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

[GH-ISSUE #5766] specify a single GPU （id=1）using Docker, Error! #3590

New Issue
Closed
opened 2026-04-12 14:19:48 -05:00 by GiteaMirror · 2 comments
GiteaMirror commented 2026-04-12 14:19:48 -05:00
Owner
Copy Link

Originally created by @catsled on GitHub (Jul 18, 2024).
Original GitHub issue: https://github.com/ollama/ollama/issues/5766

Originally assigned to: @dhiltgen on GitHub.

What is the issue?

I have 8 gpus, i want to specify the id=1's gpu in a docker,

docker run -it ... --device=/dev/dri/card1 --device=/dev/dri/renderD129 ....

Error occured!

image

when i set HIP_VISIBLE_DEVICES=1

image

image

OS

Linux

GPU

AMD

CPU

Intel

Ollama version

0.21

Originally created by @catsled on GitHub (Jul 18, 2024). Original GitHub issue: https://github.com/ollama/ollama/issues/5766 Originally assigned to: @dhiltgen on GitHub. ### What is the issue? I have 8 gpus, i want to specify the id=1's gpu in a docker, `docker run -it ... --device=/dev/dri/card1 --device=/dev/dri/renderD129 ....` Error occured! ![image](https://github.com/user-attachments/assets/ed2137b0-e0a9-4025-a55d-c57205536e39) when i set `HIP_VISIBLE_DEVICES=1` <img width="892" alt="image" src="https://github.com/user-attachments/assets/269de8c9-7d14-4b9b-a57a-75bc4f9d61f5"> ![image](https://github.com/user-attachments/assets/2a7522b4-2e77-4bf6-b0de-3318cbc038ae) ### OS Linux ### GPU AMD ### CPU Intel ### Ollama version 0.21
GiteaMirror added the dockeramdbug labels 2026-04-12 14:19:48 -05:00
GiteaMirror commented 2026-04-12 14:19:49 -05:00
Author
Owner
Copy Link

@dhiltgen commented on GitHub (Jul 22, 2024):

It's hard to follow along with the screenshots instead of text logs, but I think you're seeing 2 different problems.

I think when you filter the GPUs via docker command line flags, and set HIP_VISIBLE_DEVICES things get out of sync and we fail to work on the GPU.

I think if you omit that env var, and just filter with docer flags, then there's a gfx compatibility problem. What GPU are you trying to use, and what GFX Override are you setting?

<!-- gh-comment-id:2243995398 --> @dhiltgen commented on GitHub (Jul 22, 2024): It's hard to follow along with the screenshots instead of text logs, but I think you're seeing 2 different problems. I think when you filter the GPUs via docker command line flags, and set HIP_VISIBLE_DEVICES things get out of sync and we fail to work on the GPU. I think if you omit that env var, and just filter with docer flags, then there's a gfx compatibility problem. What GPU are you trying to use, and what GFX Override are you setting?
GiteaMirror commented 2026-04-12 14:19:50 -05:00
Author
Owner
Copy Link

@catsled commented on GitHub (Jul 25, 2024):

It's hard to follow along with the screenshots instead of text logs, but I think you're seeing 2 different problems.

I think when you filter the GPUs via docker command line flags, and set HIP_VISIBLE_DEVICES things get out of sync and we fail to work on the GPU.

I think if you omit that env var, and just filter with docer flags, then there's a gfx compatibility problem. What GPU are you trying to use, and what GFX Override are you setting?

I believe I have found the solution. The issue lies in the mapping of the GPU for Docker（Ollama must read them from 0 idx）; it can be resolved by remapping. thanks

<!-- gh-comment-id:2249530376 --> @catsled commented on GitHub (Jul 25, 2024): > It's hard to follow along with the screenshots instead of text logs, but I think you're seeing 2 different problems. > > I think when you filter the GPUs via docker command line flags, and set HIP_VISIBLE_DEVICES things get out of sync and we fail to work on the GPU. > > I think if you omit that env var, and just filter with docer flags, then there's a gfx compatibility problem. What GPU are you trying to use, and what GFX Override are you setting? I believe I have found the solution. The issue lies in the mapping of the GPU for Docker（Ollama must read them from 0 idx）; it can be resolved by remapping. thanks
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 amd bug docker
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#3590
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?