Urgent Info

  • There is no Urgent Info now.

GSIC

Addr. 2-12-1 O-okayama, Meguroku, Tokyo
152-8550 JAPAN
Email Contact this mail address.

GSIC Activities at SC16

SC16 TOP News Release Presentations Booth Talk Schedule Booth Posters Booth Map

News Release!

Visit Tokyo Tech Booth #3424 at SC16!

logo

The Tokyo Tech Booth, #3424, will showcase the latest results form TSUBAME2.5 as well as ongoing research towards TSUBAME3.0 and beyond to exascale. Tokyo Tech's TSUBAME2.5 is an upgrade to TSUBAME2.0, that was ranked 4th on the Top500 in Nov. 2010 and awarded the "Greenest Production Supercomputer in the World" from the Green500. By replacing all the 4000+ GPUs with the NVIDIA Tesla K20x, TSUBAME2.5 performance is now boosted to 5.76 Petaflops peak in double precision FP, and moreover, 17 Petaflops peak in single precision FP, making it the fastest supercomputer in Japan for that metric. We are also aggressively pursuing various R&D towards TSUBAME3.0 to be deployed 2017, and in particular TSUBAME-KFC, our prototype for advanced cooling and low power supercomputing, should be very competitive in the Nov. 2013 Green500 rankings.
Visit Tokyo Tech's booth to find out new innovations to exascale!

+ SC16 : The International Conference for High Performance Computing, Networking, Storage and Analysis
+ Conference Dates : Nov. 13 - 18, 2016
+ Exhibition Dates : Nov. 14 - 17, 2016

Prof. Matsuoka's Photo

Tokyo Tech is operating TSUBAME2.5 supercomputer, which enjoys the peak performance of 5.76PFlops with 4224 NVIDIA Tesla K20X GPUs and exhibits world's top-class power-performance ratio. Its successor, TSUBAME3.0 will be deployed in 2017, which is designed not only as a "post-peta" supercomputer with the peak performance around 20PFlops, but as a "big-data" supercomputer, with deeper storage and memory hierarchy. We are aggressively pursuing various R&D towards TSUBAME3.0 including system and software boosting big-data, deep learning, and AI applications. Visit the booth to find out!

TSUBAME ESJ

The TSUBAME e-science journal showcases many exciting ongoing projects from a wide range of disciplines. In the latest issue Vol.14 published in 29th Feb, we focus on the multi-GPU computation and large-scale parallel computations including Deep-Learning. The researchers have been able to achieve astounding breakthroughs and gain new insight into today's most challenging HPC problems, all made possible by the power of TSUBAME.



 

PageTop