Cognitive Systems Simulation, IBM F1, D3, Manyata, Bangalore - 560045
Dr. Saif Abrar earned his PhD from the Technical University of Tallinn (Estonia), after completing bachelors and masters from Aligarh Muslim University and IIT-Delhi, respectively. With close to 2 decades of industry experience, he has contributed to computing systems ranging from tiny microcontrollers to high-performance architectures. Currently he leads a team in IBM working on next-generation processors and cognitive systems. His research interests include AI systems, computer architectures, system-level design, and performance analysis, with a keen interest in upcoming disruptive technologies. In addition to corporate responsibilities, he is actively involved with academic institutes. He has driven many industry-academia projects and actively shares his experiences and knowledge with the faculty and students. Dr. Saif has delivered keynotes and lectures in numerous international conferences, drives the IP activities in his organization and evangelises AI-capable applications development.
Topic : "Supercomputing Wars: Marching Towards Exascale Computing"
This keynote will introduce the audience to the ongoing war between nations in the supercomputing arena. Traditionally, USA has dominated this domain and as of now, China has become increasingly active in the field. This has led to a race for faster supercomputers, currently performing at petaflops speed and continuously marching towards exascale computing. The keynote will discuss the evaluation of the topmost supercomputers in the world by Top500.org and its parameters of ranking. The historical rankings will be discussed, to show the strides made by various nations and how China is becoming a major player in the supercomputing domain now. After losing to China for quite some time, the USA regained the top-most position last year 2018 with its 200 petaflops SUMMIT supercomputer. Exascale is the next frontier, covered by this keynote, ushering in the new off-chip era, with advanced accelerators like GPUs and FPGAs driving modern workloads, including AI. China is planning the first exascale supercomputer by 2020 according to the head of the school of computing at the National University of Defense Technology (NUDT). The US Department of Energy and Intel announced the first exaFLOP supercomputer by the end of 2021. The European High-Performance Computing Joint Undertaking (EuroHPC JU) aims to build an exascale supercomputer by 2022/2023. The RIKEN Advanced Institute for Computational Science in Japan began planning an exascale system for 2020, intended to consume less than 30 megawatts. Finally, this keynote will introduce the audience to open cloud-based platforms, like OpenPOWER Research Facility and SuperVessel, to develop GPU optimized POWER applications for academic as well as commercial purposes.