Date: April 26, 2023 | 3 p.m. (UTC+1)

Speakers: Lubomir Riha and Ondřej Vysocký, IT4Innovations National Supercomputing Center

Moderator: Esteban Mocskos, Universidad de Buenos Aires

High-Performance Computing centers consume megawatts of electrical power, which is a limiting factor in building bigger systems on the path to exascale and post-exascale clusters. Such high power consumption leads to several challenges including robust power supply and its network, enormous energy bills, or significant CO2 emissions. To increase power efficiency, vendors accommodate various heterogeneous hardware that must be fully utilized by users’ applications, to be used efficiently. Such requirements may be hard to fulfill, which open a possibility of limiting the available resources for additional power and energy savings with no or small performance penalty.

The talk will present best practices on how to grant rights to control hardware parameters, how to measure the energy consumption of the hardware, and what can be expected from performing energy-saving activities based on hardware tuning.

About the speakers:

Lubomir Riha, Ph.D. is the Head of the Infrastructure Research Lab at IT4Innovations National Supercomputing Center. Previously he was a research scientist in the High-Performance Computing Lab at George Washington University, ECE Department. He received his Ph.D. degree in Electrical Engineering from the Czech Technical University in Prague, Czech Republic, and a Ph.D. degree in Computer Science from Bowie State University, USA. Currently, he is a local principal investigator of two EuroHPC Centers of Excellence: MAX and SPACE, and two EuroHPC projects: SCALABLE and EUPEX (designs a prototype of the European Exascale machine). Previously he was a local PI of the H2020 Center of Excellence POP2 and H2020-FET HPC READEX projects. His research interests are optimization of HPC applications, energy-efficient computing, acceleration of scientific and engineering applications using GPU and many-core accelerators, and parallel and distributed rendering.

Ondrej Vysocky is a Ph.D. candidate at VSB – Technical University of Ostrava, Czech Republic and at the same time he works at IT4Innovations in Infrastructure Research Lab. His research is focused on energy efficiency in high-performance computing. He was an investigator of the Horizon 2020 READEX project which dealt with the energy efficiency of parallel applications using dynamic tuning. Since that time, he develops a MERIC library, a runtime system for energy measurement and hardware parameters tuning during a parallel application run. Using the library he is an investigator of several H2020 projects including Performance Optimisation and Productivity (POP2), or European Pilot for Exascale (EUPEX). He is also a member of the PowerStack initiative, which works on a holistic, extensible, and scalable approach of power management.