The Engineering Department at Asco Industries uses a bullx supercomputer. For Bull, this represents its third HPC customer in the Belgian aerospace industry, along with Sonaca and Ceanero.
Asco – the Belgium-based company that has specialized for almost 60 years in manufacturing spare parts and materials for the aeronautical industry – has chosen a bullx supercomputer to run its simulations.
For its commercial and R&D projects for the aerospace industry, the design office at Asco has to carry out many extremely complex calculations. However, for a number of years it did not have enough processing power for the job. “Some calculations were taking over 20 hours to run,” recalls Marc Mineur, Group Leader Stress Engineering at Asco. “What’s more, the existing configuration didn’t allow us to instigate calculations remotely, for example from a partner or customer site.”
Asco chose the bullx platform from Bull for its performance and effectiveness.
“We were looking for enough processing power, without needing to over-specify the system,” Marc Mineur adds. “In this respect, the modular nature of bullx is a real advantage.” Asco was also impressed by Bull’s sector knowledge: two other Belgian companies in the aerospace sector, Sonaca and Cenaero, have also chosen Bull solutions.
Serviware – Bull’s subsidiary specializing in High-Performance Computing (HPC) solutions – is responsible for the installation and maintenance of the bullx system.
A significant proportion of calculations now run at least three times faster than before. Asco has also been able to carry out simulations which would have been impossible on its previous system (for example, to determine the impact of bird strikes on aircraft components).
“We are delighted that Asco has put its trust in our bullx supercomputer,” comments Saskia Van Uffelen, CEO of Bull Belgium. “This clearly demonstrates that our offerings are very well suited to the aerospace sector, with its exponentially growing demand for processing power.”
The current bullx configuration consists of 4 CPUs (with 12GB of RAM and 2 external disks of 300GB) to manage the calculations, along with 16 CPUs (with 48GB of RAM, 256GB of SSD storage and 5 additional SSD scratch disks of 256GB each) to actually run them. The entire system runs under RedHat Linux.
“For the time being, we have enough processing power; but we can easily increase it if we expand our engineering team,” concludes Marc Mineur.