CCMS

Center for Computational Materials Science
Institute for Materials Research,
Tohoku University

Supercomputing Consortium for Computational Materials Science (SCCMS)

1. Outline

From fiscal 2015, as a project related to the three fields of material science, physical science, and molecular science, a review committee on social and scientific issues that should be focused on post-K computer priority issues (Priority Issue (5) Development of new fundamental technologies for high-efficiency energy creation, conversion / storage and use, Priority Issue (7) Creation of new fundamental devices and high-performance materials to support next-generation industries) and post-K computer exploratory challenges (Frontiers of Basic Science: Challenging the Limits – exploring the extremes through multi-physics multi-scale simulations) was established, and furthermore, as a building of consortia for the development of human resources in science and technology project of JST, Professional development Consortium for Computational Materials Scientists and Element Strategy Initiative: To Form Core Research Centers have also been started.

Institute for Materials Research of Tohoku University operates "Supercomputing Consortium for Computational Materials Science" with the Institute for Solid State Physics of the University of Tokyo and Institute for Molecular Science of National Institutes of Natural Sciences, for the purpose of supporting projects that aim at massively parallel computing of high commonality in above three fields and of encouraging further efforts to enhance parallel computing to each research fields. In FY2020 and beyond, we will continue to achieve this objective by providing computational resources and exchanging information on computational technologies and results to support the promotion of large-scale projects that focus on computational materials science.

2. Available computing resources

Supercomputing system of Center for Computational Materials Science, Institute for Materials Research, Tohoku University (HPE Cray XD220v, HPE Scale-up Server 3200, and HPE Cray XD670)

System Name Supercomputer
Model HPE Cray XD220v HPE Scale-up Server 3200 HPE Cray XD670
total nodes

General node : 120
Large memory node : 7

2 nodes

11 nodes
CPU

Intel Xeon Platinum 8480+

  1. Frequency:
     2.0 GHz
  2. Number of cores:
     56 cores
  3. Number of CPUs:
     2 CPUs/node

Intel Xeon Platinum 8490H

  1. Frequency:
     1.9 GHz
  2. Number of cores:
     60 cores
  3. Number of CPUs:
     16 CPUs/node

Intel Xeon Platinum 8480+

  1. Frequency:
     2.0 GHz
  2. Number of cores:
     56 cores
  3. Number of CPUs:
     2 CPUs/node
Accelerator - -

NVIDIA H100 80GB SXM5

  1. Performance:
     33.5 TFLOPS(FP64)
  2. Number of GPUs:
     8 GPUs/node
  3. Memory bandwidth:
     3.352 GB/s
memory size 512 GiB/node
2.0 TiB/node		 4.0 TiB/node 1.0 TiB/node
Theoretical peak performance 4.052 PFLOPS
(CPU : 1.104 PFLOPS, GPU : 2.948 PFLOPS)

2.1 Available node hours

Up to 10% of total node hours of our supercomputer (HPE Cray XD220v, HPE Scale-up Server 3200, and HPE Cray XD670) in total of SCCMS.

3. Research projects and Qualification for use

3.1 Research projects

  1. The project must be a large-scale project supported by the national government or other public organizations, and materials science research using computational methods must be a major part of the theme of the project.
  2. The user group must be able to present its research results at the annual meeting and cover the travel expenses for the meeting.

3.2 Qualification for use

Researchers belonging to national, public, and private universities and national and public research institutes who are promoting the above research projects.

  1. The principal researchers and co-investigators of post-K computer priority issue (5)
  2. The principal researchers and co-investigators of post-K computer priority issue (7)
  3. The principal researchers and co-investigators of post-K computer exploratory challenges
  4. Researchers for the next generation of PCoMS
  5. The principal researchers of Element Strategy Initiative

4. Application method and inspection for SCCMS

4.1 Application method

See project application of SCCMS in the preceding page.

4.2 inspection

The node hour for application would be decided by management committee of SCCMS.

5. Usage term and fee

5.1 Usage term

Until March 31, 2026

5.2 Usage fee

The computing resource would be managed based on distributed node hour. Usage fee would not be collected.

6. Usage report

Submission of usage report to CCMS would be required at the beginning of the next fiscal year. Information of the node hour used in each subject would be reported to the management committee of SCCMS.

7. Publication of research outcome

  1. Please include proper acknowledgement text to CCMS for research papers and other publications containing results obtained on MASAMUNE-IMR.
  2. Send one copy of the paper to CCMS, IMR

Ex. We acknowledge Center for Computational Materials Science, Institute for Materials Research, Tohoku University for the use of MASAMUNE-IMR (Project No.K****).