Approximate FIR Filter Adders for Low-Power Accuracy Control
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional FIR filter circuit designs face challenges in reducing power consumption while maintaining acceptable accuracy, as they often focus on reducing adder steps or using approximate adders without an automated synthesis flow, leading to increased energy efficiency but with errors along the critical path.
Innovation Solution
The method involves replacing conventional adders with approximate adders and employing a synthesis flow that divides the numeric columns of these adders into accurate and inaccurate parts, optimizing the boundary position for low energy consumption and configuring the approximate bits to maintain appropriate accuracy and reduce energy consumption.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional adders are used in FIR filter design, then accuracy is maintained, but power consumption increases
Solution Approach 1:
The adder is segmented into two distinct parts: an accurate part that processes significant bits and an inaccurate part that processes less significant bits. This segmentation allows different portions of the addition operation to be handled with different levels of precision, reducing overall power consumption while maintaining necessary accuracy for the FIR filter output.
Solution Approach 2:
Different parts of the adder are assigned different quality levels of computation. The accurate part uses full precision addition for bits that significantly impact the output, while the inaccurate part uses approximate addition for bits with lesser impact. This local differentiation of computational quality optimizes the trade-off between accuracy and power consumption.
2Loss of energy
If approximate adders are used to reduce power consumption, then energy efficiency improves, but errors along the critical path increase
Solution Approach 1:
The critical path operations are performed using the accurate part of the adder, ensuring that operations affecting timing and reliability are executed with full precision. Non-critical path operations can utilize the inaccurate part, allowing power reduction without compromising the reliability of critical signal paths.
Solution Approach 2:
By separating the adder into accurate and inaccurate parts, the design can selectively apply approximation only where it does not critically affect the output. The accurate part handles operations that would otherwise propagate errors through the critical path, while the inaccurate part handles operations where approximation is tolerable.
3Device complexity
If the number of adder steps is reduced to simplify FIR filter design, then device complexity decreases, but manufacturing precision may be compromised
Solution Approach 1:
Multiple addition operations are merged into a single approximate adder structure that processes multiple inputs simultaneously. The accurate and inaccurate parts work together in parallel to produce the final sum, reducing the total number of sequential adder steps while maintaining acceptable precision through the combination of different precision levels.
Solution Approach 2:
The precision parameter of the adder is changed dynamically based on the significance of the input bits. By adjusting the precision level according to bit position and input characteristics, the design reduces the number of required adder steps while maintaining manufacturing precision where it matters most.
Data Source
AI summary
A finite impulse response (FIR) filter circuit design method using approximate computing, the FIR filter circuit design method including: replacing adders of the FIR filter with approximate adders; and performing a synthesis work according to a set approximate synthesis flow when the replacing of the adders of the FIR filter are replaced with the approximate adders is performed, wherein, in the approximate synthesis flow, a numeric column of each of the approximate adders is divided into an accurate part and an inaccurate part, and a numeric column of the inaccurate part is approximated. In the FIR filter, conventional adders/subtractors are replaced with addition/subtraction having an automated synthesis flow so that energy consumption can be reduced.