github-starred/ollama
2
0
Fork 0
mirror of https://github.com/ollama/ollama.git synced 2026-05-07 08:30:05 -05:00
Code Issues 14.5k Packages Projects Releases 120 Wiki Activity

[PR #13735] [MERGED] MLX - dynamic loading of mlx-c #14368

New Issue
Closed
opened 2026-04-13 00:51:54 -05:00 by GiteaMirror · 0 comments
GiteaMirror commented 2026-04-13 00:51:54 -05:00
Owner
Copy Link

📋 Pull Request Information

Original PR: https://github.com/ollama/ollama/pull/13735
Author: @dhiltgen
Created: 1/15/2026
Status: ✅ Merged
Merged: 1/17/2026
Merged by: @dhiltgen

Base: main ← Head: mlx_dlopen

📝 Commits (3)

📊 Changes

24 files changed (+9043 additions, -228 deletions)

View changed files

📝 Dockerfile (+6 -12)
➕ MLX_VERSION (+1 -0)
📝 README.md (+2 -2)
📝 scripts/build_darwin.sh (+9 -11)
📝 x/imagegen/cmd/engine/generate.go (+4 -4)
📝 x/imagegen/cmd/engine/main.go (+5 -1)
📝 x/imagegen/mlx/compile.go (+1 -1)
➕ x/imagegen/mlx/doc.go (+6 -0)
➕ x/imagegen/mlx/generate_wrappers.go (+439 -0)
➕ x/imagegen/mlx/mlx.c (+5786 -0)
📝 x/imagegen/mlx/mlx.go (+55 -190)
➕ x/imagegen/mlx/mlx.h (+2337 -0)
➕ x/imagegen/mlx/mlx_dynamic.c (+144 -0)
➕ x/imagegen/mlx/mlx_dynamic.h (+29 -0)
📝 x/imagegen/mlx/mlx_test.go (+21 -0)
📝 x/imagegen/models/qwen_image/pipeline_test.go (+21 -0)
📝 x/imagegen/models/qwen_image/qwen_image.go (+0 -1)
📝 x/imagegen/models/qwen_image_edit/rope_test.go (+22 -0)
📝 x/imagegen/nn/nn_test.go (+22 -0)
📝 x/imagegen/runner/runner.go (+6 -0)

...and 4 more files

📄 Description

Create a wrapper layer to indirect the dependency on mlx-c so the main ollama binary does not have a load-time dependency on mlx-c, mlx, and on linux, cuda. Lazy load the library via dlopen so we can adjust the path to ensure the dependencies are found and fail gracefully if not present.

This enables us to move back to shipping a single ollama CLI executable on Mac/Linux with MLX enabled, and safeguard against potential missing libraries by deferring library dependency resolution.

Verified on Mac and Linux. The Windows impl is untested for now.

🔄 This issue represents a GitHub Pull Request. It cannot be merged through Gitea due to API limitations.

## 📋 Pull Request Information **Original PR:** https://github.com/ollama/ollama/pull/13735 **Author:** [@dhiltgen](https://github.com/dhiltgen) **Created:** 1/15/2026 **Status:** ✅ Merged **Merged:** 1/17/2026 **Merged by:** [@dhiltgen](https://github.com/dhiltgen) **Base:** `main` ← **Head:** `mlx_dlopen` --- ### 📝 Commits (3) - [`ffdc257`](https://github.com/ollama/ollama/commit/ffdc257bef698bd28f10f34279076ea2a0eed7ff) MLX - dynamic loading of mlx-c - [`1c2d073`](https://github.com/ollama/ollama/commit/1c2d073750295aef05836b111f5b0b9fae2a79e8) review comments - [`ffff6e3`](https://github.com/ollama/ollama/commit/ffff6e37ac800ca59194fb26ff29392f2d224de4) fix broken tests ### 📊 Changes **24 files changed** (+9043 additions, -228 deletions) <details> <summary>View changed files</summary> 📝 `Dockerfile` (+6 -12) ➕ `MLX_VERSION` (+1 -0) 📝 `README.md` (+2 -2) 📝 `scripts/build_darwin.sh` (+9 -11) 📝 `x/imagegen/cmd/engine/generate.go` (+4 -4) 📝 `x/imagegen/cmd/engine/main.go` (+5 -1) 📝 `x/imagegen/mlx/compile.go` (+1 -1) ➕ `x/imagegen/mlx/doc.go` (+6 -0) ➕ `x/imagegen/mlx/generate_wrappers.go` (+439 -0) ➕ `x/imagegen/mlx/mlx.c` (+5786 -0) 📝 `x/imagegen/mlx/mlx.go` (+55 -190) ➕ `x/imagegen/mlx/mlx.h` (+2337 -0) ➕ `x/imagegen/mlx/mlx_dynamic.c` (+144 -0) ➕ `x/imagegen/mlx/mlx_dynamic.h` (+29 -0) 📝 `x/imagegen/mlx/mlx_test.go` (+21 -0) 📝 `x/imagegen/models/qwen_image/pipeline_test.go` (+21 -0) 📝 `x/imagegen/models/qwen_image/qwen_image.go` (+0 -1) 📝 `x/imagegen/models/qwen_image_edit/rope_test.go` (+22 -0) 📝 `x/imagegen/nn/nn_test.go` (+22 -0) 📝 `x/imagegen/runner/runner.go` (+6 -0) _...and 4 more files_ </details> ### 📄 Description Create a wrapper layer to indirect the dependency on mlx-c so the main ollama binary does not have a load-time dependency on mlx-c, mlx, and on linux, cuda. Lazy load the library via dlopen so we can adjust the path to ensure the dependencies are found and fail gracefully if not present. This enables us to move back to shipping a single `ollama` CLI executable on Mac/Linux with MLX enabled, and safeguard against potential missing libraries by deferring library dependency resolution. Verified on Mac and Linux. The Windows impl is untested for now. --- <sub>🔄 This issue represents a GitHub Pull Request. It cannot be merged through Gitea due to API limitations.</sub>
GiteaMirror added the pull-request label 2026-04-13 00:51:54 -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 pull-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#14368
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?