Prophesy: automating the modeling process

Performance models provide significant insight into the performance relationships between an application and the system, either parallel or distributed, used for execution. The development of models often requires significant time, sometimes in the range of months, to develop; this is especially the case for detailed models. This paper presents our approach to reducing the time required for model development. We present the concept of an automated model builder within the Prophesy infrastructure, which also includes automated instrumentation and extensive databases for archiving the performance data. In particular, we focus on the automation of the development of analytical performance models. The concepts include the automation of some well-established techniques, such as curve fitting, and a new technique that develops models as a composition of other models of core components or kernels in the application.

[1]  Csaba Andras Moritz,et al.  LoGPC: Modeling Network Contention in Message-Passing Programs , 2001, IEEE Trans. Parallel Distributed Syst..

[2]  William E. Johnston,et al.  Grids as production computing environments: the engineering aspects of NASA's Information Power Grid , 1999, Proceedings. The Eighth International Symposium on High Performance Distributed Computing (Cat. No.99TH8469).

[3]  B. Miller,et al.  The Paradyn Parallel Performance Measurement Tools , 1995 .

[4]  Michael W. Berry,et al.  Public international benchmarks for parallel computers: PARKBENCH committee: Report-1 , 1994 .

[5]  David H. Bailey,et al.  The Nas Parallel Benchmarks , 1991, Int. J. High Perform. Comput. Appl..

[6]  Thomas Fahringer.  Toward symbolic performance prediction of parallel programs , 1996, Proceedings of International Conference on Parallel Processing.

[7]  Anthony Skjellum,et al.  Using MPI - portable parallel programming with the message-parsing interface , 1994 .

[8]  Satish K. Tripathi,et al.  On Performance Prediction of Parallel Computations with Precedent Constraints , 2000, IEEE Trans. Parallel Distributed Syst..

[9]  David G. Messerschmitt,et al.  SPAR: A New Architecture for Large Finite Element Computations , 1995, IEEE Trans. Computers.

[10]  Ramesh Subramonian,et al.  LogP: towards a realistic model of parallel computation , 1993, PPOPP '93.

[11]  Barton P. Miller,et al.  The Paradyn Parallel Performance Measurement Tool , 1995, Computer.

[12]  Xin Li,et al.  Design and Development of the Prophesy Performance Database for Distributed Scientific Applications , 2001, PPSC.

[13]  Pankaj Mehra,et al.  Performance measurement, visualization and modeling of parallel and distributed programs using the AIMS toolkit , 1995, Softw. Pract. Exp..

[14]  Alan Jay Smith,et al.  Machine Characterization Based on an Abstract High-Level Language Machine , 1989, IEEE Trans. Computers.

[15]  Report,et al.  Public International Benchmarks for Parallel Computers , 1993 .

[16]  Brian H. Larose.  The development and implementation of a performance database server , 1993, CS / Technical report / Knoxville / University of Tennessee / Computer Science Department.

[17]  David R. Kincaid,et al.  Algorithm 586: ITPACK 2C: A FORTRAN Package for Solving Large Sparse Linear Systems by Adaptive Accelerated Iterative Methods , 1982, TOMS.

[18]  Ian T. Foster,et al.  Resource co-allocation in computational grids , 1999, Proceedings. The Eighth International Symposium on High Performance Distributed Computing (Cat. No.99TH8469).

[19]  Warren Smith,et al.  A Resource Management Architecture for Metacomputing Systems , 1998, JSSPP.

[20]  D.A. Reed,et al.  Scalable performance analysis: the Pablo performance analysis environment , 1993, Proceedings of Scalable Parallel Libraries Conference.

[21]  Richard T. Snodgrass,et al.  A relational approach to monitoring complex systems , 1988, TOCS.

[22]  V. Taylor,et al.  DESIGN AND IMPLEMENTATION OF PROPHESY AUTOMATIC INSTRUMENTATION AND DATA ENTRY SYSTEM , 2001 .

[23]  Gheith A. Abandah,et al.  Characterizing Distributed Shared Memory Performance: A Case Study of the Convex SPP1000 , 1998, IEEE Trans. Parallel Distributed Syst..

[24]  Valerie E. Taylor,et al.  Performance Coupling: A Methodology for Predicting Application Performance Using Kernel Performance , 1999, PPSC.

[25]  Alan Jay Smith,et al.  Machine Characterization BASed on an Abstract High Level Machine , 1989 .

[26]  Dennis Gannon,et al.  SIEVE: A Performance Debugging Environment for Parallel Programs , 1993, J. Parallel Distributed Comput..

[27]  Jack J. Dongarra,et al.  A Scalable Cross-Platform Infrastructure for Application Performance Tuning Using Hardware Counters , 2000, ACM/IEEE SC 2000 Conference (SC'00).

[28]  Graham R. Nudd,et al.  PACE: A Toolset to Investigate and Predict Performance in Parallel Systems , 1996 .

[29]  Stephen F. Lundstrom,et al.  Predicting Performance of Parallel Computations , 1990, IEEE Trans. Parallel Distributed Syst..

[30]  Francine Berman,et al.  Application-Level Scheduling on Distributed Heterogeneous Networks , 1996, Proceedings of the 1996 ACM/IEEE Conference on Supercomputing.