D. Arapattu
D. Gannon
ABSTRACT
In this paper we describe an interactive tool designed for performance prediction of parallel programs. Static performance prediction, in general, is a very difficult task. In order to avoid some inherent problems, we concentrate on reasonably structured scientific programs. Our prediction system, which is built as a sub-system of a larger interactive environment, uses a parser, dependence analyzer, database and an X-window based front end in analyzing programs. The system provides the user with execution times of different sections of programs. When there are unknowns involved, such as number of processors or unknown loop bounds, the output is an algebraic expression in terms of these variables. We propose a simple analytical model as an attempt to predict performance degradation due to data references in hierarchical memory systems. The predicted execution times of some Lawrence Livermore loop kernels are given together with the experimental values obtained by executing the loops on Alliant FX/8.
- 1.BBN ADVANCED COMPUTERS INC. Mach 100 Software Tools, May 1989.Google Scholar
- 2.BHUYAN, L., YANG, Q., AND ARGAWAL, D. Performance of multiprocessor interconnection networks. IEEE Computer (Feb. 1988), 25-37. Google ScholarDigital Library
- 3.COFFMAN, E., AND DENNING, P. Operating Systems Theory. Prentice-Hall, 1973. Google ScholarDigital Library
- 4.GALLIVAN, K., GANNON, D., JALBY, W., MAL- ONY, A., AND WIJSHOFF, H. Behavioral characterization of multiprocessor memory systems: A case study. In Proceedings of ACM SIGMETRICS Conference on Measuring and Modeling Computer Systems (Berkeley, CA, May 1989), pp. 79-88. Google ScholarDigital Library
- 5.GALLIVAN, K., JALBY, W., AND GANNON, D. On the problem of optimizing data transfers for complex memory systems. In Proceedings of International Conference on Supercomputing (St. Malo, France, July 1988), pp. 238-253. Google ScholarDigital Library
- 6.GALLIVAN, K., JALBY, IV., MALONY, A., AND WIJSHOFF, H. Performanc:e prediction of loop constructs on multiprocessor hierarchical-memory systems. In Proceedings of Third International Conference on Supercomputing (Crete, Greece, June 1989), pp. 433-442. Google ScholarDigital Library
- 7.GANNON, D., ATAPATTU? D., LEE, M.-H., AND SHEI, B. A software tool for building supercomputer applications. In Parallel Computations and Their impact on Mechanics, A. K. Noor, Ed. The American Society of Mechanical Engineers, New York, 1987.Google Scholar
- 8.GANNON, D., GUARNA, V., GAUR, Y., AND JABLONOWSKI, J. A software tool for programming parallel systems. In Proceedings of Supercomputing 88 (Orlando, Florida, Nov. 1988), IEEE, ACM SIGARCH, pp. 3-11.Google Scholar
- 9.KRUSKAL, C., AND SNIR, M. The performance of multistage interconnection networks for multiprocessors. IEEE Transactions on Computers C-32 (Dec. 1983), 1091-1098.Google ScholarDigital Library
- 10.MCMAHON, F. The livermore fortran kernels: A computer test of the numerical performance range. Tech. Rep. UCRL-53745, Lawrence Livermore National Laboratory., Dec. 1986.Google Scholar
- 11.PHISTER, G., AND NORTON, A. Hot spot contention and combining in multistage interconnection networks. In Proceedings of International Conference on parallel Proces.sing (1985), pp. 790-797.Google Scholar
- 12.SETHI, A., AND DEO, N. Interference in multiprocessor systems with localized memory access probabilities. IEEE Transactions on Computers C-28, 2 (Feb. 1979).Google Scholar
- 13.YEN, D., PATEL, J., AIVD DAVIDSON, E. Memory interference in synchronous multiprocessor systems. IEEE Transactions on Computers C-31 (Nov. 1982), 1116-1121.Google ScholarDigital Library
Index Terms
- Building analytical models into an interactive performance prediction tool
Recommendations
Performance Prediction of PVM Programs
IPPS '96: Proceedings of the 10th International Parallel Processing SymposiumAs workstation clusters gain popularity as a parallel computing platform, there is an increasing need for performance tools that support these platforms. Performance prediction is important for performance analysis of scalable parallel applications. ...
PAS2P tool, parallel application signature for performance prediction
PARA'10: Proceedings of the 10th international conference on Applied Parallel and Scientific Computing - Volume Part IAccurate prediction of parallel applications' performance is becoming increasingly complex. We seek to characterize the behavior of message-passing applications by extracting a signature to predict the performance in different target systems. We have ...
Comments