DocumentCode
451249
Title
SCALLOP: A Highly Scalable Parallel Poisson Solver in Three Dimensions
Author
Balls, Gregory T. ; Baden, Scott B. ; Colella, Phillip
Author_Institution
University of California, San Diego
fYear
2003
fDate
15-21 Nov. 2003
Firstpage
23
Lastpage
23
Abstract
SCALLOP is a highly scalable solver and library for elliptic partial differential equations on regular block-structured domains. SCALLOP avoids high communication overheads algorithmically by taking advantage of the locality properties inherent to solutions to elliptic PDEs. Communication costs are small, on the order of a few percent of the total running time on up to 1024 processors of NPACI´s and NERSC´s IBM Power-3 SP sytems. SCALLOP trades off numerical overheads against communication. These numerical overheads are independent of the number of processors for a wide range of problem sizes. SCALLOP is implicitly designed for infinite domain (free space) boundary conditions, but the algorithm can be reformulated to accommodate other boundary conditions. The SCALLOP library is built on top of the KeLP programming system and runs on a variety of platforms.
Keywords
algorithms; computation-intensive applications; parallel and distributed; program optimization and performance programming; Boundary conditions; Computer science; Costs; Drives; Laboratories; Libraries; Partial differential equations; Permission; Poisson equations; Scalability; algorithms; computation-intensive applications; parallel and distributed; program optimization and performance programming;
fLanguage
English
Publisher
ieee
Conference_Titel
Supercomputing, 2003 ACM/IEEE Conference
Print_ISBN
1-58113-695-1
Type
conf
DOI
10.1109/SC.2003.10047
Filename
1592926
Link To Document