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 #5411] Support hot-plugging exernal GPU(s) #3386

New Issue
Open
opened 2026-04-12 14:01:13 -05:00 by GiteaMirror · 3 comments
GiteaMirror commented 2026-04-12 14:01:13 -05:00
Owner
Copy Link

Originally created by @headcr4sh on GitHub (Jul 1, 2024).
Original GitHub issue: https://github.com/ollama/ollama/issues/5411

Originally assigned to: @dhiltgen on GitHub.

What is the issue?

My setup:

  • Notebook: (Framework 13") with an AMD Ryzen 7840U / Radeon 780M GPU
  • eGPU (connected via USB3/4 encolosure "Razer Core X Chroma"): AMD Radeon RX 7600 XT
  • OS: Windows 11 / Fedora 41

I started Ollama when the eGPU was still attached and was able to load the gemma2:9b model (100% GPU utlization). Then I ejected the eGPU and Ollama reports errors:

Error: llama runner process has terminated: exit status 0xc0000409 error:
Cannot read C:\Users\<MY_USER_NAME>\AppData\Local\Programs\Ollama\rocm\/rocblas/library/TensileLibrary.dat:
No such file or directory for GPU arch : gfx1103

Re-starting Ollama fixes the issue.

I assume that support for "hot-pluggable GPUs" (eGPUs) is not [yet?] implemented.

OS

  • Windows
  • Linux

GPU

AMD

CPU

AMD

Ollama version

0.1.48

Originally created by @headcr4sh on GitHub (Jul 1, 2024). Original GitHub issue: https://github.com/ollama/ollama/issues/5411 Originally assigned to: @dhiltgen on GitHub. ### What is the issue? My setup: - Notebook: (Framework 13") with an AMD Ryzen 7840U / Radeon 780M GPU - eGPU (connected via USB3/4 encolosure "Razer Core X Chroma"): AMD Radeon RX 7600 XT - OS: Windows 11 / Fedora 41 I started Ollama when the eGPU was still attached and was able to load the `gemma2:9b` model (100% GPU utlization). Then I ejected the eGPU and Ollama reports errors: ``` Error: llama runner process has terminated: exit status 0xc0000409 error: Cannot read C:\Users\<MY_USER_NAME>\AppData\Local\Programs\Ollama\rocm\/rocblas/library/TensileLibrary.dat: No such file or directory for GPU arch : gfx1103 ``` Re-starting Ollama fixes the issue. I assume that support for "hot-pluggable GPUs" (eGPUs) is not [yet?] implemented. ### OS * Windows * Linux ### GPU AMD ### CPU AMD ### Ollama version 0.1.48
GiteaMirror added the feature requestwindows labels 2026-04-12 14:01:13 -05:00
GiteaMirror commented 2026-04-12 14:01:14 -05:00
Author
Owner
Copy Link

@dhiltgen commented on GitHub (Jul 2, 2024):

We don't currently support hot swapping GPUs.

<!-- gh-comment-id:2204330780 --> @dhiltgen commented on GitHub (Jul 2, 2024): We don't currently support hot swapping GPUs.
GiteaMirror commented 2026-04-12 14:01:15 -05:00
Author
Owner
Copy Link

@goukisun commented on GitHub (Sep 5, 2024):

What is the issue?

My setup:

  • Notebook: (Framework) with and AMD Ryzen 7840U / Radeon 780M GPU
  • eGPU (connected via USB3/4 encolosure): AMD Radeon RX 7600 XT
  • OS: Windows 11

I started Ollama when the eGPU was still attached and was able to load the gemma2:9b model (100% GPU utlization). Then I ejected the eGPU and Ollama reports errors:

Error: llama runner process has terminated: exit status 0xc0000409 error:
Cannot read C:\Users\<MY_USER_NAME>\AppData\Local\Programs\Ollama\rocm\/rocblas/library/TensileLibrary.dat:
No such file or directory for GPU arch : gfx1103

Re-starting Ollama fixes the issue.

I assume that support for "hot-pluggable GPUs" (eGPUs) is not [yet?] implemented.

OS

Windows

GPU

AMD

CPU

AMD

Ollama version

0.1.48

hi, how is the inference speed when use rx 7600?
and for you error message, rx 7600 is gfx1102, rx 780m igpu is gfx1103

<!-- gh-comment-id:2331108740 --> @goukisun commented on GitHub (Sep 5, 2024): > ### What is the issue? > My setup: > > * Notebook: (Framework) with and AMD Ryzen 7840U / Radeon 780M GPU > * eGPU (connected via USB3/4 encolosure): AMD Radeon RX 7600 XT > * OS: Windows 11 > > I started Ollama when the eGPU was still attached and was able to load the `gemma2:9b` model (100% GPU utlization). Then I ejected the eGPU and Ollama reports errors: > > ``` > Error: llama runner process has terminated: exit status 0xc0000409 error: > Cannot read C:\Users\<MY_USER_NAME>\AppData\Local\Programs\Ollama\rocm\/rocblas/library/TensileLibrary.dat: > No such file or directory for GPU arch : gfx1103 > ``` > > Re-starting Ollama fixes the issue. > > I assume that support for "hot-pluggable GPUs" (eGPUs) is not [yet?] implemented. > > ### OS > Windows > > ### GPU > AMD > > ### CPU > AMD > > ### Ollama version > 0.1.48 hi, how is the inference speed when use rx 7600? and for you error message, rx 7600 is gfx1102, rx 780m igpu is gfx1103
GiteaMirror commented 2026-04-12 14:01:15 -05:00
Author
Owner
Copy Link

@headcr4sh commented on GitHub (Jan 10, 2025):

I just changed the title of the PR as lack of hot-plugging support is currently not something limited to Windows. I am using a Linux/Windows dual-boot setup and Ollama is not able to handle eGPU (dis)connects on either one of them.

<!-- gh-comment-id:2582457634 --> @headcr4sh commented on GitHub (Jan 10, 2025): I just changed the title of the PR as lack of hot-plugging support is currently not something limited to Windows. I am using a Linux/Windows dual-boot setup and Ollama is not able to handle eGPU (dis)connects on either one of them.
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 feature request windows
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#3386
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?