Question about running user-defined applications on cluster
 

I have written a user-defined solver and some utilities. They work perfectly on the workstations. They are compiled by wmake, with the executable path set to $FOAM_USER_APPBIN, something like a line in the Make/files file:

EXE = $(FOAM_USER_APPBIN)/appName

Now I need to run the simulations on cluster. My supervisor told me that cluster cannot "sense"my user directories, so that it cannot access these user-defined applications. If we compile these applications on the "top level", it will take time to verifiy that it is not risky to do so because it impacts every user.

I hope you get my question (despite those informal terms because I have no experience in cluster). Any idea or comment is welcomed! Thanks for the help!

Best regards,
Jing

That is hardcore: somehow the case files must get onto the cluster, right? Put the solver into it and run it (this assumes that the solver is on the case-directory): "./jingFoam". Of course this assumes that binaries from your workstation work on the cluster. If you have problems with the dynamic library you could try compiling the solver like this "wmake exe jingFoam". This builds a HUGE binary (because the required dynamic libraries are statically linked into it). Note: I have never seriously used such binaries. So your milage may vary