Skip to content

Fix multi-host profiling and resolve profiler tracer collisions - #455

Open
richaguptaa17 wants to merge 1 commit into
AI-Hypercomputer:mainfrom
richaguptaa17:fix-multihost-profiling
Open

Fix multi-host profiling and resolve profiler tracer collisions#455
richaguptaa17 wants to merge 1 commit into
AI-Hypercomputer:mainfrom
richaguptaa17:fix-multihost-profiling

Conversation

@richaguptaa17

Copy link
Copy Markdown
  • Enable ML Diagnostics profiler on secondary TPU worker processes by removing the jax.process_index() == 0 lock from _ml_diagnostics_profiler_enabled().
  • Refactor Profiler class in max_utils.py with active state tracking (_active). When both enable_ml_diagnostics and enable_profiler are True, prioritize ML Diagnostics and skip the standalone JAX profiler with a log warning to avoid RuntimeError tracer collisions.

- Enable ML Diagnostics profiler on secondary TPU worker processes by removing the jax.process_index() == 0 lock from _ml_diagnostics_profiler_enabled().
- Refactor Profiler class in max_utils.py with active state tracking (_active). When both enable_ml_diagnostics and enable_profiler are True, prioritize ML Diagnostics and skip the standalone JAX profiler with a log warning to avoid RuntimeError tracer collisions.
@richaguptaa17
richaguptaa17 requested a review from entrpn as a code owner July 31, 2026 11:33
@gemini-code-assist

Copy link
Copy Markdown

Caution

The consumer version of Gemini Code Assist on GitHub has been sunset. All code review activity has officially ceased.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant