A design study of the EARTH multiprocessor

Multithreaded node architectures have been proposed for future multiprocessor systems. However, some open issues remain: can eecient multithreading support be provided in a multiprocessor machine such that it is capable of tolerating synchronization and communication latencies, with little intrusion on the performance of sequentially-executed code? Also, how much (quantitatively) does such non-intrusive multithreading support contribute to the scalable parallel performance in the presence of increasing interprocessor communication and synchronization demands? In this paper, we describe the design of EARTH (EE-cient Architecture for Running THreads), which attempts to address the above issues. Each processor in EARTH has an oo-the-shelf RISC processor for executing threads, and an ASIC Synchronization Unit (SU) supporting dataaow-like thread synchronizations, scheduling, and remote memory requests. In preparation for an implementation of the SU, we have emulated a basic EARTH model on MANNA 2.0, an existing multiprocessor whose hardware connguration closely matches EARTH. This EARTH-MANNA emulation testbed has been fully functional, enabling us to experiment with large-scale benchmarks with impressive speed. With this platform, we demonstrate that multithread-ing support can be eeciently implemented (with little em-ulation overhead) in a multiprocessor without a major impact on uniprocessor performance. Also, we give our rst quantitative indications of how much the basic multithread-ing support can help in tolerating increasing communica-tion/synchronization demands.