コンテンツ

  1. HOME
  2. Services
  3. Supercomputer
  4. Application Procedures
  5. Service Selection Guide

Service Selection Guide

This page provides overviews of the service courses and guides users to select the course that fits the purpose of use. For details of the system specifications, please refer to the "Supercomputer(Services)". For details of the service courses or the usage fees, please refer to the "Rules for Usage Fees of Academic Center for Computing and Media Studies (ACCMS)"(*Japanese Version Only).


1. Selecting the Service Course

The Service Course consists of the Entry, Personal, Group, and Private Cluster Course. Select one that fits your purpose. For the Entry Course, you can use the service as soon as you get your user ID, however for other courses, you must apply within the application period.

For the usage fees and the available computing resources of each course, please refer to Usage Fees.

Please note that the application for the service course is open only for a fixed period of time except for the entry course.We will make an announcement on the Service Application Timeline and E-mail Newsletter when the application for the service course is open.

CourseDescriptionElapsed Time for Batch※Application
Entry
  • All users who do not belong to the other courses (users who do not apply for the service course.
  • Users can login to the system B.
  • Users can use the minimum (0.5 nodes).
  • Users can use large format printer
 1 hour Acceptable all year round
Personal
  • This course is for individuals.
  • Multiple users of the same course share one batch queue.
  • Computing resources are provided on a best effort basis.
 168 hours Open application system

Group
  • This course is for using in the laboratories and the collaborative research groups.(available for individuals)
  • The UNIX group and the dedicated batch job queue are assigned for each group.
  • Additional group members are provided free of charge according to the number of applied nodes.
    (For users who are linked to the payment manager of the group course)
    (5,000 yen/person when exceeding the free number)

*Supplemental information on the provision of computing resources

  • Fourth Priority(Type A0,B0,C0,G0)
    • Computing nodes are provided on a best-effort basis up to the number of applied nodes (Computing node are not guaranteed).
    • Computing nodes are not guaranteed, but improve the job execution by allocating a variable number of nodes with priority.
    • Range of variation is assumed 1/6 to 1/12 of the number of applied nodes.(Scheduled to operate with review on a timely basis)
  • Third Priority(Type A2,B2)
    • Guarantee 1/4 of the number of applied nodes. The rest is provided on a system availability basis.
  • Second Priority(Type A1,B1,C1,G1)
    • Guarantee 1/2 of the number of applied nodes. The rest is provided on a system availability basis.
  • Top Priority(Type A3,B3)
    • Guarantee the number of applied nodes.

336 hours
(168 hours for fourth priority)

 Open application system
Private Cluster
  • This course provides an environment that is logically independent of the supercomputer and similar to a laboratory cluster.
  • By assigning a global IP address to the login node, you can access to the cluster directly.
  • The batch processing system is not available for this course.
  • Administrator authority(Root authority)cannot be obtained.
  • Limited to small-scale use because the number of computing nodes are limited.
 ― Open application system

※ The elapsed time for batch is the upper limit of the elapsed time per one batch job. The upper limit of the number of executable jobs is not set.

2. Selecting system (For Group and Personal Course)

The supercomputer system consists of four types of systems: System A, B, C, and G. Please refer to the characteristics and the available software of each system and select one of the systems. For the system specifications, please refer to the "Supercomputer System".

System Characteristics

SystemOperation SystemProcessorCharacteristics
System A
(Camphor3)		 RHEL8 ※1
 x86_64
    • Successor to former System A (Camphor2)
    • Massive parallel computing is available
    • Latest Intel Xeon Processor is equipped
    • HBM2e memory is equipped
    • Fat tree Network
    • high-speed interconnect (Infiniband NDR400) is equipped
    System B
    (Laurel3)
    • Successor to former System B (Laurel2)
    • High affinity with a PC cluster
    • Latest Intel Xeon Processor is equipped
    • Fat tree Network
    • ISV application is available
      System C
      (Cinnamon3)
      • Successor to former System C(Cinnamon2)
      • Latest Intel Xeon Processor is equipped
      • Memory of 2TB per node is equipped
      • Small configuration of 16 nodes
      • Fat tree Network
      • ISV application is available
      System G
      (Gardenia)
      • Equipped with AMD EPYC processor
      • Equipped with four NVIDIA A100 (GPGPU) per node
      • Small configuration of 16 nodes
      • Machine learning / deep learning software are available
      • Usage fees are set per GPU
      • ※1 RHEL8: Red Hat Enterprise Linux 8 (Computing nodes,Login nodes)

