Squash: Approximate Square-Accumulate With Self-Healing

Syed Ghayoor Abbas Gillani (Corresponding Author), M. A. Hanif, M. Krone, Sabih H. Gerez, M. Shafique, Andre B.J. Kokkeler

Research output: Contribution to journalArticleAcademicpeer-review

5 Citations (Scopus)
33 Downloads (Pure)

Abstract

Approximate computing strives to achieve the highest performance-, area-, and power-efficiency for a given quality constraint and vice versa. Conventional approximate design methodology restricts the introduction of errors to avoid a high loss in quality. However, this limits the computing efficiency and the number of pareto-optimal design alternatives for a quality-efficiency tradeoff. This paper presents a novel self-healing (SH) methodology for an approximate square-accumulate (SAC) architecture. SAC refers to a hardware architecture that computes the inner product of a vector with itself. SH exploits the algorithmic error resilience of the SAC structure to ensure an effective quality-efficiency tradeoff, wherein the squarer is regarded as an approximation stage, and the accumulator as a healing stage. We propose to deploy an approximate squarer mirror pair , such that the error introduced by one approximate squarer mirrors the error introduced by the other, i.e., the errors generated by the approximate squarers are approximately the additive inverse of each other. This helps the healing stage (accumulator) to automatically average out the error originated in the approximation stage, and thereby to minimize the quality loss. For random input vectors, SH demonstrates up to 25% and 18.6% better area and power efficiency, respectively, with a better quality output than the conventional approximate computing methodology. As a case study, SH is applied to one of the computationally expensive components (SAC) of the radio astronomy calibration application, where it shows up to 46.7% better quality for equivalent computing efficiency as that of conventional methodology.
Original languageEnglish
Pages (from-to)49112-49128
Number of pages17
JournalIEEE Access
Volume6
DOIs
Publication statusPublished - 30 Aug 2018

Fingerprint

Radio astronomy
Mirrors
Calibration
Hardware
Optimal design

Keywords

  • Approximate computing, approximate multiplier, approximate squarer, multiply-accumulate, radio astronomy, self-healing, square-accumulate

Cite this

Gillani, Syed Ghayoor Abbas ; Hanif, M. A. ; Krone, M. ; Gerez, Sabih H. ; Shafique, M. ; Kokkeler, Andre B.J. / Squash : Approximate Square-Accumulate With Self-Healing. In: IEEE Access. 2018 ; Vol. 6. pp. 49112-49128.
@article{2b61cf8a356d4badae416de1cb270cab,
title = "Squash: Approximate Square-Accumulate With Self-Healing",
abstract = "Approximate computing strives to achieve the highest performance-, area-, and power-efficiency for a given quality constraint and vice versa. Conventional approximate design methodology restricts the introduction of errors to avoid a high loss in quality. However, this limits the computing efficiency and the number of pareto-optimal design alternatives for a quality-efficiency tradeoff. This paper presents a novel self-healing (SH) methodology for an approximate square-accumulate (SAC) architecture. SAC refers to a hardware architecture that computes the inner product of a vector with itself. SH exploits the algorithmic error resilience of the SAC structure to ensure an effective quality-efficiency tradeoff, wherein the squarer is regarded as an approximation stage, and the accumulator as a healing stage. We propose to deploy an approximate squarer mirror pair , such that the error introduced by one approximate squarer mirrors the error introduced by the other, i.e., the errors generated by the approximate squarers are approximately the additive inverse of each other. This helps the healing stage (accumulator) to automatically average out the error originated in the approximation stage, and thereby to minimize the quality loss. For random input vectors, SH demonstrates up to 25{\%} and 18.6{\%} better area and power efficiency, respectively, with a better quality output than the conventional approximate computing methodology. As a case study, SH is applied to one of the computationally expensive components (SAC) of the radio astronomy calibration application, where it shows up to 46.7{\%} better quality for equivalent computing efficiency as that of conventional methodology.",
keywords = "Approximate computing, approximate multiplier, approximate squarer, multiply-accumulate, radio astronomy, self-healing, square-accumulate",
author = "Gillani, {Syed Ghayoor Abbas} and Hanif, {M. A.} and M. Krone and Gerez, {Sabih H.} and M. Shafique and Kokkeler, {Andre B.J.}",
note = "Open Access",
year = "2018",
month = "8",
day = "30",
doi = "10.1109/ACCESS.2018.2868036",
language = "English",
volume = "6",
pages = "49112--49128",
journal = "IEEE Access",
issn = "2169-3536",
publisher = "IEEE",

}

Squash : Approximate Square-Accumulate With Self-Healing. / Gillani, Syed Ghayoor Abbas (Corresponding Author); Hanif, M. A.; Krone, M.; Gerez, Sabih H.; Shafique, M.; Kokkeler, Andre B.J.

