view x86/mpegvideo_mmx.c @ 10644:5da7180afadf libavcodec

refactor and optimize scalarproduct 29-105% faster apply_filter, 6-90% faster ape decoding on core2 (Any x86 other than core2 probably gets much less, since this is mostly due to ssse3 cachesplit avoidance and I haven't written the full gamut of other cachesplit modes.) 9-123% faster ape decoding on G4.

/*
 * The simplest mpeg encoder (well, it was the simplest!)
 * Copyright (c) 2000,2001 Fabrice Bellard
 *
 * Optimized for ia32 CPUs by Nick Kurshev <[e-mail]>
 * h263, mpeg1, mpeg2 dequantizer & draw_edges by Michael Niedermayer <[e-mail]>
 *
 * This file is part of FFmpeg.
 *
 * FFmpeg 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.
 *
 * FFmpeg 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 FFmpeg; 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/avcodec.h"
#include "libavcodec/dsputil.h"
#include "libavcodec/mpegvideo.h"
#include "dsputil_mmx.h"

extern uint16_t inv_zigzag_direct16[64];


static void dct_unquantize_h263_intra_mmx(MpegEncContext *s,
                                  DCTELEM *block, int n, int qscale)
{
    x86_reg level, qmul, qadd, nCoeffs;

    qmul = qscale << 1;

    assert(s->block_last_index[n]>=0 || s->h263_aic);

    if (!s->h263_aic) {
        if (n < 4)
            level = block[0] * s->y_dc_scale;
        else
            level = block[0] * s->c_dc_scale;
        qadd = (qscale - 1) | 1;
    }else{
        qadd = 0;
        level= block[0];
    }
    if(s->ac_pred)
        nCoeffs=63;
    else
        nCoeffs= s->inter_scantable.raster_end[ s->block_last_index[n] ];
//printf("%d %d  ", qmul, qadd);
__asm__ volatile(
                "movd %1, %%mm6                 \n\t" //qmul
                "packssdw %%mm6, %%mm6          \n\t"
                "packssdw %%mm6, %%mm6          \n\t"
                "movd %2, %%mm5                 \n\t" //qadd
                "pxor %%mm7, %%mm7              \n\t"
                "packssdw %%mm5, %%mm5          \n\t"
                "packssdw %%mm5, %%mm5          \n\t"
                "psubw %%mm5, %%mm7             \n\t"
                "pxor %%mm4, %%mm4              \n\t"
                ASMALIGN(4)
                "1:                             \n\t"
                "movq (%0, %3), %%mm0           \n\t"
                "movq 8(%0, %3), %%mm1          \n\t"

                "pmullw %%mm6, %%mm0            \n\t"
                "pmullw %%mm6, %%mm1            \n\t"

                "movq (%0, %3), %%mm2           \n\t"
                "movq 8(%0, %3), %%mm3          \n\t"

                "pcmpgtw %%mm4, %%mm2           \n\t" // block[i] < 0 ? -1 : 0
                "pcmpgtw %%mm4, %%mm3           \n\t" // block[i] < 0 ? -1 : 0

                "pxor %%mm2, %%mm0              \n\t"
                "pxor %%mm3, %%mm1              \n\t"

                "paddw %%mm7, %%mm0             \n\t"
                "paddw %%mm7, %%mm1             \n\t"

                "pxor %%mm0, %%mm2              \n\t"
                "pxor %%mm1, %%mm3              \n\t"

                "pcmpeqw %%mm7, %%mm0           \n\t" // block[i] == 0 ? -1 : 0
                "pcmpeqw %%mm7, %%mm1           \n\t" // block[i] == 0 ? -1 : 0

                "pandn %%mm2, %%mm0             \n\t"
                "pandn %%mm3, %%mm1             \n\t"

                "movq %%mm0, (%0, %3)           \n\t"
                "movq %%mm1, 8(%0, %3)          \n\t"

