vvenc
vvenc copied to clipboard
CMake presumes x86 SIMD and ignores ARM at all
To build the project on an ARM-based environment other an Apple (e.g. Raspberry Pi) requires manually editing CMakeLists.txt at the moment because it doesn't actually detect what platform you have and presumes x86.
VVenC is for now an encoder supporting exclusively the x86 platform. For that reason there is no need to support any other build targets. We support compilation under Mac OS X, only since Rosetta2 is such an amazing emulation environment.
We are planning to add ARM support with SIMD everywhere, analogue to VVdeC, sometime in Fall 2022. After this feature has been implemented VVenC will properly build as an ARM binary on ARM targets.
Thanks for the explanation.
Sure there is a roadmap, but still I can’t see why building on generic ARM targets has to be intentionally blocked. After all, it still works without SIMD intrinsics.
Its neither blocked or intentional. It has just never been intended to work or tested.
If you only develop on one platform for one platform your code eventually only works on that platform. That's just it.
I see. Anyways, after playing around with vvenc I believe it’s already an excellent encoder at this point and thanks for the hard work.
Thanks.
As mentioned, you can expect ARM support, with proper testing as a supported platform, after the summer.