Probabilistic Circuits Using MTJ Resistors for Low-Energy TRNG

Solution Overview

Problem

Existing random number generators (RNGs) for computing applications, such as Bayesian neural networks, consume excessive energy and have a large hardware footprint, especially when generating high-quality random numbers with specific probability distributions.

Innovation Solution

The development of true random number generator (RNG) systems that utilize autonomously learning probabilistic circuits to harness ambient noise, specifically leveraging the fluctuations in resistance of magnetic tunnel junction (MTJ) resistors, to generate high-quality random numbers with reduced hardware requirements.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If traditional CMOS-based RNG architectures are used to generate high-quality random numbers with specific probability distributions, then the quality and precision of random numbers are improved, but the hardware footprint and energy consumption increase disproportionately

Engineering Contradiction:

Improvequality of random numbersVSAvoidenergy consumption

Core Design Contradiction:

Measurement precisionVSUse of energy by moving object

Solution Approach 1:

The patent replaces traditional CMOS electronic circuits with magnetic tunnel junction (MTJ) based probabilistic circuits. The MTJ devices utilize quantum mechanical tunneling effects and magnetic hysteresis to generate intrinsic randomness, substituting the conventional electronic noise-based approach. This substitution enables high-quality random number generation with significantly reduced energy consumption and hardware footprint, as the magnetic system naturally provides the required stochastic behavior without extensive post-processing.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

Solution Approach 2:

The patent changes the fundamental operating parameters by using magnetic resistance states (high and low resistance states of MTJ) instead of traditional voltage or current levels. The probabilistic circuits leverage the resistance fluctuation and switching behavior of MTJ devices to generate random bits directly, eliminating the need for energy-intensive CMOS logic circuits and post-processing operations required to achieve similar quality random numbers.

Inventive Principle:

Principle #35Parameter changes

2Measurement precision

If traditional CMOS-based RNG architectures are used to generate high-quality random numbers with specific probability distributions, then the quality and precision of random numbers are improved, but the hardware footprint increases

Engineering Contradiction:

Improvequality of random numbersVSAvoidhardware footprint

Core Design Contradiction:

Measurement precisionVSArea of stationary object

Solution Approach 1:

The patent replaces traditional CMOS electronic circuits with magnetic tunnel junction (MTJ) based probabilistic circuits. The MTJ devices utilize quantum mechanical tunneling effects and magnetic hysteresis to generate intrinsic randomness, substituting the conventional electronic noise-based approach. This substitution enables high-quality random number generation with significantly reduced energy consumption and hardware footprint, as the magnetic system naturally provides the required stochastic behavior without extensive post-processing.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

Solution Approach 2:

The MTJ-based probabilistic circuits serve multiple functions simultaneously: they generate intrinsic randomness through quantum tunneling, perform probabilistic logic operations, and directly output random numbers with desired probability distributions. This multi-functionality eliminates the need for separate CMOS circuits for random bit generation and post-processing stages, reducing the overall hardware footprint while maintaining high quality random number generation.

Inventive Principle:

Principle #6Universality (Multi-functionality)

3Use of energy by moving object

If autonomously learning probabilistic circuits are used to generate random numbers, then hardware footprint and energy consumption are reduced, but the circuit complexity increases

Engineering Contradiction:

Improveenergy consumptionVSAvoidcircuit complexity

Core Design Contradiction:

Use of energy by moving objectVSDevice complexity

Solution Approach 1:

The probabilistic circuits are designed to autonomously learn and adapt their behavior to generate random numbers with specific probability distributions. The circuits self-adjust their internal parameters through built-in learning mechanisms, eliminating the need for external control logic or complex post-processing circuits. This self-service capability reduces overall system complexity while maintaining low energy consumption, as the circuits automatically optimize their operation without requiring additional energy-intensive control infrastructure.

Inventive Principle:

Principle #25Self-service

4Use of energy by moving object

If autonomously learning probabilistic circuits are used to generate random numbers, then hardware footprint and energy consumption are reduced, but the manufacturing complexity increases

Engineering Contradiction:

Improveenergy consumptionVSAvoidmanufacturing complexity

Core Design Contradiction:

Use of energy by moving objectVSEase of manufacture

Solution Approach 1:

The patent changes the fundamental operating parameters by using magnetic resistance states (high and low resistance states of MTJ) instead of traditional voltage or current levels. The probabilistic circuits leverage the resistance fluctuation and switching behavior of MTJ devices to generate random bits directly, eliminating the need for energy-intensive CMOS logic circuits and post-processing operations.

Inventive Principle:

Principle #35Parameter changes

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 enables the generation of high-quality random numbers with smaller hardware footprints compared to traditional CMOS-based solutions, and eliminates the need for post-processing, thereby reducing energy consumption and improving efficiency.

Implementation Method 1

leverage the fluctuations in resistance of magnetic tunnel junction (MTJ) resistors

Methodology Applied

Scientific EffectResistance fluctuation: Electrical Resistance

Implementation Method 2

individual p-bits comprise: a spin-orbit torque (SOT) layer; a magnetic tunnel junction (MTJ) resistor

Methodology Applied

Scientific EffectSpin-orbit torque:

Implementation Method 3

magnetic tunnel junction (MTJ) resistor

Methodology Applied

Scientific EffectMagnetic tunneling:

Data Source

PatentUS20250173124A1Generalized distribution true random number generator (TRNG) with autonomously learning probabilistic circuits

Publication Date: 2025.05.29 INTEL CORP

AI summary

Apparatus and methods for true random number generation (RNG) with a target probability distribution with autonomously learning probabilistic circuits. The apparatus utilizes, for individual probabilistic bits (p-bits), a magnetic tunnel junction (MTJ) resistor. The apparatus also uses circuitry to harness the fluctuating resistance of the MTJ resistor to generate high-quality random numbers. The hardware footprint depends on the precision required and is smaller than an equally precise or high-quality RNG implemented using CMOS hardware. Post-processing is not required on the generated random numbers.