M-Star CFD Version 4.0.x Notes

Summary

M-Star CFD version 4.0 mainly brings upgrades to our CUDA toolkit, build systems, and packages.

Version 4.0 introduces major changes, including:

  • Upgrade to CUDA 12.9

  • Increase in the minimum NVidia driver version to 525

  • Increase in the minimum NVidia GPU compute capability to 6.0 (See here for impacted GPUs)

  • Removal of OS support for RHEL/Oracle 7

  • Addition of OS support for RHEL/Oracle 10

  • Support for Python versions 3.9–3.13 on all platforms, including Windows

  • Upgrade of third-party software to the latest versions

  • Upgrade of compilers and build systems

  • Inclusion of NCCL support in all solver builds, with NCCL libraries bundled in the distribution

  • Removal of select “solver-only” builds

  • Packaging updates: Linux .tar.gz and Windows .zip distributions file now inlcude a top-level directory

Operating Systems supported (x86 64bit):

  • Ubuntu 20, 22, 24

  • RHEL/Oracle Linux 8, 9, 10

  • Microsoft Windows 10+

Operating Systems supported (arm64/aarch64), Solver only:

  • Ubuntu 22, 24

What is not changing:

  • The 4.0.x release line will continue to use the existing CAD kernel (OpenCASCADE)

  • All .msb and solver input files created with M-Star CFD 3.x will remain compatible with version 4.0

Notes for Windows Users

  • Download and install the latest Visual C++ Redistributable x64 package here

  • Update your NVidia driver to version 525 or later here

CUDA 12.9 Upgrade Notes

CUDA 12.9 brings a number of changes to our builds. One impact is that some GPUs will no longer run with 4.x. Starting from M-Star CFD version 4.0, GPUs must have a compute capability 6.0 or greater. Previous M-Star 3.x versions required a compute capability of 3.5. Refer to this page to see impacted GPUs (3.5–5.0 are removed).

You must upgrade your NVidia driver to 525 or later. M-Star is always compatible with the latest NVidia driver.

M-Star 4.x will now have native support for the latest Nvidia architectures (Pascal, Volta, Turing, Ampere, Ada, Hopper Blackwell). This will speed up the startup time of the solver for most users.

CUDA 12.9 modernizes our Nvidia compiler toolset. One of the benefits of this is the upgrade to various CUDA libraries and algorithms. Some users may see performance improvement on certain models.

GPU Compute Capability changes to 6.0+

M-Star CFD 4.x will now require GPU compute capability 6.0 or later. Previous M-Star 3.x versions supported compute capability 3.5+. To see the impacted GPUs, refer to this page.

Compatibility with Nvidia drivers older than 525 (Linux only)

If you are unable to update your NVidia driver on Linux to 525 or later, you may be able to use the “cuda-compat-12-9” package to get M-Star running. Refer to the instructions here to install the cuda-compat-12-9 package to set up your environment.

Linux Platform Instructions

  1. Install the latest NVidia driver on ALL platforms.

  2. Follow the instructions below for your operating system.

  3. Download the appropriate M-Star CFD linux package for your platform.

  4. Expand the M-star CFD package.

  5. Source the mstar.sh script in the M-Star install folder.

  6. Set up the M-Star license.

  7. M-Star is now ready for use.

Note

Most users will not need to compile openmpi from source anymore. You can use the openmpi package that your operating system provides.


RHEL 8, Oracle 8

Note that you will need to add the openmpi lib path to your LD_LIBRARY_PATH.

dnf -y install  \
                gtk3 \
                freetype \
                libXmu \
                mesa-libGL \
                openmpi \
                libxkbcommon-x11 libxkbcommon \
                pcre2-utf32 openmpi libglvnd-egl \
                python3.11 python3.11-pip


export LD_LIBRARY_PATH=/lib64/openmpi/lib

RHEL 9, Oracle 9

Note that you will need to add the openmpi lib path to your LD_LIBRARY_PATH.

dnf -y install  \
                gtk3 \
                freetype \
                libXmu \
                mesa-libGL \
                openmpi \
                libxkbcommon-x11 libxkbcommon \
                pcre2-utf32 openmpi libglvnd-egl \
                python3.11 python3.11-pip

export LD_LIBRARY_PATH=/lib64/openmpi/lib

RHEL 10, Oracle 10

This distribution is compiled with CUDA 13.0.

On RHEL 10 and Oracle 10 Linux, this distribution has the following minimum requirements:

  • Minimum NVidia driver version: 580

  • Minimum GPU Compute Capability: 7.5

Note that you will need to add the openmpi lib path to your LD_LIBRARY_PATH.

dnf -y install  \
                gtk3 \
                freetype \
                libXmu \
                mesa-libGL \
                openmpi \
                libxcrypt-compat \
                libxkbcommon-x11 libxkbcommon \
                pcre2-utf32 openmpi libglvnd-egl \
                python3.12 python3.12-pip

export LD_LIBRARY_PATH=/lib64/openmpi/lib

Ubuntu 20

apt-get install -y \
    libgtk-3-0 \
    libxmu6 \
    wget  \
    libfreetype6 \
    software-properties-common \
    openmpi-bin \
    libxkbcommon0 \
    libxkbcommon-x11-0 \
    libgl1 \
    python3.9 python3.9-venv

Ubuntu 22

apt-get install -y \
    libgtk-3-0 \
    libxmu6 \
    wget  \
    libfreetype6 \
    software-properties-common \
    openmpi-bin \
    libxkbcommon0 \
    libxkbcommon-x11-0 \
    libgl1 \
    python3 python3-venv

Ubuntu 24

apt-get install -y \
   libgtk-3-0 \
   libxmu6 \
   wget  \
   libfreetype6 \
   software-properties-common \
   openmpi-bin \
   libxkbcommon0 \
   libxkbcommon-x11-0 \
   libgl1 \
   python3 python3-venv