Optimizing Mutex Granularity

Refactor lock-heavy critical sections to increase parallelism and reduce wait times.