On Wednesday, March 2, 2016, Ralph Castain <r...@open-mpi.org> wrote:
> Perhaps we can all meet in the middle: > > * remove the enable-debug-by-default logic > > I have no strong opinion about that > * have mpirun -version clearly state that it is a debug build and include > Jeff’s warning about debug builds being used for performance testing > +1 I’m increasingly feeling that we shouldn’t output that message every time > someone executes a debug-based operation, even if we add a param to turn > off the warning. > +1 > > On Mar 2, 2016, at 5:48 AM, Gilles Gouaillardet < > gilles.gouaillar...@gmail.com > <javascript:_e(%7B%7D,'cvml','gilles.gouaillar...@gmail.com');>> wrote: > > Ralph, > > that means that when a developer truse to reproduce exactly what an end > user did, he/she will get a different behavior because only one of them is > known. imho, that sounds a bit too crazy. > > what about an other approach : > have mpirun --version (and MPI api if any) clearly state this is a debug > build ? > I mean a "serious" benchmark should list the third party libs and their > versions, so that could be enough. > > Cheers, > > Gilles > > On Wednesday, March 2, 2016, Ralph Castain <r...@open-mpi.org > <javascript:_e(%7B%7D,'cvml','r...@open-mpi.org');>> wrote: > >> What about this crazy idea? We already have .opal_unignore that looks at >> the username. Well, what if we did the same thing here? Have autogen.pl >> look at the username - if it is a known developer, then enable debug. If >> not, then disable it. >> >> I am just concerned that we are going to spend a bunch of time responding >> to annoyed/concerned users about this warning. Every time we ask someone to >> run a debug build to diagnose a problem, or someone builds a debug build >> for their own diagnostic purposes, they are going to get an unnecessary >> warning. We already have had to contend with the annoyed users dealing with >> auto-binding because we are trying to respond to the benchmarkers who >> refuse to make an effort to compare apples to apples - I would advise >> against extending that practice. >> >> I personally don’t mind having to explicitly enable-debug as I build from >> a platform file, but making it more convenient for those developers who >> don’t would be nice. >> >> > On Mar 2, 2016, at 4:51 AM, Jeff Squyres (jsquyres) <jsquy...@cisco.com> >> wrote: >> > >> > On Mar 2, 2016, at 6:30 AM, Mark Santcroos <mark.santcr...@rutgers.edu> >> wrote: >> >> >> >>> On 02 Mar 2016, at 5:06 , Gilles Gouaillardet <gil...@rist.or.jp> >> wrote: >> >>> what about *not* issuing this warning if OpenMPI is built from git ? >> >>> that would be friendlier for OMPI developers, >> >>> and should basically *not* affect endusers, since they would rather >> build OMPI from a tarball. >> >> >> >> VPATH builds aren't debug builds by default, or are they? >> > >> > Correct -- VPATH builds are *not* debug-by-default builds. >> > >> > Just to be clear: the debug-by-default builds are activated when a .git >> directory is seen in the build directory. This is for people who git clone >> and do a non-VPATH build. Such people are typically developers. >> > >> > And just to make *that* more clear: building from a tarball will not >> have a .git directory, so you get an optimized build by default. >> > >> > -- >> > Jeff Squyres >> > jsquy...@cisco.com >> > For corporate legal information go to: >> http://www.cisco.com/web/about/doing_business/legal/cri/ >> > >> > _______________________________________________ >> > devel mailing list >> > de...@open-mpi.org >> > Subscription: http://www.open-mpi.org/mailman/listinfo.cgi/devel >> > Link to this post: >> http://www.open-mpi.org/community/lists/devel/2016/03/18664.php >> >> _______________________________________________ >> devel mailing list >> de...@open-mpi.org >> Subscription: http://www.open-mpi.org/mailman/listinfo.cgi/devel >> Link to this post: >> http://www.open-mpi.org/community/lists/devel/2016/03/18665.php > > _______________________________________________ > devel mailing list > de...@open-mpi.org <javascript:_e(%7B%7D,'cvml','de...@open-mpi.org');> > Subscription: http://www.open-mpi.org/mailman/listinfo.cgi/devel > Link to this post: > http://www.open-mpi.org/community/lists/devel/2016/03/18666.php > > >
