[mpiwg-tools] Prefix of PMPI replacement interface

[Mohror, Kathryn]

I'd be okay with either "callback" or "hook" names for the interface. I suppose if I had to pick a preference, I'd go with "hook". The concept of "hook" seems more general to me than "callback". Then I'd choose the MPI_H prefix version over MPI_Hook, with the reason being the 32 character limit for function names in the MPI Standard.

Or another idea: MPI_I for MPI Interception Interface? Or Introspection Interface? It kind of rolls off the tongue... for me anyway. MP Eye Eye.

Kathryn

_________________________________________________________________
Kathryn Mohror, kathryn at llnl.gov, http://scalability.llnl.gov/ 
Scalability Team @ Lawrence Livermore National Laboratory, Livermore, CA, USA


> -----Original Message-----
> From: mpiwg-tools [mailto:mpiwg-tools-bounces at lists.mpi-forum.org] On
> Behalf Of Todd Gamblin
> Sent: Tuesday, September 29, 2015 9:48 PM
> To: mpiwg-tools at lists.mpi-forum.org
> Subject: Re: [mpiwg-tools] Prefix of PMPI replacement interface
> 
> FWIW, I agree with Marc-Andre.
> 
> Would it be ridiculous to just call the new functionality "MPI Hooks", based
> on what might be the most widely used similar interface, "malloc hooks"?
> This gets at the callback aspect that Marc mentions, but I think it also makes it
> pretty clear what it's for.
> 
> Thoughts?  Are the MPI_H or MPI_Hook_ prefixes taken?
> 
> -Todd
> 
> 
> 
> 
> On 9/29/15, 7:07 AM, "mpiwg-tools on behalf of Marc-Andre Hermanns"
> <mpiwg-tools-bounces at lists.mpi-forum.org on behalf of
> hermanns at jara.rwth-aachen.de> wrote:
> 
> >Dear all,
> >
> >currently the working title of the PMPI replacement interface is QMPI.
> >While this is a nice homage to a system that I mostly enjoy working on,
> >it feels to me more like an "inside joke" lacking obvious relation
> >between name and function. Already the name PMPI was misleading, as it
> >eventually was used for more than just "profiling".
> >
> >I would propose
> >
> >MPI_C_
> >
> >as a basis for discussion.
> >
> >It blends in well with the new MPI_T API prefix and the C would stand
> >for "Callback". Furthermore, it describes what it _is_ rather that
> >being presumptuous about what you do with it or who it is for (tools,
> >ft, etc.).
> >
> >What do you think? Do you have other ideas?
> >
> >I know, a concrete name is not the most pressing issue we need to
> >attend in this context, yet, in my experiences naming discussions take
> >a long time, thus if we start this now, me might come to consensus by
> >the time we have the rest of the API fleshed out. ;-)
> >
> >Cheers,
> >Marc-Andre
> >
> >--
> >Marc-Andre Hermanns
> >Jülich Aachen Research Alliance,
> >High Performance Computing (JARA-HPC)
> >Jülich Supercomputing Centre (JSC)
> >
> >Schinkelstrasse 2
> >52062 Aachen
> >Germany
> >
> >Phone: +49 2461 61 2509 | +49 241 80 24897
> >Fax: +49 2461 80 6 99753
> >www.jara.org/jara-hpc
> >email: hermanns at jara.rwth-aachen.de
> >
> 
> 
> _______________________________________________
> mpiwg-tools mailing list
> mpiwg-tools at lists.mpi-forum.org
> http://lists.mpi-forum.org/mailman/listinfo.cgi/mpiwg-tools