[Erik-Lundell]

In the Ethos-U Backend, create a delegate context that is passed to ethosu_invoke_v3 as user_args.

When ethosu_execution_stats is linked in, it registers ethosu callbacks that use the context to write to an EthosUStatsBuffer that is owned by the application and registered using set_ethosu_stats_buffer.

The buffer consists of EthosUExecutionRecords that contain data for one delegate call, including metadata to identify what npu ran the delegate.

Additional introduced API:

• EthosUBackend_npu_begin() callback
• EthosUBackend_npu_end() callback These can be used for measuring CPU time during npu dispatch.
• ethosu_execution_metrics_enabled() -> bool

Use this infra in arm_perf_monitor.

cc @digantdesai @freddan80 @per @zingo @oscarandersson8218 @mansnils @Sebastian-Larsson @robell @rascani

[pytorch-bot]

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/24432

• 📄 Preview Python docs built from this PR

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

❌ 4 New Failures, 4 Pending

As of commit 4cc369b with merge base ab68460 ():

NEW FAILURES - The following jobs have failed:

• pull / android / run-emulator (gh)
  adb: device offline
• trunk / test-cortex-m-e2e / run (mv2, cortex-m0plus) / mv2-cortex-m0plus (gh)
  undefined reference to __atomic_test_and_set'`
• trunk / test-cortex-m-e2e / run (mv3, cortex-m0plus) / mv3-cortex-m0plus (gh)
  undefined reference to __atomic_test_and_set'`
• trunk / test-cortex-m-ops / run (cortex-m0plus) / cortex-m-ops-cortex-m0plus (gh)
  undefined reference to __atomic_test_and_set'`

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

[zingo]

Ok to merge, it seems buck2 is updated. Nice.