Thank you so much for your detailed answers. I use gfortran 4.8.3, what should I do? or what is the suitable openmpi version for this version?
Thanks again Best wishes Zeinab في الخميس، 27 سبتمبر 2018 في 4:21 م تمت كتابة ما يلي بواسطة Jeff Squyres (jsquyres) via users <users@lists.open-mpi.org>: > On Sep 27, 2018, at 12:16 AM, Zeinab Salah <zeinabsa...@gmail.com> wrote: > > > > I have a problem in running an air quality model, maybe because of the > size of calculations, so I tried different versions of openmpi. > > I want to install openmpi-3.0.2 with the option of > "--with-mpi-f90-size=medium", but this option is unrecognized in the new > versions. So, what is the update option for that, or how we can control the > size of the MPI F90 module in the new versions?, > > This option did, indeed, disappear in more recent versions of Open MPI. > > The short version is that if you have a "modern" gfortran (i.e., >= v4.9) > or any other modern fortran compiler, then you will get the full "mpi" > module. > > The "mpi-f90-size" option was only necessary for older gfortran versions > that had limitations that caused us to make the user choose: do you want > small, medium, or large? > > Meaning: if your gfortran is >= v4.9 or you're using a different fortran > compiler, don't worry about this option. > > > also, does this option make the numerical models to deal with lagre data > files? > > No; this option only had to do with how many MPI API interfaces were in > the "mpi" module (i.e., if your application invokes "use mpi" instead of > "include 'mpif.h'"). Even if you chose the "small" size (meaning: many MPI > APIs were not listed in the "mpi" module), all the MPI APIs would compile > and link and work at run-time just fine. The number of interfaces in the > "mpi" module basically just means how much compile-time checking you get > when you are compiling your MPI application. > > Check out this paper ( > https://www.open-mpi.org/papers/euro-pvmmpi-2005-fortran/) if you care > about the reasons why. > > -- > Jeff Squyres > jsquy...@cisco.com > > _______________________________________________ > users mailing list > users@lists.open-mpi.org > https://lists.open-mpi.org/mailman/listinfo/users >
_______________________________________________ users mailing list users@lists.open-mpi.org https://lists.open-mpi.org/mailman/listinfo/users