Design of canonical signed digit multiplier using spurious power suppression technique adder

Abstract

Abstract Reducing power consumption is a major challenge in developing integrated processors for smart portable devices. This is particularly important for extending battery life and ensuring extended usage of the device. However, some DSP processing applications involve complex algorithms that consume more power, which poses a significant challenge in designing DSP applications for VLSI circuits. To address this issue, low-power consumption methodologies are required. Although various strategies have been developed to reduce power consumption, they have not demonstrated a significant decrease in dynamic power consumption, which is the primary factor determining the total amount of power dissipation. The focus of this research is to develop a low-power multiplier using the spurious power suppression technique (SPST), a method that divides the arithmetic unit into the most significant part (MSP) and least significant part (LSP) and turns off the MSP when it is not required for computation. This approach reduces dynamic power and overall power consumption of the VLSI combinational circuit. The proposed system also utilizes canonical signed digit (CSD) representation to further reduce power usage. The system was designed using Cadence design suite, and the results showed a significant reduction of 35.8% in power consumption for a 32-bit SPST-enabled CSD multiplier. The proposed system’s total power consumption is 0.561 mW. Additionally, the proposed system was used in a power and area-efficient 256-point FFT architecture, resulting in an 86.6% reduction in power consumption. This system is suitable for real-time applications such as systems that use orthogonal frequency division multiplexing.

Keywords

• SPST
• CSD
• FFT
• Canonical signed digit
• Most significant part
• Low power