diff options
| author | James Darnley <james.darnley@gmail.com> | 2014-08-12 23:22:02 +0200 |
|---|---|---|
| committer | Michael Niedermayer <michaelni@gmx.at> | 2014-08-13 01:14:47 +0200 |
| commit | 0081a14e7d1e44536400df19ad6c9ea8118db2e9 (patch) | |
| tree | c332e6e548a984cb078419c7bec1e56999a65d40 /libavcodec/x86/flacdsp_init.c | |
| parent | 7cafdfe9c7992e95f08b320fc7b4f1512e4e2afb (diff) | |
lavc/flacenc: add sse4 version of the 16-bit lpc encoder
From 1.8 to 2.4 times faster. Runtime is reduced by 2 to 39%. The
speed-up generally increases with compression_level.
This lpc encoder is not used with levels < 3 so it provides no speed-up
in these cases.
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
Diffstat (limited to 'libavcodec/x86/flacdsp_init.c')
| -rw-r--r-- | libavcodec/x86/flacdsp_init.c | 4 |
1 files changed, 4 insertions, 0 deletions
diff --git a/libavcodec/x86/flacdsp_init.c b/libavcodec/x86/flacdsp_init.c index a071b3d3b7..ad88e5bbd1 100644 --- a/libavcodec/x86/flacdsp_init.c +++ b/libavcodec/x86/flacdsp_init.c @@ -27,6 +27,8 @@ void ff_flac_lpc_32_sse4(int32_t *samples, const int coeffs[32], int order, void ff_flac_lpc_32_xop(int32_t *samples, const int coeffs[32], int order, int qlevel, int len); +void ff_flac_enc_lpc_16_sse4(int32_t *, const int32_t *, int, int, const int32_t *,int); + av_cold void ff_flacdsp_init_x86(FLACDSPContext *c, enum AVSampleFormat fmt, int bps) { @@ -36,6 +38,8 @@ av_cold void ff_flacdsp_init_x86(FLACDSPContext *c, enum AVSampleFormat fmt, if (EXTERNAL_SSE4(cpu_flags)) { if (bps > 16 && CONFIG_FLAC_DECODER) c->lpc = ff_flac_lpc_32_sse4; + if (bps == 16 && CONFIG_FLAC_ENCODER && CONFIG_GPL) + c->lpc_encode = ff_flac_enc_lpc_16_sse4; } if (EXTERNAL_XOP(cpu_flags)) { if (bps > 16 && CONFIG_FLAC_DECODER) |