                "add $16, %3                    \n\t"
                "jng 1b                         \n\t"
                ::"r" (block+nCoeffs), "rm"(qmul), "rm" (qadd), "r" (2*(-nCoeffs))
                : "memory"
        );
        block[0]= level;
}


static void dct_unquantize_h263_inter_mmx(MpegEncContext *s,
                                  DCTELEM *block, int n, int qscale)
{
    x86_reg qmul, qadd, nCoeffs;

    qmul = qscale << 1;
    qadd = (qscale - 1) | 1;

    assert(s->block_last_index[n]>=0 || s->h263_aic);

    nCoeffs= s->inter_scantable.raster_end[ s->block_last_index[n] ];
//printf("%d %d  ", qmul, qadd);
__asm__ volatile(
                "movd %1, %%mm6                 \n\t" //qmul
                "packssdw %%mm6, %%mm6          \n\t"
                "packssdw %%mm6, %%mm6          \n\t"
                "movd %2, %%mm5                 \n\t" //qadd
                "pxor %%mm7, %%mm7              \n\t"
                "packssdw %%mm5, %%mm5          \n\t"
                "packssdw %%mm5, %%mm5          \n\t"
                "psubw %%mm5, %%mm7             \n\t"
                "pxor %%mm4, %%mm4              \n\t"
                ASMALIGN(4)
                "1:                             \n\t"
                "movq (%0, %3), %%mm0           \n\t"
                "movq 8(%0, %3), %%mm1          \n\t"

                "pmullw %%mm6, %%mm0            \n\t"
                "pmullw %%mm6, %%mm1            \n\t"

                "movq (%0, %3), %%mm2           \n\t"
                "movq 8(%0, %3), %%mm3          \n\t"

                "pcmpgtw %%mm4, %%mm2           \n\t" // block[i] < 0 ? -1 : 0
                "pcmpgtw %%mm4, %%mm3           \n\t" // block[i] < 0 ? -1 : 0

                "pxor %%mm2, %%mm0              \n\t"
                "pxor %%mm3, %%mm1              \n\t"

                "paddw %%mm7, %%mm0             \n\t"
                "paddw %%mm7, %%mm1             \n\t"

                "pxor %%mm0, %%mm2              \n\t"
                "pxor %%mm1, %%mm3              \n\t"

                "pcmpeqw %%mm7, %%mm0           \n\t" // block[i] == 0 ? -1 : 0
                "pcmpeqw %%mm7, %%mm1           \n\t" // block[i] == 0 ? -1 : 0

                "pandn %%mm2, %%mm0             \n\t"
                "pandn %%mm3, %%mm1             \n\t"

                "movq %%mm0, (%0, %3)           \n\t"
                "movq %%mm1, 8(%0, %3)          \n\t"

                "add $16, %3                    \n\t"
                "jng 1b                         \n\t"
                ::"r" (block+nCoeffs), "rm"(qmul), "rm" (qadd), "r" (2*(-nCoeffs))
                : "memory"
        );
}


/*
  NK:
  Note: looking at PARANOID:
  "enable all paranoid tests for rounding, overflows, etc..."

#ifdef PARANOID
                if (level < -2048 || level > 2047)
                    fprintf(stderr, "unquant error %d %d\n", i, level);
#endif
  We can suppose that result of two multiplications can't be greater than 0xFFFF
  i.e. is 16-bit, so we use here only PMULLW instruction and can avoid
  a complex multiplication.
=====================================================
 Full formula for multiplication of 2 integer numbers
 which are represent as high:low words:
 input: value1 = high1:low1
        value2 = high2:low2
 output: value3 = value1*value2
 value3=high3:low3 (on overflow: modulus 2^32 wrap-around)
 this mean that for 0x123456 * 0x123456 correct result is 0x766cb0ce4
 but this algorithm will compute only 0x66cb0ce4
 this limited by 16-bit size of operands
 ---------------------------------
 tlow1 = high1*low2
 tlow2 = high2*low1
 tlow1 = tlow1 + tlow2
 high3:low3 = low1*low2
 high3 += tlow1
*/
static void dct_unquantize_mpeg1_intra_mmx(MpegEncContext *s,
                                     DCTELEM *block, int n, int qscale)
{
    x86_reg nCoeffs;
    const uint16_t *quant_matrix;
    int block0;

