Skip to content

Qwen 3.5 MoE: Add Metal build target#18881

Open
manuelcandales wants to merge 3 commits intogh/manuelcandales/175/headfrom
gh/manuelcandales/176/head
Open

Qwen 3.5 MoE: Add Metal build target#18881
manuelcandales wants to merge 3 commits intogh/manuelcandales/175/headfrom
gh/manuelcandales/176/head

Conversation

@manuelcandales
Copy link
Copy Markdown
Contributor

Adds make qwen3_5_moe-metal which builds the runner linked against the
Metal backend instead of CUDA:

  • CMakeLists.txt: conditional metal_backend vs aoti_cuda_backend linking
  • CMakePresets.json: add qwen3-5-moe-metal preset (Darwin only)
  • main.cpp: guard CUDA includes/calls behind EXECUTORCH_BUILD_CUDA,
    route T=1 prompts to decode method (prefill has min seq_len=2)
  • Makefile: add qwen3_5_moe-metal target

Authored with Claude.

[ghstack-poisoned]
@pytorch-bot
Copy link
Copy Markdown

pytorch-bot bot commented Apr 14, 2026

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/18881

Note: Links to docs will display an error until the docs builds have been completed.

❗ 1 Active SEVs

There are 1 currently active SEVs. If your PR is affected, please view them below:

❌ 1 New Failure, 2 Unrelated Failures

As of commit 598c58f with merge base 5707e2a (image):

NEW FAILURE - The following job has failed:

BROKEN TRUNK - The following jobs failed but were present on the merge base:

👉 Rebase onto the `viable/strict` branch to avoid these failures

This comment was automatically generated by Dr. CI and updates every 15 minutes.

manuelcandales added a commit that referenced this pull request Apr 14, 2026
Adds make qwen3_5_moe-metal which builds the runner linked against the
Metal backend instead of CUDA:
- CMakeLists.txt: conditional metal_backend vs aoti_cuda_backend linking
- CMakePresets.json: add qwen3-5-moe-metal preset (Darwin only)
- main.cpp: guard CUDA includes/calls behind EXECUTORCH_BUILD_CUDA,
  route T=1 prompts to decode method (prefill has min seq_len=2)
- Makefile: add qwen3_5_moe-metal target

Authored with Claude.


ghstack-source-id: 655fca1
ghstack-comment-id: 4245515800
Pull-Request: #18881
@meta-cla meta-cla bot added the CLA Signed This label is managed by the Facebook bot. Authors need to sign the CLA before a PR can be reviewed. label Apr 14, 2026
[ghstack-poisoned]
manuelcandales added a commit that referenced this pull request Apr 14, 2026
Adds make qwen3_5_moe-metal which builds the runner linked against the
Metal backend instead of CUDA:
- CMakeLists.txt: conditional metal_backend vs aoti_cuda_backend linking
- CMakePresets.json: add qwen3-5-moe-metal preset (Darwin only)
- main.cpp: guard CUDA includes/calls behind EXECUTORCH_BUILD_CUDA,
  route T=1 prompts to decode method (prefill has min seq_len=2)
- Makefile: add qwen3_5_moe-metal target

Authored with Claude.

ghstack-source-id: aa806ac
ghstack-comment-id: 4245515800
Pull-Request: #18881
[ghstack-poisoned]
manuelcandales added a commit that referenced this pull request Apr 14, 2026
Adds make qwen3_5_moe-metal which builds the runner linked against the
Metal backend instead of CUDA:
- CMakeLists.txt: conditional metal_backend vs aoti_cuda_backend linking
- CMakePresets.json: add qwen3-5-moe-metal preset (Darwin only)
- main.cpp: guard CUDA includes/calls behind EXECUTORCH_BUILD_CUDA,
  route T=1 prompts to decode method (prefill has min seq_len=2)
- Makefile: add qwen3_5_moe-metal target

Authored with Claude.

ghstack-source-id: 80a1b24
ghstack-comment-id: 4245515800
Pull-Request: #18881
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

CLA Signed This label is managed by the Facebook bot. Authors need to sign the CLA before a PR can be reviewed.

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant