diff options
| author | Christophe Gisquet <christophe.gisquet@gmail.com> | 2013-04-07 20:20:33 +0000 |
|---|---|---|
| committer | Michael Niedermayer <michaelni@gmx.at> | 2013-04-08 02:26:34 +0200 |
| commit | e2946e5c3482ab78913bfa5527a6fa9ea4e70bb4 (patch) | |
| tree | e5964637887b5189cd2b2f1f3719fcb03cc57694 /libavcodec/x86/sbrdsp_init.c | |
| parent | 1adf54de56be13a6df51c5d42f0ecf2e1ec0c648 (diff) | |
| download | ffmpeg-e2946e5c3482ab78913bfa5527a6fa9ea4e70bb4.tar.gz | |
x86: sbrdsp: implement SSE qmf_deint_bfly
From 312 to 89/68 (sse/sse2) cycles on Arrandale and Win64.
Sandybridge: 68/47 cycles.
Having a loop counter is a 7 cycle gain.
Unrolling is another 7 cycle gain.
Working in reverse scan is another 6 cycles.
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
Diffstat (limited to 'libavcodec/x86/sbrdsp_init.c')
| -rw-r--r-- | libavcodec/x86/sbrdsp_init.c | 7 |
1 files changed, 7 insertions, 0 deletions
diff --git a/libavcodec/x86/sbrdsp_init.c b/libavcodec/x86/sbrdsp_init.c index cb65a23f22..8cc37ddd4b 100644 --- a/libavcodec/x86/sbrdsp_init.c +++ b/libavcodec/x86/sbrdsp_init.c @@ -34,6 +34,8 @@ void ff_sbr_hf_gen_sse(float (*X_high)[2], const float (*X_low)[2], float bw, int start, int end); void ff_sbr_neg_odd_64_sse(float *z); void ff_sbr_qmf_post_shuffle_sse(float W[32][2], const float *z); +void ff_sbr_qmf_deint_bfly_sse(float *v, const float *src0, const float *src1); +void ff_sbr_qmf_deint_bfly_sse2(float *v, const float *src0, const float *src1); av_cold void ff_sbrdsp_init_x86(SBRDSPContext *s) { @@ -46,5 +48,10 @@ av_cold void ff_sbrdsp_init_x86(SBRDSPContext *s) s->hf_g_filt = ff_sbr_hf_g_filt_sse; s->hf_gen = ff_sbr_hf_gen_sse; s->qmf_post_shuffle = ff_sbr_qmf_post_shuffle_sse; + s->qmf_deint_bfly = ff_sbr_qmf_deint_bfly_sse; + } + + if (EXTERNAL_SSE2(mm_flags)) { + s->qmf_deint_bfly = ff_sbr_qmf_deint_bfly_sse2; } } |