[Beowulf] MPI_Alltoall
Many of your questions may have already been answered in earlier discussions or in the FAQ. The search results page will indicate current discussions as well as past list serves, articles, and papers.
Joachim Worringen joachim at ccrl-nece.deTue Apr 12 06:34:04 PDT 2005
- Previous message: [Beowulf] MPI_Alltoall
- Next message: [Beowulf] MPI_Alltoall
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
Rita Zrour wrote: > Hello I have a question, > when i do many MPI_Alltoall in my program always the > first MPI_Alltoall take too much time to be done. You should be more precise: - which MPI implementation, which interconnect? - a code and timing example would be helpful > I don't know where the first communication is always > expensive. Is that a problem of memory??????? One possible reason: some MPI implementations do not set up the "connections" between the processes until communication is actually taking place. When the first communication call is all_to_all, then this setup procedure takes some more time because each process communicates with every other one. Joachim -- Joachim Worringen - NEC C&C research lab St.Augustin fon +49-2241-9252.20 - fax .99 - http://www.ccrl-nece.de
- Previous message: [Beowulf] MPI_Alltoall
- Next message: [Beowulf] MPI_Alltoall
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
More information about the Beowulf mailing list
