Computational Complexity of Error Metrics in Approximate Computing

Keszöcze O, Soeken M, Drechsler R (2018)


Publication Language: English

Publication Type: Book chapter / Article in edited volumes

Publication year: 2018

Edited Volumes: Further Improvements in the Boolean Domain

Pages Range: Cambridge Scholars Publishing

ISBN: 978-1-5275-0371-7

Abstract

The amount of digital systems supporting our daily life is increasing continuously. Improved technical facilities for their production have led to growing challenges for engineers and scientists working in the Boolean domain. A Boolean variable can only carry two different Boolean values: FALSE or TRUE (0 or 1), and has the best interference resistance in technical systems. However, a Boolean function exponentially depends on the number of its variables. This exponential complexity is the reason for major problems in the process of design and realization of circuits. According to Moore’s Law, the complexity of digital systems approximately doubles every 18 months. This requires comprehensive knowledge and techniques to solve very complex Boolean problems. This volume represents the third book in a series that provides further insights into the Boolean domain.

Part 1 explores powerful models, methods and techniques which improve the efficiency in solving Boolean problems of extreme complexity. The universality of Boolean equations as a model to solve Non-deterministic Polynomial-time (NP) hard problems, as well as special properties of index generation functions, spectral techniques, or relational approaches, is discussed here. Both hardware devices, such as Field Programmable Gate Arrays (FPGAs) or Graphics Processing Units (GPUs), and optimized algorithms realized in software contribute to the acceleration of Boolean calculations. Part 2 contributes to the synthesis and visualization of digital circuits, and provides interesting new solutions for several types of circuits. A comprehensive collection of benchmarks supports the evolution of both existing and new synthesis approaches. The continuous reduction of the size of the transistors increases the challenges with regard to the reliability of the circuits. Part 3 describes several new approaches for the synthesis of reversible circuits. These approaches, as well as a classification of reversible functions, extend the basis of future quantum computers.

Authors with CRIS profile

Involved external institutions

How to cite

APA:

Keszöcze, O., Soeken, M., & Drechsler, R. (2018). Computational Complexity of Error Metrics in Approximate Computing. In Bernd Steinbach (Eds.), Further Improvements in the Boolean Domain. (pp. Cambridge Scholars Publishing).

MLA:

Keszöcze, Oliver, Mathias Soeken, and Rolf Drechsler. "Computational Complexity of Error Metrics in Approximate Computing." Further Improvements in the Boolean Domain. Ed. Bernd Steinbach, 2018. Cambridge Scholars Publishing.

BibTeX: Download