Approximate computing is a promising paradigm to realize fast, small, and low power characteristics, which are essential for modern applications, such as Internet of Things (IoT) devices. This paper proposes the Carry-Predicting Adder (CPredA), an approximate adder that is scalable relative to accuracy and power consumption. The proposed CPredA improves the accuracy of a previously studied adder by performing carry prediction. Detailed simulations reveal that, compared to the existing approximate adder, accuracy is improved by approximately 50% with comparable energy efficiency. Two application-level evaluations demonstrate that the proposed approximate adder is sufficiently accurate for practical use.
Toshinori SATO
Fukuoka University
Tongxin YANG
Fukuoka University
Tomoaki UKEZONO
Fukuoka University
The copyright of the original papers published on this site belongs to IEICE. Unauthorized use of the original or translated papers is prohibited. See IEICE Provisions on Copyright for details.
Copy
Toshinori SATO, Tongxin YANG, Tomoaki UKEZONO, "Trading Accuracy for Power with a Configurable Approximate Adder" in IEICE TRANSACTIONS on Electronics,
vol. E102-C, no. 4, pp. 260-268, April 2019, doi: 10.1587/transele.2018CDP0001.
Abstract: Approximate computing is a promising paradigm to realize fast, small, and low power characteristics, which are essential for modern applications, such as Internet of Things (IoT) devices. This paper proposes the Carry-Predicting Adder (CPredA), an approximate adder that is scalable relative to accuracy and power consumption. The proposed CPredA improves the accuracy of a previously studied adder by performing carry prediction. Detailed simulations reveal that, compared to the existing approximate adder, accuracy is improved by approximately 50% with comparable energy efficiency. Two application-level evaluations demonstrate that the proposed approximate adder is sufficiently accurate for practical use.
URL: https://global.ieice.org/en_transactions/electronics/10.1587/transele.2018CDP0001/_p
Copy
@ARTICLE{e102-c_4_260,
author={Toshinori SATO, Tongxin YANG, Tomoaki UKEZONO, },
journal={IEICE TRANSACTIONS on Electronics},
title={Trading Accuracy for Power with a Configurable Approximate Adder},
year={2019},
volume={E102-C},
number={4},
pages={260-268},
abstract={Approximate computing is a promising paradigm to realize fast, small, and low power characteristics, which are essential for modern applications, such as Internet of Things (IoT) devices. This paper proposes the Carry-Predicting Adder (CPredA), an approximate adder that is scalable relative to accuracy and power consumption. The proposed CPredA improves the accuracy of a previously studied adder by performing carry prediction. Detailed simulations reveal that, compared to the existing approximate adder, accuracy is improved by approximately 50% with comparable energy efficiency. Two application-level evaluations demonstrate that the proposed approximate adder is sufficiently accurate for practical use.},
keywords={},
doi={10.1587/transele.2018CDP0001},
ISSN={1745-1353},
month={April},}
Copy
TY - JOUR
TI - Trading Accuracy for Power with a Configurable Approximate Adder
T2 - IEICE TRANSACTIONS on Electronics
SP - 260
EP - 268
AU - Toshinori SATO
AU - Tongxin YANG
AU - Tomoaki UKEZONO
PY - 2019
DO - 10.1587/transele.2018CDP0001
JO - IEICE TRANSACTIONS on Electronics
SN - 1745-1353
VL - E102-C
IS - 4
JA - IEICE TRANSACTIONS on Electronics
Y1 - April 2019
AB - Approximate computing is a promising paradigm to realize fast, small, and low power characteristics, which are essential for modern applications, such as Internet of Things (IoT) devices. This paper proposes the Carry-Predicting Adder (CPredA), an approximate adder that is scalable relative to accuracy and power consumption. The proposed CPredA improves the accuracy of a previously studied adder by performing carry prediction. Detailed simulations reveal that, compared to the existing approximate adder, accuracy is improved by approximately 50% with comparable energy efficiency. Two application-level evaluations demonstrate that the proposed approximate adder is sufficiently accurate for practical use.
ER -