EUMaster4HPC Summer School 2024

The EUMaster4HPC Summer School 2024 was dedicated to the subject of High-Performance Computing in Data Science. Held from August 18 to 24 at the IT4Innovations National Supercomputing Center in Ostrava, Czech Republic, it brought together 45 students from nine European countries.

Over the week, the students participated in a variety of activities, including lectures, hands-on labs, coding challenges, quizzes, and student team competitions, learning to prepare data, understand its characteristics, and create meaningful machine learning models on HPC architectures. They also engaged in MATLAB projects, exploring large-scale HPC, deep learning, image processing, and signal analysis with real-world data on the IT4Innovations' supercomputer KAROLINA.

Reflecting on the experience, Majesa Trimmel, a student from the University of Luxembourg / Universitat Politècnica de Catalunya, shared:

"I found the Summer School to be well organized. The duration of each topic was nicely timed, which allowed us to fully absorb the material without feeling overwhelmed. The breaks kept us energized, and the social events in between were a great addition, providing an excellent opportunity to connect with fellow participants and build lasting connections. What I found particularly valuable were the sessions on R and the parallelization techniques using Python. These topics helped me gain a better understanding of scaling data analysis processes and optimizing computational tasks.

All in all, the summer school was a rewarding experience. It combined technical learning with networking opportunities, and I believe that it will benefit my future career in HPC and Data Science."

Students left the summer school with a robust set of competencies, including:

• Proficiency in using Linux-based HPC environments.

• Understanding the fundamentals of exploratory data analysis and modeling.

• Scaling data analysis processes for Big Data using R and Python.

• Building and deploying basic Machine and Deep Learning models.

• Making informed decisions between Machine Learning and Deep Learning techniques.

• Developing data processing pipelines suitable for HPC tasks.

• Implementing parallel processing to optimize Machine and Deep Learning models across multiple compute nodes and GPUs.

• Utilizing MATLAB tools for advanced HPC and data analysis tasks.