Re: [OMPI users] 50% performance reduction due to OpenMPI v 1.3.2 forcing all MPI traffic over Ethernet instead of using Infiniband

2009-06-24 Thread Noam Bernstein
On Jun 23, 2009, at 6:19 PM, Gus Correa wrote: Hi Jim, list On my OpenMPI 1.3.2 ompi_info -config gives: Wrapper extra LIBS: -lrdmacm -libverbs -ltorque -lnuma -ldl -Wl,-- export-dynamic -lnsl -lutil -lm -ldl Yours doesn't seem to have the IB libraries: -lrdmacm -libverbs So, I would gues

Re: [OMPI users] 50% performance reduction due to OpenMPI v 1.3.2forcing all MPI traffic over Ethernet instead of using Infiniband

2009-06-24 Thread Jim Kress
Noam, Gus and List, Did you statically link your openmpi when you built it? If you did (the default is NOT to do this) then that could explain the discrepancy. Jim > -Original Message- > From: users-boun...@open-mpi.org > [mailto:users-boun...@open-mpi.org] On Behalf Of Noam Bernstein

Re: [OMPI users] 50% performance reduction due to OpenMPI v 1.3.2forcing all MPI traffic over Ethernet instead of using Infiniband

2009-06-24 Thread Noam Bernstein
On Jun 24, 2009, at 11:05 AM, Jim Kress wrote: Noam, Gus and List, Did you statically link your openmpi when you built it? If you did (the default is NOT to do this) then that could explain the discrepancy. Not explicitly: env CC=gcc CXX=g++ F77=ifort FC=ifort ./configure --prefix=/shar

Re: [OMPI users] 50% performance reduction due to OpenMPI v 1.3.2forcing all MPI traffic over Ethernet instead of using Infiniband

2009-06-24 Thread Gus Correa
Hi Jim Jim Kress wrote: > Noam, Gus and List, > > Did you statically link your openmpi when you built it? If you did (the > default is NOT to do this) then that could explain the discrepancy. > > Jim No, I didn't link statically. Did you link statically? Actually, I tried to do it, and it di

Re: [OMPI users] 50% performance reduction due to OpenMPI v 1.3.2forcing all MPI traffic over Ethernet instead of using Infiniband

2009-06-24 Thread Jim Kress ORG
Well, the whole situation is really bizarre. I just uninstalled openmpi 1.3.2 on my system. Then I installed OFED 1.4.1 to see if that resolves this situation. Here's what I get: [root@master ~]# ompi_info --config Configured by: root Configured on: Wed Jun 24 11:10:00 EDT

Re: [OMPI users] 50% performance reduction due to OpenMPI v 1.3.2forcing all MPI traffic over Ethernet instead of using Infiniband

2009-06-24 Thread Jim Kress ORG
Hey Gus. I was correct. If I did: ./configure --prefix=/my/dir --with-openib=/usr --enable-static make all install then reboot and use mpi-selector to choose openmpi-1.3.2, and then: [root@master ~]# ompi_info --config Configured by: root Configured on: Wed Jun 24 18:02:0

Re: [OMPI users] 50% performance reduction due to OpenMPI v 1.3.2forcing all MPI traffic over Ethernet instead of using Infiniband

2009-06-24 Thread Gus Correa
Hi Jim, list 1) libnuma (non-uniform memory access, IIRR) is important if you have AMD Opteron (our case here), for processor and memory affinity, etc. I suppose you don't need it with Intel Xeon (pre-Nehalem, at least), but I am not positive about this (we don't have a Xeon cluster here). 2) tm

Re: [OMPI users] 50% performance reduction due to OpenMPI v 1.3.2forcing all MPI traffic over Ethernet instead of using Infiniband

2009-06-24 Thread Gus Correa
Hi Jim Jim Kress ORG wrote: > Hey Gus. I was correct. > > If I did: > > ./configure --prefix=/my/dir --with-openib=/usr --enable-static > make all install > ... > Wrapper extra LIBS: -lrdmacm -libverbs -ldl > -Wl,--export-dynamic -lnsl > -lutil -lm -ldl ... > > M

Re: [OMPI users] 50% performance reduction due to OpenMPI v 1.3.2forcing all MPI traffic over Ethernet instead of using Infiniband

2009-06-24 Thread Jim Kress ORG
> Have you investigated Jeff's question on whether the code was > compiled/linked with the same OpenMPI version (1.3.2)? > I wonder if the underlying OFED libraries must be the same as well. I was told that 1.3.2 was used. However, I have not asked about which OFED libraries were used nor have I