PERFORMANCE.md

Performance Optimizations

This document tracks performance optimizations and future improvement opportunities.

Implemented Optimizations (v0.x.x - 2025-10-31)

1. Cache extractMethodInfo Results

• Impact: 80-89% faster reflection operations
• Implementation: Global sync.Map cache for method info
• Benefit: Zero allocations for unbound method lookups

2. Cache parseSegments Results

• Impact: 10-17% faster URL generation
• Implementation: Per-context cache with RWMutex
• Benefit: Eliminated redundant pattern parsing

3. Pre-parse Route Segments

• Impact: 3% faster request handling with params
• Implementation: Store parsed segments in PageNode at Mount time
• Benefit: Zero parsing overhead in extractURLParams middleware

Benchmarks

Run benchmarks to verify performance:

# Full benchmark suite
go test -bench=. -benchmem

# Specific categories
go test -bench=BenchmarkRequestHandling -benchmem
go test -bench=BenchmarkURLGeneration -benchmem
go test -bench=BenchmarkIDGeneration -benchmem
go test -bench=BenchmarkReflection -benchmem

Future Optimization Opportunities

Medium Priority

1. Pool availableArgs Maps

   • Use sync.Pool for argument maps in method calling
   • Estimated improvement: 5-10% in request handling

2. Optimize kebabToPascal

   • Single-pass implementation without strings.Split
   • Estimated improvement: 50% in HTMX target matching

Low Priority

3. Cache Component Results

   • For pure components with no side effects
   • Requires careful API design

4. Reduce Context Value() Calls

   • Extract once at handler start
   • Pass as struct instead of multiple lookups

Performance Characteristics

Current performance (Apple M3 Ultra):

• Simple GET request: ~760ns
• Request with URL params: ~1050ns
• URL generation (1 param): ~226ns
• ID generation (unbound): ~167ns
• extractMethodInfo (cached): ~8ns

Memory per request:

• Simple GET: 1466B, 16 allocs
• With params: 2188B, 20 allocs