GPU Voltage-Frequency Scaling by Toggle Rate for Power Efficiency

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Current methods for controlling GPU power consumption result in a low performance-to-power ratio due to voltage-frequency scaling tables derived under adverse conditions, failing to adapt to varying usage scenarios.

Innovation Solution

Determine the current toggle rate of the GPU by acquiring power consumption parameters, and use a pre-trained neural network to derive a target voltage-frequency scaling relationship, optimizing operating voltage and frequency for the current scenario.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If voltage-frequency scaling table is derived under adverse conditions to ensure proper GPU function, then reliability is improved, but performance-to-power ratio deteriorates

Engineering Contradiction:

ImproveGPU function reliabilityVSAvoidperformance-to-power ratio

Core Design Contradiction:

ReliabilityVSUse of energy by moving object

Solution Approach 1:

The patent applies dynamics by transitioning from a static voltage-frequency scaling table to a dynamic adjustment mechanism. The system continuously monitors GPU toggle rates and adjusts voltage-frequency settings in real-time based on actual usage patterns, allowing the GPU to operate at optimal performance-to-power ratios rather than fixed conservative settings

Inventive Principle:

Principle #15Dynamics

Solution Approach 2:

The patent changes the parameters of voltage and frequency dynamically based on the GPU's toggle rate. By monitoring power consumption parameters and adjusting voltage-frequency scaling according to actual workload conditions, the system optimizes the balance between reliability and performance-to-power ratio

Inventive Principle:

Principle #35Parameter changes

2Device complexity

If static voltage-frequency scaling table is used to simplify control, then device complexity is reduced, but adaptability to varying usage scenarios deteriorates

Engineering Contradiction:

Improvecontrol complexityVSAvoidadaptability to usage scenarios

Core Design Contradiction:

Device complexityVSAdaptability or versatility

Solution Approach 1:

The patent implements feedback by continuously monitoring the GPU's toggle rate and power consumption parameters. This feedback loop allows the system to automatically adjust voltage-frequency settings based on actual usage patterns, enabling adaptability to varying scenarios without requiring complex manual control mechanisms

Inventive Principle:

Principle #23Feedback

Solution Approach 2:

The system performs self-service by automatically adjusting its own voltage-frequency settings based on monitored toggle rates. The GPU management mechanism autonomously optimizes performance-to-power ratio without external intervention, balancing adaptability with reasonable control complexity

Inventive Principle:

Principle #25Self-service

Data Source

PatentUS20250315096A1Power consumption control method and apparatus for graphics processing unit, device, medium and program product

Publication Date: 2025.10.09 MOORE THREADS TECH CO LTD

AI summary

The present disclosure relates to the technical field of electronic digital data processing, in particular to a method for controlling power consumption of a graphics processing unit (GPU), an apparatus, and a program product. The method comprises: acquiring a current toggle rate of a GPU; determining a target voltage-frequency scaling relationship corresponding to the GPU currently based on the current toggle rate; and determining a target operating voltage and a target operating frequency of the GPU based on the target voltage-frequency scaling relationship.