🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/12511

• 📄 Preview Python docs built from this PR

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

❌ 1 New Failure, 12 Pending

As of commit 50e4b89 with merge base de24489 ():

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

@psiddh @mcr229 I left a comment on discord, but I don't think we should do this. Does this really need to be turned on for every single build? It's also causing apple benchmark tests to be broken

@psiddh @mcr229 I left a comment on discord, but I don't think we should do this. Does this really need to be turned on for every single build? It's also causing apple benchmark tests to be broken

Yeah looks like the default is 'ON' now. Also does it mean we need to revert it on main as well ?

Moreover looks like there is already an option (cmake arg: -DEXECUTORCH_XNNPACK_ENABLE_KLEIDI) to enable / disable KleidiAI

cc @larryliu0820

@jathu hey yea this is a requirement. We should turn this on by default. If we want good performance to come out of the box rather than have users hand configure the cmake options then we need to make it default. This is also an ask by Arm folks

#12552

i have a forward fix to fix the apple builds.

#12552

i have a forward fix to fix the apple builds.

Please cherry pick that as well

#12552
i have a forward fix to fix the apple builds.

Please cherry pick that as well

yup let's leave this open until my fix lands and its cherry picked. apologies for the complexity

#12607

this is the cherry pic of the fix. they need to be landed together.