[Beowulf] MPICH question
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.
Rusty Lusk lusk at mcs.anl.govTue Feb 22 21:12:17 PST 2005
- Previous message: [Beowulf] MPICH question
- Next message: [Beowulf] The Case for an MPI ABI
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
From: John Lau <cflau at clustertech.com> Subject: [Beowulf] MPICH question Date: Tue, 22 Feb 2005 09:30:29 +0800 > Hi, > > I have a question on the number of processes spawned by MPICH. I am > using MPICH 1.2.5.2 --with-device=ch_p4. When I use mpirun -np 4 to > start a mpi process, total 8 processes will be spawned. But only 4 > processes have loadings on CPUs. I would like to know if it is the > correct behavior of MPICH? And what's the use of the 4 no-loading > process? Thank you. The other 4 processes are "listener" processes that permit dynamic creation of connections as they are needed. What you see is the correct behavior. We recommend that MPICH1 users switch to MPICH2. Regards, Rusty Lusk
- Previous message: [Beowulf] MPICH question
- Next message: [Beowulf] The Case for an MPI ABI
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
More information about the Beowulf mailing list
