System-Level PPA Trade-Off Analysis of Adder-Based MAC Architectures Using Vedic Multipliers on Zynq-7000 FPGA
International Journal of Innovative Science and Research Technology · 19.03.2026 · 1374
DOI: 10.38124/ijisrt/26mar690 ↗Kurz gefasst
Fachartikel in der Zeitschrift International Journal of Innovative Science and Research Technology, online erschienen am 19.03.2026. Verfasst von Subhadip Ghanta und 4 weiteren Autor:innen. Die Arbeit nennt 25 Literaturangaben.
Abstract (Originalsprache)
Multiply-Accumulate units are basic components of digital signal processing systems as well as embedded systems, where computational throughput as well as power constraints are of great importance. In this paper, a Power Performance Area trade-off of Vedic Multiplier-based Multiply-Accumulate units with five different adder styles, namely Ripple-Carry Adder, Carry-Skip Adder, Carry-Look-Ahead Adder, Brent-Kung Adder, and Kogge-Stone Adder, is presented. In this paper, a novel Multiply-Accumulate architecture based on a hierarchical 16×16 Vedic Multiplier with a 32-bit accumulation path is presented, which has been implemented on a Zynq-7000 FPGA platform based on the Zed Board platform. The architectures are tested with both 16-bit and 32-bit datapath configurations in terms of power consumption, propagation delay, Power Delay Product, as well as Energy Delay Product, along with hardware utilization. The experimental results show that the Kogge-Stone Adder-based Multiply Accumulate architecture has the least propagation delay of 3.2 ns as well as 5.2 ns for 16-bit as well as 32-bit datapath configurations, respectively, achieving the least PDP of 304 pJ as well as 728 pJ, respectively, compared to other adder styles. In contrast, the Ripple Carry Adder consumes the least power of 40 mW as well as 83 mW for 16-bit and 32-bit datapath configurations, respectively, utilizing the least hardware resources, but has the highest propagation delay of 12.5 ns and 21 ns, respectively. Among all architectures, the Brent-Kung Adder offers the best trade-off among power, delay, and hardware utilization. The adderbased Multiply-Accumulate architectures are suitable for implementing various high-performance as well as powerefficient applications, including digital signal processing, image processing, and machine learning accelerators.
Zitieren
APA
Ghanta, S., Shaw, S., Khan, A. B., Sarkar, D., & Das, S. (2026). System-Level PPA Trade-Off Analysis of Adder-Based MAC Architectures Using Vedic Multipliers on Zynq-7000 FPGA. International Journal of Innovative Science and Research Technology, 1374. https://doi.org/10.38124/ijisrt/26mar690
Vancouver
Ghanta S, Shaw S, Khan AB, Sarkar D, Das S. System-Level PPA Trade-Off Analysis of Adder-Based MAC Architectures Using Vedic Multipliers on Zynq-7000 FPGA. International Journal of Innovative Science and Research Technology. 2026:1374. doi:10.38124/ijisrt/26mar690
BibTeX
@article{Ghanta2026,
author = {Ghanta, Subhadip and Shaw, Sanket and Khan, Abdul Barish and Sarkar, Debjyoti and Das, Soumik},
title = {System-Level PPA Trade-Off Analysis of Adder-Based MAC Architectures Using Vedic Multipliers on Zynq-7000 FPGA},
journal = {International Journal of Innovative Science and Research Technology},
year = {2026},
pages = {1374},
doi = {10.38124/ijisrt/26mar690}
}
EUVIMED