lllyasviel / lllyasviel/stable-diffusion-webui-forge

"Memory access fault by GPU node-1 on address 0x.... Reason: Page not present or supervisor privilege." on rx 9070 xt on rocm 6.4.1

Open
#2,976 0 comments 0 reactions 0 assignees View on GitHub
Dominant language
Python
Stars
13k
Forks
1.7k
PR merge metrics
No merged PRs in 30d

Description

I am getting error on my 9070 xt that says "Memory access fault by GPU node-1 on address 0x.... Reason: Page not present or supervisor privilege." when i try to run a llm on it its not a huge llm huge a 800m parameter model and i tried smaller but it had the same issue so it not the model size ,i tried enabling and disabling the instigated graphic ,the code runs well on other systems so the code isn't the issue . i am using r7 7700x,16gb ram rx 9070xt from sapphire pure on ubuntu 22.04.05 on kernel 6.8 on the rocm 6.4.1. so how can i fix it?

Contributor guide

No contributing guide indexed for this repository

Research direction

No repository file, test, or entry point is identified. Start by reproducing an 800M-parameter LLM run on the listed Ubuntu 22.04, kernel 6.8, RX 9070 XT, and ROCm 6.4.1 environment, then compare with the reported successful systems; done means identifying a reproducible fix or a clearly documented compatibility limitation.

Written by the indexing model from the issue text.

Assessment

Tech stack
linux, ubuntu
Domain
machine-learning, operating-systems
Issue type
Bug
Difficulty
4/5
Estimated time
3-5 days
Activity status
Stale
Clarity
Needs clarification
Newbie friendliness
25/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.