diff options
author | Yunqing Wang <yunqingwang@google.com> | 2013-12-27 15:25:54 -0800 |
---|---|---|
committer | Yunqing Wang <yunqingwang@google.com> | 2014-01-31 14:44:53 -0800 |
commit | 903801f1ef7ac8d13d4f57571d048b604e8aaafd (patch) | |
tree | 23567c0947d8492ea9333ff924ed02e0d505c8bb /vp9/decoder/vp9_onyxd_int.h | |
parent | e78c174e540117dcfcdff505d38478d4ac6df844 (diff) | |
download | libvpx-903801f1ef7ac8d13d4f57571d048b604e8aaafd.tar libvpx-903801f1ef7ac8d13d4f57571d048b604e8aaafd.tar.gz libvpx-903801f1ef7ac8d13d4f57571d048b604e8aaafd.tar.bz2 libvpx-903801f1ef7ac8d13d4f57571d048b604e8aaafd.zip |
vp9 decoder: row-based multi-threaded loopfilter
Implemented parallel loopfiltering, which uses existing tile-
decoding threads. Each thread works on one row, and when that row
is loopfiltered, it moves to next unattended row. To ensure the
correct filtering order, threads are synchronized and one
superblock is filtered only if the superblocks it depends on are
filtered already.
To reduce synchronization overhead and speed up the decoder, we use
nsync > 1 for high resolution.
Performance tests:
1. on desktop:
8-tile 4k video using 8 threads, speedup: 70% - 80%
4-tile HD video using 4 threads, speedup: ~35%
2. on mobile device(Nexus 7):
4-tile 1080p video using 4 threads, speedup: 18% - 25%
4-tile 1080p video using 2 threads, speedup: 10% - 15%
Change-Id: If54b4a11960dd706c22d5ad145ad94156031f36a
Diffstat (limited to 'vp9/decoder/vp9_onyxd_int.h')
-rw-r--r-- | vp9/decoder/vp9_onyxd_int.h | 3 |
1 files changed, 3 insertions, 0 deletions
diff --git a/vp9/decoder/vp9_onyxd_int.h b/vp9/decoder/vp9_onyxd_int.h index 242c600cc..6c6c23926 100644 --- a/vp9/decoder/vp9_onyxd_int.h +++ b/vp9/decoder/vp9_onyxd_int.h @@ -14,6 +14,7 @@ #include "./vpx_config.h" #include "vp9/common/vp9_onyxc_int.h" +#include "vp9/decoder/vp9_dthread.h" #include "vp9/decoder/vp9_onyxd.h" #include "vp9/decoder/vp9_thread.h" @@ -49,6 +50,8 @@ typedef struct VP9Decompressor { VP9Worker *tile_workers; int num_tile_workers; + VP9LfSync lf_row_sync; + /* Each tile column has its own MODE_INFO stream. This array indexes them by tile column index. */ MODE_INFO **mi_streams; |