In: IEEE Access, Vol. 6, 30.08.2018, p. 49112-49128.

Research output: Contribution to journalArticleAcademicpeer-review

TY - JOUR

T1 - Squash

T2 - Approximate Square-Accumulate With Self-Healing

AU - Gillani, Syed Ghayoor Abbas

AU - Hanif, M. A.

AU - Krone, M.

AU - Gerez, Sabih H.

AU - Shafique, M.

AU - Kokkeler, Andre B.J.

N1 - Open Access

PY - 2018/8/30

Y1 - 2018/8/30

N2 - Approximate computing strives to achieve the highest performance-, area-, and power-efficiency for a given quality constraint and vice versa. Conventional approximate design methodology restricts the introduction of errors to avoid a high loss in quality. However, this limits the computing efficiency and the number of pareto-optimal design alternatives for a quality-efficiency tradeoff. This paper presents a novel self-healing (SH) methodology for an approximate square-accumulate (SAC) architecture. SAC refers to a hardware architecture that computes the inner product of a vector with itself. SH exploits the algorithmic error resilience of the SAC structure to ensure an effective quality-efficiency tradeoff, wherein the squarer is regarded as an approximation stage, and the accumulator as a healing stage. We propose to deploy an approximate squarer mirror pair , such that the error introduced by one approximate squarer mirrors the error introduced by the other, i.e., the errors generated by the approximate squarers are approximately the additive inverse of each other. This helps the healing stage (accumulator) to automatically average out the error originated in the approximation stage, and thereby to minimize the quality loss. For random input vectors, SH demonstrates up to 25% and 18.6% better area and power efficiency, respectively, with a better quality output than the conventional approximate computing methodology. As a case study, SH is applied to one of the computationally expensive components (SAC) of the radio astronomy calibration application, where it shows up to 46.7% better quality for equivalent computing efficiency as that of conventional methodology.

AB - Approximate computing strives to achieve the highest performance-, area-, and power-efficiency for a given quality constraint and vice versa. Conventional approximate design methodology restricts the introduction of errors to avoid a high loss in quality. However, this limits the computing efficiency and the number of pareto-optimal design alternatives for a quality-efficiency tradeoff. This paper presents a novel self-healing (SH) methodology for an approximate square-accumulate (SAC) architecture. SAC refers to a hardware architecture that computes the inner product of a vector with itself. SH exploits the algorithmic error resilience of the SAC structure to ensure an effective quality-efficiency tradeoff, wherein the squarer is regarded as an approximation stage, and the accumulator as a healing stage. We propose to deploy an approximate squarer mirror pair , such that the error introduced by one approximate squarer mirrors the error introduced by the other, i.e., the errors generated by the approximate squarers are approximately the additive inverse of each other. This helps the healing stage (accumulator) to automatically average out the error originated in the approximation stage, and thereby to minimize the quality loss. For random input vectors, SH demonstrates up to 25% and 18.6% better area and power efficiency, respectively, with a better quality output than the conventional approximate computing methodology. As a case study, SH is applied to one of the computationally expensive components (SAC) of the radio astronomy calibration application, where it shows up to 46.7% better quality for equivalent computing efficiency as that of conventional methodology.

KW - Approximate computing, approximate multiplier, approximate squarer, multiply-accumulate, radio astronomy, self-healing, square-accumulate

U2 - 10.1109/ACCESS.2018.2868036

DO - 10.1109/ACCESS.2018.2868036

M3 - Article

VL - 6

SP - 49112

EP - 49128

JO - IEEE Access

JF - IEEE Access

SN - 2169-3536

ER -