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 #9504] Request modular updates to reduce download size (decouple CUDA libraries from core updates) #31952

New Issue
Open
opened 2026-04-22 12:46:55 -05:00 by GiteaMirror · 2 comments
GiteaMirror commented 2026-04-22 12:46:55 -05:00
Owner
Copy Link

Originally created by @ice6 on GitHub (Mar 5, 2025).
Original GitHub issue: https://github.com/ollama/ollama/issues/9504

First, thank you for the amazing work on Ollama! 🚀

I'd like to suggest an optimization for the update mechanism. The current package size (~700MB) creates significant bandwidth consumption during updates, especially when only the core binary (~30MB) actually needs updating in many cases. The primary contributors to the package size appear to be CUDA libraries that get re-downloaded with every update regardless of version changes.

Proposed Solution:

  1. Modular Components
    Separate CUDA-related libraries from the core binary in the distribution package.

  2. Incremental Updates

    • Check existing local CUDA library versions during updates
    • Only download new components when actually required (version mismatch/new dependencies)

  3. Optional CUDA Distribution
    Provide CUDA libraries as:

    • Optional downloadable components
    • On-demand installation via ollama install-cuda [version]
    • Separate checksum-verified packages

Benefits:

  • Reduce typical update size by ~95% (from 700MB → ~30MB)
  • Save bandwidth for users with limited connections
  • Allow enterprise users to maintain internal CUDA library mirrors
  • Faster update/rollback operations
Originally created by @ice6 on GitHub (Mar 5, 2025). Original GitHub issue: https://github.com/ollama/ollama/issues/9504 First, thank you for the amazing work on Ollama! 🚀 I'd like to suggest an optimization for the update mechanism. The current package size (~700MB) creates significant bandwidth consumption during updates, especially when only the core binary (~30MB) actually needs updating in many cases. The primary contributors to the package size appear to be CUDA libraries that get re-downloaded with every update regardless of version changes. **Proposed Solution:** 1. **Modular Components** Separate CUDA-related libraries from the core binary in the distribution package. 2. **Incremental Updates** - Check existing local CUDA library versions during updates - Only download new components when actually required (version mismatch/new dependencies) 3. **Optional CUDA Distribution** Provide CUDA libraries as: - Optional downloadable components - On-demand installation via `ollama install-cuda [version]` - Separate checksum-verified packages **Benefits:** - Reduce typical update size by ~95% (from 700MB → ~30MB) - Save bandwidth for users with limited connections - Allow enterprise users to maintain internal CUDA library mirrors - Faster update/rollback operations
GiteaMirror added the feature request label 2026-04-22 12:46:55 -05:00
GiteaMirror commented 2026-04-22 12:46:56 -05:00
Author
Owner
Copy Link

@L3P3 commented on GitHub (Mar 18, 2025):

This is important, it hurts big time to download almost 2 GB for a machine that does not support CUDA anyways. 😢

<!-- gh-comment-id:2733500518 --> @L3P3 commented on GitHub (Mar 18, 2025): This is important, it hurts big time to download almost 2 GB for a machine that does not support CUDA anyways. 😢
GiteaMirror commented 2026-04-22 12:46:56 -05:00
Author
Owner
Copy Link

@ice6 commented on GitHub (Mar 18, 2025):

maybe build the code in the local machine can reduce the pain of current updating mechanism. 😆

https://github.com/ollama/ollama/blob/main/docs/development.md

<!-- gh-comment-id:2734940095 --> @ice6 commented on GitHub (Mar 18, 2025): maybe build the code in the local machine can reduce the pain of current updating mechanism. 😆 https://github.com/ollama/ollama/blob/main/docs/development.md
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#31952
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?