LovelyPanda / MatchMakingTest

Test for one of game development companies I applied to. The assignment was to improve performance of existing code performing matchmaking (matching players with the most similar parameters posible). My solution was to improve synchronization and to optimize algorithm using SSE. The result was roughly 4000 times faster than the original. There is a better solution though: to redesign and rewrite the algorythm itself. I tried using KD tree for partitioning of the space but then realised that for high number of dimensions it won't be effective.