Analysis of ALU Performance with Dynamic Voltage and Frequency Scaling Technique and Clock Gating
Abstract The Arithmetic Logic Unit (ALU) is a critical component of modern mad processors that handles arithmetic and logic operations. As power efficiency becomes more important in both high-performance and energy-constrained computing systems, Dynamic Voltage and Frequency Scaling (DVFS) and Clock Gating was emerged as approach for lowering power consumption. Therefore, this paper presenting the ALU design in both techniques, while maintaining acceptable performance. This paper shows the design finding of ALU performance with DVFS technique and Clock Gating technique. All this design was done using Synopsys simulation The study investigates how voltage and frequency variations affect the ALU’s operational efficiency, power consumption, and overall processing speed. This study compares the effectiveness of three power-saving techniques, Dynamic Voltage and Frequency Scaling (DVFS), Clock Gating and Hybrid DVFS-Clock Gating Technique against a baseline with no power-saving techniques applied, in reducing design power consumption. The design power comparison shows that while DVFS increases the power consumption slightly to 53.66 μW, Clock Gating significantly reduces it to 32.28 μW compared to the baseline consumption of 52.30 μW. Meanwhile hybrid technique DVFS and Clock Gating shows good improvement with 54% of power reduction and time saving is 21% The power reduction comparison reveals that DVFS provides only a marginal 2.59%. However, the hybrid approach increases the complexity of design verification and testing. Ensuring the correctness of power state transitions and their impact on performance is more challenging than using a single technique.
Authors
- Nafees Ahmad (ORCID: https://orcid.org/0000-0002-6202-4345)
- Noraini Marsi (ORCID: https://orcid.org/0000-0002-2973-4775)
- Hasliza Hassan (ORCID: https://orcid.org/0000-0001-6870-5398)
- Warsuzarina binti Mat Zubadi
- Mohd Shahrul Izwan Akhball
Institutions
- Malaysia Genome Institute (MY)
- Tun Hussein Onn University of Malaysia (MY)
Publication Details
- Journal
- Journal of Physics Conference Series
- Published
- 2025-05-01
- DOI
- https://doi.org/10.1088/1742-6596/3020/1/012005
- Primary Topic
- Advanced Memory and Neural Computing
- Type
- article
- Field-Weighted Citation Impact
- 0.00