cjpais / cjpais/LocalScore

Coredump when running on EndeavourOS with Ryzen AI MAX+ 395

Open
#23 1 comment 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

Dominant language
TypeScript
Stars
132
Forks
17
PR merge metrics
No merged PRs in 30d

Description

Hi, I testing localscore but it's core dumping when I run it:

./localscore-medium
import_cuda_impl: initializing gpu module...
link_cuda_dso: note: dynamically linking /home/juancho/.llamafile/v/0.9.2/ggml-rocm.so
localscore-medium: /usr/src/debug/hip-runtime/hip-runtime-clr/rocclr/os/os_posix.cpp:321: static void amd::Os::currentStackInfo(unsigned char**, size_t*): Assertion `Os::currentStackPtr() >= *base - *size && Os::currentStackPtr() < *base && "just checking"' failed.
Aborted                    (core dumped) ./localscore-medium

I'm running it on EndeavourOS with all updates on a Gmktec Evo-X2 with a AMD Ryzen AI MAX+ 395 processor. I have ROCm 6.4.3 installed, I'm new to local LLMs so I do not know what other info add that can be helpful, but let me know and I will add it!

Contributor guide

No contributing guide indexed for this repository

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

Research direction

Start by reproducing the crash with ./localscore-medium on EndeavourOS, the Ryzen AI MAX+ 395, and ROCm 6.4.3, capturing the environment and the import_cuda_impl/link_cuda_dso output. Investigate the ROCm assertion in hip-runtime and determine whether the issue is a compatibility problem; done means identifying a reproducible cause and documenting or validating a fix.

Written by the indexing model from the issue text.

Assessment

Domain
machine-learning, operating-systems
Issue type
Bug
Difficulty
4/5
Estimated time
3-5 days
Activity status
Quiet
Clarity
Needs clarification
Newbie friendliness
35/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.