diff options
| author | Reimar Döffinger <Reimar.Doeffinger@gmx.de> | 2014-11-15 18:53:58 +0100 |
|---|---|---|
| committer | Reimar Döffinger <Reimar.Doeffinger@gmx.de> | 2014-11-22 14:06:48 +0100 |
| commit | 33fc1ccb4910b05a0f64b47fc2957fd73d5806ac (patch) | |
| tree | f3a8424e7190ecce299020bd14dc399c0ea23cb6 | |
| parent | 478c61ccb2beb29d506f74a17ce9ca322698c50e (diff) | |
| download | ffmpeg-33fc1ccb4910b05a0f64b47fc2957fd73d5806ac.tar.gz | |
h264_mb: Use smaller data type for refs in await_references.
As far as I can tell the value can never go outside
the int16_t type.
And especially the cost of the initialization is
reduced quite a bit by making it smaller.
Overall decoding speedup not measurable though.
Signed-off-by: Reimar Döffinger <Reimar.Doeffinger@gmx.de>
| -rw-r--r-- | libavcodec/h264_mb.c | 4 |
1 files changed, 2 insertions, 2 deletions
diff --git a/libavcodec/h264_mb.c b/libavcodec/h264_mb.c index 7feae5761c..dd406c7068 100644 --- a/libavcodec/h264_mb.c +++ b/libavcodec/h264_mb.c @@ -49,7 +49,7 @@ static inline int get_lowest_part_list_y(H264Context *h, H264Picture *pic, int n return FFMAX(0, bottom); } -static inline void get_lowest_part_y(H264Context *h, int refs[2][48], int n, +static inline void get_lowest_part_y(H264Context *h, int16_t refs[2][48], int n, int height, int y_offset, int list0, int list1, int *nrefs) { @@ -96,7 +96,7 @@ static void await_references(H264Context *h) { const int mb_xy = h->mb_xy; const int mb_type = h->cur_pic.mb_type[mb_xy]; - int refs[2][48]; + int16_t refs[2][48]; int nrefs[2] = { 0 }; int ref, list; |