Yet another " fastest hashing algorithm"
Intrinhash is a new hashing library that claims to be the fastest for long inputs on x64 and ARM architectures. It provides 256-bit collision-free hash values and is optimized for high speed, although it is not suitable for small keys. The library has passed various benchmark tests and is currently in its prototype state, inviting user feedback for improvements.
- ▪Intrinhash is the fastest hashing library for C/C++ on x64/ARM for long inputs as of 2026.
- ▪It is designed for 256-bit collision-free hash values and utilizes AES-NI/VAES256 on x64 and NEON on ARMv8.
- ▪The library has passed multiple tests including xxHash collisionsTest and SMHasher for various bit lengths.
Opening excerpt (first ~120 words) tap to expand
intrinhash is currently (2026) the fastest quality hashing library for C/C++ on x64/ARM for long inputs. It's designed for use cases where you need 256-bit collision free hash values at high speed. It is not cryptographic and not competitive for small keys. It uses AES-NI/VAES256 on x64 and NEON on ARMv8 which covers basically every PC and smartphone from 2013 or later. Only a very slow fallback is supported for other architectures. Benchmark and quality intrinhash passes the xxHash collisionsTest (added here). It also passes SMHasher and the much harder SMHasher3 (added here) for 32, 64, 128 and 256 bit lengths, even with --extra. Below is its benchmark on a Core i7 for alignment 0.
…
Excerpt limited to ~120 words for fair-use compliance. The full article is at GitHub.