Infraestrutura
SAMPA (Advanced Analysis and Multi-Processing System) consists of a computer cluster located at the Department of Nuclear Physics of the Physics Institute at the University of São Paulo (IFUSP), funded by FAPESP and FINEP, dedicated to the intensive analysis of data and simulations of the STAR (RHIC) and ALICE (LHC) experiments, in addition to contributing to the activities of several researchers in the area at the institute and the University of Campinas (UNICAMP).
In operation since 2007 and with equipment supplied by HP, SGI, SuperMicro and Dell, currently SAMPA has the following technical features:
- 2700 CPUS's of processing, being 2376 exclusively dedicated to data analysis and processing in GRID, while the other 324 are distributed among servers, storage elements and cluster control;
- 4 GB of RAM, on average, per CPU, totaling 10 TB in the whole cluster;
- 1320 TB of storage space distributed among netwrok user areas, control, temporary space, and space dedicated to GRID data storage.
The following image represents the processing queue of the cluster; a 65x(24-48) array, where each colored cell corresponds to one job (program, simulation or analysis in process) of a specific user relative to the color.
Processing queue of the SAMPA cluster.