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 #8324] Add a CUDA+AVX2(VNNI) runner to the Docker image. #51843

New Issue
Open
opened 2026-04-28 21:03:49 -05:00 by GiteaMirror · 0 comments
GiteaMirror commented 2026-04-28 21:03:49 -05:00
Owner
Copy Link

Originally created by @x0wllaar on GitHub (Jan 6, 2025).
Original GitHub issue: https://github.com/ollama/ollama/issues/8324

Description:
I would like to ask to add a CUDA+AVX2 (maybe VNNI) model runner to the default Docker image for Ollama. I think this can help with performance in partial offload scenarios.
This should be supported at build time (#2281), but for some reason I cant find the runner in the docker image

I think that it should be possible for me to do locally by adding --build-arg CUSTOM_CPU_FLAGS=avx2,avxvnni to Docker build, but I still think that it can be beneficial to add a runner target named something like cuda_v12_moderncpu that will enable AVX2 and AVX-VNNI by default and that will build by default.

I might work on this and submit a PR.

Benefits:

  • Squeeze some more performance in partial offload

Example Use Case:
Running a 32B model on a 16GB VRAM GPU (in my case 13900HX + Laptop 4090).

Environment Variables/Configuration:
No additional configuration needed.

Related Files and Code:

  • make/cuda.make (I guess)
Originally created by @x0wllaar on GitHub (Jan 6, 2025). Original GitHub issue: https://github.com/ollama/ollama/issues/8324 **Description**: I would like to ask to add a CUDA+AVX2 (maybe VNNI) model runner to the default Docker image for Ollama. I think this can help with performance in partial offload scenarios. This should be supported at build time (#2281), but for some reason I cant find the runner in the docker image I think that it should be possible for me to do locally by adding `--build-arg CUSTOM_CPU_FLAGS=avx2,avxvnni` to Docker build, but I still think that it can be beneficial to add a runner target named something like `cuda_v12_moderncpu` that will enable AVX2 and AVX-VNNI by default and that will build by default. I might work on this and submit a PR. **Benefits**: * Squeeze some more performance in partial offload **Example Use Case**: Running a 32B model on a 16GB VRAM GPU (in my case 13900HX + Laptop 4090). **Environment Variables/Configuration**: No additional configuration needed. **Related Files and Code**: * make/cuda.make (I guess)
GiteaMirror added the feature request label 2026-04-28 21:03:49 -05:00
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#51843
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?