DocumentCode :
1667317
Title :
SPMD image processing on Beowulf clusters: directives and libraries
Author :
Oliveira, Paulo ; Du Buf, Hans
Author_Institution :
Vision Lab., Univ. do Algarve, Faro, Portugal
fYear :
2003
Abstract :
Most image processing algorithms can be parallelized by splitting parallel loops and by using very few communication patterns. Code parallelization using MPI still involves much programming overhead. In order to reduce these overheads, we first developed a small SPMD library (SPMDlib) on top of MPI. The programmer can use the library routines themselves, because they are easy to learn and to apply, even without knowing MPI. However, in order to increase user friendliness, we also develop a small set of parallelization and communication directives/pragmas (SPMDdir), together with a parser that converts these into library calls. SPMDdir is used to develop a new version of SPMDlib. This new version contains much less but generic routines that can be optimized for different network topologies. Extensions for Fortran 90/95 and C are discussed, as well as communication optimizations.
Keywords :
C language; FORTRAN; image processing; message passing; parallel algorithms; parallelising compilers; software libraries; workstation clusters; Beowulf clusters; C language; Fortran 90/95; MPI; SPMD image processing; SPMD library; SPMDdir; SPMDlib; code parallelization; communication directives/pragmas; communication optimizations; generic routines; library calls; library routines; network topologies; parallel algorithms; parser; user friendliness; Clustering algorithms; Computer science; Cost function; Distributed processing; Dynamic scheduling; Filter bank; Image processing; Libraries; Programming profession; Yarn;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Parallel and Distributed Processing Symposium, 2003. Proceedings. International
ISSN :
1530-2075
Print_ISBN :
0-7695-1926-1
Type :
conf
DOI :
10.1109/IPDPS.2003.1213419
Filename :
1213419
Link To Document :
بازگشت