Re: [OMPI devel] Common symbols warning

Dave, my bad, the error is ignored as it should. i will then close the related PR since it is now irrelevant Cheers, Gilles On 4/16/2015 12:30 AM, Dave Goodell (dgoodell) wrote: On Apr 14, 2015, at 11:02 PM, Gilles Gouaillardet wrote: Dave, my understanding is that the

Re: [OMPI devel] Common symbols warning

Ah, oh well. Its my code so I went ahead and committed that fix. -Nathan On Wed, Apr 15, 2015 at 09:02:48AM -0700, Ralph Castain wrote: >FWIW: Gilles has a pending PR that fixes all of these > https://github.com/open-mpi/ompi/pull/530 > > On Apr 15, 2015, at 8:55 AM, Nathan Hjelm

Re: [OMPI devel] Common symbols warning

FWIW: Gilles has a pending PR that fixes all of these https://github.com/open-mpi/ompi/pull/530 > On Apr 15, 2015, at 8:55 AM, Nathan Hjelm wrote: > > On Tue, Apr 14, 2015 at 08:14:09PM -0700, Ralph Castain wrote: >> Dave

Re: [OMPI devel] Common symbols warning

On Tue, Apr 14, 2015 at 08:14:09PM -0700, Ralph Castain wrote: >Dave committed this earlier today, and here is the first error report: >WARNING! Common symbols found: > comm_request.o: 0068 C ompi_comm_request_mutex > comm_request.o:

Re: [OMPI devel] [OMPI commits] Git: open-mpi/ompi branch revert-520-valgrind_cleanness created. dev-1504-g7a8a4a0

Looks like they just accidentally created a branch on the github repo. I can confirm that they indeed did not revert the commit but only fixed the relevant issue. -Nathan On Wed, Apr 15, 2015 at 07:36:16AM -0700, Ralph Castain wrote: >Sare you going to restore the rest of it? Or are

Re: [OMPI devel] Common symbols warning

On Apr 14, 2015, at 11:02 PM, Gilles Gouaillardet wrote: > Dave, > > my understanding is that the presence of common symbols should be treated as > a warning > (and hence make install should not fail) > > makes sense ? It should be a warning... are you seeing otherwise in

Re: [OMPI devel] [OMPI commits] Git: open-mpi/ompi branch revert-520-valgrind_cleanness created. dev-1504-g7a8a4a0

Hmmm... I saw that Mike closed the PR with reverting Nathan's commit and merged another one which contains just a small fix ( https://github.com/open-mpi/ompi/pull/532). I asked Nathan to look at it and if he has no objections, it resolves the problem. Best regards, Elena On Wed, Apr 15, 2015 at

Re: [OMPI devel] [OMPI commits] Git: open-mpi/ompi branch revert-520-valgrind_cleanness created. dev-1504-g7a8a4a0

S….are you going to restore the rest of it? Or are we asking Nathan to refile it without that one piece? > On Apr 15, 2015, at 7:26 AM, Elena Elkina wrote: > > Hi Ralph. > > We don't need to revert the whole commit, just to fix this small part. I > proposed a

Re: [OMPI devel] [OMPI commits] Git: open-mpi/ompi branch revert-520-valgrind_cleanness created. dev-1504-g7a8a4a0

Hi Ralph. We don't need to revert the whole commit, just to fix this small part. I proposed a fast fix for that in the PR but probably we need to fix it more intellectually. Best regards, Elena On Wed, Apr 15, 2015 at 6:08 PM, Ralph Castain wrote: > I’m really puzzled - I

Re: [OMPI devel] [OMPI commits] Git: open-mpi/ompi branch revert-520-valgrind_cleanness created. dev-1504-g7a8a4a0

I’m really puzzled - I saw where you fixed the one part of this commit that caused a problem for you. But what happened to the rest of it? Was it really necessary to revert the entire thing, or was it only that piece that caused a problem? > On Apr 15, 2015, at 5:41 AM, git...@crest.iu.edu

Re: [OMPI devel] segmentation fault

OK, Thank you for your help. I'll will do it. But, is it still possible to restart a mpi program previously checkpoint, on another different nodes ? And also, i don't know how to look at the core file from segfault to send you the backtrace. If you can guide me. Thank you. Hope to read you.

Re: [OMPI devel] Common symbols warning

Dave, my understanding is that the presence of common symbols should be treated as a warning (and hence make install should not fail) makes sense ? Cheers, Gilles On 4/15/2015 12:14 PM, Ralph Castain wrote: Dave committed this earlier today, and here is the first error report: WARNING!

Re: [OMPI devel] segmentation fault

I’m afraid I don’t know much about that code, but I can try to help. Can you look at the core file from the segfault and send the backtrace? > On Apr 14, 2015, at 8:04 AM, Cyrille DIBAMOU MBEUYO > wrote: > > Thank you for your reply. > > This error appear on the master

[OMPI devel] Common symbols warning

Dave committed this earlier today, and here is the first error report: WARNING! Common symbols found: comm_request.o: 0068 C ompi_comm_request_mutex comm_request.o: 0001 C ompi_comm_request_progress_active comm_request.o: 0210