Authors:

Ved Arora

Date Created:

Not specified

Course Title:

Professor:

Not specified

About Paper:

The remarkable expansion of high performance computing (HPC) in handling vast amounts of data has highlighted the importance of high computational power in various fields of research, spanning from quantum physics simulations to cancer research. To meet this demand, Purdue University has established an advanced HPC cluster named "Anvil," which serves as a valuable resource for thousands of researchers nationwide. With the goal of providing the best experience for users, Anvil's administration began to ask questions about how Anvil has been utilized and how they can provide better user support as well as minimize cost while optimizing performance and reliability. With these questions in mind, my partner and I decided to develop a live, interactive dashboard that could deliver real-time answers to these inquiries. With the help of XALT, a powerful tool designed to track usage on a computing cluster, we harnessed millions of data points encompassing usage on Anvil. Equipped with this extensive dataset, we applied sophisticated analytical techniques and Python workflows to generate visual responses to any questions posed by Anvil's users. To present our findings in an appealing manner, we seamlessly integrated our analytical workflows with Plotly Dash, a framework for building fully customizable data applications in Python. With the successful deployment of our comprehensive dashboard, Anvil's administrative team will gain valuable insights into the supercomputer's usage patterns, empowering them to fine-tune the cluster's performance and drive advancements in research productivity. The synergy between cutting-edge tools, analytical workflows, and live visualizations holds great potential for streamlining resource allocation, reducing costs, and maximizing Anvil's computing capabilities.

Source:

Purdue University / 2023

Topics:

No topics listed

Co-authors:

Ved Arora