github-starred/ollama
2
0
Fork 0
mirror of https://github.com/ollama/ollama.git synced 2026-05-07 16:40:08 -05:00
Code Issues 14.5k Packages Projects Releases 120 Wiki Activity

[GH-ISSUE #7902] Using GPU + single CPU or no GPU and all CPUs #51567

New Issue
Closed
opened 2026-04-28 20:35:35 -05:00 by GiteaMirror · 3 comments
GiteaMirror commented 2026-04-28 20:35:35 -05:00
Owner
Copy Link

Originally created by @WaarlandIT on GitHub (Dec 1, 2024).
Original GitHub issue: https://github.com/ollama/ollama/issues/7902

What is the issue?

Some models do not use the GPUs at all but use CPU's only and others use GPU + a single CPU.
llama3.2 uses the GPU and a single CPU core spikes to 100% and the other 11 cores are not used at all.
tinyllama uses CPU cores only and uses alle 12 cores.

Not sure if that is a bug or not but I expected that when the CPUs are used ollama will always use all cores and not just 1.

OS

Linux

GPU

Nvidia

CPU

Intel

Ollama version

0.3.13

Originally created by @WaarlandIT on GitHub (Dec 1, 2024). Original GitHub issue: https://github.com/ollama/ollama/issues/7902 ### What is the issue? Some models do not use the GPUs at all but use CPU's only and others use GPU + a single CPU. llama3.2 uses the GPU and a single CPU core spikes to 100% and the other 11 cores are not used at all. tinyllama uses CPU cores only and uses alle 12 cores. Not sure if that is a bug or not but I expected that when the CPUs are used ollama will always use all cores and not just 1. ### OS Linux ### GPU Nvidia ### CPU Intel ### Ollama version 0.3.13
GiteaMirror added the bug label 2026-04-28 20:35:35 -05:00
GiteaMirror commented 2026-04-28 20:35:36 -05:00
Author
Owner
Copy Link

@rick-github commented on GitHub (Dec 1, 2024):

Server logs would aid in debugging, specifically from when you ran llama3.2 and tinyllama. The CPU usage for llama3.2 sounds normal, as the CPU is busy waiting on the GPUs, and the CPU usage for tinyllama sounds normal for a model running only in RAM, but if llama3.2 fits on your GPU then tinyllama should as well, so that bears looking at.

<!-- gh-comment-id:2510208118 --> @rick-github commented on GitHub (Dec 1, 2024): [Server logs](https://github.com/ollama/ollama/blob/main/docs/troubleshooting.md#how-to-troubleshoot-issues) would aid in debugging, specifically from when you ran llama3.2 and tinyllama. The CPU usage for llama3.2 sounds normal, as the CPU is busy waiting on the GPUs, and the CPU usage for tinyllama sounds normal for a model running only in RAM, but if llama3.2 fits on your GPU then tinyllama should as well, so that bears looking at.
GiteaMirror commented 2026-04-28 20:35:36 -05:00
Author
Owner
Copy Link

@WaarlandIT commented on GitHub (Dec 2, 2024):

Thanks for your reply, I will try to recreate the issue and collect the log.
I have been playing around with many different models to see what was going on.

<!-- gh-comment-id:2510675341 --> @WaarlandIT commented on GitHub (Dec 2, 2024): Thanks for your reply, I will try to recreate the issue and collect the log. I have been playing around with many different models to see what was going on.
GiteaMirror commented 2026-04-28 20:35:37 -05:00
Author
Owner
Copy Link

@WaarlandIT commented on GitHub (Dec 3, 2024):

I think I might have solved the problem. I noticed that it sometimes worked as expected and sometimes it did not.
What you said made me think that the GPU might have been busy at some point when I started tinyllama and it was forced to use the CPU instead.
Ollama is also used by Home Assistant for various tasks, when I stopped that integration Ollama seemed to work fine.

<!-- gh-comment-id:2513692256 --> @WaarlandIT commented on GitHub (Dec 3, 2024): I think I might have solved the problem. I noticed that it sometimes worked as expected and sometimes it did not. What you said made me think that the GPU might have been busy at some point when I started tinyllama and it was forced to use the CPU instead. Ollama is also used by Home Assistant for various tasks, when I stopped that integration Ollama seemed to work fine.
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#51567
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?