Summary

• Optimizes renderer loops for mesh instances to reduce per-frame work and improve frame rates.

Changes

Rendering Efficiency

• Reworked mesh instance rendering loop to batch visible instances and reduce per-mesh allocations.
• Moved frustum culling earlier in the path to skip non-visible instances before per-instance processing.
• Cached frequently used transforms and bounding data to avoid repeated calculations in the hot path.
• Inlined small helpers and reused temporary vectors to minimize GC pressure.

Stability and Correctness

• Added guards to skip processing for disabled mesh instances or materials with zero alpha.
• Ensured rendering remains correct with batched processing and preserved depth sorting semantics.

Dependency Updates

• Bump PlayCanvas dependency in examples/package-lock.json from 2.8.0-dev.0 to 2.9.0-beta.1.

Test plan

• Run performance benchmarks with large meshes to verify frame-time improvements
• Compare rendered output against baseline for visual correctness
• Ensure frustum culling correctly excludes culled instances
• Update example lockfile to reflect dependency bump

🌿 Generated by Terry

ℹ️ Tag @terragon-labs to ask questions and address PR feedback

📎 Task: https://www.terragonlabs.com/task/b8f2a3b7-9092-4396-9ba9-7baa450b0bbf