BigCache

Uses a pointer-less map and byte slices to avoid garbage collection, as shown in benchmarks where GC pauses are up to 5x lower than standard maps with 20 million entries.

Employs configurable sharding to reduce lock contention, with benchmark results showing faster parallel reads and writes compared to freecache and concurrent maps.

Supports time-based eviction via LifeWindow and configurable CleanWindow intervals, allowing fine-tuned cache management without predefined size limits.

Does not require a fixed cache size upfront; can allocate additional memory when full, with an optional HardMaxCacheSize to prevent unbounded growth.