Simon Lui

Oh okay, I guess this must be distro specific. I usually pull the latest package releases from https://github.com/intel/compute-runtime/releases/ to avoid having problems with the runtime and to ensure I get...

It works in Fedora and inside Intel's Python Conda environment with the pip installation using Python 3.12 for me doing inference.

I think I found the issue. In [GLIBC 2.41 release notes](https://sourceware.org/pipermail/libc-alpha/2025-January/164378.html), the following is stated. >Deprecated and removed features, and other changes affecting compatibility: ... >* dlopen and dlmopen no...

> We didn't have enough time to have a full scope validation on the workaround. Please apply the patchelf by yourself at this time. This issue will be addressed in...

You need the latest patchelf version 0.18 to set the flag which released in 2023 so not sure how old your Ubuntu or system running is.

> It is now August 2nd with no roadmap of when we can expect IPEX 2.8. Pytorch 2.8 isn't even released yet. You actually need that to be released before...

IPEX 2.6.1 did not solve this issue, and I think enough time has passed between an IPEX release and the release of Lunar Lake to warrant trying to get this...

If I use the Nightly Pytorch wheel intended for BMG on Linux at https://github.com/intel/intel-extension-for-pytorch/issues/764#issuecomment-2723179320 on my LNL laptop, the device lost issue seems to have been fixed, but Stable Diffusion...

> This might be a known issue. When the system memory is under high pressure, it might cause this UR Error. You could witness that the system memory is almost...

> Could you take a look at how much memory it uses when the model runs? Seems that you are using Linux, you may use `xpu-smi` for example. Lunar Lake...