Connecting Large Language Model Agent to High Performance Computing Resource

2025-02-17

AI Summary

This work addresses the challenge of deep integration between large language model (LLM) agents and high-performance computing (HPC) resources. We propose a novel architecture that tightly couples the Parsl workflow system with the LangChain/LangGraph tool-calling framework, enabling LLMs to autonomously generate, schedule, and execute parallelized scientific computing tasks—such as multi-condition protein molecular dynamics simulations—directly from natural language instructions. We introduce two parallel execution paradigms tailored to distinct computational scales: queue-based concurrent invocation for local clusters and ensemble-functional supercomputing-optimized invocation for large-scale HPC systems. End-to-end validation is conducted on both a local workstation and the ALCF Polaris supercomputer. Results demonstrate that our approach enables LLMs to autonomously orchestrate heterogeneous HPC resources, manage concurrent workloads, and achieve efficient resource utilization—thereby significantly advancing automation and accessibility in scientific computing.

Technology Category

• Machine Learning: Large Multimodal Models (LMMs)
• Planning, Routing, and Scheduling: Planning with Language Models
• Multiagent Systems: Agent Communication

Application Category

• Economics, Online Markets and Human Computation: Cost models of using LLMs in production systems
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Search and Retrieval-Augmented AI: Search Tool Learning with LLM: Teaching LLMs to invoke search and make use of retrieved information

Abstract

The Large Language Model agent workflow enables the LLM to invoke tool functions to increase the performance on specific scientific domain questions. To tackle large scale of scientific research, it requires access to computing resource and parallel computing setup. In this work, we implemented Parsl to the LangChain/LangGraph tool call setup, to bridge the gap between the LLM agent to the computing resource. Two tool call implementations were set up and tested on both local workstation and HPC environment on Polaris/ALCF. The first implementation with Parsl-enabled LangChain tool node queues the tool functions concurrently to the Parsl workers for parallel execution. The second configuration is implemented by converting the tool functions into Parsl ensemble functions, and is more suitable for large task on super computer environment. The LLM agent workflow was prompted to run molecular dynamics simulations, with different protein structure and simulation conditions. These results showed the LLM agent tools were managed and executed concurrently by Parsl on the available computing resource.

Problem

• Bridging LLM agents to HPC resources
• Enabling parallel scientific computations
• Optimizing molecular dynamics simulations

Innovation

• LLM agent invokes tool functions
• Parsl bridges LLM to HPC
• Parallel execution enhances performance