Can CPU tricks beat the best sorting algorithms in a specific niche?
Sorting is everywhere, and the best general-purpose algorithms are already very fast. This project looks for a niche (a data type, size, distribution or CPU) where SIMD, branchless partitioning and cache-aware tricks give a measurable win.
| success | A reproducible benchmark showing a speedup over the current state of the art in the chosen niche. |
| correctness | Mandatory. Every variant passes the full test suite. |
| deliverable | A library and a CLI anyone can use, plus a plain summary. |