Adastra GPU Hackathon at CINES: First Session Highlights and Next Steps
The CINES recently hosted the first session of its Adastra GPU Hackathon
A strong launch for the first session
Held from March 24 to 27 at CINES, the inaugural session brought together a diverse group of experts and scientific teams over four intensive days.
This first edition gathered 6 selected projects, spanning HPC and AI, with 28 participants, including:
-
- 10 code owners
- 5 AMD experts
- 4 HPE experts
- 9 CINES experts
The hackathon focused on porting, optimizing, and scaling scientific applications on AMD GPU architectures, leveraging the capabilities of the Adastra supercomputer. Close collaboration between domain scientists and hardware/software experts enabled rapid progress, knowledge transfer, and concrete performance improvements.
This initiative is part of a broader effort led by CINES and GENCI to support the transition of scientific communities toward heterogeneous computing. By combining hands-on expertise with targeted support, the hackathon format accelerates the adaptation of key simulation and AI workloads to next-generation architectures.
Upcoming sessions
Following the success of this first session, two additional hackathons are planned:
- Second session (April 21–24, CINES)
Organized in collaboration with the Maison de la Simulation and the CEA, this session will focus on 5 projects using Kokkos, a key programming model for performance portability - Third session (June, dates TBC)
Currently being organized, this session will involve 6 new projects, continuing the momentum and expanding the community of GPU-enabled applications
With these successive sessions, the Adastra GPU Hackathon is establishing itself as a cornerstone initiative to foster collaboration between researchers, HPC centers, and technology providers. It not only delivers immediate performance gains but also contributes to building sustainable expertise around GPU programming in the European scientific ecosystem.
Why Work with us?
Joining our community means becoming part of a strategic French AI ecosystem focused on innovation, sovereignty, and ethical development.
You’ll also plug into a network of AI Factories across Europe, where collaborative projects, shared models, and coordinated research can give your work pan-European impact.
Fill in this form to contact us
Supported by



















