SQUID
-
The software update was implemented as follows.
General-Purpose CPU Nodes | GPU Nodes | Vector Nodes | |
---|---|---|---|
Compiler | Intel one API 2023.0→2023.2 | NVIDIA HPC SDK 22.11→23.11 | NEC SDK 5.0.0→5.2.0 |
Communication Library | intelMPI 2021.8.0→2021.10.0 | OpenMPI 4.1.4→4.1.5 | NEC MPI 3.2.0→3.6.0 |
Please execute the following to use the latest environment:
module load BaseCPU/2024 for General-Purpose CPU Nodes (Intel Compiler)
module load BaseGCC/2024 for General-Purpose CPU Nodes (GNU Compiler)
module load BaseGPU/2024 for GPU Nodes
module load BaseVEC/2024 for Vector Nodes
If you encounter a runtime error, please recompile with the above environment.
About data migrated from OCTOPUS
-
For those who used OCTOPUS in fiscal year 2023 and have applied for SQUID in fiscal year 2024, we are migrating data from SQUID to OCTOPUS.
Destination Directory(SQUID) |
/octfs/work /octfs/home |
---|
Due to data migration, there is a problem with the display results in the usage_view command. Please refer to the following for details.
Problems in displaying the usage_view command due to data migration
About gcc
-
gcc11 has been installed.
The default version of gcc is 8.5, so if you are using gcc11, please load the module as follows:
$ module load /system/apps/env/gcc/Compiler/gcc/11.4.0
Regarding the use of Quantum Espresso on Vector Nodes
-
Previously, when using Quantum Espresso on vector nodes, you had to enclose the command in double quotes " like this, but the double quotes are no longer necessary.
mpirun -version /opt/nec/ve/bin/mpisep.sh ${LM} "-np 80 -npool 2 -nband 1 -ntg 1 -ndiag 4 -input ausurf.in"
↓
mpirun -version /opt/nec/ve/bin/mpisep.sh ${LM} -np 80 -npool 2 -nband 1 -ntg 1 -ndiag 4 -input ausurf.in
Please refer to the following for detailed instructions on how to use QuantumESPRESSO.
About AVS
-
An OpenGL error may occur at startup due to version incompatibility between Xserver and Xclient. In such cases, please turn off OpenGL usage by specifying the "-nohw option".
About Subrequests for Parametric Requests
-
We have increased the limit for subrequests generated on the batch server from 100 to 500.
About GPU LAMMPS
-
We have released the GPU version of LAMMPS, and it is now available for use on GPU nodes.
Please refer to the following for the usage method / version details of LAMMPS.
About IDL 9.0
-
IDL 9.0 is now available for use.
*It has been confirmed that it can be launched simultaneously with the previously installed IDL 8.8.
If you are using IDL 9.0, please load the module as follows:
$ module load BaseApp/2024
$ module load IDL/9.0
$ idlde
Posted : April 17,2024