    assert(s->block_last_index[n]>=0);

    nCoeffs= s->intra_scantable.raster_end[ s->block_last_index[n] ]+1;

    if (n < 4)
        block0 = block[0] * s->y_dc_scale;
    else
        block0 = block[0] * s->c_dc_scale;
    /* XXX: only mpeg1 */
    quant_matrix = s->intra_matrix;
__asm__ volatile(
                "pcmpeqw %%mm7, %%mm7           \n\t"
                "psrlw $15, %%mm7               \n\t"
                "movd %2, %%mm6                 \n\t"
                "packssdw %%mm6, %%mm6          \n\t"
                "packssdw %%mm6, %%mm6          \n\t"
                "mov %3, %%"REG_a"              \n\t"
                ASMALIGN(4)
                "1:                             \n\t"
                "movq (%0, %%"REG_a"), %%mm0    \n\t"
                "movq 8(%0, %%"REG_a"), %%mm1   \n\t"
                "movq (%1, %%"REG_a"), %%mm4    \n\t"
                "movq 8(%1, %%"REG_a"), %%mm5   \n\t"
                "pmullw %%mm6, %%mm4            \n\t" // q=qscale*quant_matrix[i]
                "pmullw %%mm6, %%mm5            \n\t" // q=qscale*quant_matrix[i]
                "pxor %%mm2, %%mm2              \n\t"
                "pxor %%mm3, %%mm3              \n\t"
                "pcmpgtw %%mm0, %%mm2           \n\t" // block[i] < 0 ? -1 : 0
                "pcmpgtw %%mm1, %%mm3           \n\t" // block[i] < 0 ? -1 : 0
                "pxor %%mm2, %%mm0              \n\t"
                "pxor %%mm3, %%mm1              \n\t"
                "psubw %%mm2, %%mm0             \n\t" // abs(block[i])
                "psubw %%mm3, %%mm1             \n\t" // abs(block[i])
                "pmullw %%mm4, %%mm0            \n\t" // abs(block[i])*q
                "pmullw %%mm5, %%mm1            \n\t" // abs(block[i])*q
                "pxor %%mm4, %%mm4              \n\t"
                "pxor %%mm5, %%mm5              \n\t" // FIXME slow
                "pcmpeqw (%0, %%"REG_a"), %%mm4 \n\t" // block[i] == 0 ? -1 : 0
                "pcmpeqw 8(%0, %%"REG_a"), %%mm5\n\t" // block[i] == 0 ? -1 : 0
                "psraw $3, %%mm0                \n\t"
                "psraw $3, %%mm1                \n\t"
                "psubw %%mm7, %%mm0             \n\t"
                "psubw %%mm7, %%mm1             \n\t"
                "por %%mm7, %%mm0               \n\t"
                "por %%mm7, %%mm1               \n\t"
                "pxor %%mm2, %%mm0              \n\t"
                "pxor %%mm3, %%mm1              \n\t"
                "psubw %%mm2, %%mm0             \n\t"
                "psubw %%mm3, %%mm1             \n\t"
                "pandn %%mm0, %%mm4             \n\t"
                "pandn %%mm1, %%mm5             \n\t"
                "movq %%mm4, (%0, %%"REG_a")    \n\t"
                "movq %%mm5, 8(%0, %%"REG_a")   \n\t"

                "add $16, %%"REG_a"             \n\t"
                "js 1b                          \n\t"
                ::"r" (block+nCoeffs), "r"(quant_matrix+nCoeffs), "rm" (qscale), "g" (-2*nCoeffs)
                : "%"REG_a, "memory"
        );
    block[0]= block0;
}

static void dct_unquantize_mpeg1_inter_mmx(MpegEncContext *s,
                                     DCTELEM *block, int n, int qscale)
{
    x86_reg nCoeffs;
    const uint16_t *quant_matrix;

    assert(s->block_last_index[n]>=0);

    nCoeffs= s->intra_scantable.raster_end[ s->block_last_index[n] ]+1;

