Abstract | ||
---|---|---|
Highly scalable parallel computers, e.g. SCI-coupled workstation clusters, are NUMA architectures. Thus good static locality is essential for high performance and scalability of parallel programs on these machines. This paper describes novel techniques to optimize static locality at compilation time by application of data transformations and data distributions. The metric which guides the optimizations employs Ehrhart polynomials and allows to calculate the amount of static locality precisely. The effectiveness of our novel techniques has been confirmed by experiments conducted on the SCI-coupled workstation cluster of the PC2 at the University of Paderborn. |
Year | DOI | Venue |
---|---|---|
2000 | 10.1007/3-540-44520-X_53 | Euro-Par |
Keywords | Field | DocType |
novel technique,static locality,scalable parallel computer,volume driven data distribution,sci-coupled workstation cluster,numa architecture,ehrhart polynomial,good static locality,data transformation,data distribution,parallel program,parallel computer | Data processing,Locality,Data transformation (statistics),Polynomial,Computer science,Parallel computing,Workstation,Convex polytope,Multigrid method,Distributed computing,Scalability | Conference |
Volume | ISSN | ISBN |
1900 | 0302-9743 | 3-540-67956-1 |
Citations | PageRank | References |
4 | 0.47 | 7 |
Authors | ||
2 |
Name | Order | Citations | PageRank |
---|---|---|---|
Felix Heine | 1 | 18 | 5.24 |
Adrian Slowik | 2 | 39 | 4.03 |