Petrobras’ ‘Grifo 4’ cluster is now the top oil and gas performer in the latest issue of the Top500 ranking of worldwide high performance computing (HPC). The number 1 slot is held by IBM’s 16 petaflop ‘Sequoia,’ a BlueGene/Q 1.5 million core system at the Department of Energy’s Lawrence Livermore National Laboratory.
Grifo 4, built by Brazil’s Itautec, comes in at number 68 with a more modest 250 teraflops using a hybrid architecture of Nvidia Tesla 2050 graphics processing units and Intel Xeon X5670 6C CPUs.
Other oil and gas machines in the list include ENI’s 130 teraflop HPC Proliant cluster, now at number 133 and Total’s SGI Altix ICE 8200EX, a 100 teraflop machine at number 173. But both these are older machines which have been in the list for a couple of years. Few oil companies (or contractors) bother with the Top500 list. For an idea of where oil and gas HPC stands today, we have the announcement by BP (Oil ITJ April 2012) of an installed petaflop capacity in Houston (which would come in the top 20) and Total’s 2.3 SGI ICE-X (Oil ITJ February 2012) machine which is notionally at number 6.
462 out of the 500 machines on the list run some variant of Linux. Two run Windows HPC 2008 server including the Chinese ‘Dawning’ machine which continues its inexorable slide down the list, from number 10 in 2008 to number 94 today. Download the TOP500.
This article originally appeared in Oil IT Journal 2012 Issue # 6.
For more information or to comment on this topic email here.