(UG 08) Improving Computing Efficiency and Reducing Carbon Footprint for Turing Cluster

Authors: Qixing Xue

Advisors: Fabricio Murai, Ermal Toto

Category:  Undergraduate – IQP

Abstract: High performance computing (HPC) is nowadays paramount for training accurate deep learning models and creating scientific breakthroughs, meanwhile its ecological impact is gaining concerns, and user education of proper cluster usage is vital for improving energy efficiency. With one prevailing misusage being computing resource underutilization, which is less observable in HPC for lacking interactive access to compute nodes, users and admins need a way to see the problem and gain awareness, which is currently in absence. This project implements a data pipeline that encompasses data collection, cluster sampling, and email and web reporting, which has helped users to better see issues in their jobs and has aided admins in better supporting users during the half-year deployment on WPI cluster.

UN SDGs:

SDG 9 –  Industry, Innovation and Infrastructure

SDG 12 – Responsible Consumption and Production

SDG 13 – Climate Action

 

Video Presentation:

Poster Presentation:

UG 08 – Improving Computing Efficiency – Xue

Leave a Reply

Your email address will not be published. Required fields are marked *