Error Analysis for Loop Programs using Anytime Instructions in Approximate Computing

Keszöcze O, König M, Brand M, Teich J (2020)

Publication Language: English

Publication Type: Conference contribution, Original article

Publication year: 2020

Publisher: VDE Verlag

Event location: Stuttgart DE

Abstract

Approximate Computing is a novel design paradigm sacrificing computational accuracy for gains in other non-functional properties. We present a tool that allows to automatically analyze approximated loop programs for the resulting error and range of output values. Our tool allows to model non-uniform distributions for the input variables. We further support the novel concept of Anytime Instructions: Anytime Instructions encode the number of most-significant mantissa bits to be computed in floating point operations. They are typically used to achieve execution time and energy reductions. First experiments using the tool show promising results.

Authors with CRIS profile

Oliver Keszöcze Juniorprofessur für Informatik Marcel Brand Sonderforschungsbereich/Transregio 89 Invasives Rechnen Jürgen Teich Lehrstuhl für Informatik 12 (Hardware-Software-Co-Design)

Related research project(s)

None - None

How to cite

APA:

Keszöcze, O., König, M., Brand, M., & Teich, J. (2020). Error Analysis for Loop Programs using Anytime Instructions in Approximate Computing. In VDE (Eds.), Proceedings of the Methoden und Beschreibungssprachen zur Modellierung und Verifikation von Schaltungen und Systemen. Stuttgart, DE: VDE Verlag.

MLA:

Keszöcze, Oliver, et al. "Error Analysis for Loop Programs using Anytime Instructions in Approximate Computing." Proceedings of the Methoden und Beschreibungssprachen zur Modellierung und Verifikation von Schaltungen und Systemen, Stuttgart Ed. VDE, VDE Verlag, 2020.

BibTeX: Download