Abstract
High Performance Fortran (HPF) is the de facto standard language for writing data parallel programs. In case of applications that use indirect addressing on distributed arrays, HPF compilers have limited capabilities for optimizing such codes on distributed memory architectures, especially for optimizing communication and reusing communication schedules between subroutine boundaries.
This paper describes a dynamic approach for optimizing unstructured communication in codes with indirect addressing. The basic idea is that runtime data reflecting the communication patterns will be reused if possible. The user has only to specify which data in the program has to be traced for modifications. The experiments and results show the effectiveness of the chosen approach.
Chapter PDF
Keywords
- Runtime System
- Trace Attribute
- Communication Schedule
- High Performance Fortran
- Distribute Memory Architecture
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.
References
G. Agrawal and J. Saltz. Interprocedural communication optimizations for distributed memory compilation. In Language and Compilers for Parallel Computing, pages 1–16, Aug. 1994.
G. Agrawal, A. Sussman, and J. Saltz. Compiler and run-time support for structured and block-structured applications. In Supercomputing, pages 578–587. IEEE, 1993.
S. Benkner and H. Zima. Definition of HPF+ Rel.2. Technical report, HPF+ Consortium, 1997.
T. Brandes and F. Zimmermann. ADAPTOR — A Transformation Tool for HPF Programs. In K. Decker and R. Rehmann, editors, Programming Environments for Massively Parallel Distributed Systems, pages 91–96. Birkhäuser Verlag, Apr. 1994.
T. Brandes, F. Zimmermann, C. Borel, and M. Brédif. Evaluation of High Performance Fortran for an Industrial Computational Fluid Dynamics Code. In Proceedings of VECPAR 98, Porto, Portugal, June 1998. Accepted for publication.
R. Das et al. Communication optimization for irregular scientific computations on distributed memory architecturess. Journal of Parallel and Distributed Computing, (22):462–478, 1994.
C. Germain, J.Laminie, M. Pallud, and D. Etiemble. An HPF Case Study of a Domain-Decomposition Based Irregular Application. In PACT’97. LNCS, Sep 1997.
M. Gupta, E. Schonberg, and H. Srinivasan. A unified data-flow framework for optimizing communication in data-parallel programs. IEEE Trans. on Parallel and Distributed Systems, 7(7):689–704, 1996.
High Performance Fortran Forum. High Performance Fortran Language Specification. Rice Univ., Nov. 1994. Version 1.1.
High Performance Fortran Forum. High Performance Fortran Language Specification, Oct. 1997. Version 2.0.
S. Hiranandani, K. Kennedy, and C. Tseng. Compiling Fortran D for MIMD Distributed-Memory machines. CACM, 35(8):66–80, Aug. 1992.
S. Hiranandani, J. Saltz, P. Mehrotra, and H. Berryman. Performance of hashed cache migration schemes on multicomputers. Journal of Parallel and Distributed Computing, 12:415–422, 1991.
C. Koelbel, P. Mehrotra, and J. V. Rosendale. Supporting shared data structures on distributed memory architectures. In 2nd. Symp. on Principles and Practice of Parallel Programming, pages 177–186. ACM, 1990.
R. Ponnusamy, J. Saltz, and A. Choudhary. Runtime compilation techniques for data partitioning and communication schedule reuse. In ACM Int. Conf. on Supercomputing, pages 361–370, 1993.
Portland Group, Inc. PGHPF User’s Guide. Manual Release 2.2, PGI, 1997.
R. Mirchandaney et al. Principles of run-time support for parallel processing. In ACM Int. Conf. on Supercomputing, pages 140–152, 1988.
S. D. Sharma and al. Run-time and Compile-time Support for Adaptive Irregular Problems. In Supercomputing’94, pages 99–106. IEEE, 1994.
H. Zima and B. Chapman. Compiling for distributed-memory systems. Proceedings of the IEEE, Feb. 1993.
Author information
Authors and Affiliations
Editor information
Rights and permissions
Copyright information
© 1998 Springer-Verlag Berlin Heidelberg
About this paper
Cite this paper
Brandes, T., Germain, C. (1998). A tracing protocol for optimizing data parallel irregular computations. In: Pritchard, D., Reeve, J. (eds) Euro-Par’98 Parallel Processing. Euro-Par 1998. Lecture Notes in Computer Science, vol 1470. Springer, Berlin, Heidelberg. https://doi.org/10.1007/BFb0057910
Download citation
DOI: https://doi.org/10.1007/BFb0057910
Published:
Publisher Name: Springer, Berlin, Heidelberg
Print ISBN: 978-3-540-64952-6
Online ISBN: 978-3-540-49920-6
eBook Packages: Springer Book Archive