Re: [OMPI devel] sm BTL performace of the openmpi-2.0.0

2016-08-05 Thread Nathan Hjelm
Making ob1 ignore RDMA btls that are not in use for eager messages might be sufficient. Please try the following patch and let me know if it works for you. diff --git a/ompi/mca/pml/ob1/pml_ob1_rdma.c b/ompi/mca/pml/ob1/pml_ob1_rdma.c index 888e126..0c99525 100644 --- a/ompi/mca/pml/ob1/pml_ob1_r

Re: [OMPI devel] sm BTL performace of the openmpi-2.0.0

2016-08-05 Thread Nathan Hjelm
Nope. We are not going to change the flags as this will disable the blt for one-sided. Not sure what is going on here as the openib btl should be 1) not used for pt2pt, and 2) polled infrequently. The btl debug log suggests both of these are the case. Not sure what is going on yet. -Nathan > O

Re: [OMPI devel] sm BTL performace of the openmpi-2.0.0

2016-08-05 Thread r...@open-mpi.org
Perhaps those flags need to be the default? > On Aug 5, 2016, at 7:14 AM, tmish...@jcity.maeda.co.jp wrote: > > Hi Christoph, > > I applied the commits - pull/#1250 as Nathan told me and added "-mca > btl_openib_flags 311" to the mpirun command line option, then it worked for > me. I don't know

Re: [OMPI devel] sm BTL performace of the openmpi-2.0.0

2016-08-05 Thread tmishima
Hi Christoph, I applied the commits - pull/#1250 as Nathan told me and added "-mca btl_openib_flags 311" to the mpirun command line option, then it worked for me. I don't know the reason, but it looks ATOMIC_FOP in the btl_openib_flags degrades the sm/vader perfomance. Regards, Tetsuya Mishima

Re: [OMPI devel] sm BTL performace of the openmpi-2.0.0

2016-08-05 Thread Howard Pritchard
Hello Christoph The rdmacm messages while annoying are not causing the problem. If you specify tcp BTL does the BW drop disappear? Also could you post your configure options to the mail list? Thanks Howard Am Freitag, 5. August 2016 schrieb Christoph Niethammer : > Hello, > > We see the sam

Re: [OMPI devel] sm BTL performace of the openmpi-2.0.0

2016-08-05 Thread Christoph Niethammer
Hello, We see the same problem here on various machines with Open MPI 2.0.0. To us it seems that enabling the openib btl triggers bad performance for the sm AND vader btls! --mca btl_base_verbose 10 reports in both cases the correct use of sm and vader between MPI ranks - only performance differ

Re: [OMPI devel] sm BTL performace of the openmpi-2.0.0

2016-08-05 Thread Christoph Niethammer
Hello, We see the same problem here on various machines with Open MPI 2.0.0. To us it seems that enabling the openib btl triggers bad performance for the sm AND vader btls! --mca btl_base_verbose 10 reports in both cases the correct use of sm and vader between MPI ranks - only performance differ