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 #7740] Performance is decreasing #4942

New Issue
Closed
opened 2026-04-12 16:00:13 -05:00 by GiteaMirror · 2 comments
GiteaMirror commented 2026-04-12 16:00:13 -05:00
Owner
Copy Link

Originally created by @murzein on GitHub (Nov 19, 2024).
Original GitHub issue: https://github.com/ollama/ollama/issues/7740

What is the issue?

Load any model, for example gemma2:27b
OLLAMA_KEEP_ALIVE: -1

command: ollama run gemma2:27b --verbose
message: tell me about amd

1 iteration
total duration: 12.9263992s
load duration: 32.8561ms
prompt eval count: 13 token(s) <----
prompt eval duration: 73ms
prompt eval rate: 178.08 tokens/s
eval count: 480 token(s)
eval duration: 12.81s
eval rate: 37.47 tokens/s <----

2 iteration
prompt eval count: 506 token(s)
eval rate: 36.64 tokens/s

3 iteration
prompt eval count: 882 token(s)
eval rate: 36.05 tokens/s

4 iteration
prompt eval count: 1244 token(s)
eval rate: 35.81 tokens/s

5 iteration
prompt eval count: 1584 token(s)
eval rate: 35.30 tokens/s

6 iteration
prompt eval count: 1860 token(s)
eval rate: 17.04 tokens/s

Overload prompt, eval rate - 17.04 tokens/s

Reload the script
ctrl+z, and again:
command: ollama run gemma2:27b --verbose
message: tell me about amd

1 iteration
prompt eval count: 13 token(s)
eval rate: 18.01 tokens/s

Small prompt, but speed low.

I am experiencing the same issue when accessing ollama via the API.
Context is getting overflow somewhere.

OS

Linux, Windows

GPU

Nvidia, AMD

CPU

Intel

Ollama version

0.4.2

Originally created by @murzein on GitHub (Nov 19, 2024). Original GitHub issue: https://github.com/ollama/ollama/issues/7740 ### What is the issue? Load any model, for example gemma2:27b OLLAMA_KEEP_ALIVE: -1 command: **ollama run gemma2:27b --verbose** message: **tell me about amd** **1 iteration** total duration: 12.9263992s load duration: 32.8561ms prompt eval count: **13 token(s)** <---- prompt eval duration: 73ms prompt eval rate: 178.08 tokens/s eval count: 480 token(s) eval duration: 12.81s eval rate: **37.47 tokens/s** <---- **2 iteration** prompt eval count: **506 token(s)** eval rate: **36.64 tokens/s** **3 iteration** prompt eval count: **882 token(s)** eval rate: **36.05 tokens/s** **4 iteration** prompt eval count: **1244 token(s)** eval rate: **35.81 tokens/s** **5 iteration** prompt eval count: **1584 token(s)** eval rate: **35.30 tokens/s** **6 iteration** prompt eval count: **1860 token(s)** eval rate: **17.04 tokens/s** Overload prompt, eval rate - 17.04 tokens/s **Reload the script** ctrl+z, and again: command: **ollama run gemma2:27b --verbose** message: **tell me about amd** **1 iteration** prompt eval count: **13 token(s)** eval rate: **18.01 tokens/s** Small prompt, but speed low. I am experiencing the same issue when accessing ollama via the API. Context is getting overflow somewhere. ### OS Linux, Windows ### GPU Nvidia, AMD ### CPU Intel ### Ollama version 0.4.2
GiteaMirror added the bug label 2026-04-12 16:00:13 -05:00
GiteaMirror commented 2026-04-12 16:00:14 -05:00
Author
Owner
Copy Link

@rick-github commented on GitHub (Nov 19, 2024):

https://github.com/ollama/ollama/issues/7717

<!-- gh-comment-id:2485767536 --> @rick-github commented on GitHub (Nov 19, 2024): https://github.com/ollama/ollama/issues/7717
GiteaMirror commented 2026-04-12 16:00:14 -05:00
Author
Owner
Copy Link

@jessegross commented on GitHub (Nov 19, 2024):

I agree that this is likely the same as #7717 so it is best to track it there. However, @murzein I think we have a fix in main if you are able to build and try it out.

<!-- gh-comment-id:2487002945 --> @jessegross commented on GitHub (Nov 19, 2024): I agree that this is likely the same as #7717 so it is best to track it there. However, @murzein I think we have a fix in `main` if you are able to build and try it out.
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 bug
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#4942
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?