diff options
author | Rostislav Pehlivanov <atomnuker@gmail.com> | 2016-10-08 15:59:14 +0100 |
---|---|---|
committer | Rostislav Pehlivanov <atomnuker@gmail.com> | 2016-10-18 21:41:18 +0100 |
commit | d2ae5f77c61a29c3c63cc4c41c74ccfca4167649 (patch) | |
tree | bd615f22e6807e025da052c411e12e9a7be0852e /libavcodec/aacenc_ltp.c | |
parent | 3b02f6dd7be880fd6c1bcaf2fd0c1314dcee7aa6 (diff) | |
download | ffmpeg-d2ae5f77c61a29c3c63cc4c41c74ccfca4167649.tar.gz |
aacenc: add SIMD optimizations for abs_pow34 and quantization
Performance improvements:
quant_bands:
with: 681 decicycles in quant_bands, 8388453 runs, 155 skips
without: 1190 decicycles in quant_bands, 8388386 runs, 222 skips
Around 42% for the function
Twoloop coder:
abs_pow34:
with/without: 7.82s/8.17s
Around 4% for the entire encoder
Both:
with/without: 7.15s/8.17s
Around 12% for the entire encoder
Fast coder:
abs_pow34:
with/without: 3.40s/3.77s
Around 10% for the entire encoder
Both:
with/without: 3.02s/3.77s
Around 20% faster for the entire encoder
Signed-off-by: Rostislav Pehlivanov <atomnuker@gmail.com>
Tested-by: Michael Niedermayer <michael@niedermayer.cc>
Reviewed-by: James Almer <jamrial@gmail.com>
Diffstat (limited to 'libavcodec/aacenc_ltp.c')
-rw-r--r-- | libavcodec/aacenc_ltp.c | 4 |
1 files changed, 2 insertions, 2 deletions
diff --git a/libavcodec/aacenc_ltp.c b/libavcodec/aacenc_ltp.c index b9d43b470a..1bec85b9e3 100644 --- a/libavcodec/aacenc_ltp.c +++ b/libavcodec/aacenc_ltp.c @@ -190,8 +190,8 @@ void ff_aac_search_for_ltp(AACEncContext *s, SingleChannelElement *sce, FFPsyBand *band = &s->psy.ch[s->cur_channel].psy_bands[(w+w2)*16+g]; for (i = 0; i < sce->ics.swb_sizes[g]; i++) PCD[i] = sce->coeffs[start+(w+w2)*128+i] - sce->lcoeffs[start+(w+w2)*128+i]; - abs_pow34_v(C34, &sce->coeffs[start+(w+w2)*128], sce->ics.swb_sizes[g]); - abs_pow34_v(PCD34, PCD, sce->ics.swb_sizes[g]); + s->abs_pow34(C34, &sce->coeffs[start+(w+w2)*128], sce->ics.swb_sizes[g]); + s->abs_pow34(PCD34, PCD, sce->ics.swb_sizes[g]); dist1 += quantize_band_cost(s, &sce->coeffs[start+(w+w2)*128], C34, sce->ics.swb_sizes[g], sce->sf_idx[(w+w2)*16+g], sce->band_type[(w+w2)*16+g], s->lambda/band->threshold, INFINITY, &bits_tmp1, NULL, 0); |