New issue
Advanced search Search tips
Note: Color blocks (like or ) mean that a user may not be available. Tooltip shows the reason.
Starred by 8 users

Issue metadata

Status: Archived
Owner:
Last visit 17 days ago
Closed: Nov 2017
Cc:
Components:
NextAction: ----
OS: ----
Pri: 2
Type: Enhancement



Sign in to add a comment

Switch FFT to OpenMAX/pffft/Ne10

Project Member Reported by andrew@webrtc.org, Apr 21 2014

Issue description

This may consist of finalizing, testing and landing Kevin's old CL here:
http://review.webrtc.org/1819004/

Alternately we may prefer to use the OpenMAX APIs directly. Turaj is bringing in OpenMAX now for MPLC, which has x86 support.
 
Project Member

Comment 1 by andrew@webrtc.org, Apr 21 2014

This will immediately impact:
AECM (though we should probably ignore this, as it's going away)
NS fixed-point

Pieces we could move to OpenMAX in the future:
AEC (have to profile against the current SSE optimizations, and soon-to-be-started NEON optimizations.)
NS floating-point
Transient suppression

Project Member

Comment 2 by andrew@webrtc.org, May 14 2014

Cc: rtoy@google.com
Summary: Switch FFT to OpenMAX/pffft (was: Switch FFT to OpenMAX)
Actually, we should strongly consider using pffft instead. rtoy's benchmarks show it is faster than openmax on x86. Work is ongoing to improve the NEON implementation as well.
Project Member

Comment 3 by andrew@webrtc.org, May 31 2014

Labels: -Restrict-View-Talk Restrict-View-Google
Project Member

Comment 4 by bjornv@webrtc.org, Jun 2 2014

Labels: Area-SignalProcessing
Project Member

Comment 5 by andrew@webrtc.org, Jun 23 2014

Summary: Switch FFT to OpenMAX/pffft/Ne10 (was: Switch FFT to OpenMAX/pffft)
We should consider Ne10 as well:
https://github.com/projectNe10/Ne10

Yang Zhang from ARM has recently worked on improving performance:
I have updated the float/int32/int16 FFT of Ne10. Currently they are faster than OpenMax DL’s on V7. I also added the intrinsic implementation for V8 Aarch 64 mode.  And compared with C version, V8 intrinsic has better performance boost than V7 assembly.
 
Next, I’m planning to investigate the non-power-of-2 FFT.
Project Member

Comment 6 by tina.legrand@webrtc.org, Oct 30 2014

Labels: EngTriaged IceBox
Project Member

Comment 7 by bjornv@webrtc.org, Jun 9 2015

Cc: -bjornv@webrtc.org
Project Member

Comment 8 by andrew@webrtc.org, Nov 26 2015

Labels: -Restrict-View-Google
Project Member

Comment 9 by henrik.lundin@webrtc.org, Oct 5 2016

Components: Audio
Project Member

Comment 10 by henrik.lundin@webrtc.org, Oct 5 2016

Components: -SignalProcessing
Project Member

Comment 11 by tina.legrand@webrtc.org, Nov 7 2017

Status: Archived (was: Assigned)
[Bulk edit] This issue hasn't been modified the last twelve months -> archiving.

If this is still a valid issue that should be open, please reopen again.

Sign in to add a comment