        quant_matrix = s->inter_matrix;
__asm__ volatile(
                "pcmpeqw %%mm7, %%mm7           \n\t"
                "psrlw $15, %%mm7               \n\t"
                "movd %2, %%mm6                 \n\t"
                "packssdw %%mm6, %%mm6          \n\t"
                "packssdw %%mm6, %%mm6          \n\t"
                "mov %3, %%"REG_a"              \n\t"
                ASMALIGN(4)
                "1:                             \n\t"
                "movq (%0, %%"REG_a"), %%mm0    \n\t"
                "movq 8(%0, %%"REG_a"), %%mm1   \n\t"
                "movq (%1, %%"REG_a"), %%mm4    \n\t"
                "movq 8(%1, %%"REG_a"), %%mm5   \n\t"
                "pmullw %%mm6, %%mm4            \n\t" // q=qscale*quant_matrix[i]
                "pmullw %%mm6, %%mm5            \n\t" // q=qscale*quant_matrix[i]
                "pxor %%mm2, %%mm2              \n\t"
                "pxor %%mm3, %%mm3              \n\t"
                "pcmpgtw %%mm0, %%mm2           \n\t" // block[i] < 0 ? -1 : 0
                "pcmpgtw %%mm1, %%mm3           \n\t" // block[i] < 0 ? -1 : 0
                "pxor %%mm2, %%mm0              \n\t"
                "pxor %%mm3, %%mm1              \n\t"
                "psubw %%mm2, %%mm0             \n\t" // abs(block[i])
                "psubw %%mm3, %%mm1             \n\t" // abs(block[i])
                "paddw %%mm0, %%mm0             \n\t" // abs(block[i])*2
                "paddw %%mm1, %%mm1             \n\t" // abs(block[i])*2
                "paddw %%mm7, %%mm0             \n\t" // abs(block[i])*2 + 1
                "paddw %%mm7, %%mm1             \n\t" // abs(block[i])*2 + 1
                "pmullw %%mm4, %%mm0            \n\t" // (abs(block[i])*2 + 1)*q
                "pmullw %%mm5, %%mm1            \n\t" // (abs(block[i])*2 + 1)*q
                "pxor %%mm4, %%mm4              \n\t"
                "pxor %%mm5, %%mm5              \n\t" // FIXME slow
                "pcmpeqw (%0, %%"REG_a"), %%mm4 \n\t" // block[i] == 0 ? -1 : 0
                "pcmpeqw 8(%0, %%"REG_a"), %%mm5\n\t" // block[i] == 0 ? -1 : 0
                "psraw $4, %%mm0                \n\t"
                "psraw $4, %%mm1                \n\t"
                "psubw %%mm7, %%mm0             \n\t"
                "psubw %%mm7, %%mm1             \n\t"
                "por %%mm7, %%mm0               \n\t"
                "por %%mm7, %%mm1               \n\t"
                "pxor %%mm2, %%mm0              \n\t"
                "pxor %%mm3, %%mm1              \n\t"
                "psubw %%mm2, %%mm0             \n\t"
                "psubw %%mm3, %%mm1             \n\t"
                "pandn %%mm0, %%mm4             \n\t"
                "pandn %%mm1, %%mm5             \n\t"
                "movq %%mm4, (%0, %%"REG_a")    \n\t"
                "movq %%mm5, 8(%0, %%"REG_a")   \n\t"

                "add $16, %%"REG_a"             \n\t"
                "js 1b                          \n\t"
                ::"r" (block+nCoeffs), "r"(quant_matrix+nCoeffs), "rm" (qscale), "g" (-2*nCoeffs)
                : "%"REG_a, "memory"
        );
}

static void dct_unquantize_mpeg2_intra_mmx(MpegEncContext *s,
                                     DCTELEM *block, int n, int qscale)
{
    x86_reg nCoeffs;
    const uint16_t *quant_matrix;
    int block0;

    assert(s->block_last_index[n]>=0);

    if(s->alternate_scan) nCoeffs= 63; //FIXME
    else nCoeffs= s->intra_scantable.raster_end[ s->block_last_index[n] ];

