Accurate and Energy-Efficient Approximate Dadda Multiplier for Resource-Constrained Environments
Year: 2026
Type: conference-paper
Abstract: Energy-efficiency multipliers using approximate computing have been extensively investigated as an interesting paradigm for Error-Resilient Application (ERA). Approximate techniques are used to efficiently configure hardware resources while maintaining acceptable computational accuracy. Since multiplication plays a pivotal role in ERAs, herein, two approximate multipliers are designed by leveraging developed 4:2 approximate compressors with a low error rate, targeting to achieve a favorable trade-off between accuracy and energy consumption. Hence, the compressors are implemented into an 8x8 Dadda structure in two configurations, one for achieving a competitive accuracy with low power. The second aims to reduce power consumption while maintaining computational accuracy. These designs are employed in smoothing image processing, achieving high SSIM values of (0.994 and 0.941) and energy of (6.19 nJ and 3.63 nJ) for designs one and two, respectively. In short, the developed approximate multipliers are suitable for resource-limited environments, offering low power with a competitive computational accuracy.
Language: English
Subfield: Electrical and Electronic Engineering
Field: Engineering
Domain: Physical Sciences
Open Access status: hybrid
OpenAlex ID: W7204201783