ABSTRACT
As we approach exascale and start planning for beyond, the rising complexity of systems and applications demands new monitoring, analysis, and optimization approaches. This requires close coordination with the parallel programming system used, which for HPC in most cases includes MPI, the Message Passing Interface. While MPI provides comprehensive tool support in the form of the MPI Profiling interface, PMPI, which has inspired a generation of tools, it is not sufficient for the new arising challenges. In particular, it does not support modern software design principles nor the composition of multiple monitoring solutions from multiple agents or sources. We approach these gaps and present QMPI, as a possible successor to PMPI. In this paper, we present the use cases and requirements that drive its development, offer a prototype design and implementation, and demonstrate its effectiveness and low overhead.
- A. Netti, M. Müller, A. Auweter, C. Guillen, M. Ott, D. Tafani, and M. Schulz. 2019. From Facility to Application Sensor Data: Modular, Continuous and Holistic Monitoring with DCDB. (Nov 2019), 1--12.Google Scholar
- Jack Dongarra, Michael A Heroux, and Piotr Luszczek. 2016. High-performance conjugate-gradient benchmark: A new metric for ranking high-performance computing systems. The International Journal of High Performance Computing Applications 30, 1 (2016), 3--10. Google ScholarDigital Library
- Jack Dongarra, Piotr Luszczek, and Michael A Heroux. 2013. HPCG Technical Specification. Sandia National Laboratories. https://prod.sandia.gov/techlib-noauth/access-control.cgi/2013/138752.pdfGoogle Scholar
- Jonathan Eastep, Steve Sylvester, Christopher Cantalupo, Brad Geltz, Federico Ardanaz, Asma Al-Rawi, Kelly Livingston, Fuat Keceli, Matthias Maiterth, and Siddhartha Jana. 2017. Global Extensible Open Power Manager: A Vehicle for HPC Community Collaboration on Co-Designed Energy Management Solutions. In High Performance Computing, Julian M. Kunkel, Rio Yokota, Pavan Balaji, and David Keyes (Eds.). Springer International Publishing, Cham, 394--412.Google Scholar
- Edgar Gabriel, Graham E. Fagg, George Bosilca, Thara Angskun, Jack J. Dongarra, Jeffrey M. Squyres, Vishal Sahay, Prabhanjan Kambadur, Brian Barrett, Andrew Lumsdaine, Ralph H. Castain, David J. Daniel, Richard L. Graham, and Timothy S. Woodall. 2004. Open MPI: Goals, Concept, and Design of a Next Generation MPI Implementation. In Proceedings, 11th European PVM/MPI Users' Group Meeting. Budapest, Hungary, 97--104.Google Scholar
- Todd Gamblin. 2016. Measuring and Analyzing Entire HPC Centers: The Sonar Project at LLNL. Invited talk at the Tokyo Institute of Technology, Tokyo, Japan.Google Scholar
- Marc-André Hermanns, Nathan T Hjlem, Michael Knobloch, Kathryn Mohror, and Martin Schulz. 2018. Enabling callback-driven runtime introspection via MPI_T. In Proceedings of the 25th European MPI Users' Group Meeting. ACM, 8. Google ScholarDigital Library
- HPC Advisory Council. 2014. HPCG Performance Benchmark and Profiling. http://www.hpcadvisorycouncil.com/pdf/HPCG_Analysis_and_Profiling.pdfGoogle Scholar
- IBM Corporation. 2017. IBM Spectrum MPI V10.1 documentation. https://www.ibm.com/support/knowledgecenter/en/SSZTET_10.1/smpi_welcome/smpi_welcome.html. Accessed on 07.05.2019.Google Scholar
- Tanzima Islam, Kathryn Mohror, and Martin Schulz. 2016. Exploring the MPI tool information interface: features and capabilities. The International Journal of High Performance Computing Applications 30, 2 (2016), 212--222. Google ScholarDigital Library
- Robert Mijaković, Antonio Pimenta Soto, Isaías A Comprés Ureña, Michael Gerndt, Anna Sikora, and Eduardo César. 2014. Specification of periscope tuning framework plugins. (2014), 123--132.Google Scholar
- M. Schulz and B. R. De Supinski. 2006. A Flexible and Dynamic Infrastructure for MPI Tool Interoperability. (Aug 2006), 193--202. Google ScholarDigital Library
- M. Schulz and B. R. de Supinski. 2007. PNMPI tools: a whole lot greater than the sum of their parts. (Nov 2007), 1--10. Google ScholarDigital Library
- M. Schulz S.Rasmussen and K. Mohror. 2016. Allowing MPI tools builders to forget about Fortran. (2016), 208--211. Google ScholarDigital Library
- The Ohio State University. 2018. MVAPICH: MPI over InfiniBand, Omni-Path, Ethernet/iWARP, and RoCE. http://mvapich.cse.ohio-state.edu/benchmarks/. Accessed on 07.05.2019.Google Scholar
- Ulrike Yang, Robert Falgout, and Jongsoo Park. 2017. Algebraic Multigrid Benchmark, Version 00. https://www.osti.gov//servlets/purl/1389816. Accessed on 07.05.2019.Google Scholar
Index Terms
- QMPI: a next generation MPI profiling interface for modern HPC platforms
Recommendations
Problems of Parallel Realization of Algorithms Calculation of Electronic Structure of Large Molecules
DEPCOS-RELCOMEX '08: Proceedings of the 2008 Third International Conference on Dependability of Computer Systems DepCoS-RELCOMEXThe problem of the organization of effective computations for the calculation of electronic structure of large molecules using self-consistent neglect differential overlap purification method is described. A mathematical model of problem is presented. ...
OpenMP for Networks of SMPs
In this paper, we present the first system that implements OpenMP on a network of shared-memory multiprocessors. This system enables the programmer to rely on a single, standard, shared-memory API for parallelization within a multiprocessor and between ...
SPMD OpenMP versus MPI on a IBM SMP for 3 Kernels of the NAS Benchmarks
ISHPC '02: Proceedings of the 4th International Symposium on High Performance ComputingShared Memory Multiprocessors are becoming more popular since they are used to deploy large parallel computers. The current trend is to enlarge the number of processors inside such multiprocessor nodes. However a lot of existing applications are using ...
Comments