The study of nuclear matter under extreme conditions is an active field of research employing one of the largest experimental setups in the world as we can see today at LHC-CERN. One of most important goals of this research is to gain a better understanding of formation and evolution of QGP based on the information we have from anisotropic flow in ultra-relativistic heavy ion collisions.
The entire computing infrastructure used today in the simulation and data analysis from HEP experiments is using only the “traditional” configuration with large number of CPUs.
The implementation of new algorithms using GPU infrastructure will bring a change in the development of future computing cluster architecture, most probable by the usage of hybrid configurations (CPU+GPU). This will ensure a huge increase of performance at lower power consumption and at lower prices making these computing clusters more environmentally friendly.
This research will also pave the way for the future experiments that are in preparation at FAIR-GSI both in the study of collective flow of nuclear matter as well as in the development of new numerical algorithms and cluster architectures based on the versatility of GPU.