The need to support various digital signal processing (DSP)and classification applications on energy-constrained devices has steadily grown. Such applications often extensively perform matrix multiplications using fixed-point arithmetic while exhibiting tolerance for some computational errors. Hence, improving the energy efficiency of multiplications is critical. In this paper, we offer a similar speed, but with energy efficiency. The method is to collect the armature close to the closest momentum of two. An integral part of the computer, so the multiplication is eliminated, improving the speed and power consumption at a small error value. The proposed approach is to apply both signed and neglected. We offer three hardware implementations of an approximate multiplier that includes not being signed and signed for both operations. The effectiveness of this proposed multiplier is estimated by comparing its effectiveness with certain approximate and real-world by using different design parameters. In addition, the effect of the proposed approximate multipliers is examined in two applications for image processing, namely sharpness of the image. Keywords-Approximate multiplier, Energy efficiency and Power consumption ,integrated circuits, DSP I. BACKGROUND Energy minimization is one of the main design requirements in almost any electronic systems, especially the portable ones such as smart phones, tablets, and different gadgets [1]. It is highly desired to achieve this minimization with minimal performance (speed) penalty [1]. Therefore, improving the speed and power/energy-efficiency characteristics of multipliers plays a key role in improving the efficiency of processors. Many of the DSP cores implement image and video processing algorithms where final outputs are either images or videos prepared for human consumptions. This fact enables us to use approximations for improving the speed/energy efficiency. This originates from the limited perceptual abilities of human beings in observing an image or a video. In addition to the image and video processing applications, there are other areas where the exactness of the arithmetic operations is not critical to the functionality of the system (see [3], [4]). Being able to use the approximate computing provides the designer with the ability of making tradeoffs between the accuracy and the speed as well as power/energy consumption [2], [5]. Applying the approximation to the arithmetic units can be performed at different design abstraction levels including circuit, logic, and architecture levels, as well as algorithm and software layers [2]. The approximation may be performed
Multiplication is a necessary arithmetic operation in signal processing and machine learning algorithms. This paper proposes a novel method called Approximate Multiplier with Configurable Accuracy Levels (AMCAL) to obtain energy and area savings. The proposed algorithm performs the multiplication operation based on truncating the least significant bits and manipulating the remaining bits. Manipulation in the remaining input bits with a reduced bit width has been done so that the error resulting from truncating the bits is neutralized or reduced as much as possible. It has also shown that we are less concerned about the magnitude of the error injected in the input compared to other algorithms. The proposed AMCAL multiplier improves energy consumption by 86.7% compared to the Wallace multiplier, with an error margin of 0.15%. Moreover, in delay, power, and area, the proposed multiplier outperforms other approximate multipliers in the same class, such as DRUM and DSM. The AMCAL multiplier, surpasses new approximate multipliers such as DSI and TOSAM in power and area efficiency. Finally, it is shown that such a minor computational error does not affect the image quality resulting from four image processing programs, namely smoothing, sharpening, JPEG encoder and face alignment.INDEX TERMS Accuracy configurable, reconfigurable approximate multiplier, low power, neural network I. INTRODUCTION
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.