Increasing System Reliability Through Resource Management
Xiaobo Sharon Hu
Professor
Dept. of Computer Science and Engineering
University of Notre Dame
2:30 in ITEB 336 on Friday March 25
As CMOS technology continues its downward scaling trends, increase in power density leads to elevated operating temperature and frequent temperature variations, which accelerates chip wear-out. Such accelerated wear-outs eventually result in permanent faults occurring earlier and reduce processor lifetime. It has been shown that a 10–15
C difference in operating temperature can result in a 2X difference in the lifespan of a device. Furthermore, decreased feature size and operating voltage make a processor more vulnerable to transient faults, thus degrade soft-error reliability. Though chip packaging and cooling solutions can be employed to help alleviate these challenges, such solutions can be prohibitively expensive. Improving overall system reliability through resource management (such as task assignment and scheduling, and dynamic voltage scaling) can provide some powerful alternatives.
In this talk, I will first present a comprehensive framework for online task assignment and scheduling to increase system reliability through wear compensation. I will then introduce a novel analytical approach for calculating the mean time to failure due to transient faults. This analytical approach opens the door to effectively trade off reliability due to transient faults and permanent faults. I will end the talk with a solution to the problem of maximizing system availability while satisfying a given throughput constraint and considering both permanent and transient faults.
Speaker Bio:
Sharon Hu is a professor in the department of Computer Science and Engineering at the University of Notre Dame, Notre Dame, Indiana, USA. Her research interests include low-power system design, circuit and architecture design with emerging technologies, hardware/software co-design and real-time embedded systems. She has published more than 250 papers in these areas, and received the Best Paper Award from the Design Automation Conference in 2001 and from the IEEE Symposium on Nanoscale Architectures in 2009. She is the Program Chair of Design Automation Conference in 2016. She also served as Associate Editor for IEEE Transactions on VLSI, ACM Transactions on Design Automation of Electronic Systems, ACM Transactions on Embedded Computing Systems. Sharon Hu is a Fellow of IEEE.