sorry for posting and asking so much, but does anybody also have problems finding MPI, as is described here: https://gitlab.kitware.com/cmake/cmake/issues/18570
We have a CentOS system with custom gcc and Intel compilers. It seems that everything is found correctly, but I still get this error message when configuring with cmake 3.11 or 3.13:
Could NOT find MPI (missing: MPI_Fortran_FOUND) (found version “3.1”)
Addendum: I also posted this issue on the cmake mailing list, but no response either. As it appeared that all entries, i.e. compiler, libs etc., where found correctly and only the flag MPI_Fortran_FOUND was not set to true, I manually set it to true, and the configuration and generation went alright. I could compile two versions, one with OpenGL, one with osmesa for use with catalyst. I can use ParaView in parallel with many servers and one client, as well as for in situ with FORTRAN simulation code. So there must be something wrong with cmake, but at least – with this workaround – it works again.
I am having the same issue. How do you know that all compilers and libs were found correctly? Also, how did you manually set MPI_Fortran_FOUND to true?
I saw their correct location in the CMakeCache.txt file. That looked all normal. The Cmake configuration also tells you were the error occurred. I looked into that file, and manually set MPI_Fortran_FOUND to true, hard coded. Not nice, but after that it configured, generated, and everything compiled, and I can use ParaView as client/server, as well as for in-situ visualization. As our system is on maintenance at the moment, I can not send you the actual file. Drop me a line at email@example.com next week, and I can send it to you, if you’re not getting it working.