bradywatkinson / comp6771ass4 Goto Github PK

The project was to create the fastest possible implementation of a bucket sort using multiple threads. The sorting was done based on the first digit, than the value, i.e. 1000 comes before 100200 comes before 2. The amount of threads was given at runtime, thus the algorithm must be adaptable for one to potentially a lot of threads. Optimisations include binary searching for the first digit, rolling work scheduler with minimal locking and back insertion.