Rename CMake EXECUTORCH_BUILD_KERNELS_CUSTOM to EXECUTORCH_BUILD_EXTENSION_LLM

[GregoryComer]

Rename EXECUTORCH_BUILD_KERNELS_CUSTOM to EXECUTORCH_BUILD_EXTENSION_LLM. This improves clarity for users and prevents us from needing to keep EXECUTORCH_BUILD_KERNELS_CUSTOM post-API stability guarantees.

Note that the two CI failures are due to running CI from a fork.

[pytorch-bot]

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/5743

• 📄 Preview Python docs built from this PR

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:

• PyTorch Testing Nodes Undergoing ROCm 6.2.1 Upgrades

❌ 2 New Failures, 1 Unrelated Failure

As of commit 22d6aeb with merge base 905b88c ():

NEW FAILURES - The following jobs have failed:

• Apple / build-demo-ios / macos-job (gh)
  RuntimeError: Command bash /Users/runner/work/_temp/exec_script failed with exit code 1
• Apple / upload-frameworks-ios (gh)
  ##[error]Credentials could not be loaded, please check your action inputs: Could not load credentials from any providers

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

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

• trunk / test-coreml-delegate / macos-job (gh) (trunk failure)
  RuntimeError: Command bash /Users/runner/work/_temp/exec_script failed with exit code 1

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

[facebook-github-bot]

@GregoryComer has imported this pull request. If you are a Meta employee, you can view this diff on Phabricator.

[mcr229]

hmm makes sense, looking at
DEXECUTORCH_BUILD_KERNELS_OPTIMIZED=ON
and
DEXECUTORCH_BUILD_KERNELS_CUSTOM=ON

seems hard to tell the difference between what is 'optimized' and what is 'custom'

[shoumikhin]

If we continue to apply this renaming to the rest of the kernels custom instances, does that mean we're gonna distribute extension llm package eventually on apple platforms together with the executorch runtime?
Please check out what it looks like now for the end users https://pytorch.org/executorch/main/apple-runtime.html
I guess kernels customer are in fact some kernels for llms?

[GregoryComer]

@shoumikhin Yeah, that makes sense to me. If we rename them to llm kernels, do I just need to update the frameworks build as well?