[Beowulf] Re: MPI Implementations for SMP use
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.
Isaac Dooley idooley2 at uiuc.eduSun Dec 19 15:27:07 PST 2004
- Previous message: [Beowulf] Regarding Benchmarks
- Next message: [Beowulf] Re: MPI Implementations for SMP use
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
Charm++ and AMPI(an adaptive MPI implementation) also do this(http://charm.cs.uiuc.edu). They conceptually run multiple Virtual Processors on each node, and they have both local and remote delivery options. There are some real benefits to this. On a technical level they can use our own user level threads, which I believe are faster than pthreads. Isaac Dooley >But thats just my intuition speaking! Perhaps there are MPI >implementations that are competitive with threads since they can very >efficiently handle this "two-level" nature of communication: keeping it >local for processes on the same node while still *simultaneously* taking >care of the over-a-network bits for the rest. >Ed
- Previous message: [Beowulf] Regarding Benchmarks
- Next message: [Beowulf] Re: MPI Implementations for SMP use
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
More information about the Beowulf mailing list
