Trainings organization
Most of our courses are organised online.
¶
Courses are primarily designed to be completed in just one day. This format is considered a more effective way of learning than spreading the training over multiple days. The courses focus mainly on parallel programming on CPUs and GPUs and the acceleration of scientific software on supercomputers. Examples include OpenMP, MPI, OpenACC, CUDA, and OpenMP Offloading, using programming languages such as C/C++ and FORTRAN. Additionally, we also focus on Python, R, Matlab, and Julia for parallelizing applications. Regarding software acceleration, courses on GROMACS, CFD and FEM software, material science, bioinformatics, and related topics will be offered.
- The one-day training consists of three stages:
- Preparation: Pre-preparation to follow up on the lecture and practicals
- Lecture: Introduction and focus on the course topic
- Practicals: Hands-on session
¶
Bootcamps are usually one- or two-day events designed to teach scientists and researchers how to quickly start accelerating codes on modern processors (such as GPUs). Participants will be introduced to available libraries, programming models, and platforms. They will learn the basics of parallel programming (CPU, GPU, and hybrid) through extensive hands-on collaboration using real-life codes based on the parallel programming model.
¶
Hackathons are held over a longer duration, typically lasting one or two months. During the hackathons, participants focus primarily on HPC-related problems (HPC, AI, and HPDA) from industry, particularly local Luxembourg industries. Participants will be divided into multiple groups, and each group will be assigned mentors from the organizers (such as Nvidia) as well as experts from the Supercomputing NCC Luxembourg. At the end of the event, a winner will be selected, and a prize will be awarded for outstanding contributions.
