Modified Micropipline Architecture for Synthesizable Asynchronous FIR Filter Design

TL;DR

This paper presents a synthesizable asynchronous FIR filter design based on a modified micropipeline architecture, compatible with standard design tools, demonstrating energy efficiency and superior performance over synchronous counterparts.

Contribution

It introduces a novel asynchronous FIR filter design using a modified micropipeline architecture compatible with conventional synchronous design flows.

Findings

Hardware prototype verified on FPGA

Correct functionality demonstrated

Superior performance over synchronous FIR filters

Abstract

The use of asynchronous design approaches to construct digital signal processing (DSP) systems is a rapidly growing research area driven by a wide range of emerging energy constrained applications such as wireless sensor network, portable medical devices and brain implants. The asynchronous design techniques allow the construction of systems which are samples driven, which means they only dissipate dynamic energy when there processing data and idle otherwise. This inherent advantage of asynchronous design over conventional synchronous circuits allows them to be energy efficient. However the implementation flow of asynchronous systems is still difficult due to its lack of compatibility with industry-standard synchronous design tools and modelling languages. This paper devises a novel asynchronous design for a finite impulse response (FIR) filter, an essential building block of DSP systems, which is synthesizable and suitable for implementation using conventional synchronous systems design flow and tools. The proposed design is based on a modified version of the micropipline architecture and it is constructed using four phase bundled data protocol. A hardware prototype of the proposed filter has been developed on an FPGA, and systematically verified. The results prove correct functionality of the novel design and a superior performance compared to a synchronous FIR implementation. The findings of this work will allow a wider adoption of asynchronous circuits by DSP designers to harness their energy and performance benefits.