github-starred/ollama
2
0
Fork 0
mirror of https://github.com/ollama/ollama.git synced 2026-05-06 08:02:14 -05:00
Code Issues 14.5k Packages Projects Releases 120 Wiki Activity

[PR #14994] [CLOSED] ggml-metal: probe tensor support with Metal 4 #20228

New Issue
Closed
opened 2026-04-16 07:30:50 -05:00 by GiteaMirror · 0 comments
GiteaMirror commented 2026-04-16 07:30:50 -05:00
Owner
Copy Link

📋 Pull Request Information

Original PR: https://github.com/ollama/ollama/pull/14994
Author: @lentil32
Created: 3/21/2026
Status: ❌ Closed

Base: main ← Head: fix/m5-metal-mixed-bf16-f16

📝 Commits (1)

  • 1e89733 ggml-metal: probe mixed bf16/f16 tensor support

📊 Changes

1 file changed (+52 additions, -0 deletions)

View changed files

📝 ml/backend/ggml/ggml/src/ggml-metal/ggml-metal-device.m (+52 -0)

📄 Description

Metal source compilation defaults to an older language mode unless MTLCompileOptions.languageVersion is set. On Apple M5 systems, that causes the tensor support probes to fail even when the f16 and bfloat tensor paths are valid.

Compile tensor probes as Metal 4 when tensor support is available, and keep the mixed bfloat/half probe that disables tensor mode before the full library is compiled on systems that reject that combination.

Tested on an Apple M5 Pro MacBook:

  • f16 tensor probe succeeds
  • bfloat tensor probe succeeds
  • mixed bfloat/half probe fails and disables tensor mode as expected
  • glm-4.7-flash loads and responds successfully

Related: #14432
Related: #13460

Fixes #13867.

🔄 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/14994 **Author:** [@lentil32](https://github.com/lentil32) **Created:** 3/21/2026 **Status:** ❌ Closed **Base:** `main` ← **Head:** `fix/m5-metal-mixed-bf16-f16` --- ### 📝 Commits (1) - [`1e89733`](https://github.com/ollama/ollama/commit/1e8973359dd6fa0acfbf1b5cf6e2e6a3a3a8430b) ggml-metal: probe mixed bf16/f16 tensor support ### 📊 Changes **1 file changed** (+52 additions, -0 deletions) <details> <summary>View changed files</summary> 📝 `ml/backend/ggml/ggml/src/ggml-metal/ggml-metal-device.m` (+52 -0) </details> ### 📄 Description Metal source compilation defaults to an older language mode unless `MTLCompileOptions.languageVersion` is set. On Apple M5 systems, that causes the tensor support probes to fail even when the `f16` and `bfloat` tensor paths are valid. Compile tensor probes as Metal 4 when tensor support is available, and keep the mixed `bfloat`/`half` probe that disables tensor mode before the full library is compiled on systems that reject that combination. Tested on an Apple M5 Pro MacBook: - `f16` tensor probe succeeds - `bfloat` tensor probe succeeds - mixed `bfloat`/`half` probe fails and disables tensor mode as expected - `glm-4.7-flash` loads and responds successfully Related: #14432 Related: #13460 Fixes #13867. --- <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 2026-04-16 07:30:50 -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#20228
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?