The EuroCC & CASTIEL2 “Training Sprint” initiative is designed to quickly disseminate the expertise and abilities of CoEs. These training events target individuals from academia and industry who stand to gain from high-performance computing.
March 10 – 12, 2025
Data analytics for engineering data using machine learning
This course organized by EXCELLERAT-p2 CoE, in collaboration with: NCC Germany
This three-day online workshop addresses the preparation, analysis and interpretation of numerical simulation data by machine learning methods. Besides the introduction of the most important concepts like clustering, dimensionality reduction, visualization and prediction, this course provides several practical hands-on tutorials using the python libraries numpy, scikit-learn and pytorch as well as the SCAI DataViewer.
January 25, 2025
Extrae and Paraver: Profiling Weather and Climate Applications
This course organized by ESiWACE3 CoE, in collaboration with: NCC Austria
The National Competence Centre for HPC in Austria is organizing this one-day training together with ESiWACE3, the Centre of Excellence in simulation of weather and climate in Europe. With expertise in earth system modelling, software design, and high-performance computing, ESiWACE3 is supporting scientists on the path to exascale computing.
This course is about the BSC performance analysis tools, which are a set of different tools to support developers on the evaluation, tuning, and optimization of their codes. There are two main applications: Extrae, devoted to generate trace-files for a post-mortem analysis without modifying the source code, and Paraver, which is a flexible performance analyzer of the Extrae traces. They support different programming models such as C/C++, Fortran, Python, MPI, and OpenMP, as well as tracing of hardware counters, user functions, etc. In this course we will demonstrate the use of the tools in profiling weather and climate models, as an example, but they can be applied to any field of HPC.
The course consists of a theoretical part and hands-on labs for participants to practice how to analyze an Extrae trace with Paraver. We will use a climate application as a demonstration, focusing on typical problems arising to numerical models and computational fluid dynamic applications.
