Alright, so it's follow-up time. After looking over which of the 25 runs finished (4 of them did), and moving them to a separate directory, I re-submitted the runs to the queue. This time, apparently 7 of them have decided to stick and do a run. No guarantees on them finishing, but it looks promising as the error typically evolves for me after the wavefunction collection (wfc) call in the output. So the question now becomes: if the input files are working properly when they decide to stick, what could be causing them to not stick? The outdir is the same for every input file, so I suppose it could be just too busy with the calls from all of the different runs. Has anyone else had issues with this? Moreover, is there a limit to how many runs I should direct to a specific outdir? Would it be better to assign a unique outdir for each run? Thanks! I appreciate all of the help.
Most Sincerely, Patrick University of Notre Dame On Sun, Mar 7, 2010 at 4:29 PM, Patrick Holvey <pholvey at gmail.com> wrote: > Paolo (and everyone else), > > Thanks for the quick reply. Just a few questions. If the error in davcio > results solely from a bad outdir, why did the error resolve itself in my > original input file when I changed the celldms to correspond with Linh's > suggestions? Further, since the runs worked on 3 of the 25 submitted input > files, doesn't that show that there is no error in writing to the > destination? The only reason I ask is that I'm almost completely positive > that I'm not running into disk limitations (although I will check for sure > on monday), and if that's not the issue, then I need to hammer out what is > really causing this. > > Thanks again, > > > Patrick > University of Notre Dame > > On Sun, Mar 7, 2010 at 6:19 AM, Paolo Giannozzi <giannozz at > democritos.it>wrote: > >> Patrick, >> >> there is nothing wrong in your data (or, more exactly, nothing >> obviously wrong). In any event, nothing in your data can cause >> an "error in davcio", apart from a bad value of "outdir". See here: >> http://www.quantum-espresso.org/user_guide/node46.html >> and here for parallel execution, in sec.3.3.1: >> http://www.quantum-espresso.org/user_guide/node17.html . >> >> I run one of your data file and everything seems ok. You may want to >> try to find optimal values of -npool and -ndiag to achieve maximum >> speed, but this has nothing to do with your problem, which is related >> uniquely to the place where you are trying to write files. >> >> P. >> >> >> >> >> --- >> Paolo Giannozzi, Dept of Physics, University of Udine >> via delle Scienze 208, 33100 Udine, Italy >> Phone +39-0432-558216, fax +39-0432-558222 >> >> >> >> >> _______________________________________________ >> Pw_forum mailing list >> Pw_forum at pwscf.org >> http://www.democritos.it/mailman/listinfo/pw_forum >> >> > > > -- > Patrick Holvey > pholvey at nd.edu > 442 Siegfried Hall > Notre Dame, IN 46556 > Cell: (865)-659-9908 > -- Patrick Holvey pholvey at nd.edu 442 Siegfried Hall Notre Dame, IN 46556 Cell: (865)-659-9908 -------------- next part -------------- An HTML attachment was scrubbed... URL: http://www.democritos.it/pipermail/pw_forum/attachments/20100307/043c535f/attachment.htm
