Abstract
In a new approach to making computers more efficient, called "inexact," "probabilistic," or "approximate" computing, errors are not avoided; they are welcomed. Some call it "living dangerously."
- Chakrapani, L., George, J., Marr, B., Akgul, B., Palem, P. Probabilistic design: a survey of probabilistic CMOS technology and future directions for terascale IC design, VLSI-SoC: Research Trends in VLSI and Systems on Chip, December 2007. http://www.ece. rice.edu/~al4/visen/2008springer.pdfGoogle Scholar
- European Semiconductor Industry Association, Japan Electronics and Information Technology Industries Association, Korean Semiconductor Industry Association, Taiwan Semiconductor Industry Association, U.S. Semiconductor Industry Association International roadmap for semiconductors, 2011. http://www.itrs.net/Links/2011ITRS/ Home2011.htmGoogle Scholar
- Lala, Parag K. Self-checking and fault-tolerant digital design, Elsevier Science, The Morgan Kaufmann Series in Computer Architecture and Design, June 2000. http://www.elsevier. com/wps/find/bookdescription.cws_ home/677912/description#description Google ScholarDigital Library
- Palem, K., Chakrapani, L., Kedem, Z., Lingamneni, A., Muntimadugu, K. Sustaining Moore's law in embedded computing through probabilistic and approximate design: retrospects and prospects, International Conference on Compilers, Architectures, and Synthesis for Embedded Systems, Grenoble, France, October 11-26, 2009 http://www.ece.rice. edu/~al4/visen/2009cases.pdf Google ScholarDigital Library
- Shanbhag, N.R., Mitra, S., de Veciana, G., Orshansky, M., Marculescu, R., Roychowdhury, J., Jones, D., Rabaey, J.M. The search for alternative computational paradigms, Design & Test of Computers, IEEE, July-Aug. 2008 http://dl.acm.org/ citation.cfm?id=1440404.1440427 Google ScholarDigital Library
Index Terms
- Inexact design: beyond fault-tolerance
Recommendations
Inexact Newton Dogleg Methods
The dogleg method is a classical trust-region technique for globalizing Newton's method. While it is widely used in optimization, including large-scale optimization via truncated-Newton approaches, its implementation in general inexact Newton methods ...
Inexact Successive quadratic approximation for regularized optimization
Successive quadratic approximations, or second-order proximal methods, are useful for minimizing functions that are a sum of a smooth part and a convex, possibly nonsmooth part that promotes regularization. Most analyses of iteration complexity focus on ...
An Inexact Proximal-Type Algorithm in Banach Spaces
In this paper, we investigate the strong convergence of an inexact proximal-point algorithm. It is known that the proximal-point algorithm converges weakly to a solution of a maximal monotone operator, but fails to converge strongly. Solodov and Svaiter ...
Comments