Photo Justin kauffman on Unsplash
Welcome back to the COOP blog!
The past 18 months have brought big changes for the COOP team, and we’re now set for new successes in High-Performance Computing (HPC).
Gabriel Staffelbach moved to ONERA in February 2024, and continues to collaborate with Cerfacs on sustainable programming. At the same time, Corentin Lapeyre joined NVIDIA to get a broader perspective on scientific computing and AI. Pavanakumar Mohanamuraly moved back to India where he is now working at G.E. in Bangalore, India.
As the team evolved, new faces joined the journey. Mohamed Ghenai took over Gabriel’s work on continuous HPC technology watch, while Luciano Drozda stepped into Corentin’s AI-focused role within the Algo Team. Leadership also saw changes—by the end of 2024, Prof. Ulrich Ruede concluded his tenure as Algo-COOP leader. The teams have since split back into Algo, led by Luc Giraud, and COOP, now headed by Michael Rudgyard. Looking ahead, Prasanna Balaji will join COOP in April to continue Pavanakumar’s work.
With this new dynamic, COOP is now focusing on two key areas:
Advanced programming for HPC
We are dedicated to optimizing software for cutting-edge HPC systems. Our work spans:
- Software porting & optimization : with a strong focus on AVBP -the Large Eddy Reactive simulation solver backed by Energy & Safety Team at CERFACS- and responding to partner requests (notably SAFRAN).
- HPC software life-cycle management : leveraging tools to analyze and modernize legacy codes.
- Next-generation HPC development : pushing advancements in mesh manipulation, partitioning, and adaptation through proof-of-concept coding and continuous technology watch.
Our strategy is to lead with research, build software platforms for future HPC systems, and shift from a reactive to a proactive approach. While AVBP remains central, our focus is on its next-generation evolution (AVBP-NG), expanding our collaborations beyond Energy & Safety Team to engage with other CERFACS teams and partners.
HPC Production Environments
We continue to refine the infrastructure surrounding HPC applications, with efforts in:
- Pre- & post-processing techniques : a critical component of SAFRAN’s AVBP workflows, though the demand on our small team limits proactive development.
- Application workflows : streamlining and optimizing how HPC applications are deployed and executed.
To sustain and expand this work, we plan to strengthen the team, encourage broader use of these tools for advanced research, and enhance collaboration across CERFACS teams, codes, and partners.
Stay tuned for more updates as we continue to push the boundaries of high-performance computing. You will hear a lot about code transformation and code metrics this year!