AI-Assisted Parallel Debugging and Performance Tuning

Abstract

Authors: R. Karthik Narayan, Meera Singh, Anupam Yadav, S. Lavanya Rao

Abstract: Parallel computing has become the backbone of modern high-performance systems, enabling faster execution of scientific simulations, data analytics, artificial intelligence workloads, and real-time applications. However, debugging and performance tuning of parallel programs remain extremely challenging due to non-determinism, complex synchronization patterns, and massive concurrency. Traditional debugging tools and performance profilers often require expert knowledge and manual effort, which limits productivity and scalability. Recently, artificial intelligence (AI) and machine learning (ML) techniques have emerged as promising solutions to assist developers in identifying bugs, analyzing performance bottlenecks, and recommending optimizations in parallel programs. This paper presents a comprehensive review of AI-assisted parallel debugging and performance tuning techniques. We discuss the challenges inherent in parallel software development, survey existing AI-based approaches for bug detection, root-cause analysis, and performance optimization, and analyze their applicability to shared-memory, distributed, and heterogeneous computing environments. Furthermore, we highlight recent trends, open research issues, and future directions in this rapidly evolving domain. The study aims to provide researchers and practitioners with a clear understanding of how AI can improve reliability, efficiency, and developer productivity in parallel computing systems.

Keywords: Parallel Computing, Debugging, Performance Tuning, Artificial Intelligence, Machine Learning, High-Performance Computing