Kostas Siozios

Assistant Professor - Physics Department - AUTH

Kostas Siozios received his Diploma, Master and Ph.D. Degree in Electrical and Computer Engineering from the Democritus University of Thrace, Greece, in 2001, 2003 and 2009, respectively. Currently, he is Assistant Professor at Department of Physics, Aristotle University of Thessaloniki. His research interests include Hardware Accelerators, Reconfigurable Architecture, Low-Power Design, Resource Allocation, Computer-Aided Design (CAD) algorithms for physical design and Decision-Making Algorithms. He has published more than 140 papers (in peer-reviewed journals and conferences), with more than 900 citations according to G&S (h-index=15). Also, he has contributed in 5 books of Kluwer and Springer. The last years he was involved as Project Coordinator, Technical Manager or Principal Investigator in 23 research projects funded from the European Commission (EC), European Space Agency (ESA), as well as the Greek Government and Industry.

The subject of his speech...

High-Performance Computing in the Era of Big-Data Analytics

Modern trends in processing large volumes of data render the redefinition of the term "High-Performance Computing (HPC)" necessary. Contrary to the past when the term was synonymous with supercomputers for solving problems of high computational complexity, additional mechanisms are now required to efficiently manage the volume of our data (Big Data) as well as to process artificial intelligence applications. A critical factor in this problem continues to be the efficient (in terms of Performance / Watt / €) implementation of calculations. As long as there is no new technological approach that will continue to deliver the expected gains from technology’s (exponential) scaling, the research community has turned to the use of specialized hardware known as hardware accelerators. Reconfigurable Field-Programmable Gate Arrays (FPGAs) are the dominant platform for implementing such accelerators and they are used in the majority of Data Centers (eg Amazon, Microsoft, Google, etc.). The current challenges of the High Performance Computer (HPC) industry in the processing of large volumes of data with an emphasis on re-engineered architectures (FPGAs) will be presented during the talk.