Advanced Certificate in Developing High-Performance Software for HPC Clusters

The 'Advanced Certificate in Developing High-Performance Software for HPC Clusters' programme is designed for experienced software developers, researchers, and engineers with a foundational understanding of high-performance computing (HPC) who seek to enhance their skills in developing, optimizing, and deploying software applications for HPC clusters. This comprehensive programme equips learners with advanced knowledge in parallel programming, performance analysis, and the use of specialized tools and frameworks that are critical for high-performance software development.

Key skills and knowledge developed through this programme include proficiency in parallel programming paradigms such as MPI and OpenMP, understanding of optimization techniques for HPC environments, and hands-on experience with performance analysis tools like Intel VTune and NVIDIA Nsight. Learners will also gain expertise in managing and utilizing HPC resources effectively, including familiarity with cluster architecture, job submission workflows, and best practices for scaling applications to achieve peak performance.

The programme has a significant career impact, preparing graduates to lead projects in developing software for complex scientific, engineering, and industrial applications that require high computational power. Graduates are well-equipped to join roles such as HPC software developers, performance engineers, and technical leads in research institutions, universities, and leading tech companies, contributing to advancements in fields like climate modeling, computational biology, and artificial intelligence.

1. 1. Introduction to High-Performance Computing (HPC): Learners will study the fundamentals of HPC, including architectures, performance metrics, and basic parallel programming concepts. They will gain the ability to understand and analyze HPC systems and start writing simple parallel programs.
2. 2. Parallel Programming Models: This module explores various parallel programming models and frameworks (e.g., MPI, OpenMP, GPU programming) used in HPC. Learners will develop skills in designing and implementing parallel algorithms.
3. 3. Performance Tuning and Optimization: Learners will learn techniques for optimizing software performance, including profiling tools, memory management, and algorithm optimization. Practical skills in enhancing the efficiency of HPC applications will be developed.
4. 4. Cluster Management and Administration: This module covers the management and administration of HPC clusters, including system setup, user management, and maintenance. Learners will gain the skills needed to manage HPC resources effectively.
5. 5. Distributed Memory Programming with MPI: Focusing on MPI, learners will study distributed memory programming techniques and develop the ability to write and optimize MPI-based applications for HPC clusters.
6. 6. Shared Memory Programming with OpenMP: This module delves into shared memory programming using OpenMP, teaching learners how to write efficient multi-threaded applications that can take advantage of multi-core processors.
7. 7. GPU Programming with CUDA and OpenCL: Learners will learn how to leverage GPUs for high-performance computing through CUDA and OpenCL, developing skills to write and optimize GPU-accelerated applications.
8. 8. Advanced Parallel Algorithms: This module covers advanced parallel algorithms and data structures, providing learners with a deeper understanding of parallel computing and the ability to design and implement complex parallel applications.
9. 9. Advanced HPC Tools and Libraries: Learners will explore advanced tools and libraries commonly used in HPC, such as PETSc, Trilinos, and others, and develop skills in using these tools to solve complex computational problems.
10. 10. Project and Case Studies in HPC: In this capstone module, learners will apply their knowledge through real-world projects and case studies, working on advanced HPC challenges and developing a portfolio of HPC applications.