◆SC18 Exhibition will be held in November 12-15. Please stop by our booth #1619 CCS/JCAHPC, The University of Tsukuba #939 ITC/JCAHPC, The University of Tokyo
◆We will have several booth talk sessions in each University's booth at SC18!
Booth Talk Schedule
Nov. 12 (Mon)
|19:15-19:30||Kohei Fujita Earthquake Research Institute, The University of Tokyo *Accelerating nonlinear dynamic implicit finite-element earthquake simulation on Oakforest-PACS|
|19:30-19:45||Gerhard Wellein Friedrich-Alexander Universit?t Erlangen-N?rnberg *Hardware efficient Chebyshev Filter Diagonalization at Scale|
|19:45-20:00||Yasuhiro Idomura Japan Atomic Energy Agency (JAEA) *Communication avoiding matrix solvers for extreme scale nuclear CFD simulations on Oakforest-PACS|
|20:00-20:15||Richard Gerber NERSC/Lawrence Berkeley National Laboratory *Overview of the Recent Achievements by the Cori System|
|20:15-20:30||Daisuke Takahashi Center for Computational Sciences, University of Tsukuba *Automatic Tuning for Parallel FFTs on Intel Xeon Phi Clusters|
Nov. 13 (Tue)
|13:00-13:15||Ryuichi Sakamoto Information Technology Center, The University of Tokyo * Production Hardware Overprovisioning: Real-world Performance Optimization using an Extensible Power-aware Resource Management Framework|
|13:15-13:30||Yohei Miki Information Technology Center, The University of Tokyo * Performance evaluation of gravitational octree code on Volta GPU|
Nov. 15 (Thu)
|10:30-10:45||Osni Marques Lawrence Berkeley National Laboratory (LBNL) *A Methodology for Batching Matrix Kernels in HPC Applications|
|10:45-11:00||Kengo Nakajima Information Technology Center, The University of Tokyo *Recent Activities of ITC/U.Tokyo in JHPCN towards Exascale/Post-Moore Era|
Nov. 13 (Tue)
|12:00-12:15||Daisuke Takahashi Center for Computational Sciences, University of Tsukuba *Automatic Tuning for Parallel FFTs on Intel Xeon Phi Clusters|
|12:15-12:30||Kengo Nakajima Information Technology Center, The University of Tokyo *Parallel Multigrid Method on the Oakforest-PACS|
|12:30-12:45||Osamu Tatebe Center for Computational Sciences, University of Tsukuba *Storage System of the Oakforest-PACS|
Nov. 14 (Wed)
|11:15-11:30||Kazutomo Yoshii Mathematics and Computer Science, Argonne National Laboratory *The path toward heterogeneous, reconfigurable high-performance computing|
|11:30-11:45||Taisuke Boku Center for Computational Sciences, University of Tsukuba *Cygnus: GPU+FPGA cluster for multi-hybrid supercomputing|
|11:45-12:00||Jinpil Lee Programming Environment Research Team, RIKEN C-CCS *Research on a C Compiler for Stream Processing on FPGAs|
Joint Center for Advanced High Performance Computing (JCAHPC) was established in 2013 under agreement between the Center for Computational Sciences at the University of Tsukuba (CCS) and the Information Technology Center at the University of Tokyo (ITC), in order to design, operate and manage a next-generation supercomputer system provided by both two universities. JCAHPC promotes advanced computational sciences, and contributes to the fields of academia, science, and technology.
The members of JCAHPC design and develop a large-scale HPC system in which an operating system, programming languages, numerical libraries, and etc. for many-core architectures are exploited. Such research and development will proceed in establishing alliances with other organizations at the heart of the system software. Based on this design and development, two institutes will jointly procure, operate, and manage the new supercomputer. The machine which has 25 Peta FLOPS of peak performance, is scheduled to be operated around December 2016. After the system is operational, both institutions will use the system proportionally based on computation volume, operate their portions separately, and offer services to users. This arrangement not only reduces administrative costs, but also enables computing at a larger-scale than the case where each institution independently owns a supercomputer system. This effort represents the first attempt to create such a facility in Japan, as well as the first attempt in Japan to jointly operate and manage a supercomputer.
The new system is named “Oakforest-PACS” with 8,208 of next-generation manycore processors developed by Intel Corporation, and is planned to start its full operation on December 1st, 2016.