Memory Resource Allocation Using Neural Workload Prediction

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Mobile computing devices face challenges in optimizing computing resources due to dynamic application and workload demands, leading to strained resources, battery drain, and thermal issues, especially with advancements in telecommunications networks increasing data traffic and complexity.

Innovation Solution

Implementing a neural network to predict future resource requirements based on autocorrelation time intervals, using memory device parameters to proactively allocate computing resources, including memory and processing resources, optimizing performance, device lifetime, and power consumption.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If computing resources are allocated dynamically to handle increasing data traffic and workload demands, then device performance and adaptability are improved, but resource strain and power consumption increase

Engineering Contradiction:

Improveadaptability to workload demandsVSAvoidpower consumption

Core Design Contradiction:

Adaptability or versatilityVSUse of energy by moving object

Solution Approach 1:

The patent applies preliminary action by proactively allocating computing resources based on predicted future workload demands rather than reacting to current demands. The memory system predicts resource requirements ahead of time and allocates resources in advance, preventing resource strain before it occurs. This approach allows the system to handle increasing data traffic and workload demands effectively while avoiding the energy waste associated with reactive resource allocation and the performance degradation caused by resource contention.

Inventive Principle:

Principle #10Preliminary action

2Productivity

If more computing resources are allocated to handle peak workloads, then device performance is improved, but battery life and device lifetime deteriorate

Engineering Contradiction:

Improvedevice performanceVSAvoidbattery life

Core Design Contradiction:

ProductivityVSDuration of action of moving object

Solution Approach 1:

The system performs preliminary action by predicting future workload requirements and allocating resources in advance, avoiding the need to over-provision resources continuously. This predictive approach allows the system to allocate resources only when and where needed, maintaining high device performance during peak workloads while extending battery life by avoiding unnecessary resource allocation during low-demand periods.

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The patent applies parameter changes by dynamically adjusting resource allocation parameters based on predicted workload characteristics. The system changes allocation parameters proactively rather than maintaining fixed or reactive allocation, optimizing the balance between device performance and battery life by matching resource allocation to actual predicted needs rather than peak theoretical demands.

Inventive Principle:

Principle #35Parameter changes

3Productivity

If resource allocation is optimized for current workloads, then immediate performance is improved, but ability to handle future workload variations deteriorates

Engineering Contradiction:

Improveimmediate performanceVSAvoidability to handle future workloads

Core Design Contradiction:

ProductivityVSAdaptability or versatility

Solution Approach 1:

The system resolves this contradiction by performing preliminary action through predictive resource allocation. Instead of optimizing solely for current workloads, the memory system predicts future workload patterns and allocates resources in advance, ensuring both immediate performance requirements are met and future workload variations can be handled effectively. This proactive approach maintains adaptability while delivering consistent performance.

Inventive Principle:

Principle #10Preliminary action

4Productivity

If reactive resource allocation is used to respond to workload demands, then resource utilization efficiency is improved, but response time to workload changes deteriorates

Engineering Contradiction:

Improveresource utilization efficiencyVSAvoidresponse time to workload changes

Core Design Contradiction:

ProductivityVSLoss of time

Solution Approach 1:

The patent resolves this contradiction by applying preliminary action through predictive resource allocation. The memory system predicts future workload demands and allocates resources in advance, eliminating the time delay inherent in reactive allocation systems. This approach maintains high resource utilization efficiency while significantly improving response time to workload changes, as resources are already allocated and ready when needed rather than being allocated after demand is detected.

Inventive Principle:

Principle #10Preliminary action

Data Source

PatentUS12536040B2Data sequence prediction and resource allocation

Publication Date: 2026.01.27 MICRON TECHNOLOGY INC

AI summary

A method for data sequence prediction and resource allocation includes determining, by a memory system, a plurality of resource parameters associated with operation of the memory system and determining respective time intervals associated with usage patterns corresponding to the memory system, the respective time intervals being associated with one or more sets of the plurality of resource parameters. The method further includes determining, using the plurality of resource parameters, one or more weights for hidden layers of a neural network for the respective time intervals associated with the usage patterns and allocating computing resources within the memory system for use in execution of workloads based on the determined one or more weights for hidden layers of the neural network.