[mpiwg-abi] MPI_Fint in the ABI

[Jeff Hammond]

First, I withdraw my idea for MPI_Abi_sizes since MPI_Type_size already works for all of that.  We just need Advice to Users about it.

Second, I think we just need to say that the f2c/c2f symbols, while part of the C API, are part of the Fortran ABI and therefore must match the way the Fortran library was compiled.  This means that these symbols have to come from the libmpifort_abi.so (not previously discussed, but obvious), not libmpi.so, and therefore this scheme works with implementations that provide both LP64 and ILP64 Fortran, as Intel MPI does.

We will need to add Advice to Implementers about this, that ensure MPI_Type_size(MPI_INTEGER) does the right thing.

We also need Advice to Users saying that MPI_Fint is not part of the C ABI and is not present in the ABI mpi.h, but rather the user is required to query the type size at runtime and use the appropriate C type.

This is Option 2, with more detail.

Jeff


On 18. Aug 2023, at 21.04, Lisandro Dalcin <dalcinl at gmail.com> wrote:

External email: Use caution opening links or attachments



On Fri, 18 Aug 2023 at 17:28, Erik Schnetter <schnetter at gmail.com<mailto:schnetter at gmail.com>> wrote:
We can introduce a new type `MPIABI_Fint` which is always equal to a C
`int`. In the MPI ABI we only use this type.


But then we will need new MPIABI_XXX_{c2f|f2c}  converters, right? And keep MPI_Fint (and current converters) for backward compatibility?

--
Lisandro Dalcin
============
Senior Research Scientist
Extreme Computing Research Center (ECRC)
King Abdullah University of Science and Technology (KAUST)
http://ecrc.kaust.edu.sa/
--
mpiwg-abi mailing list
mpiwg-abi at lists.mpi-forum.org
https://lists.mpi-forum.org/mailman/listinfo/mpiwg-abi

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mpi-forum.org/pipermail/mpiwg-abi/attachments/20230819/fe6a6ec5/attachment-0001.html>