Author Affiliation

Department of Electrical and Computer Engineering, Old Dominion University

Faculty Advisor/Mentor

Masha Sosonkina

Location

Virginia Modeling, Analysis and Simulation Center, Room 1201

Conference Title

Modeling, Simulation and Visualization Student Capstone Conference 2023

Conference Track

Data Science

Document Type

Paper

Abstract

This work explores collecting performance metrics and leveraging the output for prediction on a memory-intensive parallel image classification algorithm - Inception v3 (or "Inception3"). Experimental results were collected by nvidia-smi on a computational node DGX-1, equipped with eight Tesla V100 Graphic Processing Units (GPUs). Time series analysis was performed on the GPU utilization data taken, for multiple runs, of Inception3’s image classification algorithm (see Figure 1). The time series model applied was Seasonal Autoregressive Integrated Moving Average Exogenous (SARIMAX).

Keywords:

Time series, Machine learning, GPU utilization, Performance measures

Start Date

4-20-2023

End Date

4-20-2023

DOI

10.25776/0gj4-bb15

Share

COinS
 
Apr 20th, 12:00 AM Apr 20th, 12:00 AM

GPU Utilization: Predictive SARIMAX Time Series Analysis

Virginia Modeling, Analysis and Simulation Center, Room 1201

This work explores collecting performance metrics and leveraging the output for prediction on a memory-intensive parallel image classification algorithm - Inception v3 (or "Inception3"). Experimental results were collected by nvidia-smi on a computational node DGX-1, equipped with eight Tesla V100 Graphic Processing Units (GPUs). Time series analysis was performed on the GPU utilization data taken, for multiple runs, of Inception3’s image classification algorithm (see Figure 1). The time series model applied was Seasonal Autoregressive Integrated Moving Average Exogenous (SARIMAX).