SGER: Understanding Unreliable Computation: Theory and Practice

William Marsh Rice University, Houston TX

Investigators

Abstract

CCF - 0635205 Title: Understanding Unreliable Computation: Theory and Practice PI: Mohanram, Kartik Inst: William Marsh Rice University ABRSTRACT Low reliability and performance variability are intrinsic to future electronic technologies, and loom large as potential show-stoppers to the commercial success of these exciting technology alternatives. The proposed research involves complementary research in the theory and practice of computation using unreliable circuits. It is proposed to investigate (i) information-theoretic, circuit-structure-independent limits for computation in the presence of noise and (ii) techniques for reliability estimation that are accurate, robust, and scalable with design complexity. The outcomes of this research will help bridge the gap between the theory and practice of reliable computation using noisy circuits. Besides advancing fault-tolerant design, this will contribute to our understanding of the capacity of Boolean circuits and provide metrics for the maximum rate at which circuits can be operated reliably. The proposed research thus has the potential to further our understanding of the theory and practice of reliable and robust circuit design, and accelerate the evolution to new computing paradigms. The models and software resulting from this effort will have significant scientific value within the broader research community, and it is proposed to develop resources for their distribution. Through (i) an introductory undergraduate class in nano-electronics and (ii) a graduate class that delves into the technologies that will define the first generation of nano-computing systems, it is proposed to enhance awareness and understanding beyond traditional intra- and inter-disciplinary boundaries.