Sequential vs Parallel Parsing
Throughput comparison of sequential and parallel (Rayon) batch parsing across different dataset sizes. The crossover point shows where parallelism starts to pay off despite thread synchronization overhead.
Scaling: Linear vs Branched Chains
Parse time comparison between PEG (linear chain:
OCCOCCOCC…O) and Teflon/PTFE (branched:
C(F)(F)C(F)(F)…) at varying chain lengths.
Helps detect O(n²) algorithmic regressions and measure
branch parsing overhead.
Hückel Validation Overhead
Measures the cost of aromaticity validation (ring detection + π-electron
counting + 4n+2 check) on top of parsing. The huckel-validation
feature flag adds this check to parse() automatically.
Molecules without aromatic rings should show near-zero overhead.
Reference Molecules
Parse time for 5 representative molecules of increasing complexity, useful for cross-language parser comparison (RDKit, OpenBabel, etc.).