      Available Software

      TypeSoftwareSystem ASystem BSystem CSystem G
      Compiler Intel oneAPI + + + ―
      NVIDIA HPC SDK ― ― ― +
      GCC + + + +
      MPI Intel MPI + + + ―
      OpenMPI + + + +
      Library IntelMKL + + + ―
      NAG ― + + ―
      IMSL ― + + ―
      Software Application AVS/Express ― AU AU ―
      ENVI/IDL ― AU AU ―
      Tecplot360 ― KU KU ―
      MSC Adams ― AU AU ―
      MSC Nastran ― AU AU ―
      MSC Patran ― AU AU ―
      MSC Marc ― AU AU ―
      MSC Marc Mentat ― AU AU ―
      LS-DYNA ― + + ―
      ANSYS ― KU KU ―
      MOPAC ― AU AU ―
      Gaussian16 + + + ―
      GaussView + + + ―
      Maple ― KU KU ―
      Mathematica ― KU KU ―
      MATLAB ― KU KU ―
      • +: Available for all users(Academic users,Private organizations)
      • AU: Available for academic users only
      • KU:Available for Kyoto University members only
      • ―:Not available

      3. Selecting type(For Group course)

      Group course provides four different types of the respective allocation of computing resources that are usually guaranteed based on the contracted resources.The following table shows the fees of each type. For the units of node and socket that can be applied, please refer to "Regulations for Usage Fees of Academic Center for Computing and Media Studies (ACCMS), Kyoto University" in "Usage Fees".

      Table of the Fee of Node/Socket

      SystemTypeAllocation Rate#coreTFlopsMEM
      (GB)      		DISK
      (TB)

      Annual Usage Fee
      (Yen/node)
      System A
      (Node)      		 Fourth Priority None TBA
      Third Priority 1/4
      Second Priority 1/2
      Top Priority 1/1
      System B
      (Node)      		 Fourth Priority None 112 5.91 128 6.4 88,000
      Third Priority 1/4 9.6 132,000
      Second Priority 1/2 16.0 220,000
      Top Priority 1/1 16.0 330,000
      System C
      (Node)      		 Fourth Priority None 112 5.91 2048 6.4 96,000
      Second Priority 1/2 16.0 240,000
      System G
      (GPU)      		 Fourth Priority None 16

      CPU:
      2.66
      GPU:
      312(half-precision)

      		 128 6.4 64,000
      Second Priority 1/2 16.0 160,000

      *The minimum unit for the "Third priority type" of Systems A and B and the "Second priority type" of System G is "2".
      *"Fourth priority type" is a best-effort service that does not guarantee computing resources.

      4. Other Services

      In addition to the selected course (Entry, Personal, Group or Private Cluster), the following additional services are available.

      ServiceUsage FeeDescriptionApplication
      Large-scale job 80,000/week～ Highly parallelized jobs are available by a week(7 days) basis. Open application system
      Additional storage capacity
      (High capacity storage)      		 10,000/year～ Additional storage capacity can be added in increments of
      10TB when the standard storage capacity provided is insufficient.      		 Open application system
      Additional storage capacity
      (High speed storage)      		 10,000/year～ High-speed storage consisting of SSDs can be added in increments of
      2TB.      		 Open application system
      Licensing service 20,000yen/year This service allows you to install applications on a PC in your laboratory.
      • ENVI/IDL
      		 Acceptable all year round
      Large format printer No additional fee Large format printer is available for printing
      (large format posters) for poster sessions at academic conferences.      		 No need
       

      Top of Page

      Copyright © Institute for Information Management and Communication, Kyoto University, all rights reserved.