The Ins and Outs of the World's Fastest Supercomputer

While processors are the obvious place to look when attempting to understand how the world's fastest supercomputers can deliver such speed, two very crucial factors that contribute to performance are the storage resource and the file system.

The architects of Lawrence Livermore National Laboratory's Sequoia supercomputer — currently the world's fastest supercomputer — needed a storage and file system that could top Japan's 10-petaflop K Computer.

LLNL selected NetApp's High Performance Computing Solution for Lustre, which combines the open-source Lustre parallel file system with NetApp's scalable rack storage solutions.

Thanks in part to the new upgrade, Sequoia can now process roughly 1.3 terabytes of data per second — resulting in a 16.3 petaflop peak performance capability.

In this video produced by InsideHPC, Bryan Berezdivin from NetApp describes the company's recent storage deployment for Sequoia.

NetApp has a track record of building out storage facilities for life science informatics cores, including the Duke Institute for Genome Sciences and Policy and the Stanford Genome Technology Center at Stanford University.

Matthew Dublin is a senior writer at Genome Technology.