This research introduces a framework that connects Large Language Model agents to high-performance computing (HPC) resources, enabling AI systems to tackle complex scientific problems at scale.

• Implements Parsl within LangChain/LangGraph to enable parallel execution of tool functions
• Enables LLM agents to access and utilize distributed computing resources automatically
• Demonstrates practical applications in scientific domains requiring intensive computation
• Successfully tested on both local workstations and HPC environments

This innovation significantly enhances LLMs' ability to solve computationally intensive scientific problems by providing seamless access to powerful computing infrastructure—bridging the gap between AI capability and computational requirements.

Connecting Large Language Model Agent to High Performance Computing Resource