1 / 12

Performance Technology for Scalable Parallel Systems

Performance Technology for Scalable Parallel Systems. Allen D. Malony Department of Computer and Information Science University of Oregon. Performance Technology. Data mining Models Expert systems. Performance Technology. Experiment management Performance storage. Performance Technology.

dawson
Télécharger la présentation

Performance Technology for Scalable Parallel Systems

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. Performance Technologyfor Scalable Parallel Systems Allen D. Malony Department of Computer and Information Science University of Oregon

  2. PerformanceTechnology • Data mining • Models • Expert systems PerformanceTechnology • Experimentmanagement • Performancestorage PerformanceTechnology • Instrumentation • Measurement • Analysis • Visualization Performance Technology • Tools for performance problem solving • Empirical-based performance optimization process • Performance technology concerns PerformanceTuning hypotheses Performance Diagnosis properties Performance Experimentation characterization Performance Observation

  3. TAU Parallel Performance System Project • Tuning and Analysis Utilities (15+ year project effort) • Performance system framework for HPC systems • Integrated, scalable, and flexible • Target parallel programming paradigms • Integrated toolkit for performance problem solving • Instrumentation, measurement, analysis, and visualization • Portable performance profiling and tracing facility • Performance data management and data mining • Partners • LLNL, ANL, LANL • Research Centre Jülich, TU Dresden • http://tau.uoregon.edu

  4. TAU Parallel Performance System Goals • Portable (open source) parallel performance system • Computer system architectures and operating systems • Different programming languages and compilers • Multi-level, multi-language performance instrumentation • Flexible and configurable performance measurement • Support for multiple parallel programming paradigms • Multi-threading, message passing, mixed-mode, hybrid, object oriented, component-based • Support for performance mapping • Integration of leading performance technology • Scalable (very large) parallel performance analysis

  5. Performance Monitoring TAU Performance System Components Performance Data Mining TAU Architecture Program Analysis PDT PerfExplorer Parallel Profile Analysis PerfDMF ParaProf TAUoverSupermon

  6. TAU Instrumentation and Measurement

  7. TAU Analysis

  8. TAU on HPC Platforms with Intel Processors • ARL (JVN / MJM, x86_64 Linux NetworX) • 14.7 TF / 52.8 TF, 2048 / 4400 processors • ARFL (Hawk, Eagle ia64 SGI Altix) • 59 TF, 9216 processors • NCSA (Abe, x86_64 Dell) • 89.47 TF, 9600 cores • NASA (Columbia, ia64 SGI Altix) • 60.96 TF, 10240 processors • MHPCC (Jaws, x86_64 Dell) • 60 TF, 5120 processors • TACC (Lonestar, x86_64 Dell) • 62 TF, 5840 processors

  9. TAU on Leadership Class Facilities and TeraGrid • Argonne National Laboratory • IBM BG/P • 111 TF • 32768 processors • Oak Ridge National Laboratory • Cray XT-4 • 119 TF peak • 23416 cores (AMD Opteron) • Texas Advanced Computing Center • Sun Blade 8000 • 504 TF peak • 62976 cores (AMD Opteron)

  10. Recent Funding • A. Malony, S. Shende, N. Nystrom, S. Moore, R. Kufrin, SDCI HPC Improvement: High-Productivity Performance Engineering (Tools, Methods, Training) for NSF HPC Applications, NSF Software Development for Cyberinfrastructure (SDCI), 11/1/2007-10/31/2010. • A. Malony, S. Shende, Knowledge-based Parallel Performance Technology, DOE Office of Science, 9/1/2007-8/31/2010. • P. Beckman, A. Malony, Extreme Performance Scalable Operating Systems, DOE Office of Science, 12/1/04-1/31/08. • A. Malony, S. Shende, Application-Specific Performance Technology for Productive Parallel Computing, DOE Office of Science, 5/1/05-4/30/08. • S. McKee, A. Malony, G. Tyson, ST-HEC: Collaborative Research: Scalable, Interoperable Tools to Support Autonomic Optimization of High-End Applications, NSF High-End Computing (HEC), 11/1/04-10/31/07. • A. Malony, Multi-core Parallel Programming and Performance Tools, Intel equipment grant, 9/15/2006. • A. Malony, M. Sottile, Multi-core Parallel Programming, Intel equipment grant, 11/1/2007.

  11. Intel Contacts • Justin Rattner, Intel Senior Fellow, Vice-President • Director, Corporate Technology Group • In 1988 Rattner (at Intel Scientific Computers) suggested implementing a performance monitor for the iPSC/2 hypercube • A. Malony, D. Reed, “A Hardware-Based Performance Monitor for theIntel iPSC/2 Hypercube,” ICS 1990. • David Kuck, Intel Fellow • Software and Solutions Group • Director, Parallel and Distributed Solutions Division • Worked for Kuck at the Center for Supercomputing Research and Development, University of Illinois, Urbana-Champaign • Tim Mattson, Senior Research Scientist • Computational Software Laboratory • Werner Krotz-Vogel • Technical Marketing Engineer, Intel Cluster Tools

  12. Support Acknowledgements • Department of Energy (DOE) • Office of Science • ASCR, Argonne National Lab • ASC/NNSA • University of Utah ASC/NNSA Level 1 • ASC/NNSA, Lawrence Livermore National Lab • Department of Defense (DoD) • HPC Modernization Office (HPCMO) • NSF Software Development for Cyberinfrastructure • Los Alamos National Laboratory • Research Centre Juelich, TU Dresden • ParaTools, Inc.

More Related