github-starred/ollama
2
0
Fork 0
mirror of https://github.com/ollama/ollama.git synced 2026-05-21 05:01:48 -05:00
Code Issues 16.9k Packages Projects Releases 122 Wiki Activity

[GH-ISSUE #7321] Support loading the same model more than once #66708

New Issue
Closed
opened 2026-05-04 07:54:56 -05:00 by GiteaMirror · 1 comment
GiteaMirror commented 2026-05-04 07:54:56 -05:00
Owner
Copy Link

Originally created by @jfwreinhardt on GitHub (Oct 22, 2024).
Original GitHub issue: https://github.com/ollama/ollama/issues/7321

Originally assigned to: @dhiltgen on GitHub.

Are there any plans to support loading the same model more than once?

On a CUDA based system with multiple GPUs, I have observed that performance decreases for each new concurrent prompt against the same model. To put it another way, we see higher tokens/s for sending a prompt to four different models concurrently than if we send four prompts to the same model concurrently.

It would be helpful from a performance perspective if ollama spawned a new runner when OLLAMA_NUM_PARALLEL was reached, rather than placing all the prompts in a queue to wait on a single runner.

Originally created by @jfwreinhardt on GitHub (Oct 22, 2024). Original GitHub issue: https://github.com/ollama/ollama/issues/7321 Originally assigned to: @dhiltgen on GitHub. Are there any plans to support loading the same model more than once? On a CUDA based system with multiple GPUs, I have observed that performance decreases for each new concurrent prompt against the same model. To put it another way, we see higher tokens/s for sending a prompt to four different models concurrently than if we send four prompts to the same model concurrently. It would be helpful from a performance perspective if ollama spawned a new runner when OLLAMA_NUM_PARALLEL was reached, rather than placing all the prompts in a queue to wait on a single runner.
GiteaMirror added the feature request label 2026-05-04 07:54:56 -05:00
GiteaMirror commented 2026-05-04 07:55:00 -05:00
Author
Owner
Copy Link

@dhiltgen commented on GitHub (Oct 22, 2024):

More advanced scheduling support for the same model is tracked in #3902

<!-- gh-comment-id:2429872852 --> @dhiltgen commented on GitHub (Oct 22, 2024): More advanced scheduling support for the same model is tracked in #3902
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 feature request
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#66708
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?