Chad Junkermeier wrote: > After looking at the examples and what has been written in the forums > my best guess as to the general form of the file to run pp.x (before > running average.x) is given here. > > > &inputPP > outdir='/workspace/junky/WFC_2/tmp', > plot_num=11 > filplot = 'WFC_2.2.pot' > prefix='WFC_2' > / > &plot > iflag=3, > output_format=3, > nx=10, > ny=10, > nz =5, > / > > My problem is that this is taking a really long time to run; much, much > longer than the corresponding singlepoint SCF calculation ran using the > same number of processors. At this point, the pp.x has been running for > 43 hours. I can't believe that this is normal. Here are the last > couple of lines from the pp.x output file: > > > Writing data to file WFC_2.2.pot > Reading data from file WFC_2.2.pot > > > What is a normal length of time for this to run (I am running on 48 > processors)? Is my input for pp.x really screwed up? Is pp.x just > sitting there eating up computer time and not computing anything?
Has the WFC_2.2.pot file been produced? If so, maybe the problem is that the second part of the pp run (that corresponding to the &plot namelist) should be performed on a single processor and in the past it used to get stuck when it was run in parallel (maybe this has been fixed in the meanwhile...) I suggest to first run pp.x in parallel leaving the &plot namelist blank, and then run again pp.x in serial specifying the plot part. HTH GS > > Thank you for your help. > -- o ------------------------------------------------ o | Gabriele Sclauzero, PhD Student | | c/o: SISSA & CNR-INFM Democritos, | | via Beirut 2-4, 34014 Trieste (Italy) | | email: sclauzer at sissa.it | | phone: +39 040 3787 511 | | skype: gurlonotturno | o ------------------------------------------------ o
