Performance evaluation of multithreading in concurrent programs

Concurrent programming may allow the programmer to create fast responsive systems without the cost overhead of full parallelism. One method of achieving concurrency is to use multiple threads of execution in a single address space. By declaring many threads within the confines of a single process, a programmer can achieve potential parallelism at low overhead. Can concurrent implementation always outperform sequential implementation on a uniprocessor? This project involves both sequentially and concurrently implementing a sort algorithm, called Bucket Sort. The core research aims to investigate the performance factors by several variations of the number of threads and problem size. Experimental results are presented on both implementations, and the design considerations of concurrent programming are also studied.