Keynote Talk: Wednesday November 18th, 2020
Variation-Tolerant & Error-Resilient Many-Core SoCs with Fine-Grain Power Management
Dr. Vivek De
Intel Fellow at Intel Corporation. IEEE Fellow
Abstract: Many-core system-on-chip (SoC) architecture & design challenges & opportunities spanning edge devices to cloud computing systems in scaled CMOS process are presented. Key techniques for robust and variation-tolerant logic, embedded memory arrays and on-die interconnect fabrics are discussed. Fine-grain multi-voltage design and power management techniques, featuring integrated voltage regulators for wide dynamic voltage-frequency operating range and flexible platform power control across multi-threaded high-throughput nearthreshold voltage (NTV) to single-threaded burst performance modes, are elucidated. Smart variation-aware workload mapping, runtime self-adaptation and error detection & recovery schemes to mitigate impacts of process-voltage-temperature (PVT) variations & aging, and achieve maximum performance under stringent thermal and energy constraints, are presented. Latest advances in design and process/package for realization of monolithic & heterogeneous 2D/3D-integrated compact, efficient, low supply noise, fine-grain, high-bandwidth & fastresponse power converters & voltage regulators, essential for implementing intelligent system-level power management and adaptation schemes across hardware and software, are also highlighted. Real SoC examples are used to demonstrate leading-edge practical systems.
Vivek De is an Intel Fellow and Director of Circuit Technology Research in Intel Labs. He is responsible for providing strategic technical directions for long term research in future circuit technologies and leading energy efficiency research across the hardware stack. He has 295 publications in refereed international conferences and journals with a citation H-index of 79, and 227 patents issued with 32 more patents filed (pending). He received an Intel Achievement
Award for his contributions to an integrated voltage regulator technology. He is the recipient of the 2019 IEEE Circuits and System Society (CASS) Charles A. Desoer Technical Achievement Award for “pioneering contributions to leading-edge performance and energy-efficient microprocessors & many-core system-on-chip (SoC) designs” and the 2020 IEEE Solid-State Circuits Society (SSCS) Industry Impact Award for “seminal impact and distinctive contributions
to the field of solid-state circuits and the integrated circuits industry”. He received a Best Paper Award at the 1996 IEEE International ASIC Conference, and nominations for Best Paper Awards at the 2007 IEEE/ACM Design Automation Conference (DAC) and 2008 IEEE/ACM International Conference on Computer-Aided Design (ICCAD). He also co-authored a paper nominated for the Best Student Paper Award at the 2017 IEEE International Electron Devices
Meeting (IEDM). One of his publications was recognized in the 2013 IEEE/ACM Design Automation Conference (DAC) as one of the “Top 10 Cited Papers in 50 Years of DAC”.
Another one of his publications received the “Most Frequently Cited Paper Award” in the IEEE Symposium on VLSI Circuits at its 30th Anniversary in 2017. He received the 2017
Distinguished Alumnus Award from the Indian Institute of Technology (IIT) Madras. He received a B.Tech from IIT Madras, India, a MS from Duke University, Durham, North Carolina, and a
PhD from Rensselaer Polytechnic Institute, Troy, New York, all in Electrical Engineering. He is a Fellow of the IEEE.
Keynote Talk: Thursday November 19th, 2020
From Deep Scaling To Deep Intelligence
Dr. Rajiv Joshi
Watson Research Center IBM
Abstract: Moore’s law driving the advancement in the semiconductor industry over decades has been coming to a screeching halt and many researchers are convinced that it is almost dead. After revival and promise of artificial intelligence (AI) due to increased computational performance and memory bandwidth aided by Moore’s law, there is overwhelming enthusiasm in researchers for increasing the pace of VLSI industry. AI uses many neural network techniques for computation which involves training and inference. The advancement in AI requires energy efficient, low power hardware systems. This is more so for servers, main processors, Internet of Things (IoT) and System on chip (SOC) applications and newer applications in cognitive computing.
In the light of AI, this talk focuses on important circuit techniques for lowering power, improving performance and functionality in nanoscale VLSI design in the midst of variability. The same techniques can be used for AI specific accelerators. Accelerator development for reduction in power and throughput improvement for both edge and data-centric accelerators compared to GPUs used for convolutional Neural (CNN) and Deep Neural (DNN) Networks are described. The talk covers memory (volatile and nonvolatile) solutions for CNN/DNN applications at extremely low Vmin. The talk also focuses on in-memory computation. Binary and analog applications using non-volatile memories (NVM) are illustrated. Accelerator architectures for bitwise convolution that features massive parallelism with high energy efficiency are described for both binary and analog memories. In our earlier work, numerical experiment results show that the binary CNN accelerator on a digital ReRAM-crossbar achieves a peak throughput of 792 GOPS at the power consumption of 4.5 mW, which is 1.61 times faster and 296 times more energy-efficient than a high-end GPU. Finally, the talk summarizes challenges and future directions for circuit applications for edge and data-centric accelerators.
Dr. Rajiv Joshi
Dr. Rajiv V. Joshi is a research staff member and key technical lead at T. J.Watson research center, IBM. He received his B.Tech I.I.T (Bombay, India), M.S (M.I.T) and Dr. Eng. Sc. (Columbia University). His novel interconnects processes and structures for aluminum, tungsten and copper technologies which are widely used in IBM for various technologies from sub-0.5μm to 14nm. He has led successfully predictive failure analytic techniques for yield prediction and also the technology-driven SRAM at IBM Server Group. He has extensively worked on novel memory designs. He commercialized these techniques. He received 3 Outstanding Technical Achievement (OTAs), 3 highest Corporate Patent Portfolio awards for licensing contributions, holds 58 invention plateaus and has over 225 US patents and over 350 including international patents. He has authored and co-authored over 190 papers. He has given over 45 invited/keynote talks and given several Seminars. He is awarded prestigious IEEE Daniel Noble award for 2018. He received the Best Editor Award from IEEE TVLSI journal. He is recipient of 2015 BMM award. He is inducted into New Jersey Inventor Hall of Fame in Aug 2014 along with pioneer Nicola Tesla. He is a recipient of 2013 IEEE CAS Industrial Pioneer award and 2013 Mehboob Khan Award from Semiconductor Research Corporation. He is a member of IBM Academy of technology. He served as a Distinguished Lecturer for IEEE CAS and EDS society. He is Distinguished visiting professor at IIT, Roorkie. He is IEEE, ISQED andWorld Technology Network fellow and distinguished alumnus of IIT Bombay. He is in the Board of Governors for IEEE CAS. He serves as an Associate Editor of TVLSI. He served on committees of ISCAS 2017, ISLPED (Int. Symposium Low Power Electronic Design), IEEE VLSI design, IEEE CICC, IEEE Int. SOI conference, ISQED and Advanced Metallization Program committees. He served as a general chair for IEEE ISLPED. He is an industry liaison for universities as a part of the Semiconductor Research Corporation. Also, he is in the industry liaison committee for IEEE CAS society.