Re: [OMPI devel] enable-smp-locks affects PSM performance

2014-11-03 Thread Jeff Squyres (jsquyres)
Sent: Monday, November 3, 2014 7:49 AM >> To: Open MPI Developers List >> Subject: Re: [OMPI devel] enable-smp-locks affects PSM performance >> >> On Nov 3, 2014, at 10:31 AM, Friedley, Andrew >> wrote: >> >>> Thanks Jeff, we're working to v

Re: [OMPI devel] enable-smp-locks affects PSM performance

2014-11-03 Thread Friedley, Andrew
devel [mailto:devel-boun...@open-mpi.org] On Behalf Of Jeff > Squyres (jsquyres) > Sent: Monday, November 3, 2014 7:49 AM > To: Open MPI Developers List > Subject: Re: [OMPI devel] enable-smp-locks affects PSM performance > > On Nov 3, 2014, at 10:31 AM, Friedley, Andrew > wrote

Re: [OMPI devel] enable-smp-locks affects PSM performance

2014-11-03 Thread Jeff Squyres (jsquyres)
On Nov 3, 2014, at 10:31 AM, Friedley, Andrew wrote: > Thanks Jeff, we're working to verify. I don't mind the slower behavior on > the trunk; we are only concerned with the stable release series. Enabling > thread safety options on the trunk/master is no problem here. > > Did the 'more expen

Re: [OMPI devel] enable-smp-locks affects PSM performance

2014-11-03 Thread Friedley, Andrew
lease series as well? Andrew > -Original Message- > From: devel [mailto:devel-boun...@open-mpi.org] On Behalf Of Jeff > Squyres (jsquyres) > Sent: Saturday, November 1, 2014 4:49 AM > To: Open MPI Developers List > Subject: Re: [OMPI devel] enable-smp-locks affects PSM perfor

Re: [OMPI devel] enable-smp-locks affects PSM performance

2014-11-01 Thread Jeff Squyres (jsquyres)
PI. We’d like performance to be good by default > for anyone compiling their own Open MPI (and using PSM). > > Andrew > > From: devel [mailto:devel-boun...@open-mpi.org] On Behalf Of Howard Pritchard > Sent: Thursday, October 30, 2014 8:53 AM > To: Open MPI Developers

Re: [OMPI devel] enable-smp-locks affects PSM performance

2014-10-30 Thread Jeff Squyres (jsquyres)
On Oct 30, 2014, at 12:19 PM, Ralph Castain wrote: > I believe we did this in the trunk to try and force thread-safety > implementation, but I don’t believe it was intended to transition to the > release branch. The only thread-related requirement on the release branch is > that libevent be co

Re: [OMPI devel] enable-smp-locks affects PSM performance

2014-10-30 Thread Friedley, Andrew
-boun...@open-mpi.org] On Behalf Of Howard Pritchard Sent: Thursday, October 30, 2014 8:53 AM To: Open MPI Developers Subject: Re: [OMPI devel] enable-smp-locks affects PSM performance Hi Andrew, In your distribution of ompi, do you include versions of ompi to support different MPI thread

Re: [OMPI devel] enable-smp-locks affects PSM performance

2014-10-30 Thread Ralph Castain
I believe we did this in the trunk to try and force thread-safety implementation, but I don’t believe it was intended to transition to the release branch. The only thread-related requirement on the release branch is that libevent be configured with thread-support. Anyone know of a reason why mu

Re: [OMPI devel] enable-smp-locks affects PSM performance

2014-10-30 Thread Howard Pritchard
Hi Andrew, In your distribution of ompi, do you include versions of ompi to support different MPI thread safetylevels? In particular, do you include a library which supports MPI_THREAD_MULTIPLE? Just trying to better understand the requirements of you ompi package in terms of MPI thread safety.

[OMPI devel] enable-smp-locks affects PSM performance

2014-10-30 Thread Friedley, Andrew
Hi, I'm reporting a performance (message rate 16%, latency 3%) regression when using PSM that occurred between OMPI v1.6.5 and v1.8.1. I would guess it affects other networks too, but I haven't tested. The problem stems from the --enable-smp-locks and --enable-opal-multi-threads options. --e