Resistive Memory Crosspoint Matrix for Algebraic Problem Solving

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing mathematical calculation circuits employing resistive memories face limitations in the type of computational operations they can perform and the computational load required to solve algebraic problems, particularly in solving systems of equations and calculating eigenvectors.

Innovation Solution

A mathematical problem solving circuit is developed using a crosspoint matrix of analog resistive memories and operational amplifiers, which configures conductance values to represent mathematical problem elements, allowing for the solution of square systems of equations, matrix inversion, and eigenvector calculation through voltage measurements, enabling efficient computation without requiring multiple iterations or complex operations.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If iterative numerical techniques are used to solve systems of equations with resistive memories, then the system can handle complex algebraic problems, but the computational load increases and requires several iterations to obtain convergence

Engineering Contradiction:
Improveability to solve algebraic problemsVSAvoidcomputational load
Core Design Contradiction:
Adaptability or versatilityVSProductivity

Solution Approach 1:

The patent replaces iterative numerical computation (mechanical/electronic switching and sequential processing) with a direct physical measurement approach. By configuring the resistive memory crosspoint matrix to represent the algebraic problem and using voltage measurements at column nodes to directly obtain solutions, the system eliminates the need for iterative numerical techniques, thereby reducing computational load while maintaining the ability to solve complex algebraic problems

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

2Adaptability or versatility

If resistive memories are configured to represent mathematical problem elements, then the circuit can solve various algebraic problems including square systems of equations and eigenvector calculation, but the device complexity increases

Engineering Contradiction:
Improvetype of executable computational operationsVSAvoidcircuit configuration
Core Design Contradiction:
Adaptability or versatilityVSDevice complexity

Solution Approach 1:

The patent implements a universal computational framework where a single resistive memory crosspoint matrix configuration can represent different types of algebraic problems (square systems of equations, eigenvector calculations, etc.). By appropriately configuring the conductance values in the matrix and applying suitable input voltages, the same hardware structure solves multiple mathematical problems, thereby increasing adaptability without proportionally increasing device complexity

Inventive Principle:
Principle #6Universality (Multi-functionality)

Solution Approach 2:

The patent utilizes parameter changes in the resistive memory elements (conductance values) to represent different mathematical problem elements. By programmatically setting the conductance values of individual memory cells to correspond to matrix elements or other problem parameters, the system can reconfigure the same hardware to solve different algebraic problems, achieving versatility through parameter reconfiguration rather than structural complexity

Inventive Principle:
Principle #35Parameter changes

3Measurement precision

If multiple iterations are performed to achieve convergence in solving systems of equations, then the solution accuracy improves, but the time required for computation increases

Engineering Contradiction:
Improvesolution accuracyVSAvoidcomputation time
Core Design Contradiction:
Measurement precisionVSLoss of time

Solution Approach 1:

The patent replaces the time-consuming iterative numerical process with a direct physical measurement. By configuring the resistive memory matrix to embody the mathematical problem and measuring voltages that directly represent the solution, the system obtains results in a single computational pass rather than through multiple iterations, thereby significantly reducing computation time while maintaining solution accuracy through proper circuit design and measurement techniques

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

Applied Scientific Principles

This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.

Function Achieved in This Case

The circuit provides an efficient and approximate solution to algebraic problems, such as solving square systems of equations and calculating eigenvectors, by configuring resistive memories and operational amplifiers to represent and solve mathematical problems, reducing computational complexity and achieving quick convergence of results.

Implementation Method 1

a crosspoint matrix MG including a plurality of row conductors Li, a plurality of column conductors Cj, and a plurality of analog resistive memories Gij each connected between a respective row conductor Li and a respective column conductor Cj

Methodology Applied
Scientific EffectElectrical conduction: Conduction (electrical)

Data Source

PatentEP3688622B1Mathematical problem solving circuit comprising resistive elements
Publication Date: 2022.08.10 POLITECNICO DI MILANO
  • EP3688622B1 patent drawingFigure 1
  • EP3688622B1 patent drawingFigure 2
  • EP3688622B1 patent drawingFigure 3

AI summary

It is described a mathematical solving circuit (100) comprising: a crosspoint matrix (MG) including a plurality of row conductors (L±), a plurality of column conductors (Cj) and a plurality of analog resistive memories (Gij), each connected between a row conductor and a column conductor; a plurality of operational amplifiers (OA±) each having: a first input terminal (INu) connected to a respective row conductor (Li), a second input terminal (IN2i) connected to a ground terminal (GR) at least one operational amplifier (OAi) of the plurality being such to take the respective first input terminal (INu) to a virtual ground.