Enhancing Code Quality and Efficiency in Programming Education and HPC

The current research landscape in programming education and high-performance computing (HPC) is witnessing significant advancements, particularly in the areas of code quality improvement, automatic code translation, and logic error detection. In programming education, there is a growing focus on identifying and addressing anti-patterns in student code, which not only enhances code quality but also prepares students for real-world coding challenges. This trend underscores the need for targeted educational interventions to improve coding practices among novices. In the realm of HPC, the integration of High-Level Synthesis (HLS) with FPGA technology is emerging as a promising solution for high-speed data processing, addressing the increasing demands of particle physics experiments. Additionally, the development of advanced neural models like CodeRosetta is pushing the boundaries of unsupervised code translation, particularly in translating between standard programming languages and their HPC extensions, showcasing improvements in both translation accuracy and compilation efficiency. Furthermore, the combination of logic-based techniques with Large Language Models (LLMs) is revolutionizing automatic debugging and repair of logic programs, offering new tools like FormHe that significantly enhance fault detection and program repair capabilities. These innovations collectively highlight a shift towards more automated, efficient, and accurate solutions in both educational and practical coding environments.

Sources

Anti-patterns in Students' Conditional Statements

Architectural Solutions for High-Speed Data Processing Demands of CERN LHC Detectors with FPGA and High-Level Synthesis

CodeRosetta: Pushing the Boundaries of Unsupervised Code Translation for Parallel Programming

Combining Logic with Large Language Models for Automatic Debugging and Repair of ASP Programs

Logic Error Localization in Student Programming Assignments Using Pseudocode and Graph Neural Networks

A test-free semantic mistakes localization framework in Neural Code Translation

Repository-Level Compositional Code Translation and Validation

Leveraging Large Language Models for Code Translation and Software Development in Scientific Computing

Built with on top of