Run-Time Exploitation of Application Dynamism for Energy-Efficient Exascale Computing (READEX)

Efficiently utilizing the resources provided on current petascale and future exascale systems will be a challenging task, potentially causing a large amount of underutilized resources and wasted energy. A promising potential to improve efficiency of HPC applications stems from the significant degree of dynamic behavior, e.g., run-time alternation in application resource requirements in HPC workloads. Manually detecting and leveraging this dynamism to improve performance and energy-efficiency is a tedious task that is commonly neglected by developers. However, using an automatic optimization approach, application dynamism can be analyzed at design-time and used to optimize system configurations at run-time. The European Union Horizon 2020 READEX project will develop a tools-aided scenario based auto-tuning methodology to exploit the dynamic behavior of HPC applications to achieve improved energy-efficiency and performance. Driven by a consortium of European experts from academia, HPC resource providers, and industry, the READEX project aims at developing the first of its kind generic framework for split design-time run-time automatic tuning for heterogeneous system at the exascale level.

[1]  Luca Benini,et al.  The ANTAREX approach to autotuning and adaptivity for energy efficient HPC systems , 2016, Conf. Computing Frontiers.

[2]  Francky Catthoor,et al.  Exploration of energy efficient memory organisations for dynamic multimedia applications using system scenarios , 2013, Des. Autom. Embed. Syst..

[3]  Henk Corporaal,et al.  System-scenario-based design of dynamic embedded systems , 2009, TODE.

[4]  Aleksandar Jemcov,et al.  OpenFOAM: A C++ Library for Complex Physics Simulations , 2007 .

[5]  Michael Gerndt,et al.  PAThWay: Performance Analysis and Tuning Using Workflows , 2013, 2013 IEEE 10th International Conference on High Performance Computing and Communications & 2013 IEEE International Conference on Embedded and Ubiquitous Computing.

[6]  Anna Sikora,et al.  AutoTune: A Plugin-Driven Approach to the Automatic Tuning of Parallel Applications , 2012, PARA.

[7]  Franz Franchetti,et al.  Automatic Application Tuning for HPC Architectures (Dagstuhl Seminar 13401) , 2013, Dagstuhl Reports.

[8]  Chantal Ykman-Couvreur,et al.  Systematic Methodology for Real-Time Cost-Effective Mapping of Dynamic Concurrent Task-Based Systems on Heterogenous Platforms , 2007 .

[9]  Wolfgang E. Nagel,et al.  HDEEM: High Definition Energy Efficiency Monitoring , 2014, 2014 Energy Efficient Supercomputing Workshop.

[10]  Emilio Luque,et al.  Modeling Master/Worker applications for automatic performance tuning , 2006, Parallel Comput..

[11]  Cédric Augonnet,et al.  PEPPHER: Efficient and Productive Usage of Hybrid Computing Systems , 2011, IEEE Micro.

[12]  Chun Chen,et al.  A scalable auto-tuning framework for compiler optimization , 2009, 2009 IEEE International Symposium on Parallel & Distributed Processing.

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

[14]  Jirí Jaros,et al.  Many Core Acceleration of the Boundary Element Method , 2015, HPCSE.