github-starred/ollama
2
0
Fork 0
mirror of https://github.com/ollama/ollama.git synced 2026-05-05 23:53:43 -05:00
Code Issues 14.5k Packages Projects Releases 120 Wiki Activity

[GH-ISSUE #3509] Can Ollama use both CPU and GPU for inference? #2162

New Issue
Closed
opened 2026-04-12 12:24:12 -05:00 by GiteaMirror · 3 comments
GiteaMirror commented 2026-04-12 12:24:12 -05:00
Owner
Copy Link

Originally created by @OPDEV001 on GitHub (Apr 6, 2024).
Original GitHub issue: https://github.com/ollama/ollama/issues/3509

What are you trying to do?

May I know whether ollama support to mix CPU and GPU together for running on windows? I know my hardware is not enough for ollama, but I still want to use the part ability of GPU.
But I checked the parameter information from link below, I still can not mix CPU&GPU, most load by CPU.
https://github.com/ollama/ollama/blob/main/docs/modelfile.md

If I put all load to GPU, it will say "Out of VRam", :) you know it.

I am guessing if it have, maybe like we can specify GPU sharing part load， and CPU on most load?

Thanks,

How should we solve this?

Please see content.

What is the impact of not solving this?

If not, all load on GPU will crash.

Anything else?

Thanks for all

Originally created by @OPDEV001 on GitHub (Apr 6, 2024). Original GitHub issue: https://github.com/ollama/ollama/issues/3509 ### What are you trying to do? May I know whether ollama support to mix CPU and GPU together for running on windows? I know my hardware is not enough for ollama, but I still want to use the part ability of GPU. But I checked the parameter information from link below, I still can not mix CPU&GPU, most load by CPU. https://github.com/ollama/ollama/blob/main/docs/modelfile.md If I put all load to GPU, it will say "Out of VRam", :) you know it. I am guessing if it have, maybe like we can specify GPU sharing part load， and CPU on most load? Thanks, ### How should we solve this? Please see content. ### What is the impact of not solving this? If not, all load on GPU will crash. ### Anything else? Thanks for all
GiteaMirror added the question label 2026-04-12 12:24:12 -05:00
GiteaMirror commented 2026-04-12 12:24:15 -05:00
Author
Owner
Copy Link

@navr32 commented on GitHub (Apr 8, 2024):

You must provide the log off ollama serve to see how ollama have tried to share the models in both memory.

<!-- gh-comment-id:2042381262 --> @navr32 commented on GitHub (Apr 8, 2024): You must provide the log off ollama serve to see how ollama have tried to share the models in both memory.
GiteaMirror commented 2026-04-12 12:24:16 -05:00
Author
Owner
Copy Link

@OPDEV001 commented on GitHub (Apr 8, 2024):

I can run ollama serve in cmd line, it is the log what you want?

<!-- gh-comment-id:2042679587 --> @OPDEV001 commented on GitHub (Apr 8, 2024): I can run ollama serve in cmd line, it is the log what you want?
GiteaMirror commented 2026-04-12 12:24:17 -05:00
Author
Owner
Copy Link

@pdevine commented on GitHub (Apr 12, 2024):

Hi @OPDEV001 , this is how Ollama works by default. It will try to offload as many layers of the LLM as possible onto the GPU, and then if it doesn't fit the other layers will run on the CPU.

I'll go ahead and close the issue, but feel free to keep commenting.

<!-- gh-comment-id:2052609678 --> @pdevine commented on GitHub (Apr 12, 2024): Hi @OPDEV001 , this is how Ollama works by default. It will try to offload as many layers of the LLM as possible onto the GPU, and then if it doesn't fit the other layers will run on the CPU. I'll go ahead and close the issue, but feel free to keep commenting.
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 question
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#2162
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?