Marcus Asteborg
Marcus Asteborg
Sorry, what I meant was. If you use the same build options for opus lib only and use same/similar input can you reproduce the same memory corruption with Opus demo...
You might wanna measure heap usage on X86 and ensure that you have enough heap on ESP32 for Opus to work. Also there is an espressif engineer that seems to...
can you do a run with fixed_point enabled?
And I guess you can --disable-intrinsic to see if the perf comes from more neon optimizations in fixed point or if the fixed point implementation is more efficient running on...
Then run floating point with disabled intrinsics. Are you also running complexity 0? The test includes both encoder and decoder?
There is for sure improvements that can be done here no doubt. But it would be good to understand how it happened by providing the CMake options set and what...
Also what version of Opus (what is the latest commit)
@ZeroAbility There was bugs in regarding to avx flags in the first iteration of cmake, there is fixes here: https://github.com/xiph/opus/commit/927de8453c502586c03e25c169ec08f2a93ebc02 Can you upgrade to latest Opus master and see if...
So the PRESUME was added here mostly for when the target machine was known to support the intrinsic version but is not the same as the buildmachine. Think separate buildsystem...
@davidebeatrici just to clarify what do want added to CMake here?