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 #2993] Ollama only runs off CPU in Ubuntu #63875

New Issue
Closed
opened 2026-05-03 15:17:34 -05:00 by GiteaMirror · 3 comments
GiteaMirror commented 2026-05-03 15:17:34 -05:00
Owner
Copy Link

Originally created by @uansah on GitHub (Mar 7, 2024).
Original GitHub issue: https://github.com/ollama/ollama/issues/2993

I am trying to run Dolphin Mistral on CPU and RAM (no GPU) I have 188gb available but it uses 100% of the CPU at full capacity. Before this I was running it on a virtual machine with the same OS and it used RAM.

Originally created by @uansah on GitHub (Mar 7, 2024). Original GitHub issue: https://github.com/ollama/ollama/issues/2993 I am trying to run Dolphin Mistral on CPU and RAM (no GPU) I have 188gb available but it uses 100% of the CPU at full capacity. Before this I was running it on a virtual machine with the same OS and it used RAM.
GiteaMirror commented 2026-05-03 15:17:39 -05:00
Author
Owner
Copy Link

@ArchitW commented on GitHub (Mar 8, 2024):

2nd to this.

neofetch

image

LLM: tinyllama:latest
prompt: what is the meaning of life

top
image

<!-- gh-comment-id:1985446624 --> @ArchitW commented on GitHub (Mar 8, 2024): 2nd to this. **neofetch** ![image](https://github.com/ollama/ollama/assets/4264602/1ff547f8-fca5-4648-8874-bd07b6a9eb52) `LLM: tinyllama:latest` `prompt: what is the meaning of life` **top** ![image](https://github.com/ollama/ollama/assets/4264602/3564d946-6a9b-45a6-b9a7-a38952023932)
GiteaMirror commented 2026-05-03 15:17:40 -05:00
Author
Owner
Copy Link

@aosan commented on GitHub (Mar 11, 2024):

CPU would be the biggest performance limitation, even if the model can fit in RAM. In my case, any model fitting in the vRAM of my GPU is fast. Any model not fitting in the vRAM is considerably slower.

For example, a simple question with a small model with GPU and fitting in vRAM can output 50-60 tokens/s. The same question with large models fitting only in system RAM and using CPU can output only 2-3 tokens/s.

<!-- gh-comment-id:1988643105 --> @aosan commented on GitHub (Mar 11, 2024): CPU would be the biggest performance limitation, even if the model can fit in RAM. In my case, any model fitting in the vRAM of my GPU is fast. Any model not fitting in the vRAM is considerably slower. For example, a simple question with a small model with GPU and fitting in vRAM can output 50-60 tokens/s. The same question with large models fitting only in system RAM and using CPU can output only 2-3 tokens/s.
GiteaMirror commented 2026-05-03 15:17:41 -05:00
Author
Owner
Copy Link

@jmorganca commented on GitHub (Mar 11, 2024):

Hi there, CPU is definitely going to be the bottleneck here. I'm not sure if there's much we can do regarding this issue other than continue making Ollama faster 😊 . Thanks so much for the issue!

<!-- gh-comment-id:1989548986 --> @jmorganca commented on GitHub (Mar 11, 2024): Hi there, CPU is definitely going to be the bottleneck here. I'm not sure if there's much we can do regarding this issue other than continue making Ollama faster 😊 . Thanks so much for the issue!
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
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#63875
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?