Skip to content
This repository

HTTPS clone URL

Subversion checkout URL

You can clone with HTTPS or Subversion.

Download ZIP
tree: 0bba26466f
Fetching contributors…

Octocat-spinner-32-eaf2f5

Cannot retrieve contributors at this time

file 174 lines (160 sloc) 5.233 kb
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173
/*
* FFT/MDCT transform with Extended 3DNow! optimizations
* Copyright (c) 2006-2008 Zuxy MENG Jie, Loren Merritt
*
* This file is part of Libav.
*
* Libav is free software; you can redistribute it and/or
* modify it under the terms of the GNU Lesser General Public
* License as published by the Free Software Foundation; either
* version 2.1 of the License, or (at your option) any later version.
*
* Libav is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
* Lesser General Public License for more details.
*
* You should have received a copy of the GNU Lesser General Public
* License along with Libav; if not, write to the Free Software
* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
*/

#include "libavutil/x86_cpu.h"
#include "libavcodec/dsputil.h"
#include "fft.h"

DECLARE_ALIGNED(8, static const unsigned int, m1m1)[2] = { 1U<<31, 1U<<31 };

#ifdef EMULATE_3DNOWEXT
#define PSWAPD(s,d)\
"movq "#s","#d"\n"\
"psrlq $32,"#d"\n"\
"punpckldq "#s","#d"\n"
#define ff_fft_calc_3dn2 ff_fft_calc_3dn
#define ff_fft_dispatch_3dn2 ff_fft_dispatch_3dn
#define ff_fft_dispatch_interleave_3dn2 ff_fft_dispatch_interleave_3dn
#define ff_imdct_calc_3dn2 ff_imdct_calc_3dn
#define ff_imdct_half_3dn2 ff_imdct_half_3dn
#else
#define PSWAPD(s,d) "pswapd "#s","#d"\n"
#endif

void ff_fft_dispatch_3dn2(FFTComplex *z, int nbits);
void ff_fft_dispatch_interleave_3dn2(FFTComplex *z, int nbits);

void ff_fft_calc_3dn2(FFTContext *s, FFTComplex *z)
{
    int n = 1<<s->nbits;
    int i;
    ff_fft_dispatch_interleave_3dn2(z, s->nbits);
    __asm__ volatile("femms");
    if(n <= 8)
        for(i=0; i<n; i+=2)
            FFSWAP(FFTSample, z[i].im, z[i+1].re);
}

void ff_imdct_half_3dn2(FFTContext *s, FFTSample *output, const FFTSample *input)
{
    x86_reg j, k;
    long n = s->mdct_size;
    long n2 = n >> 1;
    long n4 = n >> 2;
    long n8 = n >> 3;
    const uint16_t *revtab = s->revtab;
    const FFTSample *tcos = s->tcos;
    const FFTSample *tsin = s->tsin;
    const FFTSample *in1, *in2;
    FFTComplex *z = (FFTComplex *)output;

    /* pre rotation */
    in1 = input;
    in2 = input + n2 - 1;
#ifdef EMULATE_3DNOWEXT
    __asm__ volatile("movd %0, %%mm7" ::"r"(1U<<31));
#endif
    for(k = 0; k < n4; k++) {
        // FIXME a single block is faster, but gcc 2.95 and 3.4.x on 32bit can't compile it
        __asm__ volatile(
            "movd %0, %%mm0 \n"
            "movd %2, %%mm1 \n"
            "punpckldq %1, %%mm0 \n"
            "punpckldq %3, %%mm1 \n"
            "movq %%mm0, %%mm2 \n"
            PSWAPD( %%mm1, %%mm3 )
            "pfmul %%mm1, %%mm0 \n"
            "pfmul %%mm3, %%mm2 \n"
#ifdef EMULATE_3DNOWEXT
            "movq %%mm0, %%mm1 \n"
            "punpckhdq %%mm2, %%mm0 \n"
            "punpckldq %%mm2, %%mm1 \n"
            "pxor %%mm7, %%mm0 \n"
            "pfadd %%mm1, %%mm0 \n"
#else
            "pfpnacc %%mm2, %%mm0 \n"
#endif
            ::"m"(in2[-2*k]), "m"(in1[2*k]),
              "m"(tcos[k]), "m"(tsin[k])
        );
        __asm__ volatile(
            "movq %%mm0, %0 \n\t"
            :"=m"(z[revtab[k]])
        );
    }

    ff_fft_dispatch_3dn2(z, s->nbits);

#define CMUL(j,mm0,mm1)\
"movq (%2,"#j",2), %%mm6 \n"\
"movq 8(%2,"#j",2), "#mm0"\n"\
"movq %%mm6, "#mm1"\n"\
"movq "#mm0",%%mm7 \n"\
"pfmul (%3,"#j"), %%mm6 \n"\
"pfmul (%4,"#j"), "#mm0"\n"\
"pfmul (%4,"#j"), "#mm1"\n"\
"pfmul (%3,"#j"), %%mm7 \n"\
"pfsub %%mm6, "#mm0"\n"\
"pfadd %%mm7, "#mm1"\n"

    /* post rotation */
    j = -n2;
    k = n2-8;
    __asm__ volatile(
        "1: \n"
        CMUL(%0, %%mm0, %%mm1)
        CMUL(%1, %%mm2, %%mm3)
        "movd %%mm0, (%2,%0,2) \n"
        "movd %%mm1,12(%2,%1,2) \n"
        "movd %%mm2, (%2,%1,2) \n"
        "movd %%mm3,12(%2,%0,2) \n"
        "psrlq $32, %%mm0 \n"
        "psrlq $32, %%mm1 \n"
        "psrlq $32, %%mm2 \n"
        "psrlq $32, %%mm3 \n"
        "movd %%mm0, 8(%2,%0,2) \n"
        "movd %%mm1, 4(%2,%1,2) \n"
        "movd %%mm2, 8(%2,%1,2) \n"
        "movd %%mm3, 4(%2,%0,2) \n"
        "sub $8, %1 \n"
        "add $8, %0 \n"
        "jl 1b \n"
        :"+r"(j), "+r"(k)
        :"r"(z+n8), "r"(tcos+n8), "r"(tsin+n8)
        :"memory"
    );
    __asm__ volatile("femms");
}

void ff_imdct_calc_3dn2(FFTContext *s, FFTSample *output, const FFTSample *input)
{
    x86_reg j, k;
    long n = s->mdct_size;
    long n4 = n >> 2;

    ff_imdct_half_3dn2(s, output+n4, input);

    j = -n;
    k = n-8;
    __asm__ volatile(
        "movq %4, %%mm7 \n"
        "1: \n"
        PSWAPD((%2,%1), %%mm0)
        PSWAPD((%3,%0), %%mm1)
        "pxor %%mm7, %%mm0 \n"
        "movq %%mm1, (%3,%1) \n"
        "movq %%mm0, (%2,%0) \n"
        "sub $8, %1 \n"
        "add $8, %0 \n"
        "jl 1b \n"
        :"+r"(j), "+r"(k)
        :"r"(output+n4), "r"(output+n4*3),
         "m"(*m1m1)
    );
    __asm__ volatile("femms");
}
Something went wrong with that request. Please try again.