Error in writing HDF5 outputs about velociraptor-stf HOT 5 CLOSED

@ymbahe can you point me to the full log file? This is a recent check I introduced while working on #88, but there clearly is some mistake for some some unaccounted for cases.

Note that it should be possible to work around this by building with -DCMAKE_BUILD_TYPE=Release.

from velociraptor-stf.

Hi @rtobar -- the full log file is at /cosma7/data/dp004/dc-bahe1/BH_REPOS/stdres_base2_velcut0p5/logs/vr.3651220.out (and .../vr.3651220.err for the error file). The log file actually has a warning message at the top:

MPI startup(): PMI server not found. Please set I_MPI_PMI_LIBRARY variable if it is not a singleton case.

Not sure whether this is related though. I'll try your workaround later; I found an older binary file in the meantime which seems to work too.

from velociraptor-stf.

@ymbahe please try the new issue-113 branch, I believe that should fix things for you.

As expected, this seems to have just been a problem in an assertion and not in actual code, and which was triggered only by non-Release builds when compiling against a non-parallel HDF5 library (which is your case, based on the contents of your logfile). So please give the new branch a try, and after confirmation I'll merge to the master branch.

from velociraptor-stf.

Hi @rtobar, thanks for the fix. Yes, it looks like this works fine, so it can be merged to the master.

from velociraptor-stf.

Great, I've merged the fix to the master branch now.

from velociraptor-stf.