Abstract:

Key Value Store (KVS) provides a highly scalable means to store and retrieve distributed data over a network.   In datacenters, high performance KVS allow large numbers of machines to share data by reading and writing key/value pairs over high-speed Ethernet.  Algo-Logic has implemented a scaled-up KVS using Field Programmable Gate Array (FPGA) logic that achieved record-setting low latency, high throughput, and low power consumption.   In this talk, it will be shown how this FPGA KVS was scaled out to accelerate machine learning for self-driving cars using a Markov Decision Process (MDP).   Parallel systems were put together with the FPGA KVS to scale up machine learning and perform real-time decision making for 30 self-driving cars in a simulated highway driving environment.     

Bio:

John W. Lockwood is an expert in building FPGA-accelerated applications. He is CEO of Algo-Logic Systems, Inc. and has founded three companies in the areas of low latency networking, Internet security, and electronic commerce.  In industry, he worked at the National Center for Supercomputing Applications (NCSA), AT&T Bell Laboratories, IBM, and Science Applications International Corp (SAIC).  In academia, he managed the NetFPGA program at Stanford University from 2007 to 2009 and grew the Beta program 10 to 1,021 cards deployed worldwide.   As a tenured professor, he created and led the Reconfigurable Network Group within the Applied Research Laboratory at Washington University in St. Louis. He has published over 100 papers and patents on topics related to networking with FPGAs and served as served as principal investigator on dozens of federal and corporate grants. He holds BS, MS, PhD degrees in Electrical and Computer Engineering from the University of Illinois at Urbana/Champaign and is a member of IEEE, ACM, and Tau Beta Pi.

Watch the seminar on our YouTube channel:

Part 1: https://youtu.be/YNZK8V0r0uQ

Part 2: https://youtu.be/tF8Li59qUjg