High performance computing for subatomic physics
In the world of subatomic physics, computers are used to investigate complex issues about particle and astroparticle physics. A lot of computing capacity is demanded from such computer systems. These supercomputers are used in the National Institute for Subatomic Physics (Nikhef).
The research within Nikhef focuses on particle and astroparticle physics. Scientists and technicians work together on research into the smallest building blocks of matter and their interplay of forces. These tiny particles are studied both in collision processes at the large particle accelerators of, for example, CERN, and in interactions of highly energetic particles of cosmic origin, for example in the atmosphere or seawater.
Nikhef supports its experiments with IT components that have to be able to deliver a high performance. This regularly touches the limits of what is technically possible. Recently, a test was carried out with a standard IBM Power Linux server with an IBM Power Processor. The results in the field of parallel processing are impressive and enable Nikhef, for example, to perform real-time parallel data comparisons in the processor. With these results in mind, Nikhef decided to build a Linux cluster with Power systems, which serves as a data retrieval system for the various environments that need to be processed on the existing supercomputer. Use is made of the new Linux OpenPOWER systems. For more information: The OpenPOWER Foundation.
IBM OpenPOWER Linux
BPSolutions has developed this in close cooperation with IBM Netherlands and uses the IBM Power (Unix, IBM i) technology. At Nikhef, BPSolutions has implemented an HPC Cluster on IBM OpenPOWER Linux. It is a 250TB large cluster with forty Power8 cores. The strength of this solution lies in the use of standard components, combined with the ultimate performance of the IBM Power8 processor.
More information about high performance computing
Do you have questions about high performance computing? Please contact us. We are happy to help you.