All posts by editor_7

Shared Use Computers HOKUSAI
Announcement about exemption from the usage fee for the Data Science Infrastructure as a power-saving measure

Exemption from the usage fee for the Data Science Infrastructure as a power-saving measure

In order to save electricity throughout RIKEN, we promote the migration of computing resources such as computers and storage space currently in use at each laboratory to the Data Science Infrastructure HOKUSAI SailingShip (SS). To promote the migration, the SS usage fee for the migration will be exempted.

Subject
  • Servers and storage currently in use at all RIKEN sites to be migrated to SS between December 2022 and February 2023.
    • The migration period shall be the minimum necessary and in principle within one month.
  • Only those in use with the power on will be considered, not including notebook PCs.
    • Those that are turned off for power-saving measures, will be included in the scope.
Computing Resources for Exempting Usage Fees
  • From the start of the migration to the end of March 2024, the usage fees will be exempted.
  • In principle, 2 vCPU per physical core of the CPU of the computing resource to be migrated will be exempted.
    • The upper limit per laboratory is 10 SS nodes (920 vCPU).
  • The usage fee will be waived for the effective capacity of the storage space to be migrated.
    • The upper limit per laboratory is 100 TB.
Application Procedure
  1. Send an email to E-mail: with the application details shown below.
    • Receive acceptance and guidance
  2. Migration work to SS
    • Project creation at SS
    • System construction at SS
    • Data migration to SS
    • Contact us after the migration is completed
  3. Confirmation of termination of use of computing resources
    • On-site confirmation by Information Systems Division
Application Details
  1. Applicant name
  2. Payment representative
    • Must be the same as the payment representative of the SS project.
    • Permission must be obtained in advance and included in the CC of the application email.
  3. Affiliation of the payment representative
    • Center name, laboratory name, etc.
  4. Computing resources to be turned off after the migration
    • Overview of Computing Resource Usage
    • Total number of cores of servers
      • Breakdown for each computing machine, including CPU type
    • Effective capacity of storage
  5. Migration period
    • In principle, within 1 month.
    • If your system has a certain size and you are uncertain about the feasibility of the migration, please contact us.
      • Describe your system and your concerns about the transition.
Example of Application
  1. Applicant name: Taro Riken
  2. Payment representative: Jiro Wako
  3. Affiliation of the payment representative: XX center XX laboratory
  4. Computing resources to be turned off after the migration
    Overview of Computing Resource Usage: For storage and analysis of XX experimental data
    Total number of cores of servers: 360 core
    1 login node: Intel Xeon Gold 6148 (20 cores) x 2 CPUs
    8 compute nodes: Intel Xeon Gold 6148(20 cores) x 2 CPUs
    Storage effective capacity: 10 TB(NetApp EXXXX 2 nodes)

  5. Migration period: December 21 – January 20
Others
  • SS Usage Information
  • Support for PIaaS usage in SS will be available by the end of December early January 2023.
    • Combination of offering a minimum of VM images and a playbook of Ansible (IT automation tool).
    • Account management (LDAP), file server (samba, NFS), remote desktop (rdp, vnc)
    • PC cluster construction (slurm, singularity)
Contact

E-mail:

News & Topics
[Please fill out the questionnaire] Questionnaire regarding the use of Fugaku

September 15, 2022
Information Systems Division, R-IH

Please help us by filling out this questionnaire in advance of preparing for the Fugaku use support.

Purpose of the questionnaire

We are considering establishing a comprehensive framework for RIKEN researchers to promote research and development and improve operational efficiency, not only within RIKEN, but also by utilizing a variety of external and other paid computing resources and services that are appropriate for their own research and work.

This initiative aims to establish a consistent framework for the use of computational resources and services, not only within RIKEN but also from a variety of external resources and services, based on the direction set by President Gokami, who assumed his new position this fiscal year. In this fiscal year, as a trial for the use of paid computational resources and services, we expect to initially subsidize the usage fees from the President’s Discretionary Funds.

As one of this year’s trials, RIKEN is considering actively adopting and subsidizing the use of Fugaku for simulations and data processing that RIKEN expects to produce results as soon as possible. Although the details of the subsidy may vary depending on the total number of proposals submitted, at this time we are assuming that the full amount of the subsidy will be provided for this trial use. We are conducting this survey as a preliminary study for this purpose. We would appreciate your active participation in this survey.

  • Target: RIKEN employee (for visiting staff, please consult with your supervisor (PI))
  • Deadline: 6:00 p.m., September 26, 2022
  • Contact: isys-hpc [at] ml.riken.jp
Shared Use Computers HOKUSAI
Notice of HOKUSAI Maintenance

Supercomputer HOKUSAI BigWaterfall(HBW) and Data Science Infrastructure(HSS) will be stopped in the schedule below due to the planned power outage in Wako campus. At that time, regular maintenance will also be performed.

Downtime(HOKUSAI BW and SS):
from 5:00 p.m. on October 7 to 6:00 p.m. on October 11

For BW, note that job execution control will be performed before the downtime: Jobs that are expected not to end until the start time of the downtime will be not started. After resuming operation, the unstarted jobs will be started.

For SS, please stop VMs (instances) before the service is stopped. Any VMs that are running when the service is stopped will be forcibly terminated. VMs will not start automatically after the service is restarted, so please start it by yourself.