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 #7472] GPU on WIndows #4750

New Issue
Closed
opened 2026-04-12 15:41:40 -05:00 by GiteaMirror · 4 comments
GiteaMirror commented 2026-04-12 15:41:40 -05:00
Owner
Copy Link

Originally created by @godlatro on GitHub (Nov 2, 2024).
Original GitHub issue: https://github.com/ollama/ollama/issues/7472

What is the issue?

I have latest Ollama desktop
Nvidia 3060
Windows 10

Try to use any models, CPU/GPU loading ~70%/20%
I load many models one by one.
I unload the extra ones with ollama stop model

Almost all models work terribly slowly. In 90% of cases, the models can't even finish writing their answer and are interrupted every time when using the web interface open-webui on docker.
But when I running Ubuntu 24 on this computer, it shows 100% loading my GPU and all models work perfectly fast.
How can I make Windows also use 100% GPU like in ubuntu?

OS

Windows

GPU

Nvidia

CPU

Intel

Ollama version

0.1.32

Originally created by @godlatro on GitHub (Nov 2, 2024). Original GitHub issue: https://github.com/ollama/ollama/issues/7472 ### What is the issue? I have latest Ollama desktop Nvidia 3060 Windows 10 Try to use any models, CPU/GPU loading ~70%/20% I load many models one by one. I unload the extra ones with ```ollama stop model``` Almost all models work terribly slowly. In 90% of cases, the models can't even finish writing their answer and are interrupted every time when using the web interface open-webui on docker. But when I running Ubuntu 24 on this computer, it shows 100% loading my GPU and all models work perfectly fast. How can I make Windows also use 100% GPU like in ubuntu? ### OS Windows ### GPU Nvidia ### CPU Intel ### Ollama version 0.1.32
GiteaMirror added the bug label 2026-04-12 15:41:40 -05:00
GiteaMirror commented 2026-04-12 15:41:42 -05:00
Author
Owner
Copy Link

@rick-github commented on GitHub (Nov 2, 2024):

Other applications in Windows may be using the GPU, leaving less room for ollama. Server logs may help with diagnosis, along with the output of nvidia-smi.

<!-- gh-comment-id:2453218890 --> @rick-github commented on GitHub (Nov 2, 2024): Other applications in Windows may be using the GPU, leaving less room for ollama. [Server logs](https://github.com/ollama/ollama/blob/main/docs/troubleshooting.md#how-to-troubleshoot-issues) may help with diagnosis, along with the output of `nvidia-smi`.
GiteaMirror commented 2026-04-12 15:41:42 -05:00
Author
Owner
Copy Link

@godlatro commented on GitHub (Nov 3, 2024):

It seems that this is not a problem of ollama. It is open-webui that loads models differently than ollama. In the console, ollama loads the model in the GPU

<!-- gh-comment-id:2453566616 --> @godlatro commented on GitHub (Nov 3, 2024): It seems that this is not a problem of ollama. It is open-webui that loads models differently than ollama. In the console, ollama loads the model in the GPU
GiteaMirror commented 2026-04-12 15:41:43 -05:00
Author
Owner
Copy Link

@godlatro commented on GitHub (Nov 3, 2024):

ollama for windows loads GPU 100% in terminal like in linux, but some models like starcoder2 not work or say nothing useful

<!-- gh-comment-id:2453567116 --> @godlatro commented on GitHub (Nov 3, 2024): ollama for windows loads GPU 100% in terminal like in linux, but some models like starcoder2 not work or say nothing useful
GiteaMirror commented 2026-04-12 15:41:43 -05:00
Author
Owner
Copy Link

@rick-github commented on GitHub (Nov 3, 2024):

open-webui is probably using a bigger context window than the default one of 2048 tokens.

starcoder2 is generally not a chat model, it's used for FIM (fill in the middle) for completing code. starcoder2:instruct has a chat mode, but only about code.

<!-- gh-comment-id:2453580531 --> @rick-github commented on GitHub (Nov 3, 2024): open-webui is probably using a bigger context window than the default one of 2048 tokens. starcoder2 is generally not a chat model, it's used for FIM (fill in the middle) for completing code. starcoder2:instruct has a chat mode, but only about code.
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#4750
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?