github-starred/ollama
2
0
Fork 0
mirror of https://github.com/ollama/ollama.git synced 2026-05-07 16:40:08 -05:00
Code Issues 14.5k Packages Projects Releases 120 Wiki Activity

[GH-ISSUE #15487] 0.20.5: Metal shader compilation fails on M5 Max (MTLGPUFamilyMetal4) — MPPTensorOpsMatMul2d type mismatch #35661

New Issue
Open
opened 2026-04-22 20:20:00 -05:00 by GiteaMirror · 7 comments
GiteaMirror commented 2026-04-22 20:20:00 -05:00
Owner
Copy Link

Originally created by @dextersecret on GitHub (Apr 10, 2026).
Original GitHub issue: https://github.com/ollama/ollama/issues/15487

the GPU family: MTLGPUFamilyApple10 (1010) block, the MPPTensorOpsMatMul2dImpl.h:3266:5: error: static_assert failed block, and the Ollama version.
NOTE: it works on 0.18.3 on the same hardware

Originally created by @dextersecret on GitHub (Apr 10, 2026). Original GitHub issue: https://github.com/ollama/ollama/issues/15487 the GPU family: MTLGPUFamilyApple10 (1010) block, the MPPTensorOpsMatMul2dImpl.h:3266:5: error: static_assert failed block, and the Ollama version. NOTE: it works on 0.18.3 on the same hardware
GiteaMirror commented 2026-04-22 20:20:04 -05:00
Author
Owner
Copy Link

@thejohnnybot commented on GitHub (Apr 11, 2026):

same

<!-- gh-comment-id:4227932915 --> @thejohnnybot commented on GitHub (Apr 11, 2026): same
GiteaMirror commented 2026-04-22 20:20:05 -05:00
Author
Owner
Copy Link

@edmava8 commented on GitHub (Apr 11, 2026):

Same

<!-- gh-comment-id:4228063273 --> @edmava8 commented on GitHub (Apr 11, 2026): Same
GiteaMirror commented 2026-04-22 20:20:06 -05:00
Author
Owner
Copy Link

@Winddfall commented on GitHub (Apr 12, 2026):

same

<!-- gh-comment-id:4230897724 --> @Winddfall commented on GitHub (Apr 12, 2026): same
GiteaMirror commented 2026-04-22 20:20:06 -05:00
Author
Owner
Copy Link

@illimit3 commented on GitHub (Apr 12, 2026):

same

<!-- gh-comment-id:4231239879 --> @illimit3 commented on GitHub (Apr 12, 2026): same
GiteaMirror commented 2026-04-22 20:20:07 -05:00
Author
Owner
Copy Link

@alanhere commented on GitHub (Apr 13, 2026):

Confirming same issue on M5 MacBook Air
I can reproduce this on a standard M5 MacBook Air (24 GB unified memory) running Ollama 0.20.6 attempting to load gemma4:e4b.

Environment:
Hardware: Apple MacBook Air, M5, 24 GB
GPU family: MTLGPUFamilyApple10 / MTLGPUFamilyMetal4
macOS: Tahoe 26.3.1
Ollama: 0.20.6
Model: gemma4:e4b

Report
The device reports MTLGPUFamilyMetal4 (5002) and the crash report — Metal shader compilation fails with:
static_assert failed due to requirement '__is_same_v<bfloat, half>'
"Input types must match cooperative tensor types"
followed by:
ggml_metal_init: error: failed to initialize the Metal library
ggml_backend_metal_device_init: error: failed to allocate context
GGML_ASSERT(backend) failed
SIGABRT: abort

<!-- gh-comment-id:4237738385 --> @alanhere commented on GitHub (Apr 13, 2026): **Confirming same issue on M5 MacBook Air** I can reproduce this on a standard M5 MacBook Air (24 GB unified memory) running Ollama 0.20.6 attempting to load gemma4:e4b. **Environment:** Hardware: Apple MacBook Air, M5, 24 GB GPU family: MTLGPUFamilyApple10 / MTLGPUFamilyMetal4 macOS: Tahoe 26.3.1 Ollama: 0.20.6 Model: gemma4:e4b **Report** The device reports MTLGPUFamilyMetal4 (5002) and the crash report — Metal shader compilation fails with: static_assert failed due to requirement '__is_same_v<bfloat, half>' "Input types must match cooperative tensor types" followed by: ggml_metal_init: error: failed to initialize the Metal library ggml_backend_metal_device_init: error: failed to allocate context GGML_ASSERT(backend) failed SIGABRT: abort
GiteaMirror commented 2026-04-22 20:20:07 -05:00
Author
Owner
Copy Link

@marcolussetti commented on GitHub (Apr 14, 2026):

Same behavior reported by @alanhere on Ollama 0.20.7 as well.

Environment:
Hardware: Apple MacBook Pro, M5 Pro, 64 GB
GPU family: MTLGPUFamilyApple10 / MTLGPUFamilyMetal4
macOS: Tahoe 26.3.2
Ollama: 0.20.7
Model: gemma4:31b

<!-- gh-comment-id:4245675455 --> @marcolussetti commented on GitHub (Apr 14, 2026): Same behavior reported by @alanhere on Ollama 0.20.7 as well. Environment: Hardware: Apple MacBook Pro, M5 Pro, 64 GB GPU family: MTLGPUFamilyApple10 / MTLGPUFamilyMetal4 macOS: Tahoe 26.3.2 Ollama: 0.20.7 Model: gemma4:31b
GiteaMirror commented 2026-04-22 20:20:08 -05:00
Author
Owner
Copy Link

@PureBlissAK commented on GitHub (Apr 18, 2026):

🤖 Automated Triage & Analysis Report

Issue: #15487
Analyzed: 2026-04-18T18:21:09.079399

Analysis

  • Type: unknown
  • Severity: medium
  • Components: unknown

Implementation Plan

  • Effort: medium
  • Steps:

This issue has been triaged and marked for implementation.

<!-- gh-comment-id:4274307802 --> @PureBlissAK commented on GitHub (Apr 18, 2026): <!-- ollama-issue-orchestrator:v1 issue:15487 --> ## 🤖 Automated Triage & Analysis Report **Issue**: #15487 **Analyzed**: 2026-04-18T18:21:09.079399 ### Analysis - **Type**: unknown - **Severity**: medium - **Components**: unknown ### Implementation Plan - **Effort**: medium - **Steps**: *This issue has been triaged and marked for implementation.*
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
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#35661
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?