SGI powers Brazil’s ‘digital tank’ for offshore structure testing

16 processor Altix and 2.4 TB storage to power Petrobras’ MPS-based simulations.

The University of São Paulo, Brazil reports a ‘100 fold’ speed increase since it replaced its home-grown compute cluster with a 16 processor, SGI Altix shared memory machine. The computer system is used by the university and in Petrobras to speed analysis and development of offshore oil and gas production systems.

The Tank

The SGI Altix has 16 Itanium 2 processors and runs Novell SUSE Linux Enterprise 9 and SGI ProPack 4. USP also deployed a 2.4 TB SGI InfiniteStorage disk array. The system will be housed at the ‘Numerical Offshore Tank’ (TPN) laboratory at the University whose main focus is the development and analysis of Petrobras’ deepwater offshore production systems.


TPN sytem engineer Antonio Augusto Russo said, ‘We need very large amounts of memory to run our codes which scale very well on the Altix’ shared memory architecture. We also selected the Altix because of its OpenMP API which helped a lot with code development. We even use SGI’s robust storage for our financial data.’


Researchers in computational fluid dynamics (CFD) use the Moving Particle Semi-explicit (MPS) Linux toolkit to solve problems like sloshing water and fluid-solid interactions. The TPN ‘tank’ extends the MPS code to simulate floating oil production platforms.

Click here to comment on this article

Click here to view this article in context on a desktop

© Oil IT Journal - all rights reserved.