Konstantin Karganov wrote:
>>The Open MPI Team will be at SC this year.
>>I am co-hosting a BOF entitled "Why MPI Makes You Scream! And how can
>>we simplify parallel debugging?"
>By the way (though it may be offtopic here)...
>Does anybody know some statistics on MPI usage. e.i. :
>- percentage of C, C++, Fortran MPI programs
>- percentage of SPMD and MPMD programs
>- MPI function an function groups usage (how many people use
>topologies, asynchronous IO, communicator attributes etc.)
>- average number of nodes for running and for debugging
>I suppose this can be interesting for all MPI community as a valuable
>feedback from those who use MPI for real applications. There was a
>survey on SC'93 on the usage of debugging tools (
>http://citeseer.ist.psu.edu/pancake94what.html ) and It might be
>possible (if its not too late now) to do something alike.
There has been at least one project analysing these kind of things, look
for papers about automatic MPI counter profiling. These were studies
about the usage of MPI at two different HPC centers.
>users mailing list
Department of Computer Science email:gabriel_at_[hidden]
University of Houston http://www.cs.uh.edu/~gabriel
Philip G. Hoffman Hall, Room 524 Tel: +1 (713) 743-3857
Houston, TX-77204, USA Fax: +1 (713) 743-3335