github-starred/ollama
2
0
Fork 0
mirror of https://github.com/ollama/ollama.git synced 2026-05-06 08:02:14 -05:00
Code Issues 14.5k Packages Projects Releases 120 Wiki Activity

[GH-ISSUE #6913] On multi-GPU inference speed limited by performance of single CPU core #30133

New Issue
Closed
opened 2026-04-22 09:36:25 -05:00 by GiteaMirror · 1 comment
GiteaMirror commented 2026-04-22 09:36:25 -05:00
Owner
Copy Link

Originally created by @jonathankfmn on GitHub (Sep 23, 2024).
Original GitHub issue: https://github.com/ollama/ollama/issues/6913

What is the issue?

image
image

I have a server with 4 RTX 4090 GPUs. When I run a model, all GPUs are utilized between 10-25%, while the CPU is at 100%, but only using one thread. All layers are loaded into the GPUs.

I already tried, the num_thread parameter that does not change anything.

I believe the single thread process in the CPU is limiting the response speed. Is there a way to improve the speed? By changing the configuration?

OS

Linux

GPU

Nvidia

CPU

AMD

Ollama version

0.3.11

Originally created by @jonathankfmn on GitHub (Sep 23, 2024). Original GitHub issue: https://github.com/ollama/ollama/issues/6913 ### What is the issue? ![image](https://github.com/user-attachments/assets/77256c03-4d29-45a7-96d9-2f01901a782c) ![image](https://github.com/user-attachments/assets/f82028c3-e240-47a6-833c-519117caeec3) I have a server with 4 RTX 4090 GPUs. When I run a model, all GPUs are utilized between 10-25%, while the CPU is at 100%, but only using one thread. All layers are loaded into the GPUs. I already tried, the `num_thread` parameter that does not change anything. I believe the single thread process in the CPU is limiting the response speed. Is there a way to improve the speed? By changing the configuration? ### OS Linux ### GPU Nvidia ### CPU AMD ### Ollama version 0.3.11
GiteaMirror added the performancenvidiabugamd labels 2026-04-22 09:36:25 -05:00
GiteaMirror commented 2026-04-22 09:36:27 -05:00
Author
Owner
Copy Link

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

https://github.com/ggerganov/llama.cpp/issues/8684

<!-- gh-comment-id:2368072595 --> @rick-github commented on GitHub (Sep 23, 2024): https://github.com/ggerganov/llama.cpp/issues/8684
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 amd bug nvidia performance
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#30133
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?