Sistemas Informáticos Europeos, has successfully finished installing the ICCRAM cluster, belonging to the University of Burgos.
With this expansion of the computing capacity of this Institute, the research team of Dr. Santiago Cuesta is provided with a computing power of more than 200 CPU cores and a total of 5.700 GPU cores. It is also equipped with several TB of storage capacity, to safely store the data of the results of the research of this group, which investigates within the project Icarus
The SIE Ladon equipment implemented can use all of its computing capacity, since they are interconnected through a 56 Gb/s Mellanox Infiniband network with low latency.
All the implementation has been carried out within the ecosystem LadonOS and it provides the new implementations that have been incorporated in terms of image deployment, cluster management and monitoring in Ganglia, Workload manager SLURM 16 (queue system) and other improvements.
As part of the maintenance, SIE in the coming months will implement new solutions such as Centos 7.3 and the new version of RCUDA, which will allow virtualizing the GPUs to obtain even greater performance.
At the platform level, Intel solutions have been integrated, of which SIE is HPC Data Center Specialist,
to solve the CPU implementation with the new Broadwell and capacity of 512 GB per node. For the implementation of GPUs, the SIE Ladon 8GPUs platform has been used, based on Gigabyte G250-G50 of which SIE is Gigabyte System Integrator Spain and that allow ICCRAM to grow from the 5.760 cores it already has to 23.040 cores, occupying a space of only 2U in its rack.