With memories increasingly occupying a significant fraction of a chip’s real-estate, and the critical need to provision reliable, high-performance memory bandwidth for emerging applications, there is a need to build multi-layer hardware/software stacks that can adapt and opportunistically exploit manufacturing, operational and environmental variations in order to meet system performance/responsiveness, minimize power consumption, and increase system lifetime. The overall memory hierarchy is also highly vulnerable to the adverse effects of variability and operational stress.
This talk presents the concept of variability-aware memory management for nanoscale computing systems, building on the efforts of the NSF Variability Expeditions Project. After describing the challenges for dependability across the memory hierarchy, we show how to opportunistically exploit hardware variations in on-chip and off-chip memory at the system level through the deployment of variation-aware software stacks. Specific approaches to be presented include: hardware- and software-assisted memory virtualization for achieving dependable application execution; OS-level exploitation of DRAM power variation to save energy; incorporating heterogeneous memory organizations (e.g., SRAMs, Non-Volatile Memories) to increase dependability and enhance system objectives (e.g., performance, energy, system lifetime), and semantic retention of application intent to enhance dependability through exploitation of application semantics across multiple abstraction levels, including applications, compilers, run-time systems, and hardware platforms.
Nikil D. Dutt is a Chancellor’s Professor at the University of California, Irvine, with academic appointments in the CS, EECS, and Cognitive Sciences departments. He received a B.E.(Hons) in Mechanical Engineering from the Birla Institute of Technology and Science, Pilani, India in 1980, an M.S. in Computer Science from the Pennsylvania State University in 1983, and a Ph.D. in Computer Science from the University of Illinois at Urbana-Champaign in 1989. He is affiliated with the following Centers at UCI: Center for Embedded Computer Systems (CECS), Center for Cognitive Neuroscience and Engineering (CENCE), California Institute for Telecommunications and Information Technology (Calit2), the Center for Pervasive Communications and Computing (CPCC), and the Laboratory for Ubiquitous Computing and Interaction (LUCI).
Professor Dutt’s research interests are in embedded systems, electronic design automation, computer architecture, optimizing compilers, system specification techniques, distributed systems, formal methods, and brain-inspired architectures and computing. He is a coauthor of seven books. Professor Dutt’s research has been recognized by Best Paper Awards at the following conferences: CHDL’89, CHDL’91, VLSI Design 2003, CODES+ISSS 2003, CNCC 2006, ASPDAC 2006, IJCNN 2009, and DATE 2012; and Best Paper Award Nominations at: WASP 2004, DAC 2005, VLSI Design 2006, and CASES 2011. He has also received a number of departmental and campus awards for excellence in teaching at UC Irvine.
Professor Dutt currently serves as Associate Editor of ACM Transactions on Embedded Computer Systems (TECS) and of IEEE Transactions on VLSI Systems (TVLSI). He served as Editor-in-Chief of ACM Transactions on Design Automation of Electronic Systems (TODAES) between 2004-2008. He was an ACM SIGDA Distinguished Lecturer during 2001-2002, and an IEEE Computer Society Distinguished Visitor for 2003-2005. He has served on the steering, organizing, and program committees of several premier CAD and Embedded System conferences and workshops. His recent major conference activity includes: ESWeek Steering Committee Chair and TPC Co-Chair DAC-2010/2011. He currently serves on, or has served on the ACM Publications Board, the advisory boards of ACM SIGBED, ACM SIGDA, and IFIP WG 10.5. He is a Fellow of the IEEE, an ACM Distinguished Scientist, and an IFIP Silver Core awardee.
This talk is organized by the Embedded Computing Systems Group at the Institute of Computer Engineering and the Institute of Computer Technology (ICT).