Description | I have a custom multiphase solver written in OpenFOAM 5.x. I have used on a workstation with Debian 11 compiled with GCC 8. Recently a new workstation was bought equipped with Debian 12 and GCC 12. I compiled the same version of OpenFOAM with the same solver on the new machine and tried to run a case that I know to be working. Interestingly, the first GAMG iteration looks exactly the same, with the same residuals down to the last digit, but on the new machine the second iterations diverges completely to infinity. The only difference that might have been introduced, apart from hardware, is compiler versions (both were run serial so no MPI involved I think). |
---|