mev-beta/.qwen/config/optimization.md

Qwen Code Performance Optimization Settings

Workflow Preferences

• Always commit changes: Use git commit -am "math: descriptive message" for mathematical implementations
• Branch naming: Use prefixes (math-sqrt-price, algo-liquidity-calc, perf-uniswap)
• Context management: Focus on mathematical precision and performance
• Parallel processing: Leverage Go's concurrency patterns for independent calculations

File Organization Preferences

• Mathematical functions: Place in pkg/uniswap/ or pkg/math/
• Test files: Place alongside source files with _test.go suffix
• Documentation: Inline comments explaining mathematical formulas
• Precision libraries: Use github.com/holiman/uint256 for uint256 arithmetic

Performance Monitoring

# Enable metrics endpoint for performance tracking
export METRICS_ENABLED="true"
export METRICS_PORT="9090"

# Monitor memory usage of mathematical calculations
go tool pprof http://localhost:9090/debug/pprof/heap

# Monitor CPU usage of mathematical functions
go tool pprof http://localhost:9090/debug/pprof/profile?seconds=30

# Run benchmarks for mathematical functions
go test -bench=. -benchmem ./pkg/uniswap/...

Precision Requirements

• Uint256 Arithmetic: Use github.com/holiman/uint256 for all uint256 calculations
• Floating Point: Use math/big for floating-point calculations when needed
• Rounding: Implement proper rounding strategies for financial calculations
• Overflow Handling: Handle overflow and underflow conditions properly

Optimization Focus Areas

1. Mathematical Computation Efficiency

   • Minimize computational overhead in pricing functions
   • Optimize sqrtPriceX96 to price conversions
   • Efficient tick calculations

2. Memory Allocation Reduction

   • Object pooling for frequently created mathematical objects
   • Pre-allocation of slices and buffers
   • Minimize garbage collection pressure

3. Algorithmic Optimization

   • Mathematical formula simplification
   • Lookup table implementation for repeated calculations
   • Caching strategies for expensive computations