Summary: The Architectural Costs of Streaming I/O:
A Comparison of Workstations, Clusters, and SMPs
Remzi H. Arpaci­Dusseau, Andrea C. Arpaci­Dusseau,
David E. Culler, Joseph M. Hellerstein, and David A. Patterson
Computer Science Division
University of California, Berkeley
fremzi, dusseau, culler, jmh, pattersong@cs.berkeley.edu
Abstract
We investigate resource usage while performing streaming I/O by con­
trasting three architectures, a single workstation, a cluster, and an SMP,
under various I/O benchmarks. We derive analytical and empirically­
based models of resource usage during data transfer, examining the
I/O bus, memory bus, network, and processor of each system. By
investigating each resource in detail, we assess what comprises a well­
balanced system for these workloads.
We find that the architectures we study are not well balanced for
streaming I/O applications. Across the platforms, the main limitation
to attaining peak performance is the CPU, due to lack of data locality.
Increasing processorperformance (especially with improvedblock op­
eration performance) will be of great aid for these workloads in the

Source: Arpaci-Dusseau, Remzi - Department of Computer Sciences, Department of Computer Sciences, University of Wisconsin at Madison

Collections: Computer Technologies and Information Sciences