    if (n < 4)
        block0 = block[0] * s->y_dc_scale;
    else
        block0 = block[0] * s->c_dc_scale;
    quant_matrix = s->intra_matrix;
__asm__ volatile(
                "pcmpeqw %%mm7, %%mm7           \n\t"
                "psrlw $15, %%mm7               \n\t"
                "movd %2, %%mm6                 \n\t"
                "packssdw %%mm6, %%mm6          \n\t"
                "packssdw %%mm6, %%mm6          \n\t"
                "mov %3, %%"REG_a"              \n\t"
                ASMALIGN(4)
                "1:                             \n\t"
                "movq (%0, %%"REG_a"), %%mm0    \n\t"
                "movq 8(%0, %%"REG_a"), %%mm1   \n\t"
                "movq (%1, %%"REG_a"), %%mm4    \n\t"
                "movq 8(%1, %%"REG_a"), %%mm5   \n\t"
                "pmullw %%mm6, %%mm4            \n\t" // q=qscale*quant_matrix[i]
                "pmullw %%mm6, %%mm5            \n\t" // q=qscale*quant_matrix[i]
                "pxor %%mm2, %%mm2              \n\t"
                "pxor %%mm3, %%mm3              \n\t"
                "pcmpgtw %%mm0, %%mm2           \n\t" // block[i] < 0 ? -1 : 0
                "pcmpgtw %%mm1, %%mm3           \n\t" // block[i] < 0 ? -1 : 0
                "pxor %%mm2, %%mm0              \n\t"
                "pxor %%mm3, %%mm1              \n\t"
                "psubw %%mm2, %%mm0             \n\t" // abs(block[i])
                "psubw %%mm3, %%mm1             \n\t" // abs(block[i])
                "pmullw %%mm4, %%mm0            \n\t" // abs(block[i])*q
                "pmullw %%mm5, %%mm1            \n\t" // abs(block[i])*q
                "pxor %%mm4, %%mm4              \n\t"
                "pxor %%mm5, %%mm5              \n\t" // FIXME slow
                "pcmpeqw (%0, %%"REG_a"), %%mm4 \n\t" // block[i] == 0 ? -1 : 0
                "pcmpeqw 8(%0, %%"REG_a"), %%mm5\n\t" // block[i] == 0 ? -1 : 0
                "psraw $3, %%mm0                \n\t"
                "psraw $3, %%mm1                \n\t"
                "pxor %%mm2, %%mm0              \n\t"
                "pxor %%mm3, %%mm1              \n\t"
                "psubw %%mm2, %%mm0             \n\t"
                "psubw %%mm3, %%mm1             \n\t"
                "pandn %%mm0, %%mm4             \n\t"
                "pandn %%mm1, %%mm5             \n\t"
                "movq %%mm4, (%0, %%"REG_a")    \n\t"
                "movq %%mm5, 8(%0, %%"REG_a")   \n\t"

                "add $16, %%"REG_a"             \n\t"
                "jng 1b                         \n\t"
                ::"r" (block+nCoeffs), "r"(quant_matrix+nCoeffs), "rm" (qscale), "g" (-2*nCoeffs)
                : "%"REG_a, "memory"
        );
    block[0]= block0;
        //Note, we do not do mismatch control for intra as errors cannot accumulate
}

static void dct_unquantize_mpeg2_inter_mmx(MpegEncContext *s,
                                     DCTELEM *block, int n, int qscale)
{
    x86_reg nCoeffs;
    const uint16_t *quant_matrix;

    assert(s->block_last_index[n]>=0);

    if(s->alternate_scan) nCoeffs= 63; //FIXME
    else nCoeffs= s->intra_scantable.raster_end[ s->block_last_index[n] ];

        quant_matrix = s->inter_matrix;
__asm__ volatile(
                "pcmpeqw %%mm7, %%mm7           \n\t"
                "psrlq $48, %%mm7               \n\t"
                "movd %2, %%mm6                 \n\t"
                "packssdw %%mm6, %%mm6          \n\t"
                "packssdw %%mm6, %%mm6          \n\t"
                "mov %3, %%"REG_a"              \n\t"
                ASMALIGN(4)
                "1:                             \n\t"
                "movq (%0, %%"REG_a"), %%mm0    \n\t"
                "movq 8(%0, %%"REG_a"), %%mm1   \n\t"
                "movq (%1, %%"REG_a"), %%mm4    \n\t"
                "movq 8(%1, %%"REG_a"), %%mm5   \n\t"
                "pmullw %%mm6, %%mm4            \n\t" // q=qscale*quant_matrix[i]
                "pmullw %%mm6, %%mm5            \n\t" // q=qscale*quant_matrix[i]
                "pxor %%mm2, %%mm2              \n\t"
                "pxor %%mm3, %%mm3              \n\t"
                "pcmpgtw %%mm0, %%mm2           \n\t" // block[i] < 0 ? -1 : 0
                "pcmpgtw %%mm1, %%mm3           \n\t" // block[i] < 0 ? -1 : 0
                "pxor %%mm2, %%mm0              \n\t"
                "pxor %%mm3, %%mm1              \n\t"
                "psubw %%mm2, %%mm0             \n\t" // abs(block[i])
                "psubw %%mm3, %%mm1             \n\t" // abs(block[i])
                "paddw %%mm0, %%mm0             \n\t" // abs(block[i])*2
                "paddw %%mm1, %%mm1             \n\t" // abs(block[i])*2
                "pmullw %%mm4, %%mm0            \n\t" // abs(block[i])*2*q
                "pmullw %%mm5, %%mm1            \n\t" // abs(block[i])*2*q
                "paddw %%mm4, %%mm0             \n\t" // (abs(block[i])*2 + 1)*q
                "paddw %%mm5, %%mm1             \n\t" // (abs(block[i])*2 + 1)*q
                "pxor %%mm4, %%mm4              \n\t"
                "pxor %%mm5, %%mm5              \n\t" // FIXME slow
                "pcmpeqw (%0, %%"REG_a"), %%mm4 \n\t" // block[i] == 0 ? -1 : 0
                "pcmpeqw 8(%0, %%"REG_a"), %%mm5\n\t" // block[i] == 0 ? -1 : 0
                "psrlw $4, %%mm0                \n\t"
                "psrlw $4, %%mm1                \n\t"
                "pxor %%mm2, %%mm0              \n\t"
                "pxor %%mm3, %%mm1              \n\t"
                "psubw %%mm2, %%mm0             \n\t"
                "psubw %%mm3, %%mm1             \n\t"
                "pandn %%mm0, %%mm4             \n\t"
                "pandn %%mm1, %%mm5             \n\t"
                "pxor %%mm4, %%mm7              \n\t"
                "pxor %%mm5, %%mm7              \n\t"
                "movq %%mm4, (%0, %%"REG_a")    \n\t"
                "movq %%mm5, 8(%0, %%"REG_a")   \n\t"

                "add $16, %%"REG_a"             \n\t"
                "jng 1b                         \n\t"
                "movd 124(%0, %3), %%mm0        \n\t"
                "movq %%mm7, %%mm6              \n\t"
                "psrlq $32, %%mm7               \n\t"
                "pxor %%mm6, %%mm7              \n\t"
                "movq %%mm7, %%mm6              \n\t"
                "psrlq $16, %%mm7               \n\t"
                "pxor %%mm6, %%mm7              \n\t"
                "pslld $31, %%mm7               \n\t"
                "psrlq $15, %%mm7               \n\t"
                "pxor %%mm7, %%mm0              \n\t"
                "movd %%mm0, 124(%0, %3)        \n\t"

                ::"r" (block+nCoeffs), "r"(quant_matrix+nCoeffs), "rm" (qscale), "r" (-2*nCoeffs)
                : "%"REG_a, "memory"
        );
}

static void  denoise_dct_mmx(MpegEncContext *s, DCTELEM *block){
    const int intra= s->mb_intra;
    int *sum= s->dct_error_sum[intra];
    uint16_t *offset= s->dct_offset[intra];

    s->dct_count[intra]++;

    __asm__ volatile(
        "pxor %%mm7, %%mm7                      \n\t"
        "1:                                     \n\t"
        "pxor %%mm0, %%mm0                      \n\t"
        "pxor %%mm1, %%mm1                      \n\t"
        "movq (%0), %%mm2                       \n\t"
        "movq 8(%0), %%mm3                      \n\t"
        "pcmpgtw %%mm2, %%mm0                   \n\t"
        "pcmpgtw %%mm3, %%mm1                   \n\t"
        "pxor %%mm0, %%mm2                      \n\t"
        "pxor %%mm1, %%mm3                      \n\t"
        "psubw %%mm0, %%mm2                     \n\t"
        "psubw %%mm1, %%mm3                     \n\t"
        "movq %%mm2, %%mm4                      \n\t"
        "movq %%mm3, %%mm5                      \n\t"
        "psubusw (%2), %%mm2                    \n\t"
        "psubusw 8(%2), %%mm3                   \n\t"
        "pxor %%mm0, %%mm2                      \n\t"
        "pxor %%mm1, %%mm3                      \n\t"
        "psubw %%mm0, %%mm2                     \n\t"
        "psubw %%mm1, %%mm3                     \n\t"
        "movq %%mm2, (%0)                       \n\t"
        "movq %%mm3, 8(%0)                      \n\t"
        "movq %%mm4, %%mm2                      \n\t"
        "movq %%mm5, %%mm3                      \n\t"
        "punpcklwd %%mm7, %%mm4                 \n\t"
        "punpckhwd %%mm7, %%mm2                 \n\t"
        "punpcklwd %%mm7, %%mm5                 \n\t"
        "punpckhwd %%mm7, %%mm3                 \n\t"
        "paddd (%1), %%mm4                      \n\t"
        "paddd 8(%1), %%mm2                     \n\t"
        "paddd 16(%1), %%mm5                    \n\t"
        "paddd 24(%1), %%mm3                    \n\t"
        "movq %%mm4, (%1)                       \n\t"
        "movq %%mm2, 8(%1)                      \n\t"
        "movq %%mm5, 16(%1)                     \n\t"
        "movq %%mm3, 24(%1)                     \n\t"
        "add $16, %0                            \n\t"
        "add $32, %1                            \n\t"
        "add $16, %2                            \n\t"
        "cmp %3, %0                             \n\t"
            " jb 1b                             \n\t"
        : "+r" (block), "+r" (sum), "+r" (offset)
        : "r"(block+64)
    );
}

static void  denoise_dct_sse2(MpegEncContext *s, DCTELEM *block){
    const int intra= s->mb_intra;
    int *sum= s->dct_error_sum[intra];
    uint16_t *offset= s->dct_offset[intra];

    s->dct_count[intra]++;

    __asm__ volatile(
        "pxor %%xmm7, %%xmm7                    \n\t"
        "1:                                     \n\t"
        "pxor %%xmm0, %%xmm0                    \n\t"
        "pxor %%xmm1, %%xmm1                    \n\t"
        "movdqa (%0), %%xmm2                    \n\t"
        "movdqa 16(%0), %%xmm3                  \n\t"
        "pcmpgtw %%xmm2, %%xmm0                 \n\t"
        "pcmpgtw %%xmm3, %%xmm1                 \n\t"
        "pxor %%xmm0, %%xmm2                    \n\t"
        "pxor %%xmm1, %%xmm3                    \n\t"
        "psubw %%xmm0, %%xmm2                   \n\t"
        "psubw %%xmm1, %%xmm3                   \n\t"
        "movdqa %%xmm2, %%xmm4                  \n\t"
        "movdqa %%xmm3, %%xmm5                  \n\t"
        "psubusw (%2), %%xmm2                   \n\t"
        "psubusw 16(%2), %%xmm3                 \n\t"
        "pxor %%xmm0, %%xmm2                    \n\t"
        "pxor %%xmm1, %%xmm3                    \n\t"
        "psubw %%xmm0, %%xmm2                   \n\t"
        "psubw %%xmm1, %%xmm3                   \n\t"
        "movdqa %%xmm2, (%0)                    \n\t"
        "movdqa %%xmm3, 16(%0)                  \n\t"
        "movdqa %%xmm4, %%xmm6                  \n\t"
        "movdqa %%xmm5, %%xmm0                  \n\t"
        "punpcklwd %%xmm7, %%xmm4               \n\t"
        "punpckhwd %%xmm7, %%xmm6               \n\t"
        "punpcklwd %%xmm7, %%xmm5               \n\t"
        "punpckhwd %%xmm7, %%xmm0               \n\t"
        "paddd (%1), %%xmm4                     \n\t"
        "paddd 16(%1), %%xmm6                   \n\t"
        "paddd 32(%1), %%xmm5                   \n\t"
        "paddd 48(%1), %%xmm0                   \n\t"
        "movdqa %%xmm4, (%1)                    \n\t"
        "movdqa %%xmm6, 16(%1)                  \n\t"
        "movdqa %%xmm5, 32(%1)                  \n\t"
        "movdqa %%xmm0, 48(%1)                  \n\t"
        "add $32, %0                            \n\t"
        "add $64, %1                            \n\t"
        "add $32, %2                            \n\t"
        "cmp %3, %0                             \n\t"
            " jb 1b                             \n\t"
        : "+r" (block), "+r" (sum), "+r" (offset)
        : "r"(block+64)
    );
}

