Portability of the new version about flusi HOT 2 CLOSED

well, it is there since summer. the motivation was that for the vortexstreet simulation, the number n=nx_ny_nz is larger than the standart integer. therefore the code failed.

I see you already resolved to problem. actually, as you know, integer(kind=8) is not standaridzed and thus compiler dependent. the iso_fortran_stuff was an effort made by other people to increase portability. A pity that some compilers do not have it.

I'll be more careful in the future. it would be nice to have acces to your machine for testing. I already use gfortran+mpich, ifort+mpich, ifort+openmpi, IBMXL to check. none of these complained

from flusi.

You could use Jenkins to test that it compiles (and the unit tests pass) on a variety of architectures.

from flusi.