From Prompts to Performance: Evaluating LLMs for Task-based Parallel Code Generation

TL;DR

This paper evaluates how well large language models generate task-based parallel code across different frameworks and input prompts, revealing their strengths and limitations in high-performance computing contexts.

Contribution

It systematically assesses LLMs' ability to produce correct and scalable parallel code from various prompts and frameworks, highlighting areas for improvement.

Findings

LLMs can generate correct parallel code for simple problems

Performance varies significantly with problem complexity and framework

Certain frameworks are more amenable to LLM-generated code

Abstract

Large Language Models (LLM) show strong abilities in code generation, but their skill in creating efficient parallel programs is less studied. This paper explores how LLMs generate task-based parallel code from three kinds of input prompts: natural language problem descriptions, sequential reference implementations, and parallel pseudo code. We focus on three programming frameworks: OpenMP Tasking, C++ standard parallelism, and the asynchronous many-task runtime HPX. Each framework offers different levels of abstraction and control for task execution. We evaluate LLM-generated solutions for correctness and scalability. Our results reveal both strengths and weaknesses of LLMs with regard to problem complexity and framework. Finally, we discuss what these findings mean for future LLM-assisted development in high-performance and scientific computing.