Big Data and CFD Simulation on TSUBAME 2.0
Date : 2 July 2013, Tuesday
Time: 10 AM
Venue: Exploration Theatre, Level 4, Matrix Building, Biopolis, Singapore
Welcome Address by Prof Tan Tin Wee, Chairman, A*CRC
Title: Extreme Big Data and Resilience in Tsubame2.0 towards 3.0
Speaker: Prof Satoshi Matsuoka
Tokyo Institute of Technology
Supercomputers often stress their FLOPS as their primary benefit. While this would be true in a classical sense, there is growing need for very fast I/O capabilities in handling large quantities of data,, or namely, “Big Data”. Many are lead to believe that current-day cloud infrastructures are more suitable for big data processing compared to supercomputers, but such is simply not true considering the current day technologies as well as future technological trend trajectories. Tsubame2.0, Tokyo Tech.’s petascale supercomputer is touted often for its FLOPS and greenness, but the other highlighted characteristics is that it is likely the world’s first supercomputer to facilitated fast I/O for both resilience and big data processing. Currently, various researches are being undertaken in our group to further enhance these properties, such that Tsubame2.5, a 6 petaflop update to Tsubame 2.0 in mid-2013, and Tsubame3.0, a 25 petaflop machine being planned for late 2015, would be considered as “big data supercomputers”.
Satoshi Matsuoka is a Professor at the Global Scientific Information and Computing Center of Tokyo Institute of Technology (GSIC). He is the leader of TSUBAME series of supercomputers, which became the 4th fastest in the world on the Top500 and awarded the "Greenest Production Supercomputer in the World" by the Green 500 in November, 2010 and June 2011. He has also co-lead the Japanese national grid project NAREGI during 2003-2007, and is currently leading various projects such as the JST-CREST Ultra Low Power HPC and JSPS Billion-Scale Supercomputer Resilience. He has authored over 500 papers according to Google Scholar, and has chaired many ACM/IEEE conferences, including the Technical Papers Chair, Community Chair, and the upcoming Program Chair for Supercomputing Conferences 09, 11 and 13 respectively. He is a fellow of ACM and European ISC, and has won many awards including the JSPS Prize from the Japan Society for Promotion of Science in 2006, awarded by his Highness Prince Akishinomiya, the ACM Gordon Bell Prizes for 2011, and the Commendation for Science and Technology by the Minister of Education, Culture, Sports, Science and Technology in 2012.
Title: A Turbulent Air Flow Simulation in Metropolitan Tokyo for 10km x 10km area with 1m resolution and Several Peta-scale Real-world Simulations on TSUBAME 2.0
Speaker: Prof Takayuki Aoki
Global Scientific Information and Computing Center, Tokyo Institute of technology
Turbulent modeling is a key issue of CFD (Computational Fluid Dynamics), since most flow phenomena become turbulent with higher Reynolds number. We have developed a CFD code based on Lattice Boltzmann Method with a LES (Large-Eddy Simulation) model. The dynamic Smagorinsky model is often used, however it requires costly average operations for wide area to determine the model constant. We applied the coherent-structure Smagorinsky model which is able to determine the model constant locally. We study a turbulent air flows in a metropolitan Tokyo 10km x 10km area with 1-m resolution taking account for the real building data.
We also demonstrate several stencil applications carried out on the whole TSUBAME 2.0 system, which currently has 2.4 PFLOPS of the peak performance at the Tokyo Institute of Technology. One of them is high resolution meso-scale atmosphere model ASUCA that is being developed by the Japan Meteorological Agency (JMA) for the purpose of the next-generation weather forecasting service. We have succeeded in a weather prediction with 500-m resolution (cf. Current JMA weather forecast uses 5km mesh). We also talk about a phase-field simulation to develop new materials by studying the dendritic solidification of Al-Si alloy with 2.0 PFLOPS in single precision, which is 44.5% of the peak performance.
Takayuki Aoki received a BSc in Applied Physics (1983), an MSc in Energy Science and Dr. Sci (1989) from Tokyo Institute of Technology, has been a professor in Tokyo Institute of Technology since 2001 and the deputy director of the Global Scientific Information and Computing Center since 2009. He received the Achievement Awards from Japan Society of Mechanical Engineers, Japan Society for Industrial and Applied Mathematics and many awards and honors in GPU computing, scientific visualization, and others. His team got the Gordon Bell Prize, Special Achievement in Scalability and Time-to-Solution in 2011. He was also recognized as a CUDA fellow (currently 11 fellows in the world) by NVIDIA in 2012.
Prof Aoki Presentation File
No of Participants: 99