Hybrid Multithreaded Architecture with Symmetric Multiprocessors

Multithreading is effective to solve latency and synchronization problems which are fundamental issues in parallel processing. One of multithreaded computers, DAVRID has the computational model which is based on dataflow model together with advantages of von Neumann model. In this paper, we introduce some problems of DAVRID through performance evaluation and propose new architecture, MSA(Hybrid Multithreaded Symmetric Multiprocessor Architecture). DAVRID effectively solves latency and synchronization problem in the view of computational model, but has difficulty in exploiting locality in the view of architecture. MSA pursuits efficient parallel processing in the way it inherits the computational model of DAVRID and has symmetric multiprocessor architecture. Simulation result shows that the execution time of MSA is remarkably reduced compared with that of DAVRID, and also high variance of utilization among units is reduced and overall utilization is increased.