Modelling of ASCI High Performance Applications Using PACE

There is a wide range of models being developed for the performance evaluation of parallel and distributed systems. This has become an important area of research especially with the development of dynamic processing capabilities promised with Computational GRIDs [3]. A performance modelling approach described in this paper is based on a layered framework of the PACE methodology. In this system, the model described by a Performance Specification Language (PSL) provides the capability for rapid calculation of relevant performance information without sacrificing accuracy of predictions. An example of the performance evaluation of an ASCI kernel application, Sweep3D, is used to illustrate the approach. The validation of the model is shown for a crossplatform analysis on two parallel and distributed architectures with different problem sizes. Results show that a reasonable accuracy (approximately 10% error at most) can be obtained with a rapid evaluation time (typically less than 2s).

[1]  Charles McMillan.  ASCI applications , 1997, Proceedings. The Sixth IEEE International Symposium on High Performance Distributed Computing (Cat. No.97TB100183).

[2]  J. Schopf,et al.  Structural Prediction Models for High-Performance Distributed Applications , 1997 .

[3]  Monica S. Lam,et al.  Maximizing Multiprocessor Performance with the SUIF Compiler , 1996, Digit. Tech. J..

[4]  Joel H. Saltz,et al.  A Performance Prediction Framework for Data Intensive Applications on Large Scale Parallel Machines , 1998, LCR.

[5]  Graham R. Nudd,et al.  Analytical Modeling of Set-Associative Cache Behavior , 1999, IEEE Trans. Computers.

[6]  Connie U. Smith,et al.  Performance Engineering of Software Systems , 1990, SIGMETRICS Perform. Evaluation Rev..

[7]  Sarita V. Adve,et al.  RSIM Reference Manual: Version 1.0 , 1997 .

[8]  Reda A. Ammar,et al.  Micro Time Cost Analysis of Parallel Computations , 1991, IEEE Trans. Computers.

[9]  Mary K. Vernon,et al.  LoPC: modeling contention in parallel algorithms , 1997, PPOPP '97.

[10]  Graham R. Nudd,et al.  Pace—A Toolset for the Performance Prediction of Parallel and Distributed Systems , 2000, Int. J. High Perform. Comput. Appl..

[11]  Mary K. Vernon,et al.  Poems: end-to-end performance design of large parallel adaptive computational systems , 1998, WOSP '98.

[12]  Mineo Takai,et al.  Parssec: A Parallel Simulation Environment for Complex Systems , 1998, Computer.

[13]  Ami Marowka,et al.  The GRID: Blueprint for a New Computing Infrastructure , 2000, Parallel Distributed Comput. Pract..

[14]  Graham R. Nudd,et al.  Application Execution Steering using On-the-Fly Performance Prediction , 1998, HPCN Europe.