This cluster provides software through Environment Modules using Lmod.
Modules let you quickly enable specific software stacks (e.g., CUDA, GROMACS) without changing the system installation for everyone.
Modules help users to:
On this cluster, modules are shared via NFS and are available on login and compute nodes.
Tip: The list may grow over time. Use
module availto see the latest.
| Module name | Purpose (short) | Typical users |
|---|---|---|
cuda/12.2 |
NVIDIA CUDA Toolkit 12.2 runtime + nvcc paths |
GPU workloads, GPU builds |
gromacs/2023.3-cpu |
GROMACS 2023.3 CPU-only (AVX2) | MD on CPU partitions |
gromacs/2023.3-cuda12.2-avx2 |
GROMACS 2023.3 GPU (CUDA 12.2) build (SM 86/89) | MD on GPU partitions |
python-test/1.0 |
Simple test module (points to system Python) | Module system testing |
module avail
module avail gromacs
module spider cuda
module load cuda/12.2
module load gromacs/2023.3-cpu
module list
module unload gromacs/2023.3-cpu
module purge # unload everything
module show gromacs/2023.3-cuda12.2-avx2
Always initialize the environment and clean modules before loading what you need:
source /etc/profile
module purge
module load <your-module-here>
Note: Some modules automatically load dependencies.
Example:gromacs/2023.3-cuda12.2-avx2will loadcuda/12.2for you.
module avail or module spider <name> and check spelling/version.cuda/<version> module exists and is loaded (or use a module that depends on it).module --ignore_cache avail or module --ignore_cache load <module>.