MACISH: Designing Approximate MAC Accelerators with Internal-Self-Healing

Syed Ghayoor Abbas Gillani, M.A. Hanif, Bart Verstoep, Sabih H. Gerez, M. Shafique, Andre B.J. Kokkeler

Research output: Contribution to journalArticleAcademicpeer-review

11 Downloads (Pure)

Abstract

Approximate computing studies the quality-efficiency trade-off to attain a best-efficiency (e.g., area, latency, and power) design for a given quality constraint and vice versa. Recently, self-healing methodologies for approximate computing have emerged that showed an effective quality-efficiency tradeoff as compared to the conventional error-restricted approximate computing methodologies. However, state-of-the-art self-healing methodologies are constrained to highly parallel implementations with similar modules (or parts of a datapath) in multiples of two and for square-accumulate functions through the pairing of mirror versions to achieve error cancellation. In this article, we propose a novel methodology for InternalSelf-Healing (ISH) that allows exploiting self-healing within a computing element internally without requiring a paired, parallel module, which extends the applicability to irregular/asymmetric datapaths while relieving the restriction of multiples of two for modules in a given datapath, as well as going beyond
square functions. We employ our ISH methodology to design an approximate multiply-accumulate (xMAC), wherein the multiplier is regarded as an approximation stage and the accumulator as a healing stage. We propose to approximate a recursive multiplier in such a way that a near-to-zero average error is achieved for a given input distribution to cancel out the error at an accurate accumulation stage. To increase the efficacy of such a multiplier, we propose a novel 2 × 2 approximate multiplier design that alleviates the overflow problem within an n × n approximate recursive multiplier. The proposed ISH methodology shows a more effective quality-efficiency trade-off for an xMAC as compared to the conventional error-restricted methodologies for random inputs and for radio-astronomy calibration processing (up to 55% better quality output for equivalent-efficiency designs).
Original languageEnglish
JournalIEEE Access
DOIs
Publication statusAccepted/In press - 2019

Fingerprint

Particle accelerators
Radio astronomy
Mirrors
Calibration
Processing

Keywords

  • Approximate computing
  • multiply-accumulate (MAC) accelerator
  • internal-self-healing methodology
  • radio astronomy processing
  • power efficiency

Cite this

Gillani, Syed Ghayoor Abbas ; Hanif, M.A. ; Verstoep, Bart ; Gerez, Sabih H. ; Shafique, M. ; Kokkeler, Andre B.J. / MACISH: Designing Approximate MAC Accelerators with Internal-Self-Healing. In: IEEE Access. 2019.
@article{dfaba7f4ffcd40a88ed85b73c7a1601a,
title = "MACISH: Designing Approximate MAC Accelerators with Internal-Self-Healing",
abstract = "Approximate computing studies the quality-efficiency trade-off to attain a best-efficiency (e.g., area, latency, and power) design for a given quality constraint and vice versa. Recently, self-healing methodologies for approximate computing have emerged that showed an effective quality-efficiency tradeoff as compared to the conventional error-restricted approximate computing methodologies. However, state-of-the-art self-healing methodologies are constrained to highly parallel implementations with similar modules (or parts of a datapath) in multiples of two and for square-accumulate functions through the pairing of mirror versions to achieve error cancellation. In this article, we propose a novel methodology for InternalSelf-Healing (ISH) that allows exploiting self-healing within a computing element internally without requiring a paired, parallel module, which extends the applicability to irregular/asymmetric datapaths while relieving the restriction of multiples of two for modules in a given datapath, as well as going beyondsquare functions. We employ our ISH methodology to design an approximate multiply-accumulate (xMAC), wherein the multiplier is regarded as an approximation stage and the accumulator as a healing stage. We propose to approximate a recursive multiplier in such a way that a near-to-zero average error is achieved for a given input distribution to cancel out the error at an accurate accumulation stage. To increase the efficacy of such a multiplier, we propose a novel 2 × 2 approximate multiplier design that alleviates the overflow problem within an n × n approximate recursive multiplier. The proposed ISH methodology shows a more effective quality-efficiency trade-off for an xMAC as compared to the conventional error-restricted methodologies for random inputs and for radio-astronomy calibration processing (up to 55{\%} better quality output for equivalent-efficiency designs).",
keywords = "Approximate computing, multiply-accumulate (MAC) accelerator, internal-self-healing methodology, radio astronomy processing, power efficiency",
author = "Gillani, {Syed Ghayoor Abbas} and M.A. Hanif and Bart Verstoep and Gerez, {Sabih H.} and M. Shafique and Kokkeler, {Andre B.J.}",
year = "2019",
doi = "10.1109/ACCESS.2019.2920335",
language = "English",
journal = "IEEE Access",
issn = "2169-3536",
publisher = "IEEE",

}

MACISH: Designing Approximate MAC Accelerators with Internal-Self-Healing. / Gillani, Syed Ghayoor Abbas; Hanif, M.A.; Verstoep, Bart; Gerez, Sabih H.; Shafique, M.; Kokkeler, Andre B.J.

