VibeCodeHPC: An Agent-Based Iterative Prompting Auto-Tuner for HPC Code Generation Using LLMs
This addresses the challenge of efficient HPC code optimization for developers, though it appears incremental as it builds on existing multi-agent and LLM methods.
The paper tackled the problem of automatically tuning HPC programs for code generation by proposing VibeCodeHPC, a multi-agent LLM-based system with iterative prompting, which achieved higher-quality code generation per unit time compared to a solo-agent configuration in a case study converting CPU-based matrix-matrix multiplication to GPU code.
We propose VibeCodeHPC, an automatic tuning system for HPC programs based on multi-agent LLMs for code generation. VibeCodeHPC tunes programs through multi-agent role allocation and iterative prompt refinement. We describe the system configuration with four roles: Project Manager (PM), System Engineer (SE), Programmer (PG), and Continuous Delivery (CD). We introduce dynamic agent deployment and activity monitoring functions to facilitate effective multi-agent collaboration. In our case study, we convert and optimize CPU-based matrix-matrix multiplication code written in C to GPU code using CUDA. The multi-agent configuration of VibeCodeHPC achieved higher-quality code generation per unit time compared to a solo-agent configuration. Additionally, the dynamic agent deployment and activity monitoring capabilities facilitated more effective identification of requirement violations and other issues.