Hi Peter,
Thank you for your question and interest in SimVascular! Just a clarifying question: are you planning on using the SimVascular front-end GUI on the cluster? Or do you just plan to run flow simulations on it? It is quite common for users to do the model construction and meshing in the front-end GUI using the pre-built versions on their desktop machines, then using the cluster to run the flow simulations.
I ask this because it is much easier to compile the flow solver on clusters than it is to compile the entire front-end GUI on clusters (since the GUI uses lots of packages like QT that may not be standard on Linux computing clusters). The Github link to JUST the solver source files is here:
https://github.com/SimVascular/svSolver, and instructions on how to compile are given below:
1. First, make sure you load appropriate compilers. We recommend gcc/gfortran compilers and the OpenMPI implementation of the MPI standard. Intel compilers and MPICH is also possible. If your cluster uses a module-based way of organizing software libraries, you will have to use commands like:
module load gcc
module load openmpi
2. Clone the code repository into your cluster: git clone
https://github.com/SimVascular/svSolver.git svsolver
3. Move into the build folder: cd svsolver/BuildWithMake
4. Download VTK binaries: ./get-vtk-binaries.sh centos_6
5. Set parameters for cluster installation: cp SampleOverrides/centos_6/global_overrides.mk . (<---- notice period!)
5a. Open the global_overrides.mk in a text editor (i.e. vim or emacs) and edit your MPI choice. The default is "NOMPI", but you can change the flag according to the MPI implementation you have (OpenMPI or MPICH).
6. Compile the code: make
Hope this helps! Let us know if you are able to compile the flow solver on the cluster.