brsh1

I think it's somehow related to it being on linux because it works for me with rx580/570 on windows

should it work for gfx900?

now I wonder if I use the gpu as passtrough on my xen server should I expect pcie atomic issues cause of virtualization layer?

the reason I am asking is I have this error cant initialize nvml. when trying the version you suggested.

just to make sure what rocm version should I be using? 5.6?

should I be using: HSA_OVERRIDE_GFX_VERSION=10.3.0 for gfx900? torch.cuda.is_available reports true, but mnist example, stable diffusion both fail to run. stuck with 100% cpu until process killed by me. any ideas?