PNMPI tools: a whole lot greater than the sum of their parts

ABSTRACT

P^NMPI extends the PMPI profiling interface to support multiple concurrent PMPI-based tools by enabling users to assemble tool stacks. We extend this basic concept to include new services for tool interoperability and to switch between tool stacks dynamically. This allows P^NMPI to support modules that virtualize MPI execution environments within an MPI job or that restrict the application of existing, unmodified tools to a dynamic subset of MPI calls or even call sites.

Further, we extend P^NMPI to platforms without dynamic linking, such as BlueGene/L, and we introduce an extended performance model along with experimental data from microbenchmarks to show that the performance overhead on any platform is negligible. More importantly, we provide significant new MPI tool components that are sufficient to compose interesting MPI tools. We present three detailed P^NMPI usage scenarios that demonstrate that it significantly simplifies the creation of application-specific tools.

REFERENCES

• 1: Accelerated Strategic Computing Initiative. The ASCI sweep3d benchmark code. http://www.llnl.gov/asci_benchmarks/asci/limited/sweep3d/, December 1995.
• 2: R. Bell, A. Malony, and S. Shende. ParaProf: A Portable, Extensible, and Scalable Tool for Parallel Performance Profile Analysis. In Proceedings of the International Conference on Parallel and Distributed Computing (Euro-Par 2003), pages 17--26, August 2003.
• 3: H. Brunst, D. Kranzlmüller, and W. Nagel. Tools for Scalable Parallel Program Analysis - Vampir NG and DeWiz. The International Series in Engineering and Computer Science, Distributed and Parallel Systems, 777:92--102, 2005.
• 4: Robert D. Falgout , Ulrike Meier Yang, hypre: A Library of High Performance Preconditioners, Proceedings of the International Conference on Computational Science-Part III, p.632-641, April 21-24, 2002
• 5: Francois Gygi , Erik W. Draeger , Martin Schulz , Bronis R. de Supinski , John A. Gunnels , Vernon Austel , James C. Sexton , Franz Franchetti , Stefan Kral , Christoph W. Ueberhuber , Juergen Lorenz, Large-scale electronic structure calculations of high-Z metals on the BlueGene/L platform, Proceedings of the 2006 ACM/IEEE conference on Supercomputing, November 11-17, 2006, Tampa, Florida [doi>10.1145/1188455.1188502]
• 6: B. Krammer, M. Müller, and M. Resch. Runtime Checking of MPI Applications with MARMOT. In Mini-Symposium on Tools Support for Parallel Programming at ParCo 2005, September 2005.
• 7: G. Kumfert, J. Leek, and T. Epperly. Babel remote method invocation. In Proceedings of the 21st International Parallel and Distributed Processing Symposium, March 2007.
• 8: T. Ludwig, R. Wismüller, V. Sunderam, and A. Bode. OMIS --- On-line Monitoring Interface Specifi cation (Version 2.0), volume 9 of LRR-TUM Research Report Series. Shaker Verlag, Aachen, Germany, 1997. ISBN 3-8265-3035-7.
• 9: J. May, D. Jefferson, N. Barton, R. Becker, J. Knap, G. Kumfert, J. Leek, and J. Tannahill. Introducing Cooperative Parallelism. Presented at the CCA Forum, presentation available at http://www.cca-forum.org/download/mtg/2007--01/may-coop-cca.ppt, January 2007.
• 10: National Center for Atmospheric Research (NCAR). Community Climate System Model (CCSM). http://www.ccsm.ucar.edu/, 2006.
• 11: Martin Schulz , Bronis R. de Supinski, A Flexible and Dynamic Infrastructure for MPI Tool Interoperability, Proceedings of the 2006 International Conference on Parallel Processing, p.193-202, August 14-18, 2006 [doi>10.1109/ICPP.2006.6]
• 12: J. S. Vetter and C. Chambreau. mpiP: Lightweight, Scalable MPI Profiling. http://www.llnl.gov/CASC/mpip/, April 2005.
• 13: Jeffrey S. Vetter , Bronis R. de Supinski, Dynamic software testing of MPI applications with umpire, Proceedings of the 2000 ACM/IEEE conference on Supercomputing (CDROM), p.51-es, November 04-10, 2000, Dallas, Texas, United States
• 14: Richard Vuduc , Martin Schulz , Dan Quinlan , Bronis de Supinski , Andreas Sæbjørnsen, Improving distributed memory applications testing by message perturbation, Proceedings of the 2006 workshop on Parallel and distributed systems: testing and debugging, July 17-17, 2006, Portland, Maine, USA [doi>10.1145/1147403.1147409]
• 15: R. Wismüller. Interoperable Laufzeit-Werkzeuge für parallele und verteilte Systeme. Inaugural dissertation (Habilitation), Fakultät für Informatik, Technische Universität München, München, Germany, August 2001.