Skip to content

Add ETDump tracing to CpuBackend (#23926) - #23926

Open
JakeStevens wants to merge 1 commit into
pytorch:mainfrom
JakeStevens:export-D125525223
Open

JakeStevens wants to merge 1 commit into
pytorch:mainfrom
JakeStevens:export-D125525223

Conversation

@JakeStevens

@JakeStevens JakeStevens commented Oct 16, 2026 •

Copy link
Copy Markdown
Contributor

Summary:

CPUPlan records one ETDump delegate event per step, named provider/implementation:first_node[+N], so every provider is traced without provider changes. The XNNPACK provider enables XNNPACK runtime profiling when a tracer is attached and logs XNNPACK:<operator> #<n> events inside its step, with each operator's input and output shapes as JSON delegate metadata.

Shapes come from XNNPACK's internal runtime (src/xnnpack/subgraph.h). That header dependency and the code that reads it are only compiled when ET_EVENT_TRACER_ENABLED is defined. runner writes an ETDump when ET_CPU_ETDUMP is set.

Authored with Claude Code.

Differential Revision: D125525223

@pytorch-bot

pytorch-bot Bot commented Oct 16, 2026 •

Copy link
Copy Markdown

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/23926

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

❌ 1 New Failure

As of commit 39fcadc with merge base acb20a1 (image):

NEW FAILURE - The following job has failed:

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

@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 Oct 16, 2026
@meta-codesync

meta-codesync Bot commented Oct 16, 2026

Copy link
Copy Markdown
Contributor

@JakeStevens has exported this pull request. If you are a Meta employee, you can view the originating Diff in D125525223.

@github-actions

Copy link
Copy Markdown

This PR needs a release notes: label

If your change should be included in the release notes (i.e. would users of this library care about this change?), please use a label starting with release notes:. This helps us keep track and include your important work in the next release notes.

To add a label, you can comment to pytorchbot, for example
@pytorchbot label "release notes: none"

For more information, see
https://github.com/pytorch/pytorch/wiki/PyTorch-AutoLabel-Bot#why-categorize-for-release-notes-and-how-does-it-work.

@meta-codesync meta-codesync Bot changed the title Add ETDump tracing to CpuBackend Add ETDump tracing to CpuBackend (#23926) Oct 16, 2026
JakeStevens added a commit to JakeStevens/executorch that referenced this pull request Oct 16, 2026
Summary:

`CPUPlan` records one ETDump delegate event per step, named `provider/implementation:first_node[+N]`, so every provider is traced without provider changes. The XNNPACK provider enables XNNPACK runtime profiling when a tracer is attached and logs `XNNPACK:<operator> #<n>` events inside its step, with each operator's input and output shapes as JSON delegate metadata.

Shapes come from XNNPACK's internal runtime (`src/xnnpack/subgraph.h`). That header dependency and the code that reads it are only compiled when `ET_EVENT_TRACER_ENABLED` is defined. `runner` writes an ETDump when `ET_CPU_ETDUMP` is set.

Authored with Claude Code.

Differential Revision: D125525223
JakeStevens added a commit to JakeStevens/executorch that referenced this pull request Oct 27, 2026
Summary:

`CPUPlan` records one ETDump delegate event per step, named `provider/implementation:first_node[+N]`, so every provider is traced without provider changes. The XNNPACK provider enables XNNPACK runtime profiling when a tracer is attached and logs `XNNPACK:<operator> #<n>` events inside its step, with each operator's input and output shapes as JSON delegate metadata.

Shapes come from XNNPACK's internal runtime (`src/xnnpack/subgraph.h`). That header dependency and the code that reads it are only compiled when `ET_EVENT_TRACER_ENABLED` is defined. `runner` writes an ETDump when `ET_CPU_ETDUMP` is set.

Authored with Claude Code.

Differential Revision: D125525223
JakeStevens added a commit to JakeStevens/executorch that referenced this pull request Oct 27, 2026
Summary:

`CPUPlan` records one ETDump delegate event per step, named `provider/implementation:first_node[+N]`, so every provider is traced without provider changes. The XNNPACK provider enables XNNPACK runtime profiling when a tracer is attached and logs `XNNPACK:<operator> #<n>` events inside its step, with each operator's input and output shapes as JSON delegate metadata.

Shapes come from XNNPACK's internal runtime (`src/xnnpack/subgraph.h`). That header dependency and the code that reads it are only compiled when `ET_EVENT_TRACER_ENABLED` is defined. `runner` writes an ETDump when `ET_CPU_ETDUMP` is set.

Authored with Claude Code.

Differential Revision: D125525223
Summary:

`CPUPlan` records one ETDump delegate event per step, named `provider/implementation:first_node[+N]`, so every provider is traced without provider changes. The XNNPACK provider enables XNNPACK runtime profiling when a tracer is attached and logs `XNNPACK:<operator> #<n>` events inside its step, with each operator's input and output shapes as JSON delegate metadata.

Shapes come from XNNPACK's internal runtime (`src/xnnpack/subgraph.h`). That header dependency and the code that reads it are only compiled when `ET_EVENT_TRACER_ENABLED` is defined. `runner` writes an ETDump when `ET_CPU_ETDUMP` is set.

Authored with Claude Code.

Differential Revision: D125525223

This branch has not been deployed

No deployments
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. meta-exported

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant