Show HN: QuadSort, Esoteric Fast Sort

Uses the wolfsort benchmark, which uses a time-based seed for the pseudo-random-number generator &quot;rand&quot;. These are therefore unreproducible results (for the randomized inputs).<p>The results are very interesting and generally follow the rule that bigger hand-optimized code can be faster code.

TLDR: A visualisation of how QuadSort adaptively combines various novel strategies to achieve the best performance on a wide variety of inputs: <a href="https:&#x2F;&#x2F;github.com&#x2F;scandum&#x2F;quadsort#visualization">https:&#x2F;&#x2F;github.com&#x2F;scandum&#x2F;quadsort#visualization</a><p>Benchmarks of how it beats qsort, Timsort, pdqsort etc.: <a href="https:&#x2F;&#x2F;github.com&#x2F;scandum&#x2F;quadsort#benchmark-quadsort-vs-stdstable_sort-vs-timsort">https:&#x2F;&#x2F;github.com&#x2F;scandum&#x2F;quadsort#benchmark-quadsort-vs-st...</a><p>Is it called esoteric just because it&#x27;s complex?

looks awesome. making this the default sort would surely save lots of co_2.

Are there any other examples of such hand crafted algorithms that perform better than theoretically best algorithms?

&gt; WSL 2 gcc version 7.5.0 (Ubuntu 7.5.0-3ubuntu1~18.04)<p>Why not run the benchmarks in an environment that matters? Latest stable gcc is 12 and I wouldn&#x27;t think anybody runs production code on WSL.