Visit ShanghaiTech University | 中文 | How to find us
HOME > News and Events > Events
Physics-Based Electron-migration Modeling and Cross-Layer Reliability Management
Date: 2015/11/25             Browse: 661

Speaker: Sheldon Tan

Time: Nov 25, 4:00pm - 5:00pm.

Location: Room 310, Teaching Center, Zhangjiang Campus


Reliability has become a significant challenge for design of current nanometer integrated circuits (ICs). It was expected that the future chips will show signs of reliability-induced age much faster than the previous generations.

In the first part of this talk, I will first present a new physics-based electromigration (EM) model and a novel EM assessment technique for full-chip power delivery networks of VLSI systems. The new model, which consists of void nucleation and growth phases, is much more predictable for different stressing conditions and the new IR-drop based assessment technique can naturally account the essential redundancy in the power-ground (p/g) networks. Then I will present a novel physics-based EM model for multi-terminal interconnect trees. The new model was obtained by solving coupled stress differential equations for analytic solutions for the first time. I will also present recovery effect modeling for the physics-based EM models by considering time-varying current densities and temperatures. We show that the EM recovery effect can be quite significant even under unidirectional current loads and the recovery effects can be leveraged to extend the lifetime of EM-stressed wires.

In the second part of the talk, I will first present a new approach for system-level reliability management technique for multi/many core microprocessors. In the new approach, the EM mean time to failure (MTTF) at the system level is modeled as a resource, which is abstracted from the new physics-based EM model at the chip physical level. On top of the new EM model, we propose a novel task migration method to explicitly balance consumption of EM resources for all the cores. The new method will lead to the equal chance of failure of these cores, which will maximize the life time of the whole multi/many-core system. Then I will present a new learning-based new dynamic reliability management technique for emerging dark-silicon many-core computing systems based on the new physics-based EM model.


Dr. Sheldon Tan is a Professor in the Department of Electrical Engineering, University of California, Riverside, CA.  He is the Associate Director of Compute Engineering Program (CEN) and cooperative faculty member in the Department of Computer Science and Engineering at UCR. Dr. Sheldon Tan received his B.S. and M.S. degrees in electrical engineering from Fudan University, Shanghai, China in 1992 and 1995, respectively and the Ph.D. degree in electrical and computer engineering from the University of Iowa, Iowa City, in 1999.  He is a Chaired Guest Professor of  Shanghai Jiaotong University and a Guest Professor of University of Electronic Science and Technology of China, China. His research interests include VLSI reliability modeling, optimization and management at circuit and system levels, thermal modeling, optimization and dynamic thermal management for many-core processors,  statistical modeling, simulation and optimization of mixed-signal/RF/analog circuits, parallel circuit simulation techniques based on GPU and multicore systems.

Dr. Tan received the Best Paper Award from DAC in 1999 and the Best Paper Award from ICCD 2007. He also received Best Paper Award nominations from DAC’05, DAC’09, DAC’14 and ASPDAC’15. Dr. Tan received NSF CAREER Award in 2004.  He has co-authored four books and over 230 publications. He now is serving as an Associate Editor for three journals: IEEE Transaction on VLSI Systems (TVLSI),  ACM Transaction on Design Automation of Electronic Systems (TODAE),  Integration, The VLSI Journal.               

SIST-Seminar 15049