[PR #6736] [MERGED] Verify permissions for AMD GPU #43459

Closed
opened 2026-04-24 23:05:19 -05:00 by GiteaMirror · 0 comments
Owner

📋 Pull Request Information

Original PR: https://github.com/ollama/ollama/pull/6736
Author: @dhiltgen
Created: 9/10/2024
Status: Merged
Merged: 9/11/2024
Merged by: @dhiltgen

Base: mainHead: check_rocm_permissions


📝 Commits (1)

  • 7302a4c Verify permissions for AMD GPU

📊 Changes

2 files changed (+32 additions, -0 deletions)

View changed files

📝 docs/troubleshooting.md (+11 -0)
📝 gpu/amd_linux.go (+21 -0)

📄 Description

This adds back a check which was lost many releases back to verify /dev/kfd permissions which when lacking, can lead to confusing failure modes of:
"rocBLAS error: Could not initialize Tensile host: No devices found"

This implementation does not hard fail the serve command but instead will fall back to CPU with an error log. In the future we can include this in the GPU discovery UX to show detected but unsupported devices we discovered.

Fixes #6685


🔄 This issue represents a GitHub Pull Request. It cannot be merged through Gitea due to API limitations.

## 📋 Pull Request Information **Original PR:** https://github.com/ollama/ollama/pull/6736 **Author:** [@dhiltgen](https://github.com/dhiltgen) **Created:** 9/10/2024 **Status:** ✅ Merged **Merged:** 9/11/2024 **Merged by:** [@dhiltgen](https://github.com/dhiltgen) **Base:** `main` ← **Head:** `check_rocm_permissions` --- ### 📝 Commits (1) - [`7302a4c`](https://github.com/ollama/ollama/commit/7302a4cf28bb1f38d3f86f355af2a4ce8a0e0f13) Verify permissions for AMD GPU ### 📊 Changes **2 files changed** (+32 additions, -0 deletions) <details> <summary>View changed files</summary> 📝 `docs/troubleshooting.md` (+11 -0) 📝 `gpu/amd_linux.go` (+21 -0) </details> ### 📄 Description This adds back a check which was lost many releases back to verify /dev/kfd permissions which when lacking, can lead to confusing failure modes of: "rocBLAS error: Could not initialize Tensile host: No devices found" This implementation does not hard fail the serve command but instead will fall back to CPU with an error log. In the future we can include this in the GPU discovery UX to show detected but unsupported devices we discovered. Fixes #6685 --- <sub>🔄 This issue represents a GitHub Pull Request. It cannot be merged through Gitea due to API limitations.</sub>
GiteaMirror added the pull-request label 2026-04-24 23:05:19 -05:00
Sign in to join this conversation.
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: github-starred/ollama#43459