Using Both High-Level and Low-Level Languages for Teaching Parallel Programming

Title: Using Both High-Level and Low-Level Languages for Teaching Parallel Programming

Abstract: This research investigated the use of both MPI (Message Passing Interface) and OpenMP (Open Multi-Processing) in teaching parallel programming to senior undergraduate and junior graduate students. The study discussed the advantages and disadvantages of using OpenMP over message passing methods and how these drawbacks are addressed by supplementing OpenMP with additional MPI codes and projects.

Research Question: Can using both high-level and low-level languages enhance the teaching of parallel programming to senior undergraduate and junior graduate students?

Methodology: The study used a comparative approach to evaluate the effectiveness of using both MPI and OpenMP in teaching parallel programming. The authors introduced the OpenMP standard and discussed why they chose to use it in their classes. They also discussed the advantages of using OpenMP over message passing methods and how these drawbacks are addressed by supplementing OpenMP with additional MPI codes and projects. Several projects given in their class were also described.

Results: The study found that using both MPI and OpenMP in parallel programming classes can enhance the learning experience for senior undergraduate and junior graduate students. The advantages of using OpenMP include its ease of use, reduced code length, and improved readability and maintainability. The drawbacks of using OpenMP, such as increased memory usage and the need for additional MPI projects, were addressed by supplementing OpenMP with additional MPI codes and projects.

Implications: The study suggests that using both high-level and low-level languages can provide a more comprehensive learning experience for students interested in parallel programming. It also encourages the development of supplementary materials and projects to address the drawbacks of using high-level languages like OpenMP.

Conclusion: The research concluded that using both MPI and OpenMP in parallel programming classes can enhance the learning experience for senior undergraduate and junior graduate students. The study also provided insights into how to address the drawbacks of using high-level languages like OpenMP by supplementing them with additional low-level projects and materials.

Link to Article: https://arxiv.org/abs/0110058v1 Authors: arXiv ID: 0110058v1