Hi everyone,
Maya Version:
- Maya 2022 / 2024
- Python 3
Context:
I’m working on production-style tools that process large numbers of nodes
(geometry queries, attribute reads/writes, and DAG traversal). These tools
are executed frequently, sometimes via UI callbacks.
Common advice suggests replacing maya.cmds with OpenMaya for performance,
but in real-world tools the results don’t always seem consistent.
What I’m Trying to Understand:
- When OpenMaya actually provides measurable performance benefits
- When maya.cmds is “good enough” and easier to maintain
- How much overhead comes from Python vs Maya’s internal execution
Observations:
- Some OpenMaya loops are faster, but setup and data conversion add overhead
- In mixed pipelines, the gains sometimes feel marginal
- UI-driven tools often hide the actual performance difference
What I’ve Tried:
- Profiling cmds-based implementations vs OpenMaya equivalents
- Using MFn* classes for direct node access
- Reducing string-based lookups and repeated cmds calls
Questions:
1. In which scenarios does OpenMaya truly outperform cmds in production?
2. Are there internal costs (wrapping, data conversion) that reduce gains?
3. What patterns do experienced TDs follow when deciding between the two?
Looking for real production experience and not just theoretical guidance.
Thanks in advance.
