Mercurial > libavcodec.hg
view mlib/dsputil_mlib.c @ 11352:6e0af2cfdcfe libavcodec
Do MC and IDCT in coding (hilbert) order
This increases the slice size to 64 pixels, due to having to decode an
entire chroma superblock row per slice.
This can be up to 6% slower depending on clip and CPU, but is necessary
for future optimizations that gain significantly more than was lost.
author | conrad |
---|---|
date | Wed, 03 Mar 2010 23:27:40 +0000 |
parents | 04423b2f6e0b |
children |
line wrap: on
line source
/* * Sun mediaLib optimized DSP utils * Copyright (c) 2001 Fabrice Bellard * * 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 "libavcodec/dsputil.h" #include "libavcodec/mpegvideo.h" #include <mlib_types.h> #include <mlib_status.h> #include <mlib_sys.h> #include <mlib_algebra.h> #include <mlib_video.h> /* misc */ static void get_pixels_mlib(DCTELEM *restrict block, const uint8_t *pixels, int line_size) { int i; for (i=0;i<8;i++) { mlib_VectorConvert_S16_U8_Mod((mlib_s16 *)block, (mlib_u8 *)pixels, 8); pixels += line_size; block += 8; } } static void diff_pixels_mlib(DCTELEM *restrict block, const uint8_t *s1, const uint8_t *s2, int line_size) { int i; for (i=0;i<8;i++) { mlib_VectorSub_S16_U8_Mod((mlib_s16 *)block, (mlib_u8 *)s1, (mlib_u8 *)s2, 8); s1 += line_size; s2 += line_size; block += 8; } } static void add_pixels_clamped_mlib(const DCTELEM *block, uint8_t *pixels, int line_size) { mlib_VideoAddBlock_U8_S16(pixels, (mlib_s16 *)block, line_size); } /* put block, width 16 pixel, height 8/16 */ static void put_pixels16_mlib (uint8_t * dest, const uint8_t * ref, int stride, int height) { switch (height) { case 8: mlib_VideoCopyRef_U8_U8_16x8(dest, (uint8_t *)ref, stride); break; case 16: mlib_VideoCopyRef_U8_U8_16x16(dest, (uint8_t *)ref, stride); break; default: assert(0); } } static void put_pixels16_x2_mlib (uint8_t * dest, const uint8_t * ref, int stride, int height) { switch (height) { case 8: mlib_VideoInterpX_U8_U8_16x8(dest, (uint8_t *)ref, stride, stride); break; case 16: mlib_VideoInterpX_U8_U8_16x16(dest, (uint8_t *)ref, stride, stride); break; default: assert(0); } } static void put_pixels16_y2_mlib (uint8_t * dest, const uint8_t * ref, int stride, int height) { switch (height) { case 8: mlib_VideoInterpY_U8_U8_16x8(dest, (uint8_t *)ref, stride, stride); break; case 16: mlib_VideoInterpY_U8_U8_16x16(dest, (uint8_t *)ref, stride, stride); break; default: assert(0); } } static void put_pixels16_xy2_mlib(uint8_t * dest, const uint8_t * ref, int stride, int height) { switch (height) { case 8: mlib_VideoInterpXY_U8_U8_16x8(dest, (uint8_t *)ref, stride, stride); break; case 16: mlib_VideoInterpXY_U8_U8_16x16(dest, (uint8_t *)ref, stride, stride); break; default: assert(0); } } /* put block, width 8 pixel, height 4/8/16 */ static void put_pixels8_mlib (uint8_t * dest, const uint8_t * ref, int stride, int height) { switch (height) { case 4: mlib_VideoCopyRef_U8_U8_8x4(dest, (uint8_t *)ref, stride); break; case 8: mlib_VideoCopyRef_U8_U8_8x8(dest, (uint8_t *)ref, stride); break; case 16: mlib_VideoCopyRef_U8_U8_8x16(dest, (uint8_t *)ref, stride); break; default: assert(0); } } static void put_pixels8_x2_mlib (uint8_t * dest, const uint8_t * ref, int stride, int height) { switch (height) { case 4: mlib_VideoInterpX_U8_U8_8x4(dest, (uint8_t *)ref, stride, stride); break; case 8: mlib_VideoInterpX_U8_U8_8x8(dest, (uint8_t *)ref, stride, stride); break; case 16: mlib_VideoInterpX_U8_U8_8x16(dest, (uint8_t *)ref, stride, stride); break; default: assert(0); } } static void put_pixels8_y2_mlib (uint8_t * dest, const uint8_t * ref, int stride, int height) { switch (height) { case 4: mlib_VideoInterpY_U8_U8_8x4(dest, (uint8_t *)ref, stride, stride); break; case 8: mlib_VideoInterpY_U8_U8_8x8(dest, (uint8_t *)ref, stride, stride); break; case 16: mlib_VideoInterpY_U8_U8_8x16(dest, (uint8_t *)ref, stride, stride); break; default: assert(0); } } static void put_pixels8_xy2_mlib(uint8_t * dest, const uint8_t * ref, int stride, int height) { switch (height) { case 4: mlib_VideoInterpXY_U8_U8_8x4(dest, (uint8_t *)ref, stride, stride); break; case 8: mlib_VideoInterpXY_U8_U8_8x8(dest, (uint8_t *)ref, stride, stride); break; case 16: mlib_VideoInterpXY_U8_U8_8x16(dest, (uint8_t *)ref, stride, stride); break; default: assert(0); } } /* average block, width 16 pixel, height 8/16 */ static void avg_pixels16_mlib (uint8_t * dest, const uint8_t * ref, int stride, int height) { switch (height) { case 8: mlib_VideoCopyRefAve_U8_U8_16x8(dest, (uint8_t *)ref, stride); break; case 16: mlib_VideoCopyRefAve_U8_U8_16x16(dest, (uint8_t *)ref, stride); break; default: assert(0); } } static void avg_pixels16_x2_mlib (uint8_t * dest, const uint8_t * ref, int stride, int height) { switch (height) { case 8: mlib_VideoInterpAveX_U8_U8_16x8(dest, (uint8_t *)ref, stride, stride); break; case 16: mlib_VideoInterpAveX_U8_U8_16x16(dest, (uint8_t *)ref, stride, stride); break; default: assert(0); } } static void avg_pixels16_y2_mlib (uint8_t * dest, const uint8_t * ref, int stride, int height) { switch (height) { case 8: mlib_VideoInterpAveY_U8_U8_16x8(dest, (uint8_t *)ref, stride, stride); break; case 16: mlib_VideoInterpAveY_U8_U8_16x16(dest, (uint8_t *)ref, stride, stride); break; default: assert(0); } } static void avg_pixels16_xy2_mlib(uint8_t * dest, const uint8_t * ref, int stride, int height) { switch (height) { case 8: mlib_VideoInterpAveXY_U8_U8_16x8(dest, (uint8_t *)ref, stride, stride); break; case 16: mlib_VideoInterpAveXY_U8_U8_16x16(dest, (uint8_t *)ref, stride, stride); break; default: assert(0); } } /* average block, width 8 pixel, height 4/8/16 */ static void avg_pixels8_mlib (uint8_t * dest, const uint8_t * ref, int stride, int height) { switch (height) { case 4: mlib_VideoCopyRefAve_U8_U8_8x4(dest, (uint8_t *)ref, stride); break; case 8: mlib_VideoCopyRefAve_U8_U8_8x8(dest, (uint8_t *)ref, stride); break; case 16: mlib_VideoCopyRefAve_U8_U8_8x16(dest, (uint8_t *)ref, stride); break; default: assert(0); } } static void avg_pixels8_x2_mlib (uint8_t * dest, const uint8_t * ref, int stride, int height) { switch (height) { case 4: mlib_VideoInterpAveX_U8_U8_8x4(dest, (uint8_t *)ref, stride, stride); break; case 8: mlib_VideoInterpAveX_U8_U8_8x8(dest, (uint8_t *)ref, stride, stride); break; case 16: mlib_VideoInterpAveX_U8_U8_8x16(dest, (uint8_t *)ref, stride, stride); break; default: assert(0); } } static void avg_pixels8_y2_mlib (uint8_t * dest, const uint8_t * ref, int stride, int height) { switch (height) { case 4: mlib_VideoInterpAveY_U8_U8_8x4(dest, (uint8_t *)ref, stride, stride); break; case 8: mlib_VideoInterpAveY_U8_U8_8x8(dest, (uint8_t *)ref, stride, stride); break; case 16: mlib_VideoInterpAveY_U8_U8_8x16(dest, (uint8_t *)ref, stride, stride); break; default: assert(0); } } static void avg_pixels8_xy2_mlib(uint8_t * dest, const uint8_t * ref, int stride, int height) { switch (height) { case 4: mlib_VideoInterpAveXY_U8_U8_8x4(dest, (uint8_t *)ref, stride, stride); break; case 8: mlib_VideoInterpAveXY_U8_U8_8x8(dest, (uint8_t *)ref, stride, stride); break; case 16: mlib_VideoInterpAveXY_U8_U8_8x16(dest, (uint8_t *)ref, stride, stride); break; default: assert(0); } } /* swap byte order of a buffer */ static void bswap_buf_mlib(uint32_t *dst, const uint32_t *src, int w) { mlib_VectorReverseByteOrder_U32_U32(dst, src, w); } /* transformations */ static void ff_idct_put_mlib(uint8_t *dest, int line_size, DCTELEM *data) { int i; uint8_t *cm = ff_cropTbl + MAX_NEG_CROP; mlib_VideoIDCT8x8_S16_S16 (data, data); for(i=0;i<8;i++) { dest[0] = cm[data[0]]; dest[1] = cm[data[1]]; dest[2] = cm[data[2]]; dest[3] = cm[data[3]]; dest[4] = cm[data[4]]; dest[5] = cm[data[5]]; dest[6] = cm[data[6]]; dest[7] = cm[data[7]]; dest += line_size; data += 8; } } static void ff_idct_add_mlib(uint8_t *dest, int line_size, DCTELEM *data) { mlib_VideoIDCT8x8_S16_S16 (data, data); mlib_VideoAddBlock_U8_S16(dest, (mlib_s16 *)data, line_size); } static void ff_idct_mlib(DCTELEM *data) { mlib_VideoIDCT8x8_S16_S16 (data, data); } static void ff_fdct_mlib(DCTELEM *data) { mlib_VideoDCT8x8_S16_S16 (data, data); } void dsputil_init_mlib(DSPContext* c, AVCodecContext *avctx) { c->get_pixels = get_pixels_mlib; c->diff_pixels = diff_pixels_mlib; c->add_pixels_clamped = add_pixels_clamped_mlib; c->put_pixels_tab[0][0] = put_pixels16_mlib; c->put_pixels_tab[0][1] = put_pixels16_x2_mlib; c->put_pixels_tab[0][2] = put_pixels16_y2_mlib; c->put_pixels_tab[0][3] = put_pixels16_xy2_mlib; c->put_pixels_tab[1][0] = put_pixels8_mlib; c->put_pixels_tab[1][1] = put_pixels8_x2_mlib; c->put_pixels_tab[1][2] = put_pixels8_y2_mlib; c->put_pixels_tab[1][3] = put_pixels8_xy2_mlib; c->avg_pixels_tab[0][0] = avg_pixels16_mlib; c->avg_pixels_tab[0][1] = avg_pixels16_x2_mlib; c->avg_pixels_tab[0][2] = avg_pixels16_y2_mlib; c->avg_pixels_tab[0][3] = avg_pixels16_xy2_mlib; c->avg_pixels_tab[1][0] = avg_pixels8_mlib; c->avg_pixels_tab[1][1] = avg_pixels8_x2_mlib; c->avg_pixels_tab[1][2] = avg_pixels8_y2_mlib; c->avg_pixels_tab[1][3] = avg_pixels8_xy2_mlib; c->put_no_rnd_pixels_tab[0][0] = put_pixels16_mlib; c->put_no_rnd_pixels_tab[1][0] = put_pixels8_mlib; c->bswap_buf = bswap_buf_mlib; } void MPV_common_init_mlib(MpegEncContext *s) { if(s->avctx->dct_algo==FF_DCT_AUTO || s->avctx->dct_algo==FF_DCT_MLIB){ s->dsp.fdct = ff_fdct_mlib; } if(s->avctx->idct_algo==FF_IDCT_MLIB){ s->dsp.idct_put= ff_idct_put_mlib; s->dsp.idct_add= ff_idct_add_mlib; s->dsp.idct = ff_idct_mlib; s->dsp.idct_permutation_type= FF_NO_IDCT_PERM; } }