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 #7801] GPU usage is not high, but the display memory is full #67044

New Issue
Closed
opened 2026-05-04 09:20:08 -05:00 by GiteaMirror · 2 comments
GiteaMirror commented 2026-05-04 09:20:08 -05:00
Owner
Copy Link

Originally created by @duolax on GitHub (Nov 22, 2024).
Original GitHub issue: https://github.com/ollama/ollama/issues/7801

What is the issue?

image
When running the model, the CPU usage is often full, but the GPU usage is not high. However, after checking the details, it turns out that the GPU memory is fully occupied during operation

OS

Windows

GPU

Nvidia

CPU

AMD

Ollama version

0.4.2

Originally created by @duolax on GitHub (Nov 22, 2024). Original GitHub issue: https://github.com/ollama/ollama/issues/7801 ### What is the issue? ![image](https://github.com/user-attachments/assets/1c2cf4c5-069b-4d48-8d9f-8d8d4a4ebcd6) When running the model, the CPU usage is often full, but the GPU usage is not high. However, after checking the details, it turns out that the GPU memory is fully occupied during operation ### OS Windows ### GPU Nvidia ### CPU AMD ### Ollama version 0.4.2
GiteaMirror added the bug label 2026-05-04 09:20:08 -05:00
GiteaMirror commented 2026-05-04 09:20:19 -05:00
Author
Owner
Copy Link

@rick-github commented on GitHub (Nov 23, 2024):

If you have loaded a model that does not fit 100% on the GPU, part of the model will be loaded in system RAM and the CPU is used for inference on that part of the model. When you do a completion, the part that runs on the GPU will finish quickly, and the part on the CPU will take longer. Thus the CPU will have high usage while the GPU is idle waiting for the next completion.

<!-- gh-comment-id:2495481046 --> @rick-github commented on GitHub (Nov 23, 2024): If you have loaded a model that does not fit 100% on the GPU, part of the model will be loaded in system RAM and the CPU is used for inference on that part of the model. When you do a completion, the part that runs on the GPU will finish quickly, and the part on the CPU will take longer. Thus the CPU will have high usage while the GPU is idle waiting for the next completion.
GiteaMirror commented 2026-05-04 09:20:20 -05:00
Author
Owner
Copy Link

@jmorganca commented on GitHub (Nov 23, 2024):

Thanks for the issue. @rick-github is correct – this is common if running the model requires the CPU

<!-- gh-comment-id:2495654443 --> @jmorganca commented on GitHub (Nov 23, 2024): Thanks for the issue. @rick-github is correct – this is common if running the model requires the CPU
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#67044
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?