Energy Consumption Measurement Linear Interpolation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for measuring energy consumption in computers often result in incorrect readings due to measurement device update intervals being longer than the program's behavioral changes, leading to zero energy consumption being observed between measurement updates.
Innovation Solution
An information processing apparatus that stores correspondence data between acquisition time points, cumulative energy, and memory addresses, and uses linear interpolation to calculate cumulative energy for each executed part of a program, correcting for erroneous measurements by determining time points of execution and updating energy consumption values.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If measurement data is acquired at fixed intervals by a measurement device, then the measurement process is simple and periodic, but the measurement precision deteriorates when the update interval is longer than the program's behavioral change interval, resulting in incorrect energy consumption readings
Solution Approach 1:
The system performs preliminary actions by storing the correspondence between acquisition time points, cumulative energy, and memory addresses before the measurement precision problem occurs. This allows the system to later determine accurate time points for executed parts and calculate cumulative energy through linear interpolation, preventing the loss of measurement precision that would otherwise occur due to infrequent measurement updates
Solution Approach 2:
The system introduces an intermediary mechanism by storing correspondence data (acquisition time points, cumulative energy, and memory addresses) that acts as a bridge between the coarse measurement intervals and the fine program execution intervals. This intermediary data structure enables accurate energy consumption calculation for each executed part through linear interpolation, resolving the contradiction between measurement update frequency and precision
2Device complexity
If the measurement device updates at longer intervals to reduce measurement overhead, then the measurement complexity decreases, but the reliability of energy consumption data deteriorates due to missed updates and zero energy consumption observations
Solution Approach 1:
The system performs preliminary action by pre-storing the correspondence between acquisition time points, cumulative energy values, and memory addresses at which instructions are executed. This pre-stored information enables reliable determination of time points for executed parts and accurate calculation of cumulative energy through linear interpolation, ensuring data reliability without requiring frequent measurement device updates
Solution Approach 2:
The stored correspondence data serves as an intermediary that bridges the gap between infrequent measurement updates and continuous program execution. By using this intermediary data structure, the system can reliably calculate energy consumption for each executed part even when the measurement device updates at longer intervals, thus maintaining reliability while reducing measurement system complexity
Data Source
AI summary
A storage unit stores therein first information indicating the correspondence among each acquisition time point at which a result of measuring cumulative energy consumed during the execution of a program was acquired, the cumulative energy, and the memory address of an instruction executed at the acquisition time point. The storage unit stores therein second information indicating the correspondence between each executed part of the program and a range of the memory addresses of instructions of the program. A computation unit determines time points in the execution of a predetermined executed part, on the basis of the first and second information, and calculates cumulative energy for each of the time points with linear interpolation.