Goal
Reduce the end-to-end time to import/bake the full SF map in the Unity pipeline.
Context
The current pipeline imports SF map data (elevation contours, parking regulations, etc.) and bakes chunk prefabs (with parked cars baked in; traffic spawned at runtime). Targeting the full-city extent, bake/import time is the bottleneck.
Scope of this techdesign
Evaluate concrete approaches to speed up the import/bake, e.g.:
- Caching intermediate artifacts (heightfield / contour processing already produces .heightcache files)
- Parallel / multi-threaded chunk baking
- Incremental / dirty-only rebakes
- Data format or I/O optimizations
- Reducing redundant work across chunks
Output: a technical approach doc with recommended optimizations, tradeoffs, and rough expected speedups.
Captured via /sdlc.
Goal
Reduce the end-to-end time to import/bake the full SF map in the Unity pipeline.
Context
The current pipeline imports SF map data (elevation contours, parking regulations, etc.) and bakes chunk prefabs (with parked cars baked in; traffic spawned at runtime). Targeting the full-city extent, bake/import time is the bottleneck.
Scope of this techdesign
Evaluate concrete approaches to speed up the import/bake, e.g.:
Output: a technical approach doc with recommended optimizations, tradeoffs, and rough expected speedups.
Captured via /sdlc.