EU-project EXCESS public presentation: "EXCESS: Execution Models for Energy-Efficient Computing Systems" - Prof. Christoph Kessler, Linköping University, Sweden
Faglig kontaktperson er førsteamanuensis Phuong H. Ha.
Alle interesserte er hjertelig velkomne.
Påmelding er ikke nødvendig.
Vel møtt!
-----------------------------------------------------------------------------------------------------------------------------------------
Date: Wednesday October 21st 2015, 11:00
Place: Teknologibygget 1.023AUD, Campus Tromsø
Title: EU-project EXCESS public presentation: "EXCESS: Execution Models for Energy-Efficient Computing Systems"
Lecturer: Prof. Christoph Kessler, Linköping University, Sweden
Abstract
Energy efficiency of computations and computing systems has been a key issue in power-constrained embedded systems since many years, but is more recently also becoming a main concern in general-purpose and, in particular, high-performance computing (HPC). Modern computing systems thus expose lots of hardware mechanisms and software techniques that can affect energy usage and/or performance, such as parallelizing computations over multiple cores, vectorization, off-loading suitable computations to accelerators, dynamic voltage and frequency scaling, powering down idle units, data locality optimizations, or switching between different applicable variants of algorithms, data structures, and communication / synchronization mechanisms. Today's parallel and heterogeneous computing systems already impose enough challenges in programmability, performance and code portability on the programmer. On top of that, the systematic coordination of all these knobs to achieve energy-efficient program execution calls for automated solutions, and a holistic approach involving all layers of the hardware-software stack has the best chances to leverage the full energy optimization potential in complex computing systems.
EXCESS (Execution Models for Energy-Efficient Computing Systems) is a 3-year EU FP7 research project (2013-2016) involving five partners: Chalmers University of Technology, Sweden; Linköping University, Sweden; Movidius Ltd., Ireland; Stuttgart University - HLRS, Germany; and University of Tromso, Norway. The EXCESS project is developing a holistic and generic approach to optimizing execution flow that considers both hardware and software (system software, programming model, application) aspects together. EXCESS develops novel methods and tools for the modeling of execution platforms, energy, and computations. These methods and tools are generic and support a large spectrum of systems, ranging from embedded processors (such as Movidius Myriad2) up to typical GPU-accelerated HPC server and cluster configurations. EXCESS investigates high-level, compositional programming models and global optimization techniques, and implements prototype toolchains that help to drastically simplify the development, automated optimization, and monitoring of energy-aware applications.
This talk gives a high-level overview of the EXCESS project and presents selected results.
Short CV:
Christoph Kessler is a professor for Computer Science at Linköping University, Sweden, where he leads the Programming Environment Laboratory's research group on compiler technology and parallel computing. Christoph Kessler received a PhD degree in Computer Science in 1994 from the University of Saarbrücken, Germany, and a Habilitation degree in 2001 from the University of Trier, Germany. In 2001 he joined Linköping university, Sweden, as associate professor at the programming environments lab (PELAB) of the computer science department (IDA). In 2007 he was appointed full professor at Linköping university. His research interests include parallel programming, compiler technology, code generation, optimization algorithms, and software composition. He has published two books, several book chapters and more than 100 scientific papers in international journals and conferences. His contributions include e.g. the OPTIMIST retargetable optimizing integrated code generator for VLIW and DSP processors, thePARAMAT approach to pattern-based automatic parallelization, the concept of multi-variant performance-aware parallel components for optimized composition, the PEPPHER component model and composition tool for heterogeneous multicore/manycore based systems, the SkePU library of tunable generic components for GPU-based systems, and the parallel programming languages Fork and NestStep.
