A Massively Parallel Multithreaded Architecture: DAVRID

MPAs(Massively Parallel Architectures) should address two fundamental issues for scalability: synchronization and communication latency. Dataaow ar-chitectures cause problems of excessive synchronization costs and ineecient execution of sequential programs while they ooer the ability to exploit massive parallelism inherent in programs. In contrast, MPAs based on von Neumann computational model may suffer from ineecient synchronization mechanism and communication latencies. DAVRID(DAtaaow Von Neumann RISC hybrID) is a massively parallel mul-tithreaded architecture which, by combining advantages of von Neumann model and dataaow model, preserves good single thread performance as well as tolerates latency and synchronization costs. In this paper, we describe the DAVRID architecture and evaluate it through simulation results over several benchmarks.