Dana, the only thing you didn't make clear is whether your code is using the fixed or floating data type. If it's using the floating one, it would be interesting to isolate why performance is so poor when more conventional code is generated by the compiler. I can understand float code being slightly slower than double, if the hardware FPU is implemented in double size, as it normally would be.
The floating performance as previously benchmarked is poor enough that I wondered whether there was really hardware, or whether some of that blobby code was processing float in an exception handler. On Wed, Feb 5, 2020, 8:16 PM Dana Myers <[email protected]> wrote: > On 2/5/2020 4:25 PM, Bruce Perens via Freetel-codec2 wrote: > > Bill, before you go any farther oh, you should make a floating point > benchmark. I don't believe the necessary performance is there. > > I used to think that, but then Espressif released their ESP-DSP library. I > ported my > Bell 202 modem from Cortex-M4F using CMSIS-DSP to ESP32 running at 240MHz > using ESP-DSP and see comparable performance per clock, and my actual modem > is single-threaded and thus uses only one of the ESP32's two cores. It's > conceivable > if some additional latency is tolerable and the algorithm divisible, it > could be split > over the two cores. > > Espressif offers both "ANSI C" portable functions and ESP32-specific > assembly > functions - the latter are considerably faster and what I am using. > > Cheers, > Dana K6JQ > > > > _______________________________________________ > Freetel-codec2 mailing list > [email protected] > https://lists.sourceforge.net/lists/listinfo/freetel-codec2 >
_______________________________________________ Freetel-codec2 mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/freetel-codec2
