PNNL

Abstract

We have conducted a detailed study to understand the potential of multithreaded architectures to increase the performance of data-intensive, irregular scientific applications. Our study is based on a power system state estimation application and an anomaly detection application applied to network traffic data. We also conducted a detailed evaluation of the platforms using microbenchmarks in order to gain insight into their architectural capabilities and their interaction with programming models and application software. The evaluation was performed on the Cray MTA-2 and the Sun Niagara. Our results show that irregular applications perform well on these two diverse multithreaded platforms in cases where the bandwidth provided by the underlying platform can be exploited successfully.