github-starred/ollama
2
0
Fork 0
mirror of https://github.com/ollama/ollama.git synced 2026-05-06 16:11:34 -05:00
Code Issues 14.5k Packages Projects Releases 120 Wiki Activity

[PR #2195] [MERGED] Ignore AMD integrated GPUs #21357

New Issue
Closed
opened 2026-04-19 15:35:47 -05:00 by GiteaMirror · 0 comments
GiteaMirror commented 2026-04-19 15:35:47 -05:00
Owner
Copy Link

📋 Pull Request Information

Original PR: https://github.com/ollama/ollama/pull/2195
Author: @dhiltgen
Created: 1/26/2024
Status: ✅ Merged
Merged: 1/26/2024
Merged by: @dhiltgen

Base: main ← Head: rocm_real_gpus

📝 Commits (1)

  • 9d7b5d6 Ignore AMD integrated GPUs

📊 Changes

3 files changed (+35 additions, -3 deletions)

View changed files

📝 gpu/gpu.go (+25 -1)
📝 gpu/gpu_info.h (+1 -0)
📝 gpu/gpu_info_rocm.c (+9 -2)

📄 Description

Fixes #2054

Integrated GPUs (APUs) from AMD may be reported by ROCm, but we can't run on them with our current llama.cpp configuration. These iGPUs report 512M of memory, so I've coded the check to ignore any ROCm reported GPU that has less than 1G of memory. If we detect only an integrated GPU, this will fallback to CPU mode. If we detect multiple ROCm GPUs, meaning one or more are discrete, and one is integrated, we'll now set ROCR_VISIBLE_DEVICES so we ignore the iGPU. If the user has explicitly set ROCR_VISIBLE_DEVICES we'll respect their setting.

🔄 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/2195 **Author:** [@dhiltgen](https://github.com/dhiltgen) **Created:** 1/26/2024 **Status:** ✅ Merged **Merged:** 1/26/2024 **Merged by:** [@dhiltgen](https://github.com/dhiltgen) **Base:** `main` ← **Head:** `rocm_real_gpus` --- ### 📝 Commits (1) - [`9d7b5d6`](https://github.com/ollama/ollama/commit/9d7b5d6c91b0686619010f5355d7eb44c856a95d) Ignore AMD integrated GPUs ### 📊 Changes **3 files changed** (+35 additions, -3 deletions) <details> <summary>View changed files</summary> 📝 `gpu/gpu.go` (+25 -1) 📝 `gpu/gpu_info.h` (+1 -0) 📝 `gpu/gpu_info_rocm.c` (+9 -2) </details> ### 📄 Description Fixes #2054 Integrated GPUs (APUs) from AMD may be reported by ROCm, but we can't run on them with our current llama.cpp configuration. These iGPUs report 512M of memory, so I've coded the check to ignore any ROCm reported GPU that has less than 1G of memory. If we detect only an integrated GPU, this will fallback to CPU mode. If we detect multiple ROCm GPUs, meaning one or more are discrete, and one is integrated, we'll now set `ROCR_VISIBLE_DEVICES` so we ignore the iGPU. If the user has explicitly set `ROCR_VISIBLE_DEVICES` we'll respect their setting. --- <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 2026-04-19 15:35:47 -05:00
Sign in to join this conversation.
No Branch/Tag Specified
Branches Tags
Labels
Clear labels
amd
api
app
bug
build
cli
cloud
compatibility
context-length
create
docker
documentation
embeddings
feature request
feedback wanted
good first issue
gpt-oss
gpu
harmony
help wanted
image
install
intel
js
launch
linux
macos
memory
mlx
model
needs more info
networking
nvidia
ollama.com
performance
pull-request
Mirrored from GitHub Pull Request
 python
question
registry
rendering
thinking
tools
top
vulkan
windows
wsl
No Label pull-request
Milestone
No items
No Milestone
Projects
Clear projects
No project
Assignees
Clear assignees
GiteaMirror ninjasurge
No Assignees
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: github-starred/ollama#21357
Delete Branch "%!s()"

Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?