Processing irregular codes containing arrays with multi-dimensional distributions by the PREPARE HPF compiler
The automatic parallelization of irregular codes for massively parallel systems is a challenging problem of growing importance. This paper describes techniques for processing irregular codes that are being implemented in the Prepare HPF compiler. Our work makes two contributions: (1) it constructs a library called Parti+ that is used for the runtime analysis of codes with runtime data dependent accesses to arrays with multi-dimensional distributions, and (2) it develops the appropriate compiler support allowing optimizations.
KeywordsProcessor Time Processor Array Local Segment Runtime Analysis Work Distributor
Unable to display preview. Download preview PDF.
- 1.A. Veen, and M. de Lange. Overview of the PREPARE Project. 4th International Workshop on Compilers for Parallel Computers, Delft, The Netherlands, Dec. 1993.Google Scholar
- 2.P. Brezany, M. Gerndt, V. Sipkova, and H.P. Zima. SUPERB Support for Irregular Scientific Computations. In Proc. of the SHPCC '92, Williamsburg, USA.Google Scholar
- 3.C. Koelbel. Compiling Programs for Nonshared Memory Machines. Ph.D. Dissertation, Purdue University, West Lafayette, IN, Nov. 1990.Google Scholar
- 4.J. Saltz et al. Run-time scheduling and execution of loops on message passing machines. Journal of Parallel and Distributed Computing, 1990.Google Scholar
- 5.R. Das, and J. Saltz. A manual for PARTI runtime primitives — Revision 2. Internal Research Report, University of Maryland, Dec. 1992.Google Scholar
- 6.M. Le Fur, J-L. Pazat, and F. André. Commutative Loop Nests Distribution. 4th Int. Workshop on Compilers for Par. Comp., Delft, The Netherlands, Dec. 1993.Google Scholar
- 7.R. Ponnusamy et al. A manual for the CHAOS runtime library. Technical report, University of Maryland, May 1994.Google Scholar