SIGSEGV fault for large problems

More
2 years 3 months ago #2442 by gcdiwan

JuliusZ wrote: I can confirm this issue.
I had the same problem and compiling openmpi with gcc 8.3.0 fixed it.


Ok, so the version of C or Fortran compiler used for MUMPS has no effect and it's just the openmpi that has to be compiled with gcc 8.3.0?

Please Log in or Create an account to join the conversation.

More
2 years 3 months ago - 2 years 3 months ago #2443 by JuliusZ
well, gcc contains a c (gcc) and fortran (gfortran) compiler. Ideally, everything is compiled with the same version. You can also set this by the environment variables CC, CXX, and FC for cmake.
Last edit: 2 years 3 months ago by JuliusZ.

Please Log in or Create an account to join the conversation.

More
2 years 3 months ago - 2 years 3 months ago #2444 by lkogler
Yes, you need a new MPI installation. Maybe ask you administrators if they could install it for you.

The MUMPS project does not use cmake. We cannot directly give it the same c/c++/fortran compilers that are used for NGSolve, because it would not find the necessary MPI libraries and flags. Instead, we give it the MPI wrapper-compilers (mpicc, mpicxx, mpifort). When the c/c++/fortran compiler underlying the MPI-wrappers mismatch, you run into problems like yours.

So, when your MPI is compiled with gcc8.3, MUMPS will also be compiled with that version.
Last edit: 2 years 3 months ago by lkogler.

Please Log in or Create an account to join the conversation.

Time to create page: 0.143 seconds