diff options
author | Reimar Döffinger <Reimar.Doeffinger@gmx.de> | 2021-01-16 02:03:45 +0100 |
---|---|---|
committer | Josh Dekker <josh@itanimul.li> | 2021-02-18 14:11:53 +0100 |
commit | 30f80d855bb0eb3742ac51d8900f636968d0b70b (patch) | |
tree | 87df9d0d487ec522bea842503a883a5b8f643c9f /libavcodec/hevcdsp.h | |
parent | 67c8c863c70f2989add027ffccccbe4b87988415 (diff) |
lavc/aarch64: port HEVC SIMD idct NEON
Makes SIMD-optimized 8x8 and 16x16 idcts for 8 and 10 bit depth
available on aarch64.
For a UHD HDR (10 bit) sample video these were consuming the most time
and this optimization reduced overall decode time from 19.4s to 16.4s,
approximately 15% speedup.
Test sample was the first 300 frames of "LG 4K HDR Demo - New York.ts",
running on Apple M1.
Signed-off-by: Josh Dekker <josh@itanimul.li>
Diffstat (limited to 'libavcodec/hevcdsp.h')
-rw-r--r-- | libavcodec/hevcdsp.h | 1 |
1 files changed, 1 insertions, 0 deletions
diff --git a/libavcodec/hevcdsp.h b/libavcodec/hevcdsp.h index c605a343d6..0e013a8328 100644 --- a/libavcodec/hevcdsp.h +++ b/libavcodec/hevcdsp.h @@ -129,6 +129,7 @@ void ff_hevc_dsp_init(HEVCDSPContext *hpc, int bit_depth); extern const int8_t ff_hevc_epel_filters[7][4]; extern const int8_t ff_hevc_qpel_filters[3][16]; +void ff_hevc_dsp_init_aarch64(HEVCDSPContext *c, const int bit_depth); void ff_hevc_dsp_init_arm(HEVCDSPContext *c, const int bit_depth); void ff_hevc_dsp_init_ppc(HEVCDSPContext *c, const int bit_depth); void ff_hevc_dsp_init_x86(HEVCDSPContext *c, const int bit_depth); |