Supercomputing Division, Information Technology Center, The University Tokyo
JAPANESE

For first-time visitors


Thank you for considering the supercomputer systems of the Information Technology Center, the University of Tokyo (hereafter, “the Center”). This page introduces the characteristics of the supercomputers that are currently in operation and the methods of use. If you have any questions, please contact us through the contact details provided at the bottom of this page.

Characteristics of each supercomputer system

The Center currently operates the Reedbush and Oakforest-PACS supercomputer systems. The characteristics of each system are provided here to serve as reference when you are selecting a system. As it is not possible to say, as a rule, which supercomputer is more suitable for a specific purpose, we recommend that you make use of the trial use course described later to try out both supercomputer systems for yourself. For detailed information of each supercomputer, please refer to the linked introductory pages below.

  • 　Oakbridge-CX Supercomputer System
    • Orthodox configuration that combines elements with a high market share, such as Intel CPU and Omni-path.
    • Supports processing that requires high file input / output performance by installing SSD in 128 nodes.
    • Possible to incorporate dedicated HD and cables by applying for fixed nodes
  • 　Oakforest-PACS Supercomputer System
    • Incorporates the latest elemental technology, such as Xeon Phi processor and Omni-path architecture.
    • Rich computational resources (8,208 nodes × 68 cores)
  • 　Reedbush Supercomputer System
    • Orthodox configuration that combines elements with a high market share, such as Intel CPU and InfiniBand.
    • Some compute nodes mounted with GPU (Reedbush-H, Reedbush-L)
    • Possible to incorporate dedicated HD and cables by applying for fixed nodes

Application and System

Academic use

Application System Cost Period of use Prior
sceening		 Reception
period		 Required to submit report Report
release
ITC Parallel Programming
Workshop with Trial Account		 Free One month Not available About 10 times
a year		 Not available Not available
Trial use
system 		Paid trial Free One month Not available any time Not available Not available
Personal Fee required Less than 1 year Not available any time Results report Release
Group
young researchers
and women reserchers 		General Free 6 months Available Twice a year
(Around January,July)		 Problem report /
Presentation
at the result briefing session		 Release
Intern Onece a year
(Around June)
Large-Scale HPC Challenge Free Maximum of 24 hours Available 3 times a year Report of results Release
Educational use Free Semester Not available any time Not available Not available
JHPCN Joint Research Projects Free One month Available Once a year
(About January)		 User report /
Presentation
in the symposium		 Release
HPCI Geneal Projects
Junior researcher Projects		 Free Less than 1 year Available Once a year
(Around November)		 User report /
presentation of results
(academic papers / patent etc)		 Release
※１
※１ The grace period a maximum of 3 years.

Corporate use

Application System Cost Period
of use		 Prior
sceening		 Reception
period		 Required to submit report Report
release
ITC Parallel Programming
Workshop with Trial Account		 Free One month Not available About 10 times
a year		 Not available Not available
Trial use
system 		Free tral Free One month Not available any time Not available Not available
Personal Fee required Less than 1 year Not available
※２		 any time Results report Release
※３
Group Free 1 - 3 months Available Twice a year
(Around January,July)
Fee required Less than 1 year
young researchers
and women reserchers 		Free 6 months Available Twice a year
(Around January,July)		 Problem report /
Announcement
in the result briefing session		 Release
Large-Scale HPC Challenge Free Maximum of 24 hours Available 3 times a year Report of results Release
Corporate use system Fee required Less than 1 year Available Twice a year
(Around January,July)		 Report on the results of use Release
※３
Joint reserch with
adademic institution		 Fee required
※１		 Less than 1 year Not available any time Not available Not available
JHPCN Industrial joint reseache projects Free One month Available Onece a year
(About January)		 User report /
Presentation
in the symposium		 Release
HPCI Industrial use
(Non-proprietary use)		 Free Less than 1 year Available Onece a year
(Around November)		 User report /
presentation of results
(academic papers / patent etc)		 Release
※３
※１ In the case of the collaborative investigation with the academic institution, it is an academic price.
※２ It's the condition of the application to attend a workshop.
※３ The grace period a maximum of 3 years

Methods of use

Various methods of use are available for each supercomputer. Please select the method of use that best suits your purpose.

  • 　General use

    • The personal course, for use by individual researchers, and the group course, for use by research groups, are available to users from universities and public organizations.
  • 　Corporate use

    • As a part of our social contribution initiatives, some computational resources are also offered to corporations that meet the application criteria.
  • 　Trial use course

    • This system offers the trial use of the supercomputers at no charge, or at a lower cost than normal. Corporate users may also make use of this system.
  • 　Use by youths/women

    • This program is open to public applications, and is targeted at young researchers (including students) below 40 years of age and female researchers. Selected subjects may use the computational resources they have applied for at no charge.
  • 　Educational use

    • Computational resources are offered for use in undergraduate and graduate classes. Applications from institutions other than the University of Tokyo are also accepted.
  • 　Large-Scale HPC Challenge

    • This program is open to public applications, and allows one research group to monopolize the maximum number of compute nodes of Reedbush-H and Oakforest-PACS for a maximum of 24 hours.
  • 　Use through other public application systems

    • Computational resources are offered by the Center in its position as a HPCI resource provider, and as a Joint Usage/Research Center for Interdisciplinary Large-Scale Information Infrastructures.

Application

The use of the Center’s supercomputers is restricted to eligible users of the Center, and users are required to comply with the terms of use. When applying for use, please read the following “Eligibility for use” and “Terms of use” carefully. Please also refer to the other materials where necessary.

Parallel programming workshop with trial account

Parallel programming workshops that come with a trial account, allowing participants to practice using the supercomputers, are held regularly. New users who are considering the use of the supercomputer systems (including corporate engineers and researchers) are also welcome to participate in these workshops. Through the workshop, participants can gain an idea of how to use the supercomputers through hands-on experience at using the systems, and also learn about parallelization methods. While restricted to the contents covered in the workshop, accounts that can be used for a month are also distributed to participants.
Workshops that correspond to the levels of participants are also held, starting from the introductory session, which requires no basic knowledge about parallel processing, to parallelization at the practical level of numeric calculations. The workshops are free of charge, so please utilize them actively. For details, please refer to the following link.

Inquiries

Research Support Team, Information Strategy Group, Information Systems Department, The University of Tokyo
E-mail：E-mail address