#if HAVE_SSSE3
#define HAVE_SSSE3_BAK
#endif
#undef HAVE_SSSE3
#define HAVE_SSSE3 0

#undef HAVE_SSE2
#undef HAVE_MMX2
#define HAVE_SSE2 0
#define HAVE_MMX2 0
#define RENAME(a) a ## _MMX
#define RENAMEl(a) a ## _mmx
#include "mpegvideo_mmx_template.c"

#undef HAVE_MMX2
#define HAVE_MMX2 1
#undef RENAME
#undef RENAMEl
#define RENAME(a) a ## _MMX2
#define RENAMEl(a) a ## _mmx2
#include "mpegvideo_mmx_template.c"

#undef HAVE_SSE2
#define HAVE_SSE2 1
#undef RENAME
#undef RENAMEl
#define RENAME(a) a ## _SSE2
#define RENAMEl(a) a ## _sse2
#include "mpegvideo_mmx_template.c"

#ifdef HAVE_SSSE3_BAK
#undef HAVE_SSSE3
#define HAVE_SSSE3 1
#undef RENAME
#undef RENAMEl
#define RENAME(a) a ## _SSSE3
#define RENAMEl(a) a ## _sse2
#include "mpegvideo_mmx_template.c"
#endif

void MPV_common_init_mmx(MpegEncContext *s)
{
    if (mm_flags & FF_MM_MMX) {
        const int dct_algo = s->avctx->dct_algo;

        s->dct_unquantize_h263_intra = dct_unquantize_h263_intra_mmx;
        s->dct_unquantize_h263_inter = dct_unquantize_h263_inter_mmx;
        s->dct_unquantize_mpeg1_intra = dct_unquantize_mpeg1_intra_mmx;
        s->dct_unquantize_mpeg1_inter = dct_unquantize_mpeg1_inter_mmx;
        if(!(s->flags & CODEC_FLAG_BITEXACT))
            s->dct_unquantize_mpeg2_intra = dct_unquantize_mpeg2_intra_mmx;
        s->dct_unquantize_mpeg2_inter = dct_unquantize_mpeg2_inter_mmx;

        if (mm_flags & FF_MM_SSE2) {
            s->denoise_dct= denoise_dct_sse2;
        } else {
                s->denoise_dct= denoise_dct_mmx;
        }

        if(dct_algo==FF_DCT_AUTO || dct_algo==FF_DCT_MMX){
#if HAVE_SSSE3
            if(mm_flags & FF_MM_SSSE3){
                s->dct_quantize= dct_quantize_SSSE3;
            } else
#endif
            if(mm_flags & FF_MM_SSE2){
                s->dct_quantize= dct_quantize_SSE2;
            } else if(mm_flags & FF_MM_MMX2){
                s->dct_quantize= dct_quantize_MMX2;
            } else {
                s->dct_quantize= dct_quantize_MMX;
            }
        }
    }
}