|Main Work||Institute of Mathematics for Industry
|Keyword||Mathematical Optimization Problem / Graph Analysis / Operations Research / Data Science / High Performance Computing|
Main Research Topics
1. Advanced Computing and Optimization Infrastructure for Extremely Large-Scale Graphs on Post Peta-Scale Supercomputers.
The objective of many ongoing research projects in high performance computing (HPC) areas is to develop an advanced computing and optimization infrastructure for extremely large-scale graphs on the peta-scale supercomputers.
The objective of our researches is to develop advanced computing and optimization infrastructures for extremely large-scale graphs on post peta-scale supercomputers.
2. Development of High Performance Graph Search and Optimization Library for Graph Analysis.
Large-scale graph data are divided between multiple nodes, and then, we perform graph analysis and search algorithms, such as the BFS kernel for Graph500, on multiple CPUs and GPUs.
Implementations, including communication-avoiding algorithms and techniques for overlapping computation and communication, are needed for these libraries.
We propose a hierarchal graph stores and process extremely large-scale graphs with minimum performance degradation by carefully considering the data structures of a given graph and the access patterns to both DRAM and NVM devices.
3. A Challenge to Graph 500 and Green Graph 500 benchmarks.
We explain our challenge to Graph 500 and Green Graph 500 benchmarks that are designed to measure the performance of a computer system for applications that require irregular memory and network access patterns.
The Graph500 benchmark measures the performance of any supercomputer performing a BFS (Breadth-First Search) in terms of traversed edges per second (TEPS).
In ISC14, our project team was a winner of the 8th Graph500 benchmark and 3rd Green Graph 500 benchmark.
4. High Performance Computing for Mathematical Optimization Problems.
We also present our parallel implementation for large-scale SDP (SemiDefinite Programming) problem.
We solved the largest SDP problem (which has over 2.33 million constraints), thereby creating a new world record.
Our implementation also achieved 1.713 PFlops in double precision for large-scale Cholesky factorization using 2,720 CPUs and 4,080 GPUs on the TSUBAME 2.5 supercomputer.
5. Graph Analysis and High Performance Computing Techniques for Realizing Urban OS.
We have started the research project for developing the Urban OS (Operating System) and implementing it on a large city (Fukuoka) from 2013.
The Urban OS gathers big data sets of people and transportation movements by utilizing different sensor technologies and storing them to the cloud storage system.
We have another research project whose objective is to develop advanced computing and optimization infrastructures for extremely large-scale graphs on post peta-scale supercomputers.
The Urban OS employs the graph analysis system developed by this research project and provides a feedback to a predicting and controlling center to optimize many social systems and services.