john/McRogueFace
1
0
Fork 0
Code Issues 27 Pull requests Projects Releases Wiki Activity

chore: Extend benchmark to test 5000 entities

Browse source 
Validate SpatialHash scalability with larger entity counts.
Results at 5,000 entities:
- N×N visibility: 216.9× faster (431ms → 2ms)
- Single query: 37.4× faster (0.11ms → 0.003ms)

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
This commit is contained in:
John McCardle 2025-12-28 00:46:43 -05:00
commit 366ccecb7d
1 changed files with 2 additions and 2 deletions
Download patch file Download diff file Expand all files Collapse all files

4
tests/benchmarks/entity_scale_benchmark.py
View file

@ -29,8 +29,8 @@ GRID_SIZE = (100, 100) # 10,000 cells - entities will actually see each other
# Full suite - may timeout on large counts due to O(n²) visibility # Full suite - may timeout on large counts due to O(n²) visibility
# ENTITY_COUNTS = [100, 500, 1000, 2500, 5000, 10000] # ENTITY_COUNTS = [100, 500, 1000, 2500, 5000, 10000]
# Quick suite for initial baseline (on 100x100 grid, these give densities of 1-20%) # Extended suite to validate scalability (on 100x100 grid)
ENTITY_COUNTS = [100, 500, 1000, 2000] ENTITY_COUNTS = [100, 500, 1000, 2000, 5000]
QUERY_RADIUS = 15 # Smaller radius for smaller grid QUERY_RADIUS = 15 # Smaller radius for smaller grid
MOVEMENT_PERCENT = 0.10 # 10% of entities move each frame MOVEMENT_PERCENT = 0.10 # 10% of entities move each frame
N2N_SAMPLE_SIZE = 50 # Sample size for N×N visibility test N2N_SAMPLE_SIZE = 50 # Sample size for N×N visibility test