Analog CAM Bit-Split Circuit for Higher Programmable Precision
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Analog Content Addressable Memory (aCAM) arrays face limitations in precision due to the finite and discrete programmable conductance states of memristors, restricting their ability to perform complex computations requiring higher degrees of complexity, which is a challenge in minimizing circuit hardware while maintaining computational power, efficiency, and speed.
Innovation Solution
The approach involves converting an input signal into two analog voltage signals representing the most significant and least significant bits, allowing aCAM sub-circuits to perform Boolean operations, thereby increasing the number of programmable levels from 2^M to 2^(2*M) using the same memristor precision, effectively doubling the programmable levels without significant hardware additions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional aCAM cells with discrete memristor conductance states are used, then the circuit hardware remains simple, but the measurement precision and computational complexity are limited
Solution Approach 1:
The input signal is segmented into two separate analog voltage signals representing the most significant bits and least significant bits. This segmentation allows each aCAM sub-circuit to handle a portion of the precision requirements, enabling the system to achieve higher overall precision (2^(2*M) levels) without requiring each individual memristor to provide all the precision bits, thus managing circuit complexity.
Solution Approach 2:
The patent transitions from a single-dimension precision approach (relying solely on memristor conductance states) to a two-dimension approach by separating inputs into most significant bits and least significant bits. This dimensional change in signal processing allows the system to multiply the effective precision levels while maintaining the same physical hardware constraints.
2Productivity
If more aCAM cells are added to increase computational power, then the computational capability improves, but the circuit hardware size and power consumption increase
Solution Approach 1:
The aCAM sub-circuits are designed to perform multiple functions: they can process both most significant bits and least significant bits of input signals, and they can operate in combination to achieve higher precision computations. This multi-functionality allows the same hardware blocks to be reused for different precision levels and computational tasks, increasing computational power without proportionally increasing hardware size.
3Productivity
If higher precision computations are performed, then the computational complexity increases, but the number of programmable levels remains constrained by memristor precision
Solution Approach 1:
The system performs preliminary action by converting the input signal into two separate analog voltage signals representing the most significant bits and least significant bits before processing. This pre-processing step enables the subsequent aCAM sub-circuits to work with divided precision requirements, allowing higher overall computational complexity to be achieved while each sub-circuit operates within the constraints of memristor precision.
Data Source
AI summary
Examples increase precision for aCAMs by converting an input signal (x) received by a circuit into a first analog voltage signal (V(xMSB)) representing the most significant bits of the input signal (x) and a second analog voltage signal (V(xLSB)) representing the least significant bits of the input signal (x). By dividing the input signal (x) bit-wise into the first analog voltage signal (V(xMSB)) and the second analog voltage signal (V(xLSB)), the circuit can utilize aCAM sub-circuits implementing a combination of Boolean operations to search the input signal (x) against 22*M programmable levels, where “M” represents the number of programmable bits for each aCAM sub-circuit. Thus, using similar circuit hardware, example circuits square the number of programmable levels of conventional aCAMs (which generally only have 2M programmable levels). Accordingly, examples provide new aCAMs that can carry out more complex computations than conventional aCAMs of comparable cost, size, and power consumption.