Skip to main content

【Supercomputer】Maintenance work performed during the scheduled maintenance

publication date :

The Supercomputer Systems have been updated upon completing the maintenance work performed from March 30 to April 6, 2026.

  1. Installation of New Software Versions
  2. Changes in Default Versions When Configuring Module
  3. Additional Settings for Using Intel oneAPI for NVIDIA GPUAdditional Settings for Using Intel oneAPI for NVIDIA GPU
  4. Fix for quota settings on LARGE Volume Disk

1. Installation of New Software Versions

The table below shows the new software versions installed after the last maintenance.
For details on how to use each software, please refer to the following web pages.

SoftwareNew VersionModulefile
Intel oneAPI
Intel Compiler2025.3intel/2025.3
Intel MPI Library2025.3intelmpi/2025.3
Intel VTune2025.3intel-vtune/2025.3
Intel Advisor2025.3intel-advisor/2025.3
Intel Debugger2025.3intel-debugger/2025.3
NVIDIA HPC SDK
NVIDIA HPC SDK Compiler25.11nvhpc/25.11
CUDA13.0.2cuda/13.0.2
hpcx2.25.1hpcx/2.25.1_nvhpc-25.11
CUDNN9.20.0 for 13.xcudnn/9.20.0.48_cuda-13.x
ISV Application
ENVI6.2.0envi/6.2.0
IDL9.2.0idl/9.2.0
Maple2026.0maple/2026.0
MSC Adams2025.2adams/2025.2
MSC Nastran2025.2nastran/2025.2
MSC Patran2025.2patran/2025.2
MSC Marc/Mentat2025.2marc/2025.2
Linaro Forge25.1.2linaro_forge/25.1.2
OSS
miniforge326.1.0miniforge3/26.1.0
lammps22Jul2025lammps/22Jul2025_nvhpc-23.9-openmpi
lammps/22Jul2025_intelmpi-2023.2-impi

※With the release of maple/2026.0, support for maple/2022.0 and maple/2023.0 has been discontinued.

2. Changes in Default Versions When Configuring Module

Those listed below are the default versions when configuring using the module command.

If you load without using the slash (/) and the version number of module files, the default version will be set. Also, if you use the software version besides default, specify the slash and the version number when loading.

For details on the configuration using Modules, visit the following webpages:

SoftwareDefault Version(Available Hereafter)Default Version (Previously Available)
ENVIenvi/6.2.0envi/6.1.0
IDLidl/9.2.0idl/9.1.0
Maplemaple/2026.0maple/2024.2
MATLABmatlab/R2025amatlab/R2024a
MSC Adamsadams/2025.2adams/2024.2
MSC Nastrannastran/2025.2nastran/2024.2
MSC Patranpatran/2025.2patran/2024.2
MSC Marc/Mentatmarc/2025.2marc/2024.2
Linaro Forgelinaro_forge/25.1.2linaro_forge/24.1.1

3. Additional Settings for Using Intel oneAPI for NVIDIA GPU

The version of CUDA preinstalled on the compute nodes has been updated from the 12.x series to 13.0.2.
As a result, when compiling using Intel oneAPI for NVIDIA GPUs on System G, please load and use CUDA 12.

Module NameCUDA modules that need to be loaded
intel/2023.2-gpucuda/12.2.2 or cuda/12.6.3
intel/2024.0-gpucuda/12.2.2 or cuda/12.6.3

4. Fix for quota settings on LARGE Volume Disk

We had identified an issue where storage quotas were not being correctly applied for some groups, but we have resolved this issue during our routine maintenance.
If you encounter an error indicating insufficient storage space, such as “no enough space,” while running a job or creating a file, please use the `quota` command to check your available storage space.

Please note that if there have been changes to course content or the number of nodes since last year, your available storage capacity may have changed.If you exceed your storage quota, please free up space by deleting unnecessary files.

Tag/Service