[BENCHMARKING] Scaling experiments: pattern count and text length

[hubertwojcik]

Goal

Run systematic scaling experiments varying pattern count and text size to characterize how each implementation scales.

Tasks

• Pattern count scaling: 10, 100, 500, 1000 patterns — all implementations, fixed 100 MB text
• Text length scaling: 10 MB, 100 MB, 500 MB, 1 GB — all implementations, fixed 100 patterns
• Thread/process count scaling for OpenMP and MPI: 1, 2, 4, 8, 10 cores
• Store all results in benchmarks/results/scaling_patterns.csv and benchmarks/results/scaling_text.csv

Acceptance Criteria

• All scaling experiments complete and results committed to benchmarks/results/
• No implementation crashes or produces incorrect results at any scale point