The world of computing platforms, from hardware to software and to systems, is going to change dramatically in coming years as a result of the spreading of machine learning. This talk will survey the current trends in parallel computing systems and heterogeneous platforms. Then, I will deep-dive into the domain of deep learning and its acceleration by hardware and software.
Yariv has 20 years of experience in leading software R&D of cutting-edge products in the domains of parallel and distributed systems, heterogeneous systems, and deep-learning acceleration. He is currently the head of software in NeuReality (neureality.ai), a new startup that builds a unique AI-centric deep-learning inference platform. Before NeuReality, he led the development of DL acceleration on Mobileye EyeQ platforms and Intel platforms and was a member of the Khronos consortium group which defined OpenCL (Open Compute), an open standard for parallel programming of heterogeneous systems. In Cisco, he led the development of WANOP technology for file systems protocols, integrated in Cisco routers. In IBM research, he led the innovation of new technologies for automatic provisioning of multi-tier server/network/storage. Yariv holds a PhD. in computer science from Tel-Aviv university. He has more than twenty academic publications in journals and conferences, and fourteen issued patents.