How do you effectively upgrade your university supercomputer when you have no additional power to work with? Martyn Williams writes that Satoshi Matsuoka from Tokyo Institute of Technology’s found himself faciing this challenge and the result was Tsubame 2.0, a GPU-powered super that proves that HPC can go hand-in-hand with energy efficiency.
“Our CIO said, you guys can build a great machine, but you’re not going to get any more electricity,” said Satoshi Matsuoka, director of the Global Scientific Information and Computing Centre at the university, of a discussion he had during the planning stages for Tsubame 2.0.
Working with HP and Nvidia, Matsuoka’s team architected Tsubame 2.0 to have 1,408 computing nodes, each with an HP ProLiant SL390 server with Intel Xeon processor and Nvidia Tesla GPUs. With three Tesla chips inside each of Tsubame’s 1,408 nodes and each chip has 448 processing cores, the total system comprises almost 1.9 million graphics processing cores.
Tsubame 2.0 is ranked #4 in the TOP500 with a sustained maximum performance of 1.19 Petaflops and #2 on the Green500 with an energy efficiency of 958 Megaflops per watt. It was the only computer to feature in the top five of both rankings.