Haus1

If you can include the output of `rocminfo` and `hipconfig` it may help with pointing you in the right direction. Setting AMD_LOG_LEVEL=4 can also be helpful. I'm guessing it's a...

huh, somehow I misread 6.0 as 5.6. ironically, if it's a gfx1036 that may be the solution for you https://github.com/ROCm/ROCm/discussions/2867 Xnack support typically seems to be the first thing to...

> That is also something that could be done on the Vulkan backend. I've done some rudimentary support for Unified Memory Architecture, but haven't had the time to look further...

Yeah, it seems the hardware is designed to have 64 warps resident but limited register space only allows them to track 32 threads at a time. The rest seem to...