Blitsort: A fast, in-place stable hybrid merge/quick sort

On mobile at the moment but it will be interesting to see how this compares to Glidesort [1]. Though I don’t <i>think</i> it’s been released yet.<p>[1] <a href="https:&#x2F;&#x2F;m.youtube.com&#x2F;watch?v=2y3IK1l6PI4" rel="nofollow">https:&#x2F;&#x2F;m.youtube.com&#x2F;watch?v=2y3IK1l6PI4</a>

Dumb questions from a programmer who is weak at math.<p>1) Is there a theoretical minimum assymptotic speed for sorting?<p>2) Some of these newer sorts seem to have arbitrary parameters in them (e.g. do X if size of set is smaller than Y). Are we likely to see more and more complex rule sets like that lead to increased sorting speeds?

thanks for sharing this with us. i admire this approach that builds on small improvements here and there. and these improvements are interesting in their own way. it reminds me of micro-optimizations with encoding routines. i didn&#x27;t SIMD in the code at first glance. you might gain significant speed with SIMD.

Woah, I like the solid memory guarantees. Could be useful for embedded projects.

Since the sorting scene seems to be fully assembled in this thread, I take the opportunity to ask quick question about a use case that does not seem uncommon: Is there an algorithm that is especially efficient if the array contains many sorted sequences, and still works alright for fully shuffled arrays? And are there algorithms that should be avoided in these cases.

Question: are any of these novel sorting algorithms being used in modern databases or tech stacks?

Is there a TrollSort? I&#x27;m thinking of an algorithm which initially seems to be fast and efficient but takes exponentially longer time with larger arrays and exponentially longer towards the end of sorting.

Faster than radix sort?

Heads up: No license given in the repo, be careful if you are thinking of using this for a project.<p>EDIT: Retracted -- did a search like for &quot;license&quot;, but apparently the search results omits variants like &quot;sublicense&quot; which would have caught the MIT license at the beginning of the source file: <a href="https:&#x2F;&#x2F;github.com&#x2F;scandum&#x2F;blitsort&#x2F;search?q=license" rel="nofollow">https:&#x2F;&#x2F;github.com&#x2F;scandum&#x2F;blitsort&#x2F;search?q=license</a> vs. <a href="https:&#x2F;&#x2F;github.com&#x2F;scandum&#x2F;blitsort&#x2F;search?q=sublicense" rel="nofollow">https:&#x2F;&#x2F;github.com&#x2F;scandum&#x2F;blitsort&#x2F;search?q=sublicense</a>