Applicative 2016 / Speakers

Heterogeneous Computing: Hardware and Software Perspectives

In the beginning was the single core ... Then we moved to multicore, before we are fully ready for it! Then GPUs appear in the scene, giving us very high performance for some type of applications ... What is next? How can we get more performance? The very near future will be the era of heterogeneous computing. We already have a glimpse of it now; you write code for multicore and GPUs together, right? As computer systems become more and more heterogeneous (cores of different capabilities, GPUs, application specific hardware, ...), writing efficient code for it becomes more and more challenging. What type of heterogeneity are we talking about? Why do we need this heterogeneity? How can we write software that makes the best use of that? ... These are the topics we will discuss in this talk.

About

Mohamed Zahran received his Ph.D. in Electrical and Computer Engineering from University of Maryland at College Park in 2003. He is currently a faculty member with the Computer Science Department at NYU. His research interest spans several aspects of computer architecture, such as architecture of heterogeneous systems, hardware/software interaction, and biologically- inspired architectures. His work tries to answer the following question: How can we build powerful computers, that are easy to program yet are efficient in terms of power. Zahran is a senior member of IEEE, senior member of ACM, and Sigma Xi scientific honor society.

Website: http://www.mzahran.com/