BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Events//NONSGML v1.0//EN
METHOD:PUBLISH
BEGIN:VEVENT
DTSTART;TZID="Pacific Time (US & Canada)":20181116T131000
DTEND;TZID="Pacific Time (US & Canada)":20181116T140000
SUMMARY:Special Speaker-Sudip Shekhar
LOCATION:Engineering Teaching Research Laboratory (ETRL), Pullman, WA
DESCRIPTION:Topic: Photonic ICs for datacenters\n\nOur insatiable demand for data has led to a rapid growth of datacenters. This has resulted in a need to compute, switch and route massive amount of data within servers, racks of servers within a datacenter, and between geographically co-located datacenters up to a hundred km.\n\nThis talk will first give a quick overview of photonics devices and integrated circuits, drawing several analogies from CMOS electronics circuits. Next, dierent challenges and opportunities for optical interconnects within datacenters and between datacenters will be described.\n\nOur recent results on high-speed laser drivers and avalanche photodetector based receivers in CMOS for short-to-medium reach links will be presented, followed by silicon photonic drivers for large-reach links interconnecting datacenters. Large-scale switches in the datacenters route dierent optical inputs to dierent optical outputs. However, the switching is done electronically, requiring optical-to-electrical and electrical-to-optical conversions. Challenges in all-optical routing will be described, and our results for the first multiple-wavelength broadband optical switch, as well as the largest-ever resonant switch network implemented in a silicon-photonic process will be presented.\n\nAn emerging field of research as an alternative to the centralized, universal Von-Neumann architectures is Neuromorphic photonics – scalable analog systems whose governing physical equations are isomorphic to those describing an abstract neural network model. It could potentially unlock new domains of machine intelligence by combining ultrafast speeds oered by photonics, energy efficiency oered by neuromorphic architectures, and algorithmic strengths of artificial neural network models. Our early work in this field will be presented.\n\nSudip Shekhar is an Associate Professor of Electrical and Computer Engineering at the University of British Columbia, Vancouver, BC. His group works on circuits for high-speed electrical and optical interfaces, frequency synthesizers, and wireless transceivers. Prior to UBC, he spent six years with the Intel Labs, Hillsboro, OR, working on high-speed I/O architectures. He received his B.Tech. degree from the Indian Institute of Technology, Kharagpur, and the Ph.D. degree from the University of Washington, Seattle, in 2003 and 2008, respectively. Dr. Shekhar was a recipient of the IEEE Transactions on Circuit and Systems Darlington Best Paper Award in 2010 and a co-recipient of IEEE Radio-Frequency IC Symposium Best Student Paper Award in 2015.
BEGIN:VALARM
ACTION:DISPLAY
DESCRIPTION:REMINDER
TRIGGER;RELATED=START:-PT00H15M00S
END:VALARM
END:VEVENT
END:VCALENDAR