January 13, 2026

Benchmarks or it didn’t cache

Show HN: Cachekit – High performance caching policies library in Rust

Rust’s CacheKit drops; commenters roast the demo and demand real benchmarks

TLDR: CacheKit is a new Rust tool for faster app caching, but commenters pounced on a confusing code example and demanded proper benchmark comparisons. The community wants clear naming, sane defaults, and hard numbers before buying the “high performance” pitch — because proof beats promises.

CacheKit, a new Rust library promising turbo-charged cache speed, hit Show HN and instantly turned into courtroom drama. It’s a toolkit for smarter “caching” — storing stuff so apps can grab it faster — with layered strategies and optional metrics. But the crowd didn’t clap; they cross-examined. The headline controversy? A sample snippet says “capacity of 100” while the code clearly starts the cache with “2.” Cue the chorus: “Why 100? Why not 2?” One commenter wondered if there’s a secret default. Others just called it a red flag, and the jokes wrote themselves: LRU? More like LOL-RU.

Then came the naming police. A sharp-eyed dev demanded conformity with official Rust style, insisting “LrukCache, not LRUKCache,” citing the guidelines. The bikeshedding alarm went off, but hey — it’s the internet, and casing debates are a sport.

Finally, the performance hype got grilled. Another voice asked for head-to-head numbers against rivals like Foyer and quick-cache. The vibe was pure benchmarks or it didn’t happen: people want “sane defaults,” easy memory limits, and receipts. CacheKit’s docs talk a big game about metrics and real-world tests, but until the charts drop, the community’s energy is: cool project, now prove it — with fewer confusing comments and more data.

Key Points

  • CacheKit is a Rust library providing high-performance cache replacement policies including FIFO, LRU, and LRU-K.
  • It offers tiered caching primitives and backends for in-memory and composite cache strategies.
  • Optional integration with metrics collectors like Prometheus and the Rust metrics crate is available.
  • Benchmarks are provided to compare policy performance under real-world workloads.
  • Installation is via Cargo.toml using the OxidizeLabs GitHub repository, with example LRUKCache usage shown.

Hottest takes

"Why 100? Why not 2?" — carlmr
"Should be checked against official Rust API naming guidelines" — virtualritz
"Would love to see benchmarks vs competition" — fyrn_
Made with <3 by @siedrix and @shesho from CDMX. Powered by Forge&Hive.