In: IEEE Access, 2019.

Research output: Contribution to journalArticleAcademicpeer-review

TY - JOUR

T1 - MACISH: Designing Approximate MAC Accelerators with Internal-Self-Healing

AU - Gillani, Syed Ghayoor Abbas

AU - Hanif, M.A.

AU - Verstoep, Bart

AU - Gerez, Sabih H.

AU - Shafique, M.

AU - Kokkeler, Andre B.J.

PY - 2019

Y1 - 2019

N2 - Approximate computing studies the quality-efficiency trade-off to attain a best-efficiency (e.g., area, latency, and power) design for a given quality constraint and vice versa. Recently, self-healing methodologies for approximate computing have emerged that showed an effective quality-efficiency tradeoff as compared to the conventional error-restricted approximate computing methodologies. However, state-of-the-art self-healing methodologies are constrained to highly parallel implementations with similar modules (or parts of a datapath) in multiples of two and for square-accumulate functions through the pairing of mirror versions to achieve error cancellation. In this article, we propose a novel methodology for InternalSelf-Healing (ISH) that allows exploiting self-healing within a computing element internally without requiring a paired, parallel module, which extends the applicability to irregular/asymmetric datapaths while relieving the restriction of multiples of two for modules in a given datapath, as well as going beyondsquare functions. We employ our ISH methodology to design an approximate multiply-accumulate (xMAC), wherein the multiplier is regarded as an approximation stage and the accumulator as a healing stage. We propose to approximate a recursive multiplier in such a way that a near-to-zero average error is achieved for a given input distribution to cancel out the error at an accurate accumulation stage. To increase the efficacy of such a multiplier, we propose a novel 2 × 2 approximate multiplier design that alleviates the overflow problem within an n × n approximate recursive multiplier. The proposed ISH methodology shows a more effective quality-efficiency trade-off for an xMAC as compared to the conventional error-restricted methodologies for random inputs and for radio-astronomy calibration processing (up to 55% better quality output for equivalent-efficiency designs).

AB - Approximate computing studies the quality-efficiency trade-off to attain a best-efficiency (e.g., area, latency, and power) design for a given quality constraint and vice versa. Recently, self-healing methodologies for approximate computing have emerged that showed an effective quality-efficiency tradeoff as compared to the conventional error-restricted approximate computing methodologies. However, state-of-the-art self-healing methodologies are constrained to highly parallel implementations with similar modules (or parts of a datapath) in multiples of two and for square-accumulate functions through the pairing of mirror versions to achieve error cancellation. In this article, we propose a novel methodology for InternalSelf-Healing (ISH) that allows exploiting self-healing within a computing element internally without requiring a paired, parallel module, which extends the applicability to irregular/asymmetric datapaths while relieving the restriction of multiples of two for modules in a given datapath, as well as going beyondsquare functions. We employ our ISH methodology to design an approximate multiply-accumulate (xMAC), wherein the multiplier is regarded as an approximation stage and the accumulator as a healing stage. We propose to approximate a recursive multiplier in such a way that a near-to-zero average error is achieved for a given input distribution to cancel out the error at an accurate accumulation stage. To increase the efficacy of such a multiplier, we propose a novel 2 × 2 approximate multiplier design that alleviates the overflow problem within an n × n approximate recursive multiplier. The proposed ISH methodology shows a more effective quality-efficiency trade-off for an xMAC as compared to the conventional error-restricted methodologies for random inputs and for radio-astronomy calibration processing (up to 55% better quality output for equivalent-efficiency designs).

KW - Approximate computing

KW - multiply-accumulate (MAC) accelerator

KW - internal-self-healing methodology

KW - radio astronomy processing

KW - power efficiency

UR - https://ieeexplore.ieee.org/document/8727537

U2 - 10.1109/ACCESS.2019.2920335

DO - 10.1109/ACCESS.2019.2920335

M3 - Article

JO - IEEE Access

JF - IEEE Access

SN - 2169-3536

ER -