Parallel Programming in functional F#

Author: Tengiz Merabishvili
Keywords: parallel programming, F#, functional
Annotation:

Our computers keep getting faster each year, and processors keep getting more cores. It is very important to keep our software up to date with our hardware, because otherwise progress will be slowed. Parallel algorithms are a very interesting topic, and that’s why I decided to do my research in that field. Sorting algorithms have a lot of every day usages, so I decided to research parallel and functional programming, and implement a sorting algorithm. The purpose of my research was to see how hard it is to write parallel code, how polished is F# as a language, and what benefits would I receive running my code on several processors. The results were pleasing – F#, being a pretty young language, has a lot of support and documentation. Running my code on 8 cores was more than two times faster than 4 cores, and more than 30 times faster than a single core for 100 000 elements. I shared my code on github so anyone who is interested can check it out and continue working on improving that algorithm.