[Mpi-forum] MPI Forum Survey Question
goodell at mcs.anl.gov
Fri Mar 5 22:31:32 CST 2010
My recollection was that the Survey Monkey website did have a
histogram of sorts for these results, but it was terrible in one or
more ways. I think we primarily were looking for you to take the raw
data and run it through a program that doesn't distort the data so
heavily by default. Gnuplot or similar should do the job.
For example, it may have been too big to it on one screen with the
projector we had in Atlanta. Also, I think that "0" was shown as "1"
on the plots for some reason.
There may have been more, but w/o having that website in front of me
again I'm not sure that I can recall any details.
On Mar 5, 2010, at 4:37 PM, Josh Hursey wrote:
> Jeff and I were looking through the survey notes from the last MPI
> Forum meeting and were a bit confused by the following:
> * Create a compressed version of the histograms for Q.10 to make it
> easier to analyze.
> Q.10 was:
> How much are each of the following sets of MPI functionality used in
> your MPI applications?
> - Point-to-point communications
> - Collective communications
> - Derived / complex datatypes
> - Communicators other than MPI_COMM_WORLD
> - Graph or Cartesian process topologies
> - Error handlers other than the default MPI_ERRORS_ARE_FATAL
> - Dynamic MPI processes (spawn, connect/accept, join)
> - One-sided communication
> - Generalized requests
> - Parallel I/O
> - "PMPI" profiling interface
> - MPI_THREAD_MULTIPLE (multiple threads simultaneously using MPI)
> - Multiple threads, but only one in MPI at a time
> Unfortunately my notes are woefully sparse on what was requested.
> Can anyone shed some light on this particular statement?
> mpi-forum mailing list
> mpi-forum at lists.mpi-forum.org
More information about the mpi-forum