Seminar in Computer Engineering

Bilkent University
Department of Computer Engineering
S E M I N A R

High-performance Computing and Advanced Programmer Productivity

Dr. Engin Kayraklıoğlu
Hewlett Packard Enterprise (formerly Cray)

High-performance computing (HPC) resources such as supercomputers and massively parallel architectures are getting more and more ubiquitous. Demand for HPC is driven by explosive increase in data sizes and problem complexities in many different disciplines. Because of this demand from fields that do not traditionally utilize supercomputers; many scientists, engineers and enthusiasts with limited experience and/or time budget need to use these complicated systems. Meanwhile, this demand is supplied thanks to more parallelism as postulated by Moore's Law, easier access to HPC resources via as-a-service systems such as IBM Cloud and Microsoft Azure. In this talk, we will cover these general trends after a brief introduction to general HPC concepts. Then, we will take a deeper look into what programming languages and supporting environments (i.e. profilers, auto-optimizers) can do to help both experienced and inexperienced users to program supercomputers. The talk will conclude with some examples of the ongoing efforts on the Chapel programming language at Cray/HPE.
Bio: Engin Kayraklioglu received his B.Sc. in Computer Engineering from Bilkent University in 2013. During his undergraduate studies, he did an internship at the High Performance Computing Laboratory (HPCL) at The George Washington University (GWU) in Washington, D.C. Following graduation, he started the PhD program at GWU in the same laboratory. During his doctoral studies, he did an internship at Cray Inc. working on Chapel, a parallel programming language. He finished his Ph.D. studies in 2019 and has been a software engineer in Hewlett Packard Enterprise (formerly Cray) since. His research and development interests are high-performance computing, systems programming, programming languages and programmer productivity.

DATE: 28 January 2020, Tuesday @ 13:40
PLACE: EA 409