Workshop: 8th International Workshop on Containers and New Orchestration Paradigms for Isolated Environments in HPC (CANOPIE-HPC)
Authors: Vanessa Sochat (Lawrence Livermore National Laboratory (LLNL)) and Daniel Milroy (Lawrence Livermore National Laboratory)
Abstract: Large Language Models (LLMs) are serving as a catalyst of change for research practices, influencing the daily lives of staff scientists, software engineers, and system administrators. Adopting LLM-based agents promises new degrees of autonomy, where categories of human work and decision making are replaced by goal-oriented systems. This transition necessitates novel architectural paradigms and solid understanding of the strengths and limitations of LLMs. In this work, we design agents to conduct the entire life-cycle of an HPC application experimental run in cloud -- creation and build of a container, deployment in Kubernetes, optimization, and orchestration of a scaling study. We execute the study for five well-known HPC applications to build multi-platform container images and optimize figures of merit across 21 instance types in Kubernetes. We present a successful agentic scaling study, designs that improve work time to completion, and suggested best practices for agentic design and collaboration.
Back to 8th International Workshop on Containers and New Orchestration Paradigms for Isolated Environments in HPC (CANOPIE-HPC) Archive Listing
Back to Full Workshop Archive Listing