Alessio Bazzica 4e9c5b592a RNN VAD: GRU layer optimized ... Using `VectorMath::DotProduct()` in GatedRecurrentLayer to reuse existing SIMD optimizations. Results: - When SSE2/AVX2 is avilable, the GRU layer takes 40% of the unoptimized code - The realtime factor for the VAD improved as follows - SSE2: from 570x to 630x - AVX2: from 610x to 680x This CL also improved the GRU layer benchmark by (i) benchmarking a GRU layer havibng the same size of that used in the VAD and (ii) by prefetching a long input sequence. Bug: webrtc:10480 Change-Id: I9716b15661e4c6b81592b4cf7c172d90e41b5223 Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/195545 Reviewed-by: Per Åhgren <peah@webrtc.org> Commit-Queue: Alessio Bazzica <alessiob@webrtc.org> Cr-Commit-Position: refs/heads/master@{#32803}		2020-12-08 15:37:38 +00:00
..
band_energies.dat.sha1	AGC2 RNN VAD: Spectral features internal API.	2018-05-08 11:52:32 +00:00
gru_in.dat.sha1	RNN VAD: GRU layer optimized	2020-12-08 15:37:38 +00:00
pitch_buf_24k.dat.sha1
pitch_lp_res.dat.sha1
pitch_search_int.dat.sha1	AGC2 RNN VAD: Polishing.	2018-05-15 16:41:02 +00:00
samples.pcm.sha1	AGC2 RNN VAD: Polishing.	2018-05-15 16:41:02 +00:00
vad_prob.dat.sha1	RNN VAD: clean-up unit tests	2019-04-29 12:55:02 +00:00