BitFaster.Caching

Designed for lock-free concurrency with high throughput and predictable latency, prioritizing scalability in concurrent environments, as highlighted in its philosophy.

Implements state-of-the-art policies like 2Q for LRU and W-TinyLFU for LFU, providing near-optimal hit rates that outperform traditional LRU caches, per the wiki documentation.

ConcurrentLru and ConcurrentLfu can be used just like ConcurrentDictionary with bounded size, making integration straightforward, as shown in the getting started examples.

Configurable atomic valueFactory ensures only one thread computes a value for a missing key, mitigating cache stampede issues, a key feature documented in the wiki.

Provides wrappers for IDisposable cache values with scoped lifetime management, enabling safe caching of resources that require cleanup, as detailed in the wiki.