Optimizing parallel applications for wide-area clusters
- 27 November 2002
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- No. 10637133,p. 784-790
- https://doi.org/10.1109/ipps.1998.670017
Abstract
Recent developments in networking technology cause a growing interest in connecting local area clusters of workstations over wide area links, creating multilevel clusters, or metacomputers. Often, latency and bandwidth of local area and wide area networks differ by two orders of magnitude or more. One would expect only very coarse grain applications to achieve good performance. To test this intuition, we analyze the behavior of several existing medium-grain applications on a wide-area multicluster. We find that high performance can be obtained if the programs are optimized to take the multilevel network structure into account. The optimizations reduce intercluster traffic and hide intercluster latency, and substantially improve performance on wide area multiclusters. As a result, the range of metacomputing applications is larger than previously assumed.Keywords
This publication has 9 references indexed in Scilit:
- Communication in GLOBE: an object-based worldwide operating systemPublished by Institute of Electrical and Electronics Engineers (IEEE) ,2002
- The SPLASH-2 programs: characterization and methodological considerationsPublished by Institute of Electrical and Electronics Engineers (IEEE) ,2002
- Globus: a Metacomputing Infrastructure ToolkitThe International Journal of Supercomputer Applications and High Performance Computing, 1997
- The Legion vision of a worldwide virtual computerCommunications of the ACM, 1997
- Performance of a High-Level Parallel Language on a High-Speed NetworkJournal of Parallel and Distributed Computing, 1997
- A worldwide flock of Condors: Load sharing among workstation clustersFuture Generation Computer Systems, 1996
- Myrinet: a gigabit-per-second local area networkIEEE Micro, 1995
- High performance messaging on workstationsPublished by Association for Computing Machinery (ACM) ,1995
- Directory-based cache coherence in large-scale multiprocessorsComputer, 1990