[Wien] lapw2para -qtl -c -p -band problem

Peter Blaha pblaha at zeus.theochem.tuwien.ac.at
Mon Apr 11 15:53:27 CEST 2005


Hi,
This is quite trivial:
Running lapw2 with -qtl is doing its work only on one processor. The -p 
just tells the program that it should read the parallel vectors.
Since -qtl produces (usually a large) qtl-file, it does not make too much 
sense to distribute this over the nodes and then summ up all qtl_xx files.
Furthermore, -qtl is used only "once" (i.e. not within scf cycle).

So don't worry, this is normal behaviour.

> 
> First of all, sorry for the repost but I didn't get reply last time and I'm a bit confused.
> 
> My problem is that tht within tasks: bandstructure, lapw2 is not getting sent to the nodes. When I run standard scf this works fine. I have been reading the mailing list and noticed that previously there may have been a problem when the amount of jobs changed. During the standard SCF I have 12 k points and this is distributed amongst 6 nodes equally. When I run bandstructure I have 109 k points in the klist so this is distributed over 7 jobs (18 points over 6 nodes followed by a single point over the fastest computer). When I the run lapw2cpara -qtl -c -p band testpara2 reports that the junks are distributed over 6 nodes (100 percent), but only the server is showing any work and only lapw2.def is created. Also the cpu usage is running very irradict (seen using "top"). I checked the version of lapw2para we says 1.10, so I guess that either no changes in the revision number have been made or lapw2para is never updated (which I doubt). I'm running the latest version of the Wi!
 en2k code on a linux cluster. I've tried running in completely new folders, but the problem still remains. 
> 
> I would really apreciate some form of help on this problem. 
> 
> Michael


                                      P.Blaha
--------------------------------------------------------------------------
Peter BLAHA, Inst.f. Materials Chemistry, TU Vienna, A-1060 Vienna
Phone: +43-1-58801-15671             FAX: +43-1-58801-15698
Email: blaha at theochem.tuwien.ac.at    WWW: http://info.tuwien.ac.at/theochem/
--------------------------------------------------------------------------




More information about the Wien mailing list