Repository navigation
Release Summary
This project started with the ambition of building a better alternative to sync.Pool. Along the way, a number of advanced features were added to provide tighter control over object lifecycle and memory usage—features that inevitably introduced additional latency.
Despite this tradeoff, performance has improved significantly, narrowing the original 60% performance gap with sync.Pool down to just 23.58% (which is still significant).
While this implementation isn't intended as a drop-in replacement for sync.Pool, it serves a different purpose: fine-grained control over object allocation in high-concurrency systems. Use it only if your use case truly demands such control—especially when reducing memory churn is more important than achieving the absolute lowest latency.