github-starred/ollama
2
0
Fork 0
mirror of https://github.com/ollama/ollama.git synced 2026-05-06 16:11:34 -05:00
Code Issues 14.5k Packages Projects Releases 120 Wiki Activity

[GH-ISSUE #2210] Keep models in RAM #27025

New Issue
Closed
opened 2026-04-22 03:55:32 -05:00 by GiteaMirror · 2 comments
GiteaMirror commented 2026-04-22 03:55:32 -05:00
Owner
Copy Link

Originally created by @LeoPiresDeSouza on GitHub (Jan 26, 2024).
Original GitHub issue: https://github.com/ollama/ollama/issues/2210

I am testing llama2:7b models both using ollama and calling direct from a langchain python script.
My models are stored in an Ubuntu server withu 12 cores e 36 Gb of ram, but no GPU.
When I cal the model direct from python, setting memlock parameter to true, my memory usage goes above 6Gb, but when using ollma it stays below 3Gb.
It seams that ollama is not keeping the model entirely in ram, and it is taking a long time to response.

Is there a parameter like memlock to be set in Ollama to make it use my ram extensivelly?

I have installed Ollama using curl https://ollama.ai/install.sh | sh.

Originally created by @LeoPiresDeSouza on GitHub (Jan 26, 2024). Original GitHub issue: https://github.com/ollama/ollama/issues/2210 I am testing llama2:7b models both using ollama and calling direct from a langchain python script. My models are stored in an Ubuntu server withu 12 cores e 36 Gb of ram, but no GPU. When I cal the model direct from python, setting memlock parameter to true, my memory usage goes above 6Gb, but when using ollma it stays below 3Gb. It seams that ollama is not keeping the model entirely in ram, and it is taking a long time to response. Is there a parameter like memlock to be set in Ollama to make it use my ram extensivelly? I have installed Ollama using curl https://ollama.ai/install.sh | sh.
GiteaMirror commented 2026-04-22 03:55:33 -05:00
Author
Owner
Copy Link

@easp commented on GitHub (Jan 27, 2024):

Ollama automatically unloads models from memory after 5 minutes of inactivity. That will be user-configurable in the next version 0.1.23.

Another thing to be aware of is that models are memory mapped and so they don't show up in process memory. They are instead accounted for in file cache.

<!-- gh-comment-id:1913324259 --> @easp commented on GitHub (Jan 27, 2024): Ollama automatically unloads models from memory after 5 minutes of inactivity. That will be user-configurable in the next version 0.1.23. Another thing to be aware of is that models are memory mapped and so they don't show up in process memory. They are instead accounted for in file cache.
GiteaMirror commented 2026-04-22 03:55:34 -05:00
Author
Owner
Copy Link

@pdevine commented on GitHub (Jan 28, 2024):

Going to close this since #2146 has merged.

<!-- gh-comment-id:1913742830 --> @pdevine commented on GitHub (Jan 28, 2024): Going to close this since #2146 has merged.
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#27025
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?