github-starred/ollama
2
0
Fork 0
mirror of https://github.com/ollama/ollama.git synced 2026-07-15 12:48:05 -05:00
Code Issues 16.9k Packages Projects Releases 122 Wiki Activity

[GH-ISSUE #2793] Determining Server Capacity for Hosting Ollama #1686

New Issue
Closed
opened 2026-04-12 11:39:35 -05:00 by GiteaMirror · 3 comments
GiteaMirror commented 2026-04-12 11:39:35 -05:00
Owner
Copy Link

Originally created by @juned-adenwalla on GitHub (Feb 27, 2024).
Original GitHub issue: https://github.com/ollama/ollama/issues/2793

I have a VPS server with 48GB RAM and a 12-core processor. I'm planning to host Ollama on it and develop a custom web app using Ollama's API. I'm curious about how many users my server will be able to handle simultaneously and how I can test the load balancing.

Could anyone suggest an approximate number of users my current server configuration can handle? Additionally, any guidance on testing the load balancing would be greatly appreciated.

Originally created by @juned-adenwalla on GitHub (Feb 27, 2024). Original GitHub issue: https://github.com/ollama/ollama/issues/2793 I have a VPS server with 48GB RAM and a 12-core processor. I'm planning to host Ollama on it and develop a custom web app using Ollama's API. I'm curious about how many users my server will be able to handle simultaneously and how I can test the load balancing. Could anyone suggest an approximate number of users my current server configuration can handle? Additionally, any guidance on testing the load balancing would be greatly appreciated.
GiteaMirror commented 2026-04-12 11:39:37 -05:00
Author
Owner
Copy Link

@easp commented on GitHub (Feb 28, 2024):

Ollama doesn't load balance requests. It processes each request serially in the order received.

Beyond that, CPU inference is generally quite slow for all but the smallest models. If you want to know how fast it will be, install ollama and ollama run mixtral "why is the sky blue" --verbose It'll print out the speed for prompt processing and text generation.

<!-- gh-comment-id:1968030654 --> @easp commented on GitHub (Feb 28, 2024): Ollama doesn't load balance requests. It processes each request serially in the order received. Beyond that, CPU inference is generally quite slow for all but the smallest models. If you want to know how fast it will be, install ollama and `ollama run mixtral "why is the sky blue" --verbose` It'll print out the speed for prompt processing and text generation.
GiteaMirror commented 2026-04-12 11:39:38 -05:00
Author
Owner
Copy Link

@juned-adenwalla commented on GitHub (Feb 29, 2024):

total duration: 1m5.52821173s
load duration: 1.368387941s
prompt eval count: 25 token(s)
prompt eval duration: 2.376201s
prompt eval rate: 10.52 tokens/s
eval count: 367 token(s)
eval duration: 1m1.782406s
eval rate: 5.94 tokens/s

I got this as an output, can you give me a general understand say if I install this on cloud and give it to my team how many people will be able to use it parallel.

<!-- gh-comment-id:1970646407 --> @juned-adenwalla commented on GitHub (Feb 29, 2024): total duration: 1m5.52821173s load duration: 1.368387941s prompt eval count: 25 token(s) prompt eval duration: 2.376201s prompt eval rate: 10.52 tokens/s eval count: 367 token(s) eval duration: 1m1.782406s eval rate: 5.94 tokens/s I got this as an output, can you give me a general understand say if I install this on cloud and give it to my team how many people will be able to use it parallel.
GiteaMirror commented 2026-04-12 11:39:38 -05:00
Author
Owner
Copy Link

@jmorganca commented on GitHub (May 10, 2024):

As @easp mentioned it's hard to estimate, however, this will get easier with new parallel requests as an upcoming feature. Stay tuned!

For a CPU server and a small model, it should handle at least a few clients at a time.

<!-- gh-comment-id:2103681972 --> @jmorganca commented on GitHub (May 10, 2024): As @easp mentioned it's hard to estimate, however, this will get easier with new parallel requests as an upcoming feature. Stay tuned! For a CPU server and a small model, it should handle at least a few clients at a time.
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#1686
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?