github-starred/ollama
2
0
Fork 0
mirror of https://github.com/ollama/ollama.git synced 2026-05-07 08:30:05 -05:00
Code Issues 14.5k Packages Projects Releases 120 Wiki Activity

[GH-ISSUE #5635] ollama not use all GPUs #3514

New Issue
Closed
opened 2026-04-12 14:13:02 -05:00 by GiteaMirror · 2 comments
GiteaMirror commented 2026-04-12 14:13:02 -05:00
Owner
Copy Link

Originally created by @mavershang on GitHub (Jul 11, 2024).
Original GitHub issue: https://github.com/ollama/ollama/issues/5635

Originally assigned to: @dhiltgen on GitHub.

What is the issue?

I ran ollama on a server with 4x A100. It only uses 1 of them. Is there some setting need to be changed? Thanks

image

OS

Linux

GPU

Nvidia

CPU

Intel

Ollama version

0.2.1

Originally created by @mavershang on GitHub (Jul 11, 2024). Original GitHub issue: https://github.com/ollama/ollama/issues/5635 Originally assigned to: @dhiltgen on GitHub. ### What is the issue? I ran ollama on a server with 4x A100. It only uses 1 of them. Is there some setting need to be changed? Thanks ![image](https://github.com/ollama/ollama/assets/8919917/8a49b4a4-fd85-4715-8ff0-1ea0c1276939) ### OS Linux ### GPU Nvidia ### CPU Intel ### Ollama version 0.2.1
GiteaMirror added the documentation label 2026-04-12 14:13:02 -05:00
GiteaMirror commented 2026-04-12 14:13:04 -05:00
Author
Owner
Copy Link

@Demirrr commented on GitHub (Jul 11, 2024):

I have also similar issue. Setting num_gpu:2 doesn't do anything

Edit: OLLAMA_NUM_PARALLEL=8 did the trick. Thank you!

<!-- gh-comment-id:2223623813 --> @Demirrr commented on GitHub (Jul 11, 2024): I have also similar issue. Setting [num_gpu:2](https://github.com/ollama/ollama/blob/main/docs/api.md#generate-request-with-options) doesn't do anything Edit: `OLLAMA_NUM_PARALLEL=8` did the trick. Thank you!
GiteaMirror commented 2026-04-12 14:13:04 -05:00
Author
Owner
Copy Link

@dhiltgen commented on GitHub (Jul 23, 2024):

I'll update the FAQ to make this clearer. In short, the default behavior is to use a single GPU if the model will fit as that typically yields the best performance.

<!-- gh-comment-id:2246406575 --> @dhiltgen commented on GitHub (Jul 23, 2024): I'll update the FAQ to make this clearer. In short, the default behavior is to use a single GPU if the model will fit as that typically yields the best performance.
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 documentation
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#3514
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?