Graphics Processor Power Prediction via Leakage and Switching Estimation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for adjusting power consumption in computing devices, such as smartphones, are inefficient due to the lack of real-time current data, requiring frequent and lengthy iterations to balance power consumption and performance, and are too simplistic for nuanced adjustments across different applications or scenarios.
Innovation Solution
A power consumption prediction method that estimates overall power consumption based on a single frame, incorporating leakage and switching power consumption, using factors like signal toggle rate, logic gate number, voltage, frequency, and switching power consumption per gate transition, to provide real-time adjustments.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If power consumption is measured using a power meter, then power consumption data can be obtained, but the measurement is not real-time because analog signal current cannot be tracked or detected in real-time
Solution Approach 1:
The patent creates a computational model that copies the essential characteristics of power consumption behavior. Instead of directly measuring current in real-time, the system uses a power consumption model that replicates the relationship between performance parameters and power consumption, allowing real-time estimation without physical current tracking.
Solution Approach 2:
The patent replaces the physical measurement system (power meter measuring analog current) with a computational system. The power consumption model substitutes the mechanical/electrical measurement process with mathematical calculations based on performance data, enabling real-time power consumption determination without physical current detection.
2Manufacturing precision
If engineers frequently repeat the performance testing process to balance power consumption and performance, then accurate power consumption adjustment can be achieved, but the adjustment procedure becomes rather lengthy and inefficient
Solution Approach 1:
The patent performs preliminary actions by pre-establishing the power consumption model with the relationships between performance parameters and power consumption. This preliminary work allows subsequent power consumption adjustments to be made quickly through model calculations rather than repeated full performance testing cycles, significantly improving adjustment efficiency while maintaining accuracy.
Solution Approach 2:
The patent implements a feedback mechanism where the power consumption model continuously receives performance parameter data and provides real-time power consumption estimates. This feedback loop enables engineers to make informed adjustments without repeatedly running full performance tests, reducing the number of iterations needed while maintaining adjustment precision.
3Adaptability or versatility
If power consumption is adjusted according to switching among applications or scenes, then power consumption can be managed, but the adjusting method is too simple and rough for nuanced adjustments
Solution Approach 1:
The patent applies local quality by making the power consumption model application-specific and parameter-specific. Different applications and scenarios have their own characteristic performance parameters and power consumption relationships. The model can be customized or configured for different application types, providing precise power consumption estimation tailored to each specific use case rather than a generic one-size-fits-all approach.
Data Source
AI summary
A computing device, a power consumption prediction method thereof, and a non-transitory computer-readable storage medium are provided. In one embodiment, leakage power consumption of a graphics processor is obtained. Switching power consumption data corresponding to the graphics processor running a frame of image is obtained. Switching power consumption is estimated according to the switching power consumption data. Overall power consumption of the graphics processor is obtained according to the leakage power and the switching power consumption. Overall power consumption of the graphics processor processing one frame of image is estimated based on the overall power consumption. Power consumption performance of the graphics processor is therefore predicted in real-time.