Quickstart#

NuMagSANS can be executed through the python interface, which allows users to generate configuration files and run simulations programmatically.

The repository contains a minimal example together with a small dataset.

Quick System Check#

The following command check your system.

Linux#

nvidia-smi
nvcc --version
cmake --version
g++ --version
python --version

Windows PowerShell#

nvidia-smi
nvcc --version
cmake --version
cl
python --version

HPC Example Environments#

Many HPC clusters provide software through environment modules. Module names and partitions vary between systems.

The following examples illustrate typical setups on two HPC systems. Adjust the module names according to your cluster.

Example: IRIS HPC (University of Luxembourg)#

https://hpc-docs.uni.lu

salloc -p interactive --gpus 1 --cpus-per-gpu 1 --mem 8G --time=60

module purge
module load lib/GDRCopy/2.4-GCCcore-13.2.0
module load lang/Python/3.11.5-GCCcore-13.2.0
module load system/CUDA/12.6.0
module load devel/CMake/3.27.6-GCCcore-13.2.0

Example: MPSD HPC (Hamburg)#

https://computational-science.mpsd.mpg.de/about.html

salloc -p gpu-interactive --gpus 1 --cpus-per-gpu 1 --mem 8G --time=60

module purge
mpsd-modules 25c native

module load gcc/13.2.0
module load python/3.11.7
module load cuda/12.6.2
module load cmake/3.27.9

Quick Build NuMagSANS#

The following commands

  1. download NuMagSANS from GitHub

  2. compile the C++/CUDA source code using CMake

  3. create a virtual environment

  4. install the NuMagSANS python interface

  5. run the NuMagSANS example script

Linux#

git clone https://github.com/AdamsMP92/NuMagSANS.git
cd NuMagSANS
cmake -S . -B build -DCMAKE_CUDA_ARCHITECTURES=native
cmake --build build -j
python3 -m venv .venv
source .venv/bin/activate
pip install --upgrade pip
pip install -e .
python3 -c "from NuMagSANS import NuMagSANS; print('Installation successful')"
cd example
python3 NuMagSANSrun.py

Windows PowerShell#

git clone https://github.com/AdamsMP92/NuMagSANS.git
cd NuMagSANS
cmake -S . -B build -DCMAKE_CUDA_ARCHITECTURES=native
cmake --build build -j
python -m venv .venv
.venv\Scripts\Activate.ps1
pip install --upgrade pip
pip install -e .
python -c "from NuMagSANS import NuMagSANS; print('Installation successful')"
cd example
python NuMagSANSrun.py

Example Script#

The following provides a quick overview on the NuMagSANS example that is executed in the last step of the Quick build. It demonstrates a minimal NuMagSANS workflow.

This example performs the following steps:

  1. Create a NuMagSANS interface object.

  2. Generate a temporary configuration file.

  3. Specify simulation parameters and input data.

  4. Execute the NuMagSANS simulation.

  5. Remove the temporary configuration file.

Input data are provided in the repository:

example/RealSpaceData/MagData

Simulation results are written to:

example/NuMagSANS_Output

from NuMagSANS import NuMagSANS
from pathlib import Path

BASE_DIR = Path(__file__).resolve().parent

# Create NuMagSANS interface object
sim = NuMagSANS()

# Temporary configuration file
config = BASE_DIR / "NuMagSANSInput_temp.conf"

# Write configuration file
sim.write_config(
    config,
    MagData_activate=1,
    MagDataPath=str(BASE_DIR / "RealSpaceData" / "MagData"),
    foldernameSANSData=str(BASE_DIR / "NuMagSANS_Output"),
    Fourier_Approach="atomistic",
    User_Selection=[1, 2, 3],
    Scattering_Volume_V=2.618e-24,
    enable_outputs=["SpinFlip_2D", "SpinFlip_1D"]
)

# Run NuMagSANS simulation
sim.run(config)

# Remove temporary configuration file
sim.config_clear(config)