ABSTRACT
Communication analysis of parallel applications based on event traces depends on correct matching of associated MPI send and receive events. Selective monitoring techniques, however, may result in incomplete MPI event traces and, in that case, current matching strategies fail. In this paper we introduce an additional unique identifier for each message to make MPI events distinguishable from others. Therefore, it is possible to identify missing MPI events and match all remaining MPI events correctly. An overhead study with a real-life application and a benchmark suite demonstrates the applicability and benefits of this approach.
- CRESTA. Collaborative Research into Exascale Systemware, Tools and Applications. http://cresta-project.eu.Google Scholar
- D. Eschweiler, M. Wagner, M. Geimer, A. Knüpfer, W. E. Nagel, and F. Wolf. Open Trace Format 2: The Next Generation of Scalable Trace Formats and Support Libraries. In Applications, Tools and Techniques on the Road to Exascale Computing, volume 22 of Advances in Parallel Computing, pages 481--490, 2012.Google Scholar
- M. Geimer, F. Wolf, B. J. Wylie, E. Ábrahám, D. Becker, and B. Mohr. The Scalasca Performance Toolset Architecture. Concurrency and Computation: Practice and Experience, 22(6):702--719, 2010. Google ScholarDigital Library
- B. Hess, C. Kutzner, D. van der Spoel, and E. Lindahl. GROMACS 4: Algorithms for Highly Efficient, Load-Balanced, and Scalable Molecular Simulation. Journal of Chemical Theory and Computation, 4(3):435--447, 2008.Google ScholarCross Ref
- T. Hilbrich, M. S. Müller, B. R. de Supinski, M. Schulz, and W. E. Nagel. GTI: A Generic Tools Infrastructure for Event-Based Tools in Parallel Systems. In Proceedings of the 2012 IEEE 26th International Parallel and Distributed Processing Symposium, IPDPS '12, pages 1364--1375, Washington, DC, USA, 2012. IEEE Computer Society. Google ScholarDigital Library
- T. Hilbrich, J. Protze, M. Schulz, B. R. de Supinski, and M. S. Müller. MPI Runtime Error Detection with MUST: Advances in Deadlock Detection. In Proceedings of the International Conference on High Performance Computing, Networking, Storage and Analysis, SC '12, pages 30:1--30:11, Los Alamitos, CA, USA, 2012. IEEE Computer Society. Google ScholarDigital Library
- T. Hoefler, T. Schneider, and A. Lumsdaine. Characterizing the influence of system noise on large-scale applications by simulation. In Proceedings of the 2010 ACM/IEEE International Conference for High Performance Computing, Networking, Storage and Analysis, SC '10, pages 1--11, Washington, DC, USA, 2010. IEEE Computer Society. Google ScholarDigital Library
- A. Knüpfer, H. Brunst, J. Doleschal, M. Jurenz, M. Lieber, H. Mickler, M. S. Müller, and W. E. Nagel. The Vampir Performance Analysis Tool Set. In Tools for High Performance Computing, pages 139--155. Springer, July 2008.Google ScholarCross Ref
- A. Knüpfer, C. Rössel, D. Mey, S. Biersdorff, K. Diethelm, D. Eschweiler, M. Geimer, M. Gerndt, D. Lorenz, A. Malony, W. E. Nagel, Y. Oleynik, P. Philippen, P. Saviankou, D. Schmidl, S. Shende, R. Tschüter, M. Wagner, B. Wesarg, and F. Wolf. Score-P: A Joint Performance Measurement Run-Time Infrastructure for Periscope, Scalasca, TAU, and Vampir. In H. Brunst, M. S. Müller, W. E. Nagel, and M. M. Resch, editors, Tools for High Performance Computing 2011, pages 79--91. Springer Berlin Heidelberg, 2012.Google ScholarCross Ref
- Message Passing Interface Forum. MPI: A Message-Passing Interface Standard, Version 3.0, 2012. http://www.mpi-forum.org/docs/mpi-3.0/mpi30-report.pdf.Google Scholar
- NASA Advanced Supercomputing (NAS) Division. NAS Parallel Benchmarks, Version 3.3. http://www.nas.nasa.gov/publications/npb.html.Google Scholar
- M. Schulz, G. Bronevetsky, R. Fernandes, D. Marques, K. Pingali, and P. Stodghill. Implementation and evaluation of a scalable application-level checkpoint-recovery scheme for mpi programs. In Proceedings of the 2004 ACM/IEEE conference on Supercomputing, SC '04, pages 38--, Washington, DC, USA, 2004. IEEE Computer Society. Google ScholarDigital Library
- M. Schulz, G. Bronevetsky, and B. R. Supinski. On the performance of transparent mpi piggyback messages. In Proceedings of the 15th European PVM/MPI Users' Group Meeting on Recent Advances in Parallel Virtual Machine and Message Passing Interface, pages 194--201, Berlin, Heidelberg, 2008. Springer-Verlag. Google ScholarDigital Library
- S. Shende and A. D. Malony. The TAU Parallel Performance System, SAGE Publications. International Journal of High Performance Computing Applications, 20(2):287--331, 2006. Google ScholarDigital Library
- M. Wagner, A. Knüpfer, and W. E. Nagel. Enhanced Encoding Techniques for the Open Trace Format 2. Procedia Computer Science, 9:1979--1987, 2012.Google ScholarCross Ref
- M. Wagner and W. E. Nagel. Strategies for Real-Time Event Reduction. In Euro-Par 2012: Parallel Processing Workshops, volume 7640 of Lecture Notes in Computer Science, pages 429--438. Springer Berlin Heidelberg, 2013. Google ScholarDigital Library
Recommendations
A Dedicated Message Matching Mechanism for Collective Communications
ICPP Workshops '18: Workshop Proceedings of the 47th International Conference on Parallel ProcessingThe Message Passing Interface (MPI) libraries use message queues to guarantee correct message ordering between communicating processes. Message queues are in the critical path of MPI communications and thus, the performance of message queue operations ...
Receiving message prediction method
Special issue: Parallel and distributed scientific and engineering computingThis paper proposes and evaluates the Receiving Message Prediction Method for high performance message passing. In this method, a node in the idle state predicts the next message reception, and speculatively executes the message reception and user ...
A dynamic, unified design for dedicated message matching engines for collective and point-to-point communications
Highlights- Proposing a unified message queue design that uses the PNP approach [4] for pt2pt elements along with the COL approach [28] to improve message matching ...
AbstractThe Message Passing Interface (MPI) libraries use message queues to guarantee correct message ordering between communicating processes. Message queues are in the critical path of MPI communications and thus, the performance of message ...
Comments