feat: add performance benchmarks

[kholdrex]

This PR introduces a professional-grade performance benchmarking suite for IRT Ruby, providing users with detailed insights into computational characteristics, scaling behavior, and optimization performance across all three IRT models (Rasch, 2PL, 3PL).

🎯 Key Features Added

1. Performance Benchmark Suite (benchmarks/performance_benchmark.rb)

• Execution Speed Analysis: Measures iterations per second across 5 dataset sizes (50 to 100,000 data points)
• Memory Profiling: Tracks object allocation and memory usage patterns
• Scaling Analysis: Determines computational complexity (O(n^x)) for different dataset sizes
• Missing Data Impact: Compares performance across different missing data strategies

2. Convergence Analysis Suite (benchmarks/convergence_benchmark.rb)

• Tolerance Optimization: Tests convergence behavior with different precision settings
• Learning Rate Analysis: Identifies optimal learning rates for speed vs. reliability
• Dataset Characteristics: Measures convergence under various data conditions
• Missing Data Patterns: Analyzes how missing data affects convergence