github-starred/ollama
2
0
Fork 0
mirror of https://github.com/ollama/ollama.git synced 2026-05-07 00:22:43 -05:00
Code Issues 14.5k Packages Projects Releases 120 Wiki Activity

[GH-ISSUE #5846] [Very minor] When eGPU gets disconnected, Ollama falls back to using CPU and doesn't detec when eGPU is connected again. #65684

New Issue
Closed
opened 2026-05-03 22:14:05 -05:00 by GiteaMirror · 2 comments
GiteaMirror commented 2026-05-03 22:14:05 -05:00
Owner
Copy Link

Originally created by @kosmallafilip on GitHub (Jul 22, 2024).
Original GitHub issue: https://github.com/ollama/ollama/issues/5846

What is the issue?

This is quite a big issue, but affects probably less than 0.01% of users, so I am tagging this as minor.

I am using Ollama locally on Asus ROG Flow Z13 with XG Mobile eGPU - this is Asus's proprietary eGPU technology, it's quite different than connecting with Thunderbolt, so I can't really tell if it's going to affect other types of eGPUs as well.

If, for any reason, the eGPU disconnects, Ollama falls back to using a CPU for generating messages, which is great, but when I connect it back I'd expect Ollama to detect a new GPU being connected and use it instead. This doesn't happen and I was left wondering why Ollama uses so much CPU and lags the whole system.

Due to how I use Ollama, I am still on 0.1.45, but let me know if you want me to test this the latest release.
I don't expect this to get fixed anytime soon, but I wanted to let you know it happens.

OS

Windows

GPU

Nvidia

CPU

Intel

Ollama version

0.1.45

Originally created by @kosmallafilip on GitHub (Jul 22, 2024). Original GitHub issue: https://github.com/ollama/ollama/issues/5846 ### What is the issue? This is quite a big issue, but affects probably less than 0.01% of users, so I am tagging this as minor. I am using Ollama locally on Asus ROG Flow Z13 with XG Mobile eGPU - this is Asus's proprietary eGPU technology, it's quite different than connecting with Thunderbolt, so I can't really tell if it's going to affect other types of eGPUs as well. If, for any reason, the eGPU disconnects, Ollama falls back to using a CPU for generating messages, which is great, but when I connect it back I'd expect Ollama to detect a new GPU being connected and use it instead. This doesn't happen and I was left wondering why Ollama uses so much CPU and lags the whole system. Due to how I use Ollama, I am still on 0.1.45, but let me know if you want me to test this the latest release. I don't expect this to get fixed anytime soon, but I wanted to let you know it happens. ### OS Windows ### GPU Nvidia ### CPU Intel ### Ollama version 0.1.45
GiteaMirror added the bug label 2026-05-03 22:14:05 -05:00
GiteaMirror commented 2026-05-03 22:14:13 -05:00
Author
Owner
Copy Link

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

I've just found #5411 , so I assume this will get closed as not planned, but since it's a bit different case, I am leaving this open so you're aware that it's happening

<!-- gh-comment-id:2242477767 --> @kosmallafilip commented on GitHub (Jul 22, 2024): I've just found #5411 , so I assume this will get closed as not planned, but since it's a bit different case, I am leaving this open so you're aware that it's happening
GiteaMirror commented 2026-05-03 22:14:17 -05:00
Author
Owner
Copy Link

@dhiltgen commented on GitHub (Jul 23, 2024):

Hotplug/external GPU support is tracked via #5411

<!-- gh-comment-id:2245894027 --> @dhiltgen commented on GitHub (Jul 23, 2024): Hotplug/external GPU support is tracked via #5411
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 bug
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#65684
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?