On Fri, 2019-03-15 at 18:02 +0000, Sergio None wrote: > Jeff, > > Yes, i know that C99 have these fixed standard library. > > The point is that many common used function, rand for example, return > non fixed types. And then you need doing castings, typically to more > big types. It can be a bit annoying.
I just want to add--if you use a "sysdef.h" type header file like I described a couple of days ago, you can make up your own function names and put macros in the header to map them to what is available on each machine (e.g. abs32() could be abs() or labs()) with suitable variable casts in or out as needed. If necessary, this could even handle problems like two's complement vs sign+magnitude negatives or IBM vs IEEE floating point. Generally you want to store all variables in types that you then pass as byte strings through MPI messages and store/convert to/from actual arithmetic types on each end as needed. Not "transparent" as you want, but once set up, easy to maintain and extend. George > > > > That's why I was looking for an OpenMPI way to do this transparent. > > Thank you very much for your help. > > > Regards, > > Sergio. > > > _______________________________________________ > users mailing list > users@lists.open-mpi.org > https://urldefense.proofpoint.com/v2/url?u=https-3A__lists.open-2Dmpi.org_mailman_listinfo_users&d=DwICAg&c=JeTkUgVztGMmhKYjxsy2rfoWYibK1YmxXez1G3oNStg&r=jya52EHQYs5GJJ8uBNZZeZ3OFSM3QRdLnT55J0n0u6Q&m=qgXdyQX4m9E731x3l4K8-M6c8bNrajYXFSP_E_kqlEQ&s=0z97udGAcALw9E4RPhSxVBf_adelDu9krjaSAs8DMb0&e= _______________________________________________ users mailing list users@lists.open-mpi.org https://lists.open-mpi.org/mailman/listinfo/users