| Project/Area Number |
19H04119
|
|---|
| Research Category |
Grant-in-Aid for Scientific Research (B)
|
| Allocation Type | Single-year Grants |
|---|
| Section | 一般 |
|---|
| Review Section |
Basic Section 60090:High performance computing-related
|
|---|
| Research Institution | Institute of Physical and Chemical Research |
|---|
Principal Investigator |
Domke Jens 国立研究開発法人理化学研究所, 計算科学研究センター, チームリーダー (70815480)
|
|---|
| Co-Investigator(Kenkyū-buntansha) |
遠藤 敏夫 東京工業大学, 学術国際情報センター, 教授 (80396788)
|
|---|
| Project Period (FY) |
2019-04-01 – 2024-03-31
|
| Project Status |
Completed (Fiscal Year 2023)
|
| Budget Amount *help |
¥17,160,000 (Direct Cost: ¥13,200,000、Indirect Cost: ¥3,960,000)
Fiscal Year 2023: ¥3,250,000 (Direct Cost: ¥2,500,000、Indirect Cost: ¥750,000)
Fiscal Year 2022: ¥3,250,000 (Direct Cost: ¥2,500,000、Indirect Cost: ¥750,000)
Fiscal Year 2021: ¥3,510,000 (Direct Cost: ¥2,700,000、Indirect Cost: ¥810,000)
Fiscal Year 2020: ¥3,250,000 (Direct Cost: ¥2,500,000、Indirect Cost: ¥750,000)
Fiscal Year 2019: ¥3,900,000 (Direct Cost: ¥3,000,000、Indirect Cost: ¥900,000)
|
|---|
| Keywords | HPC interconnects / routing algorithms / network design / artificial intelligance / message passing / routing / hierarchical / supercomputing |
|---|
| Outline of Research at the Start |
The research objective is the invention and development of a novel type of
algorithms, which calculate the communication paths within supercomputer
networks. These novel algorithms will be hierarchical to overcome scalability
challenges of existing algorithms, which are insufficient for future system.
|
| Outline of Final Research Achievements |
Modern society is seeking for ever-increasing compute power to serve science fields such as artificial intelligence (e.g. ChatGPT), weather forecast, airplane and supernova simulations, and medicine discovery. Certain physical limitations prevent us from making computer chips much faster, and hence supercomputers and other tightly coupled compute systems started to scale-out to larger and larger systems. The backbone of all these architectures is the interconnection network, which must be "routed" correctly to increase the effectiveness of the entire system. This process is similar to GoogleMaps telling us which roads to drive. In the supercomputers, the routing tells the messages which path to take. However, the state-of-the-art routing algorithms cannot keep pace with the future scale-out and hardware trends, and hence new algorithms have to be invented. Our project aims to design novel routing approaches for highly complex networks which connect thousands or millions of computers.
|
| Academic Significance and Societal Importance of the Research Achievements |
Our developed routing algorithms, and methods to make supercomputer interconnects faster, will help other scientists to accelerate their workflows. Meaning, with optimal routing, the supercomputers can finish more scientific simulations, and hence the scientists can get more results in shorter time.
|
