Workshop on Parallel Programming with oneAPI
Introduction: The workshop "Parallel Programming with oneAPI" was held on February 10th at the
Complutense University of Madrid, Spain. The goal of the workshop was to introduce attendees to
parallel programming and the Intel oneAPI environment for heterogeneous systems (clusters, GPUs,
etc.).
The day began in a presentation room where the attendees had the opportunity to meet each other
and have a light meal. Afterwards, an introduction to oneAPI was given. After the introduction, the
attendees moved to the computer lab where they had the opportunity to test and work with some
oneAPI codes.
Supervised Parallelization: The challenge was preceded by a short period of "supervised
parallelization", where the attendees had time to begin working on their parallelization. After this,
they continued their work offline until the agreed delivery date.
Challenge: The highlight of the workshop was the challenge, where each participant received a
sequential version of a code and was tasked with optimizing and parallelizing it to improve
performance. The attendee with the fastest code would win. The attendees were allowed to use
multi-core CPUs and GPUs, among other resources, to make their code outperform the others.
Conclusion: The "Parallel Programming with oneAPI" workshop was a great success, with nearly 30
attendees participating. The attendees left with a deeper understanding of parallel programming and
the Intel oneAPI environment. The hands-on approach allowed the attendees to apply their
knowledge and work on real-world problems. The challenge added an element of fun and
competition, making the workshop both educational and enjoyable. The workshop content can be
found at https://github.com/garsanca/FDI_SemanaInformatica23.
