diff options
| author | Ronald S. Bultje <rsbultje@gmail.com> | 2013-02-18 21:03:02 -0800 |
|---|---|---|
| committer | Michael Niedermayer <michaelni@gmx.at> | 2013-02-19 16:25:50 +0100 |
| commit | 1acd7d594c15aa491729c837ad3519d3469e620a (patch) | |
| tree | 4a515aa31c7483d8a2fdfa1d4399c9a9d2b29cae /libavcodec/h264dsp.h | |
| parent | a1f1ca96b42698002920467c0aab9e636893088a (diff) | |
| download | ffmpeg-1acd7d594c15aa491729c837ad3519d3469e620a.tar.gz | |
h264: integrate clear_blocks calls with IDCT.
The non-intra-pcm branch in hl_decode_mb (simple, 8bpp) goes from 700
to 672 cycles, and the complete loop of decode_mb_cabac and hl_decode_mb
(in the decode_slice loop) goes from 1759 to 1733 cycles on the clip
tested (cathedral), i.e. almost 30 cycles per mb faster.
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
Diffstat (limited to 'libavcodec/h264dsp.h')
| -rw-r--r-- | libavcodec/h264dsp.h | 4 |
1 files changed, 2 insertions, 2 deletions
diff --git a/libavcodec/h264dsp.h b/libavcodec/h264dsp.h index 82e147fe35..98ea15c330 100644 --- a/libavcodec/h264dsp.h +++ b/libavcodec/h264dsp.h @@ -103,8 +103,8 @@ typedef struct H264DSPContext { void (*h264_chroma_dc_dequant_idct)(int16_t *block, int qmul); /* bypass-transform */ - void (*h264_add_pixels8)(uint8_t *dst, int16_t *block, int stride); - void (*h264_add_pixels4)(uint8_t *dst, int16_t *block, int stride); + void (*h264_add_pixels8_clear)(uint8_t *dst, int16_t *block, int stride); + void (*h264_add_pixels4_clear)(uint8_t *dst, int16_t *block, int stride); } H264DSPContext; void ff_h264dsp_init(H264DSPContext *c, const int bit_depth, |