Fumiyoshi Shoji

Learn More
RIKEN and Fujitsu have been working together to develop the K computer, with the aim of beginning shared use by the fall of 2012, as a part of the High-Performance Computing Infrastructure (HPCI) initiative led by Japan’s Ministry of Education, Culture, Sports, Science and Technology (MEXT). Since the K computer involves over 80 000 compute nodes, building(More)
Real space DFT (RSDFT) is a simulation technique most suitable for massively-parallel architectures to perform first-principles electronic-structure calculations based on density functional theory. We here report unprecedented simulations on the electron states of silicon nanowires with up to 107,292 atoms carried out during the initial performance(More)
The K computer is a distributed memory super-computer system consisting of more than 800 compute nodes which is being developed by RIKEN as a Japanese national project. Its performance is aimed at achieving 10 peta-flops sustained in the LINPACK benchmark. The system is under installation and adjustment. The whole system will be operational in 2012.
This paper proposes the design of ultra scalable MPI collective communication for the K computer, which consists of 82,944 computing nodes and is the world’s first system over 10 PFLOPS. The nodes are connected by a Tofu interconnect that introduces six dimensional mesh/torus topology. Existing MPI libraries, however, perform poorly on such a direct network(More)
Single-particle coherent X-ray diffraction imaging using an X-ray free-electron laser has the potential to reveal the three-dimensional structure of a biological supra-molecule at sub-nanometer resolution. In order to realise this method, it is necessary to analyze as many as 1 × 10(6) noisy X-ray diffraction patterns, each for an unknown random target(More)
We report failure analysis for three and a half years on the K computer which is well known as one of the largest and fastest supercomputer in the world. The failures of the major components, such as CPU, DIMM, compute node, etc. as well as the system wide failures are analyzed and compared with the that of other systems. The results show that the K(More)