Mapping fine-grained power measurements to HPC application runtime characteristics on IBM POWER7

Optimization of energy consumption is a key issue for future HPC. Evaluation of energy consumption requires a fine-grained power measurement. Additional useful information is obtained when performing these measurements at component level. In this paper we describe a setup which allows to perform fine-grained power measurements up to a 1 ms resolution at component level on IBM POWER (IBM and POWER are trademarks of IBM in USA and/or other countries.) machines. We further developed a plug-in for VampirTrace that allows us to correlate these power measurements with application performance characteristics, e.g. obtained by hardware performance counters. This environment enables us to generate both power and performance profiles. Such profiles provide valuable input to develop future strategies for improving workload-driven energy usage per performance. We show in comparison with power profiles of coarser granularity that these fine-grained measurements are necessary to capture the dynamics of power switching.

[1]  Michael M. Resch,et al.  Tools for High Performance Computing - Proceedings of the 2nd International Workshop on Parallel Tools for High Performance Computing, July 2008, HLRS, Stuttgart , 2008, Parallel Tools Workshop.

[2]  Dirk Schmidl,et al.  Score-P: A Joint Performance Measurement Run-Time Infrastructure for Periscope, Scalasca, TAU, and Vampir , 2011, Parallel Tools Workshop.

[3]  Kirk W. Cameron,et al.  Energy and performance characteristics of different parallel implementations of scientific applications on multicore systems , 2011, Int. J. High Perform. Comput. Appl..

[4]  Jack J. Dongarra,et al.  Collecting Performance Data with PAPI-C , 2009, Parallel Tools Workshop.

[5]  John Shalf,et al.  Power efficiency in high performance computing , 2008, 2008 IEEE International Symposium on Parallel and Distributed Processing.

[6]  Bishop Brock,et al.  Introducing the Adaptive Energy Management Features of the Power7 Chip , 2011, IEEE Micro.

[7]  Xiaorui Wang,et al.  Server-Level Power Control , 2007, Fourth International Conference on Autonomic Computing (ICAC'07).

[8]  Dong Li,et al.  PowerPack: Energy Profiling and Analysis of High-Performance Systems and Applications , 2010, IEEE Transactions on Parallel and Distributed Systems.

[9]  Wolfgang Frings,et al.  Measuring power consumption on IBM Blue Gene/P , 2011, Computer Science - Research and Development.

[10]  Gabriel Wittum,et al.  Competence in High Performance Computing 2010 , 2012, Springer Berlin Heidelberg.

[11]  Thomas Ilsche,et al.  The VampirTrace Plugin Counter Interface: Introduction and Examples , 2010, Euro-Par Workshops.

[12]  Matthias S. Müller,et al.  The Vampir Performance Analysis Tool-Set , 2008, Parallel Tools Workshop.

[13]  Michael M. Resch,et al.  Tools for High Performance Computing 2009 , 2010 .

[14]  Luigi Brochard,et al.  Optimizing performance and energy of HPC applications on POWER7 , 2010, Computer Science - Research and Development.

[15]  Bernd Mohr,et al.  The Scalasca performance toolset architecture , 2010, Concurr. Comput. Pract. Exp..

[16]  Bernd Mohr,et al.  eeClust: Energy-Efficient Cluster Computing , 2010, CHPC.

[17]  Francisco J. Cazorla,et al.  Energy-Aware Accounting and Billing in Large-Scale Computing Facilities , 2011, IEEE Micro.

[18]  Rolf Krause,et al.  A massively parallel, multi-disciplinary Barnes-Hut tree code for extreme-scale N-body simulations , 2012, Comput. Phys. Commun..

[19]  Godehard Sutmann,et al.  Particle Based Simulations of Complex Systems with MP2C : Hydrodynamics and Electrostatics , 2010 .