30 ms·
I can attest that it works really well on my 6900xt. Compiling CUDA kernels is merely a matter of a using a #define shim. Also, provided you download the ROCm p
by zamalek 3y ago
I can attest that it works really well on my 6900xt. Compiling CUDA kernels is merely a matter of a using a #define shim. Also, provided you download the ROCm pytorch (and force compatibility with the HSA_OVERRIDE_GFX_VERSION) everything just works.