Logic-Memory Cells for Compact Stochastic Computing Circuits

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Stochastic computing is limited by the need for additional circuit elements to generate random bit streams with a desired percentage of '1' bits, leading to increased circuit size, power usage, and complexity.

Innovation Solution

The use of logic-memory cells, such as magnetic tunneling junctions and spin-Hall effect-based devices, to generate random bit streams and perform logical operations, reducing the number of circuit elements required for mathematical operations.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If additional circuit elements are used to generate random bit streams with desired percentage of '1' bits, then the accuracy of stochastic computing is improved, but the circuit size increases

Engineering Contradiction:
Improveaccuracy of stochastic computingVSAvoidcircuit size
Core Design Contradiction:
Measurement precisionVSArea of stationary object

Solution Approach 1:

The patent combines the random bit stream generation function and the logical operation function into a single logic-memory cell. The logic-memory cell stochastically transitions between low-resistance and high-resistance states to generate random bits, while simultaneously performing logical operations (AND, OR, NOT) on these bits. This merging eliminates the need for separate random number generator circuits, thereby reducing circuit size while maintaining computational accuracy.

Inventive Principle:
Principle #5Merging (Combining)

Solution Approach 2:

The logic-memory cell is designed to serve multiple functions: it acts as a random number generator by stochastically switching between resistance states, stores computational results in its memory function, and performs logical operations through controlled current flow. This multi-functionality reduces the total number of circuit elements required while preserving the accuracy of stochastic computations.

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

2Measurement precision

If additional circuit elements are used to generate random bit streams, then the accuracy of stochastic computing is improved, but the power consumption increases

Engineering Contradiction:
Improveaccuracy of stochastic computingVSAvoidpower consumption
Core Design Contradiction:
Measurement precisionVSUse of energy by stationary object

Solution Approach 1:

By merging random bit generation and logical operations into the same logic-memory cell, the patent eliminates redundant circuit elements that would otherwise consume power. The stochastic transitions occur within the intrinsic thermal noise of the device itself, requiring no additional power-hungry random number generator circuits, thus reducing overall power consumption while maintaining computational accuracy.

Inventive Principle:
Principle #5Merging (Combining)

Solution Approach 2:

The logic-memory cell utilizes its own intrinsic thermal noise and stochastic properties to generate random bits without requiring external random number generation circuits. The device serves itself by leveraging its inherent physical characteristics for random bit generation, eliminating the need for separate power-consuming randomization hardware.

Inventive Principle:
Principle #25Self-service

3Measurement precision

If additional circuit elements are used to generate random bit streams, then the accuracy of stochastic computing is improved, but the device complexity increases

Engineering Contradiction:
Improveaccuracy of stochastic computingVSAvoidcircuit complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent merges three distinct functions—random number generation, memory storage, and logical operation execution—into a single logic-memory cell structure. This consolidation dramatically simplifies the overall circuit architecture by eliminating the need for separate random number generator circuits, additional memory elements, and multiple logic gate implementations, thereby reducing device complexity while maintaining full computational capability.

Inventive Principle:
Principle #5Merging (Combining)

Solution Approach 2:

The logic-memory cell is designed as a universal component that can perform random bit generation, data storage, and various logical operations (AND, OR, NOT) within a single device structure. This multi-functionality reduces the total number of discrete components and interconnections required, significantly simplifying the overall circuit design and reducing device complexity.

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

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

This approach reduces circuit size and power consumption while maintaining accuracy in stochastic computations by using logic-memory cells to generate and execute random bit streams.

Implementation Method 1

A logic circuit includes a first magnetic tunneling junction, a second magnetic tunneling junction and a third magnetic tunneling junction

Methodology Applied
Scientific EffectMagnetic tunneling:

Implementation Method 2

spin-Hall effect-based devices to generate random bit streams

Methodology Applied
Scientific EffectSpin-Hall effect:

Data Source

PatentUS12494238B2Stochastic computing using logic-memory cells
Publication Date: 2025.12.09 REGENTS OF THE UNIVERSITY OF MINNESOTA
  • US12494238B2 patent drawing
  • US12494238B2 patent drawing
  • US12494238B2 patent drawing

AI summary

A circuit includes a first two-state device, a second two-state device and a third two-state device, each two-state device having a first resistance in a first state and a second resistance in a second state. First control elements are configured to apply a first voltage to the first two-state device to stochastically place the first two-state device in either the first state or the second state. Second control elements are configured to apply a second voltage to the second two-state device to stochastically place the second two-state device in either the first state or the second state. Third control elements are configured to send respective currents through the first two-state device and the second two-state device so as to place the third two-state device in either the first state or the second state based on the state of the first two-state device and the state of the second two-state devices.