On Mar 14, 2008, at 8:24 PM, Supalov, Alexander wrote:

It appears to me that the current proposal basically intends to pass data thru to the presumably existing C implementation layer. How will this work for an implementation where the Fortran binding is a primary one?

I believe that at least NEC has followed this route. I'm adding Hubert Ritzdorf, in case he's accidentally not involved into this discussion yet, to verify NEC's position.

I thought that NEC provided functions written in C that are callable directly via Fortran. The code in the C MPI_Send and the "Fortran" MPI_SEND was essentially the same C code -- the point of the design was the avoid a second function call from the Fortran API functions. Do we really need to worry about [real] native Fortran MPI implementations? -- Jeff Squyres Cisco Systems