Computational tools are an integral part of many areas of research nowadays. Also computational models and algorithms are becoming more diverse, detailed, and complex and thus require scientists to write efficient and parallel software to support their research. The education for preparing researchers to use computation in their work, however, commonly focuses on theoretical and methodological aspects of the computational research. Which results in a lack of understanding and practical skills to meet the requirements of modern computational research and utilizing an advanced computing infrastructure effectively. While modern computer hardware for HPC is becoming increasingly well accessible to scientists all around the globe, it also differs from the traditional simple monolithic picture of a single powerful machine through its diversity, flexibility, and variety of hardware configurations. Many scientific problems can only be properly addressed through using parallelism and understanding important technical aspects this hardware, which is frequently non-uniform in many ways: CPU architecture, memory access, network, and persistent storage for data. In the second part of the school more advanced topics will be addressed such as use of techniques of ML and QC on large scale infrastructures.
 

The Magurele Summer School for Computing in a Rapidly Evolving Society – “Parallel Algorithms and Optimizations” aims to familiarize the participants with the fundamentals of high-performance computing (HPC) within a parallel environment, making use of traditional CPU architecture, but also accelerators such as GPUs. The main courses focus on coding for HPC clusters in C and FORTRAN. The expected outcomes are related to enabling the participants to solve difficult computational problems using the most modern software and hardware tools available.