High-performance Computing

Computers can provide computational results faster than human beings; however, computers are not completely versatile machines always capable of completing such complicated computations within moments. There are numerous problems that cannot be solved within a reasonable period of time even by the world's fastest supercomputers. Our research focuses on software technology that accelerates computations for solving such problems. We work on solving applied problems using accelerators (e.g., graphics processing units (GPUs) that comprise thousands of cores) and parallel computers (e.g., clusters of computers comprising hundreds of thousand cores). We also focus on developing optimization techniques and compilation frameworks with an aim to introduce the produced technology to the real world through joint research and product development with companies.