Software
Using existing software
Scruggs contains various scientific codes and libraries that may be useful to you. Here is the current list, which can be accessed using module avail.
anaconda/3-2020.11 gaussian/g09 gsl/2.6 lammps/29Oct20 openbabel/3.1.1 openmpi/3.1.4-intel orca/5.0.3
anaconda/3-2022.05 gaussian/g16 intel/oneAPI lammps/29Oct20-intel openmpi/2.1.6 openmpi/4.1.1 orca/6.0.0
cmake/3.19 gcc/10.3 lammps/23Jun22-intel nwchem/7.0 openmpi/3.1.4 openmpi/4.1.6
If you desire to access any of these, use the module load command, for example
[vishnura😀10:14 AMscruggs:~]$ module load intel
[vishnura😀10:14 AMscruggs:~]$ module list
Currently Loaded Modulefiles:
1) rocks-openmpi 2) intel/oneAPI
To remove a module from your environment, use the module unload command
Compiling your software
If you need a library/software that is not available in the module, you are free to compile it for yourself. Scruggs has gcc and intel compilers, and the openmpi library for parallel applications. It also has cmake (PS: In my opinion, the cmake version on Scruggs is a bit outdated. I recommend getting a newer cmake version from online and buiding it).
For Python based software, you can load the anaconda module.
For GPU accelerated software, it is important for you to build your GPU code on the compute-0-6 and compute-0-7 GPU nodes directly. To do this, you can ssh to the nodes directly.
[vishnura😀10:15 AMscruggs:~]$ ssh compute-0-6
Last login: Mon Apr 14 09:39:56 2025 from scruggs.local
Rocks Compute Node
Rocks 7.0 (Manzanita)
Profile built 19:35 13-Oct-2021
Kickstarted 19:41 13-Oct-2021
[vishnura😀10:16 AMcompute-0-6:~]$ nvidia-smi
Mon Apr 14 10:24:36 2025
+-----------------------------------------------------------------------------+
| NVIDIA-SMI 470.57.02 Driver Version: 470.57.02 CUDA Version: 11.4 |
| -------------------------------+----------------------+----------------------+
| GPU Name Persistence-M| Bus-Id Disp.A | Volatile Uncorr. ECC |
| Fan Temp Perf Pwr:Usage/Cap| Memory-Usage | GPU-Util Compute M. |
| | | MIG M. |
|===============================+======================+======================|
| 0 NVIDIA A30 Off | 00000000:01:00.0 Off | 0 |
| N/A 30C P0 32W / 165W | 0MiB / 24258MiB | 0% Default |
| | | Disabled |
+-------------------------------+----------------------+----------------------+
| 1 NVIDIA A30 Off | 00000000:41:00.0 Off | 0 |
| N/A 28C P0 28W / 165W | 0MiB / 24258MiB | 0% Default |
| | | Disabled |
+-------------------------------+----------------------+----------------------+
| 2 NVIDIA A30 Off | 00000000:81:00.0 Off | 0 |
| N/A 29C P0 32W / 165W | 0MiB / 24258MiB | 0% Default |
| | | Disabled |
+-------------------------------+----------------------+----------------------+
| 3 NVIDIA A30 Off | 00000000:C1:00.0 Off | 0 |
| N/A 29C P0 31W / 165W | 0MiB / 24258MiB | 16% Default |
| | | Disabled |
+-------------------------------+----------------------+----------------------+
+-----------------------------------------------------------------------------+
| Processes: |
| GPU GI CI PID Type Process name GPU Memory |
| ID ID Usage |
|=============================================================================|
| No running processes found |
+-----------------------------------------------------------------------------+
Here the CUDA libraries are present in the deafult location (/usr/local/cuda), so “most” softwares would manage to locate it by itself. If it doesn’t, reach out for help at vishnura@illinois.edu