Timothy B. Terriberry
2014-Dec-25 22:51 UTC
[opus] [Opus][RFC][FFT][Fixed-Point][NEON] NEON-Optimize Fixed-Point FFT?
Jean-Marc Valin wrote:> There is definitely some use for a Neon fixed-point FFT. How much > exactly I'm not sure. Fixed-point is a bit more than just a fall-backWell, we use fixed-point mode by default in Firefox for both Firefox OS and Fennec (Firefox on Android). The reason is that, although there is some NEON-class hardware where float does finally appear to be a little bit faster (e.g., recent A9's), there are still plenty where it is _significantly_ slower. So if you're going to pick one version to run on many devices, fixed-point has much better worst-case performance.
Phil Wang
2014-Dec-26 03:12 UTC
[opus] [Opus][RFC][FFT][Fixed-Point][NEON] NEON-Optimize Fixed-Point FFT?
Thanks Timothy and Jean-Marc, I will start NEON optimizing fixed-point FFT. Is it int32 good enough? Benchmark data shows that FFT using int16 is much faster than FFT using int32.> -----Original Message----- > From: Timothy B. Terriberry [mailto:tterribe at xiph.org] > Sent: Friday, December 26, 2014 6:52 AM > To: Phil Wang; opus at xiph.org > Cc: Zhongwei Yao; Yang Zhang; Zhou (Joe) Yu; Steve Bannister > Subject: Re: [opus] [Opus][RFC][FFT][Fixed-Point][NEON] NEON-Optimize > Fixed-Point FFT? > > Jean-Marc Valin wrote: > > There is definitely some use for a Neon fixed-point FFT. How much > > exactly I'm not sure. Fixed-point is a bit more than just a fall-back > > Well, we use fixed-point mode by default in Firefox for both Firefox OS and > Fennec (Firefox on Android). The reason is that, although there is some NEON- > class hardware where float does finally appear to be a little bit faster (e.g., > recent A9's), there are still plenty where it is _significantly_ slower. So if > you're going to pick one version to run on many devices, fixed-point has much > better worst-case performance.-- IMPORTANT NOTICE: The contents of this email and any attachments are confidential and may also be privileged. If you are not the intended recipient, please notify the sender immediately and do not disclose the contents to any other person, use it for any purpose, or store or copy the information in any medium. Thank you. ARM Limited, Registered office 110 Fulbourn Road, Cambridge CB1 9NJ, Registered in England & Wales, Company No: 2557590 ARM Holdings plc, Registered office 110 Fulbourn Road, Cambridge CB1 9NJ, Registered in England & Wales, Company No: 2548782
Timothy B. Terriberry
2014-Dec-26 03:25 UTC
[opus] [Opus][RFC][FFT][Fixed-Point][NEON] NEON-Optimize Fixed-Point FFT?
Phil Wang wrote:> Is it int32 good enough? Benchmark data shows that FFT using int16 is much faster than FFT using int32.For kissfft, we found that we needed the input and output to be 32 bits to get enough accuracy, but that the twiddles could be 16-bit.
Maybe Matching Threads
- [RFC][FFT][Fixed-Point][NEON] NEON-Optimize
- [RFC][FFT][Fixed-Point][NEON] NEON-Optimize Fixed-Point FFT?
- [RFC][FFT][Fixed-Point][NEON] NEON-Optimize Fixed-Point FFT?
- [RFC][FFT][Fixed-Point][NEON] NEON-Optimize
- [RFC][FFT][Fixed-Point][NEON] NEON-Optimize Fixed-Point FFT?