Power Consumption Patterns Using Telemetry Data

TL;DR

The need to understand power consumption patterns is underscored and areas where stakeholders like Intel can make improvements to reduce environmental impact effectively and efficiently are identified.

Abstract

This paper examines the analysis of package power consumption using Intel's telemetry data. It challenges the prevailing belief that hardware choice is the primary determinant of a device's power consumption and instead emphasizes the significant role of user behavior. The paper includes two sections: Exploratory Data Analysis (EDA) and a linear model for power consumption. The EDA section provides valuable insights from Intel's telemetry data, comparing power consumption across countries, with a specific focus on power consumption patterns in the US and China. Our simple linear model affirms those patterns and highlight the possible importance of user behavior and its influence on power consumption. Ultimately, the paper underscores the need to understand power consumption patterns and identifies areas where stakeholders like Intel can make improvements to reduce environmental impact effectively and efficiently.

Figures (10)

• Figure 1: Illustration of the feature matrix for the linear model. Each entry (row) is a daily record from a machine (GUID day).
• Figure 2: Median User Avaerge Power (MUAP) for countries within the dataset. Each scatter point is a country and the top 10 most represented countries are shown, with the US and China starred. China's MUAP is almost double that of the US.
• Figure 3: Histogram of User Average Power (UAP) for the US (blue) and China (red) with users in all TDP levels (left) and users between [15, 45)W (right). China's distribution of UAP is shifted to the right, even after controlling for TDP. Note that each data point in this plot is an aggregate for a single user.
• Figure 4: Average power consumption at each hour of the day for the US (blue) and China (red). China's power consumption increases as the day goes by, towards the evening, US's power consumption decreases throughout the day after peaking at around 8 am.
• Figure 5: Area under the curve of time of day plot (like Figure \ref{['fig:time']}) for each day of the week. The area under the curve represents an average user's energy consumption (Wh) on each day of the week for the US (blue) and China (red). Notice the US's area is smaller on Saturdays and Sundays, whereas China does not have such weekend patterns and the area stays consistent throughout all days.