From: Jose M. de la R. T. <del...@gm...> - 2017-10-28 15:01:20
|
Hi John, Thanks for your feedback. As you already guessed, the binaries are compiled against a given version of libmpi (if I remember correctly we are using a CentOS 6 or similar). So, I think that if you build from source it should properly link to the correct libmpi version. Let us know if that work for you or any other idea. Cheers, Jose Miguel On Sat, Oct 28, 2017 at 8:12 AM, John Heumann <joh...@co...> wrote: > Greetings! > > I recently installed the v1.1_2017-06-14 binaries on a RHEL7 system. > None of the additional packages are currently installed. All of the small > tests pass except 2 which seem to be using eman components (not installed), > so I'm ignoring those. The Xmipp workflow test failed at the ML2D step, > however. The stdout log indicates failure to load libmpi.so.1. It's not > surprising that it can't find libmpi.so.1 since the current version is > libmpi.so.12.0.6. As a temporary workaround I made a symbolic link > libmpi.so.1 -> libmpi.so, which, in turn, points to libmpi.so.12.06. The > test passes after this kludge. > > Why is this test trying to load a specific, older version of the mpi > library? Isn't the core functionality in the newer versions backwards > compatible? Would I avoid this problem if I build from source myself? Or > have I configured something incorrectly? > > Thanks in advance! > > Regards, > -jh- > > -- > John M. Heumann > Department of Molecular, Cellular, and Developmental Biology > 347 UCB, University of Colorado > Boulder, CO 80309-0347 > > > ------------------------------------------------------------ > ------------------ > Check out the vibrant tech community on one of the world's most > engaging tech sites, Slashdot.org! http://sdm.link/slashdot > _______________________________________________ > scipion-users mailing list > sci...@li... > https://lists.sourceforge.net/lists/listinfo/scipion-users > |