Alexey Cherkaev <alexey.cherk...@gmail.com> writes: > Found the problem: "double precision" on my machine (presumably, any 64-bit > machine) with "-fdefault-real-8" flag will be "real*16" (something like > non-existent "Float128" in Julia).
In Fortran, if you want your code to be portable and avoid these type of issues it is much better to use selected_real_kind (see https://gcc.gnu.org/onlinedocs/gfortran/SELECTED_005fREAL_005fKIND.html) Cheers, -- Ángel de Vicente http://www.iac.es/galeria/angelv/
