BwUniCluster2.0/Maintenance/2024-05

From bwHPC Wiki
Jump to navigation Jump to search

The following changes were introduced during the maintenance interval between on 21.05.2024 (Tuesday) 08:30 and 24.05.2024 (Friday) 15:00.

We are not planning to change the host key of the system. You should not receive any warnings by your SSH client(s), but if there should be a warning or if you want to check that you are connecting to the correct system, you can verify the key hashes using the following list:

Algorithm Hash (SHA256) Hash (MD5)
RSA p6Ion2YKZr5cnzf6L6DS1xGnIwnC1BhLbOEmDdp7FA0 59:2a:67:44:4a:d7:89:6c:c0:0d:74:ba:3c:c4:63:6d
ECDSA k8l1JnfLf1y1Qi55IQmo11+/NZx06Rbze7akT5R7tE8 85:d4:d9:97:e0:f0:43:30:6e:66:8e:d0:b6:9b:85:d1
ED25519 yEe5nJ5hZZ1YbgieWr+phqRZKYbrV7zRe8OR3X03cn0 42:d2:0d:ab:87:48:fc:1d:5d:b3:7c:bf:22:c3:5f:b7

Hardware

  • All firmware versions on all components were upgraded.

Operating system

  • The operating system was upgraded from RHEL 8.6 EUS to RHEL 8.8 EUS. We recommend to re-compile all applications after the upgrade.
  • The Mellanox OFED InfiniBand Stack was updated.

Compilers, Libaries and Runtime Environments

  • Certain older compilers will be deprecated and hidden (a dot prepended):
    These compilers and Software is still available, but instead of module load compiler/gnu/9.3 will now be available as module load compiler/gnu/.9.3.
  • The compilers gnu-9.3, gnu-10.3, gnu-11.1, gnu-11.2, gnu-12.1, gnu-13.2, as well as llvm-11.0, llvm-13.0, llvm-14.0, llvm-15.0, llvm-16.0 will be deprecated and hidden. Modules depending on those as well.
  • The available (non-hidden) compilers therefore are:
    • gnu-8.5.0 (the system compiler)
    • gnu-10.2 (which a lot of modules still depend on)
    • gnu-13.3 (the new gnu default)
    • gnu-14.1 (which may require You to adapt your software, see https://gcc.gnu.org/gcc-14/porting_to.html)
    • llvm-17.0
    • llvm-18.1 (the new llvm default)
  • GNU compilers are compiled against CUDA to support parallelization for the GPU using OpenMP target and OpenACC.
  • Intel OneAPI 2023 and 2024 are now available
  • Intel parallel studio 2020 (compiler 19.1, impi 2020, mkl 2020) are deprecated

Userspace tools

Software Modules

  • CMake has been updated to 3.29.3 (the old version 3.23.3 is hidden); the old version 3.18 is still available.
  • Valgrind has been updated to 3.23 (the old version 3.22 is hidden)
  • Dwarves tool pahole has been updated to 1.26 (the old version 1.24 is hidden)
  • Open MPI has been upgraded to version 5.0.3 compiled against the compilers gnu-8.5.0 (the system default), gnu-10.2, gnu-13.2 and gnu-13.3, gnu-14.1, intel-2023.1.0, intel-2023.1.0_llvm, intel-2024.0_llvm, llvm-17.0 and llvm-18.1
    Please note: Open MPI now links against several internal libraries statically -- which adds the CUDA library as dependency. Therefore all modules load the default CUDA-12.2 module as well as add a soft-link to libcuda.so.1. You should be able to replace the CUDA version by unloading and loading your desired CUDA module.
  • The Library HDF5 has been upgraded to 1.14.4-2 for several compilers and MPI versions.

Batch system

  • The Slurm version was upgraded to version 23.11.5

Storage

  • Lustre client, BeeGFS client and Spectrum Scale client were updated.

Graphics stack

  • The NVIDIA driver was upgraded to 550.54.15
  • CUDA 12.2 is now the default version, 12.4 is available
  • NVIDIA toolkit: default is now 23.9

Containers

JupyterHub

  • Jupyterhub was upgraded to 4.1.3