High Throughput Pseudo-Random Number Generation via Automata Processor

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Current pseudo-random number generators (PRNGs) face challenges in balancing computing efficiency with the quality of randomness, often requiring impractical resources or long run times to produce high-quality randomness, and struggle with power efficiency due to traditional von Neumann architectures.

Innovation Solution

The implementation of a scalable and high-throughput PRNG using a specialized pattern-matching device with a digitized finite state machine based on probabilistic characteristics, specifically leveraging an Automata Processor (AP) to emulate non-deterministic finite automata, which provides power efficiency and performance gains by utilizing parallel Markov chains and reconfigurable state transition elements.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If traditional PRNG algorithms are used to generate high-quality random numbers, then the quality of randomness is improved, but the computing time and resource consumption increase significantly

Engineering Contradiction:

Improvequality of randomnessVSAvoidcomputing time

Core Design Contradiction:

ReliabilityVSLoss of time

Solution Approach 1:

The patent divides the random number generation process into multiple independent parallel Markov chains, each operating simultaneously to generate random numbers. This segmentation allows the system to produce high-quality randomness through multiple concurrent operations rather than sequential processing, thereby reducing total computing time while maintaining statistical quality.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The patent transitions from traditional sequential von Neumann architecture to a parallel architecture using multiple Markov chains operating in parallel dimensions. This dimensional change enables simultaneous generation of random numbers across multiple chains, significantly increasing throughput while maintaining the quality requirements through statistical testing.

Inventive Principle:

Principle #17Another dimension (Dimensionality change)

2Ease of manufacture

If traditional von Neumann architectures are used for PRNG, then implementation simplicity is maintained, but power efficiency deteriorates

Engineering Contradiction:

Improveimplementation simplicityVSAvoidpower efficiency

Core Design Contradiction:

Ease of manufactureVSUse of energy by moving object

Solution Approach 1:

The patent replaces the traditional sequential mechanical processing model of von Neumann architecture with a parallel computational model using multiple Markov chains. This substitution enables simultaneous operations across multiple chains, reducing the total energy consumption required for random number generation while maintaining implementation through standardized computational components.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

3Productivity

If PRNG output is expanded to produce many times more random output than input, then productivity is improved, but the risk of non-uniform output configurations increases

Engineering Contradiction:

ImprovethroughputVSAvoiduniformity of output

Core Design Contradiction:

ProductivityVSReliability

Solution Approach 1:

The patent incorporates statistical testing and validation mechanisms that provide feedback on the quality and uniformity of generated random numbers. This feedback loop allows the system to monitor output distributions and adjust the Markov chain parameters or reconfiguration strategies to maintain uniformity while achieving high throughput expansion.

Inventive Principle:

Principle #23Feedback

Solution Approach 2:

The patent employs reconfigurable Markov chains that can dynamically adjust their transition probabilities and state spaces based on operational requirements. This dynamic reconfiguration capability allows the system to adapt to different output requirements and maintain uniformity across expanded throughput by adjusting chain parameters in real-time.

Inventive Principle:

Principle #15Dynamics

Data Source

PatentUS9977652B2System, method, and computer-readable medium for high throughput pseudo-random number generation

Publication Date: 2018.05.22 UNIV OF VIRGINIA PATENT FOUND

AI summary

Disclosed embodiments include systems, methods, and computer-readable media for generating pseudo-random numbers. Disclosed embodiments may receive, by the at least one processor, range data indicating a range of numbers. Disclosed embodiments may generate, based on the range data and by the at least one processor, a digitized finite state machine configured to produce pseudo-random output within the range of numbers. Further, disclosed embodiments may provide, by the at least one processor to a specialized pattern-matching device, programmable instructions to implement the digitized finite state machine on the specialized pattern-matching device. Disclosed embodiments may transmit, by the at least one processor to the specialized pattern-matching device, a pseudo-random bit stream for processing by the digitized finite state machine. Disclosed embodiments may receive, by the at least one processor from the specialized pattern-matching device, pseudo-random output from the digitized finite state machine.