M-Star CFD
M-Star CFD
M-Star CFD solves advanced Lattice Boltzmann algorithms on GPUs to help users produce detailed, accurate process simulations in minutes. M-Star CFD is built to run on millions and billions of lattice grid points, so users can model free surface flows, open systems, particle growth, heat transfers and much more in mere minutes.
M-Star CFD contains three primary components: M-Star Build, M-Star Solve, and M-Star Post. M-Star Pre is the GUI used to prepare and interact with models. M-Star Solve uses input files generated from the interface to execute the simulation. M-Star Post loads, renders, and plots data generated by the solver.
System requirements
Before running the NGC M-Star CFD container please ensure your system meets the following requirements.
- One of the following container runtimes
- nvidia-docker
- Singularity >= 3.1
- One of the following NVIDIA GPU(s)
- Pascal (sm60)
- Volta (sm70)
- Ampere (sm80)
- Hopper (sm90)
x86_64
- CPU with AVX2 instruction support
- One of the following CUDA driver versions
- r520 (>=.61.05)
- >= 450.80.02
License Requirements
Use of the M-Star CFD container requires a valid license. Please submit license inquiries to the M-Star CFD support team.
Examples
Running with nvidia-docker
Throughout this example the container version will be referenced as $MSTAR_VERSION, replace this with the tag you wish to run.
Set path to license file
export mstar_LICENSE=/mstar.lic
Start an interactive session in the container and bind the license file to /mstar.lic
docker run -it --rm --gpus all -v $mstar_LICENSE:/mstar.lic --shm-size=1g --ulimit memlock=-1 nvcr.io/hpc/m-star:$MSTAR_VERSION
Run the solver
mpirun -np 8 mstar-cfd-mgpu --gpu-auto --disable-ipc -i input.xml -o out
Note: Docker < v1.40
Docker versions below 1.40 must enable GPU support with --runtime nvidia.
docker run --rm --runtime nvidia -v $mstar_LICENSE:/mstar.lic --shm-size=1g --ulimit memlock=-1 nvcr.io/hpc/m-star:$MSTAR_VERSION