>>> Can MPI be as quick as threads on a 2- or 4-core single >>> machine? > > no, but I think you are worried about something that is such a a small > percentage of compute time that I doubt that it is significant for a Go > program.
A random playout might complete in 10 microseconds. Is thread or MPI overhead going to be say less than 5% of that? If so, yes, I probably don't care. >> The usage of MPI that I've seen required explicit population of >> (user-defined) communication structures that get copied in the process >> of doing an MPI call. > > This is true and something we tried to put some wrappers around to > simplify, but this underlying fact could not be removed. Have you seen Boost.MPI? http://www.generic-programming.org/~dgregor/boost.mpi/doc/mpi/tutorial.html Is it successfully hiding the issue you are talking about? (MPI was accepted into boost, but isn't in boost 1.34, and I doubt we'll see 1.35 released for over 6 months; it seems to have a non-trivial install and build stage...) Darren _______________________________________________ computer-go mailing list [email protected] http://www.computer-go.org/mailman/listinfo/computer-go/
