@10: "madd.s %[temp0], %[temp0], %[temp12], %[temp13] \n\t" yading@10: "lwc1 %[temp8], 3080(%[v0]) \n\t" yading@10: "madd.s %[temp1], %[temp1], %[temp14], %[temp15] \n\t" yading@10: "lwc1 %[temp9], 1544(%[s0]) \n\t" yading@10: "madd.s %[temp2], %[temp2], %[temp16], %[temp17] \n\t" yading@10: "lwc1 %[temp10], 3084(%[v0]) \n\t" yading@10: "madd.s %[temp3], %[temp3], %[temp18], %[temp19] \n\t" yading@10: "lwc1 %[temp11], 1548(%[s0]) \n\t" yading@10: "lwc1 %[temp12], 3840(%[v0]) \n\t" yading@10: "lwc1 %[temp13], 1792(%[s0]) \n\t" yading@10: "lwc1 %[temp14], 3844(%[v0]) \n\t" yading@10: "lwc1 %[temp15], 1796(%[s0]) \n\t" yading@10: "lwc1 %[temp16], 3848(%[v0]) \n\t" yading@10: "lwc1 %[temp17], 1800(%[s0]) \n\t" yading@10: "lwc1 %[temp18], 3852(%[v0]) \n\t" yading@10: "lwc1 %[temp19], 1804(%[s0]) \n\t" yading@10: "madd.s %[temp0], %[temp0], %[temp4], %[temp5] \n\t" yading@10: "lwc1 %[temp4], 4096(%[v0]) \n\t" yading@10: "madd.s %[temp1], %[temp1], %[temp6], %[temp7] \n\t" yading@10: "lwc1 %[temp5], 2048(%[s0]) \n\t" yading@10: "madd.s %[temp2], %[temp2], %[temp8], %[temp9] \n\t" yading@10: "lwc1 %[temp6], 4100(%[v0]) \n\t" yading@10: "madd.s %[temp3], %[temp3], %[temp10], %[temp11] \n\t" yading@10: "lwc1 %[temp7], 2052(%[s0]) \n\t" yading@10: "madd.s %[temp0], %[temp0], %[temp12], %[temp13] \n\t" yading@10: "lwc1 %[temp8], 4104(%[v0]) \n\t" yading@10: "addiu %[dst], %[dst], 16 \n\t" yading@10: "madd.s %[temp1], %[temp1], %[temp14], %[temp15] \n\t" yading@10: "lwc1 %[temp9], 2056(%[s0]) \n\t" yading@10: "addiu %[s0], %[s0], 16 \n\t" yading@10: "madd.s %[temp2], %[temp2], %[temp16], %[temp17] \n\t" yading@10: "lwc1 %[temp10], 4108(%[v0]) \n\t" yading@10: "addiu %[v0], %[v0], 16 \n\t" yading@10: "madd.s %[temp3], %[temp3], %[temp18], %[temp19] \n\t" yading@10: "lwc1 %[temp11], 2044(%[s0]) \n\t" yading@10: "lwc1 %[temp12], 4848(%[v0]) \n\t" yading@10: "lwc1 %[temp13], 2288(%[s0]) \n\t" yading@10: "lwc1 %[temp14], 4852(%[v0]) \n\t" yading@10: "lwc1 %[temp15], 2292(%[s0]) \n\t" yading@10: "lwc1 %[temp16], 4856(%[v0]) \n\t" yading@10: "lwc1 %[temp17], 2296(%[s0]) \n\t" yading@10: "lwc1 %[temp18], 4860(%[v0]) \n\t" yading@10: "lwc1 %[temp19], 2300(%[s0]) \n\t" yading@10: "madd.s %[temp0], %[temp0], %[temp4], %[temp5] \n\t" yading@10: "lwc1 %[temp4], 0(%[v0]) \n\t" yading@10: "madd.s %[temp1], %[temp1], %[temp6], %[temp7] \n\t" yading@10: "lwc1 %[temp5], 0(%[s0]) \n\t" yading@10: "madd.s %[temp2], %[temp2], %[temp8], %[temp9] \n\t" yading@10: "lwc1 %[temp6], 4(%[v0]) \n\t" yading@10: "madd.s %[temp3], %[temp3], %[temp10], %[temp11] \n\t" yading@10: "lwc1 %[temp7], 4(%[s0]) \n\t" yading@10: "madd.s %[temp0], %[temp0], %[temp12], %[temp13] \n\t" yading@10: "lwc1 %[temp8], 8(%[v0]) \n\t" yading@10: "madd.s %[temp1], %[temp1], %[temp14], %[temp15] \n\t" yading@10: "lwc1 %[temp9], 8(%[s0]) \n\t" yading@10: "madd.s %[temp2], %[temp2], %[temp16], %[temp17] \n\t" yading@10: "lwc1 %[temp10], 12(%[v0]) \n\t" yading@10: "madd.s %[temp3], %[temp3], %[temp18], %[temp19] \n\t" yading@10: "lwc1 %[temp11], 12(%[s0]) \n\t" yading@10: "lwc1 %[temp12], 768(%[v0]) \n\t" yading@10: "lwc1 %[temp13], 256(%[s0]) \n\t" yading@10: "lwc1 %[temp14], 772(%[v0]) \n\t" yading@10: "lwc1 %[temp15], 260(%[s0]) \n\t" yading@10: "lwc1 %[temp16], 776(%[v0]) \n\t" yading@10: "lwc1 %[temp17], 264(%[s0]) \n\t" yading@10: "lwc1 %[temp18], 780(%[v0]) \n\t" yading@10: "lwc1 %[temp19], 268(%[s0]) \n\t" yading@10: "swc1 %[temp0], -16(%[dst]) \n\t" yading@10: "swc1 %[temp1], -12(%[dst]) \n\t" yading@10: "swc1 %[temp2], -8(%[dst]) \n\t" yading@10: "bne %[v0], %[v0_end], 1b \n\t" yading@10: " swc1 %[temp3], -4(%[dst]) \n\t" yading@10: "mul.s %[temp0], %[temp4], %[temp5] \n\t" yading@10: "lwc1 %[temp4], 1024(%[v0]) \n\t" yading@10: "mul.s %[temp1], %[temp6], %[temp7] \n\t" yading@10: "lwc1 %[temp5], 512(%[s0]) \n\t" yading@10: "mul.s %[temp2], %[temp8], %[temp9] \n\t" yading@10: "lwc1 %[temp6], 1028(%[v0]) \n\t" yading@10: "mul.s %[temp3], %[temp10], %[temp11] \n\t" yading@10: "lwc1 %[temp7], 516(%[s0]) \n\t" yading@10: "madd.s %[temp0], %[temp0], %[temp12], %[temp13] \n\t" yading@10: "lwc1 %[temp8], 1032(%[v0]) \n\t" yading@10: "madd.s %[temp1], %[temp1], %[temp14], %[temp15] \n\t" yading@10: "lwc1 %[temp9], 520(%[s0]) \n\t" yading@10: "madd.s %[temp2], %[temp2], %[temp16], %[temp17] \n\t" yading@10: "lwc1 %[temp10], 1036(%[v0]) \n\t" yading@10: "madd.s %[temp3], %[temp3], %[temp18], %[temp19] \n\t" yading@10: "lwc1 %[temp11], 524(%[s0]) \n\t" yading@10: "lwc1 %[temp12], 1792(%[v0]) \n\t" yading@10: "lwc1 %[temp13], 768(%[s0]) \n\t" yading@10: "lwc1 %[temp14], 1796(%[v0]) \n\t" yading@10: "lwc1 %[temp15], 772(%[s0]) \n\t" yading@10: "lwc1 %[temp16], 1800(%[v0]) \n\t" yading@10: "lwc1 %[temp17], 776(%[s0]) \n\t" yading@10: "lwc1 %[temp18], 1804(%[v0]) \n\t" yading@10: "lwc1 %[temp19], 780(%[s0]) \n\t" yading@10: "madd.s %[temp0], %[temp0], %[temp4], %[temp5] \n\t" yading@10: "lwc1 %[temp4], 2048(%[v0]) \n\t" yading@10: "madd.s %[temp1], %[temp1], %[temp6], %[temp7] \n\t" yading@10: "lwc1 %[temp5], 1024(%[s0]) \n\t" yading@10: "madd.s %[temp2], %[temp2], %[temp8], %[temp9] \n\t" yading@10: "lwc1 %[temp6], 2052(%[v0]) \n\t" yading@10: "madd.s %[temp3], %[temp3], %[temp10], %[temp11] \n\t" yading@10: "lwc1 %[temp7], 1028(%[s0]) \n\t" yading@10: "madd.s %[temp0], %[temp0], %[temp12], %[temp13] \n\t" yading@10: "lwc1 %[temp8], 2056(%[v0]) \n\t" yading@10: "madd.s %[temp1], %[temp1], %[temp14], %[temp15] \n\t" yading@10: "lwc1 %[temp9], 1032(%[s0]) \n\t" yading@10: "madd.s %[temp2], %[temp2], %[temp16], %[temp17] \n\t" yading@10: "lwc1 %[temp10], 2060(%[v0]) \n\t" yading@10: "madd.s %[temp3], %[temp3], %[temp18], %[temp19] \n\t" yading@10: "lwc1 %[temp11], 1036(%[s0]) \n\t" yading@10: "lwc1 %[temp12], 2816(%[v0]) \n\t" yading@10: "lwc1 %[temp13], 1280(%[s0]) \n\t" yading@10: "lwc1 %[temp14], 2820(%[v0]) \n\t" yading@10: "lwc1 %[temp15], 1284(%[s0]) \n\t" yading@10: "lwc1 %[temp16], 2824(%[v0]) \n\t" yading@10: "lwc1 %[temp17], 1288(%[s0]) \n\t" yading@10: "lwc1 %[temp18], 2828(%[v0]) \n\t" yading@10: "lwc1 %[temp19], 1292(%[s0]) \n\t" yading@10: "madd.s %[temp0], %[temp0], %[temp4], %[temp5] \n\t" yading@10: "lwc1 %[temp4], 3072(%[v0]) \n\t" yading@10: "madd.s %[temp1], %[temp1], %[temp6], %[temp7] \n\t" yading@10: "lwc1 %[temp5], 1536(%[s0]) \n\t" yading@10: "madd.s %[temp2], %[temp2], %[temp8], %[temp9] \n\t" yading@10: "lwc1 %[temp6], 3076(%[v0]) \n\t" yading@10: "madd.s %[temp3], %[temp3], %[temp10], %[temp11] \n\t" yading@10: "lwc1 %[temp7], 1540(%[s0]) \n\t" yading@10: "madd.s %[temp0], %[temp0], %[temp12], %[temp13] \n\t" yading@10: "lwc1 %[temp8], 3080(%[v0]) \n\t" yading@10: "madd.s %[temp1], %[temp1], %[temp14], %[temp15] \n\t" yading@10: "lwc1 %[temp9], 1544(%[s0]) \n\t" yading@10: "madd.s %[temp2], %[temp2], %[temp16], %[temp17] \n\t" yading@10: "lwc1 %[temp10], 3084(%[v0]) \n\t" yading@10: "madd.s %[temp3], %[temp3], %[temp18], %[temp19] \n\t" yading@10: "lwc1 %[temp11], 1548(%[s0]) \n\t" yading@10: "lwc1 %[temp12], 3840(%[v0]) \n\t" yading@10: "lwc1 %[temp13], 1792(%[s0]) \n\t" yading@10: "lwc1 %[temp14], 3844(%[v0]) \n\t" yading@10: "lwc1 %[temp15], 1796(%[s0]) \n\t" yading@10: "lwc1 %[temp16], 3848(%[v0]) \n\t" yading@10: "lwc1 %[temp17], 1800(%[s0]) \n\t" yading@10: "lwc1 %[temp18], 3852(%[v0]) \n\t" yading@10: "lwc1 %[temp19], 1804(%[s0]) \n\t" yading@10: "madd.s %[temp0], %[temp0], %[temp4], %[temp5] \n\t" yading@10: "lwc1 %[temp4], 4096(%[v0]) \n\t" yading@10: "madd.s %[temp1], %[temp1], %[temp6], %[temp7] \n\t" yading@10: "lwc1 %[temp5], 2048(%[s0]) \n\t" yading@10: "madd.s %[temp2], %[temp2], %[temp8], %[temp9] \n\t" yading@10: "lwc1 %[temp6], 4100(%[v0]) \n\t" yading@10: "madd.s %[temp3], %[temp3], %[temp10], %[temp11] \n\t" yading@10: "lwc1 %[temp7], 2052(%[s0]) \n\t" yading@10: "madd.s %[temp0], %[temp0], %[temp12], %[temp13] \n\t" yading@10: "lwc1 %[temp8], 4104(%[v0]) \n\t" yading@10: "madd.s %[temp1], %[temp1], %[temp14], %[temp15] \n\t" yading@10: "lwc1 %[temp9], 2056(%[s0]) \n\t" yading@10: "madd.s %[temp2], %[temp2], %[temp16], %[temp17] \n\t" yading@10: "lwc1 %[temp10], 4108(%[v0]) \n\t" yading@10: "madd.s %[temp3], %[temp3], %[temp18], %[temp19] \n\t" yading@10: "lwc1 %[temp11], 2060(%[s0]) \n\t" yading@10: "lwc1 %[temp12], 4864(%[v0]) \n\t" yading@10: "lwc1 %[temp13], 2304(%[s0]) \n\t" yading@10: "lwc1 %[temp14], 4868(%[v0]) \n\t" yading@10: "lwc1 %[temp15], 2308(%[s0]) \n\t" yading@10: "madd.s %[temp0], %[temp0], %[temp4], %[temp5] \n\t" yading@10: "lwc1 %[temp16], 4872(%[v0]) \n\t" yading@10: "madd.s %[temp1], %[temp1], %[temp6], %[temp7] \n\t" yading@10: "lwc1 %[temp17], 2312(%[s0]) \n\t" yading@10: "madd.s %[temp2], %[temp2], %[temp8], %[temp9] \n\t" yading@10: "lwc1 %[temp18], 4876(%[v0]) \n\t" yading@10: "madd.s %[temp3], %[temp3], %[temp10], %[temp11] \n\t" yading@10: "lwc1 %[temp19], 2316(%[s0]) \n\t" yading@10: "madd.s %[temp0], %[temp0], %[temp12], %[temp13] \n\t" yading@10: "addiu %[dst], %[dst], 16 \n\t" yading@10: "madd.s %[temp1], %[temp1], %[temp14], %[temp15] \n\t" yading@10: "madd.s %[temp2], %[temp2], %[temp16], %[temp17] \n\t" yading@10: "madd.s %[temp3], %[temp3], %[temp18], %[temp19] \n\t" yading@10: "swc1 %[temp0], -16(%[dst]) \n\t" yading@10: "swc1 %[temp1], -12(%[dst]) \n\t" yading@10: "swc1 %[temp2], -8(%[dst]) \n\t" yading@10: "swc1 %[temp3], -4(%[dst]) \n\t" yading@10: ".set pop \n\t" yading@10: yading@10: : [dst]"+r"(dst), [v0]"+r"(vv0), [s0]"+r"(s0), yading@10: [temp0]"=&f"(temp0), [temp1]"=&f"(temp1), [temp2]"=&f"(temp2), yading@10: [temp3]"=&f"(temp3), [temp4]"=&f"(temp4), [temp5]"=&f"(temp5), yading@10: [temp6]"=&f"(temp6), [temp7]"=&f"(temp7), [temp8]"=&f"(temp8), yading@10: [temp9]"=&f"(temp9), [temp10]"=&f"(temp10), [temp11]"=&f"(temp11), yading@10: [temp12]"=&f"(temp12), [temp13]"=&f"(temp13), [temp14]"=&f"(temp14), yading@10: [temp15]"=&f"(temp15), [temp16]"=&f"(temp16), [temp17]"=&f"(temp17), yading@10: [temp18]"=&f"(temp18), [temp19]"=&f"(temp19) yading@10: : [v0_end]"r"(v0_end) yading@10: : "memory" yading@10: ); yading@10: } yading@10: else yading@10: { yading@10: fdsp->vector_fmul (out, v , sbr_qmf_window , 64 >> div); yading@10: fdsp->vector_fmul_add(out, v + ( 192 >> div), sbr_qmf_window + ( 64 >> div), out , 64 >> div); yading@10: fdsp->vector_fmul_add(out, v + ( 256 >> div), sbr_qmf_window + (128 >> div), out , 64 >> div); yading@10: fdsp->vector_fmul_add(out, v + ( 448 >> div), sbr_qmf_window + (192 >> div), out , 64 >> div); yading@10: fdsp->vector_fmul_add(out, v + ( 512 >> div), sbr_qmf_window + (256 >> div), out , 64 >> div); yading@10: fdsp->vector_fmul_add(out, v + ( 704 >> div), sbr_qmf_window + (320 >> div), out , 64 >> div); yading@10: fdsp->vector_fmul_add(out, v + ( 768 >> div), sbr_qmf_window + (384 >> div), out , 64 >> div); yading@10: fdsp->vector_fmul_add(out, v + ( 960 >> div), sbr_qmf_window + (448 >> div), out , 64 >> div); yading@10: fdsp->vector_fmul_add(out, v + (1024 >> div), sbr_qmf_window + (512 >> div), out , 64 >> div); yading@10: fdsp->vector_fmul_add(out, v + (1216 >> div), sbr_qmf_window + (576 >> div), out , 64 >> div); yading@10: out += 64 >> div; yading@10: } yading@10: } yading@10: } yading@10: yading@10: #define sbr_qmf_analysis sbr_qmf_analysis_mips yading@10: #define sbr_qmf_synthesis sbr_qmf_synthesis_mips yading@10: yading@10: #endif /* HAVE_MIPSFPU */ yading@10: #endif /* HAVE_INLINE_ASM */ yading@10: yading@10: #endif /* AVCODEC_MIPS_AACSBR_FLOAT_H */