[Wien] Still problem with CLMSUM

Michael Gurnett michael.gurnett at kau.se
Sat Mar 11 22:00:31 CET 2006


The line that gives the crash is 

lapw0              0000000000428235  MAIN__                    472
lapw0.F


After dstart case.clmsum seem completely correct
After the 1st cycle has completed case.clmsum_old is fine

I checked the case.scfm file and found the following

:DIS  :  CHARGE DISTANCE       0.3733027
:REDuction and QMX before broyd:    1.1383    1.3646    0.0349
:PLANE:  INTERSTITAL DISTANCE   NaN     
 :Pratt iteration over-ride, DMIX   3.488558598726488E-002
       BROYD MIXING SCHEME WITH 0.035

       CHARGES OF MIXED CHARGE DENSITY
:CTO  : TOTAL   INTERSTITIAL CHARGE= NaN       

:NEC03: NUCLEAR AND ELECTRONIC CHARGE 1452.00000 NaN       NaN      

Not sure if this is an issue.

The strange thing is that these calculations work fine for GMax 16 but
gives this problem with GMax 20. As the structure contains H, I need to
increase GMax.

all case.error files are empty with the exception of lapw0.error which
has the usual helpful information of "Error in LAPW0".

I've been looking through the other files, but I haven't seen anything
that stands out as being wrong.

If someone wants to try the struct file I'll gladly post it

Michael




On Sat, 2006-03-11 at 11:58 -0600, L. D. Marks wrote:
> 1) Check all the case.error files
> 2) Read (in detail) the case.scf file to see what went on, particularly in 
> the last part from case.scfm
> 3) Run, by hand, the seperate steps to find out where things are going 
> wrong.
> 
> Some combination of this and your brain (the best computer) should lead 
> you in the right direction.
> 
> On Sat, 11 Mar 2006, Michael Gurnett wrote:
> 
> > This occured on the start of the 2nd scf in a clean start. Many good
> > ideas, but none of the 4 points are valid in this case. I've started in
> > a new folder now so that I can be sure that no remnants from previous
> > run can be causing this (even though clean_lapw was used before
> > starting)
> >
> > Michael
> >
> > On Sat, 2006-03-11 at 11:40 -0600, L. D. Marks wrote:
> >> It's not in lapw0. Many possible sources:
> >> 1) Is :ITE increasing in your case.scf?
> >> 2) Did the iteration diverge (PW too large) and/or show large angles in
> >> mixer?
> >> 3) If you were using -in1new, have the values in case.in1 gone bad; can be
> >> associated with 2), or using -in1new too soon. Does not look like you
> >> were, but you might have had values in case.in1.
> >> 4) Did the minimization just move some large distances, e.g. 0.3 a.u? If
> >> it did, the discontinuities introduced at the RMT may have been too much
> >> for your values in case.inm; -renorm helps, or just reduce the initial
> >> trust radius in case.inM with PORT.
> >>
> >> These are just a few. I think you have to track down yourself where things
> >> went wrong; it is almost certainly "user error".
> >>
> >> On Sat, 11 Mar 2006, Michael Gurnett wrote:
> >>
> >>> I'm recompiling as I type this. It still seems to be a problem with the
> >>> clmsum file much like before with the following problem
> >>>
> >>> MIXED TOTAL CHARGE DENSITY IN INTERSTITIAL   NEW TOTAL CHARGE DENSITY
> >>>
> >>>            1616829 NUMBER OF PW
> >>>       0    0    0 NaN                NaN                NaN        NaN
> >>>       0    0   -1 NaN                NaN                NaN        NaN
> >>>       0    0    1 NaN                NaN                NaN        NaN
> >>>       0    0   -2 NaN                NaN                NaN        NaN
> >>>       0    0    2 NaN                NaN                NaN        NaN
> >>>
> >>> So I'm guessing that it's not a problem with lapw0
> >>>
> >>> Michael
> >>>
> >>> On Sat, 2006-03-11 at 07:58 -0600, L. D. Marks wrote:
> >>>> You are not providing enough information
> >>>> a) Recompile lapw0 with -tracebak
> >>>> b) Look at the output, and find the relevant line in the code
> >>>> c) Look at case.scf0, case.output0 etc
> >>>> d) Look at case.clmsum
> >>>>
> >>>> On Sat, 11 Mar 2006, Michael Gurnett wrote:
> >>>>
> >>>>> Ran the new version of Wien2k. This occurs directly at the start of the
> >>>>> second cycle. What is happening
> >>>>>
> >>>>>>>> (min) status after run_lapw -I -i 40 -fc 1.0 -p \: 9 -> exit
> >>>>> lapw0              0000000000403E29  Unknown               Unknown  Unknown
> >>>>> libc.so.6          000000351E01C3CF  Unknown               Unknown  Unknown
> >>>>> lapw0              0000000000403EEA  Unknown               Unknown  Unknown
> >>>>> lapw0              00000000004282A9  Unknown               Unknown  Unknown
> >>>>> lapw0              0000000000481664  Unknown               Unknown  Unknown
> >>>>> lapw0              00000000004709CC  Unknown               Unknown  Unknown
> >>>>> lapw0              0000000000470D63  Unknown               Unknown  Unknown
> >>>>> lapw0              000000000049930C  Unknown               Unknown  Unknown
> >>>>> lapw0              00000000004B52DE  Unknown               Unknown  Unknown
> >>>>> lapw0              00000000004B757B  Unknown               Unknown  Unknown
> >>>>> Image              PC                Routine            Line        Source
> >>>>> forrtl: severe (64): input conversion error, unit 8, file /home/Michael/Wien2k/HCT-Lisqrt3/HCT-Lisqrt3.clmsum
> >>>>> in cycle 2    ETEST: 84.5785025000000000   CTEST: .3760374
> >>>>> ec cc and fc_conv 1 1 0
> >>>>>
> >>>>> _______________________________________________
> >>>>> Wien mailing list
> >>>>> Wien at zeus.theochem.tuwien.ac.at
> >>>>> http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien
> >>>>>
> >>>>
> >>>> Note: if you have an old email address for me, please note that "nwu" has
> >>>> been changed to "northwestern".
> >>>> -----------------------------------------------
> >>>> Laurence Marks
> >>>> Department of Materials Science and Engineering
> >>>> MSE Rm 2036 Cook Hall
> >>>> 2220 N Campus Drive
> >>>> Northwestern University
> >>>> Evanston, IL 60201, USA
> >>>> Tel: (847) 491-3996 Fax: (847) 491-7820
> >>>> email: L-marks at northwestern dot edu
> >>>> http://www.numis.northwestern.edu
> >>>> -----------------------------------------------
> >>>>
> >>>> _______________________________________________
> >>>> Wien mailing list
> >>>> Wien at zeus.theochem.tuwien.ac.at
> >>>> http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien
> >>>>
> >>>
> >>> _______________________________________________
> >>> Wien mailing list
> >>> Wien at zeus.theochem.tuwien.ac.at
> >>> http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien
> >>>
> >>
> >> Note: if you have an old email address for me, please note that "nwu" has
> >> been changed to "northwestern".
> >> -----------------------------------------------
> >> Laurence Marks
> >> Department of Materials Science and Engineering
> >> MSE Rm 2036 Cook Hall
> >> 2220 N Campus Drive
> >> Northwestern University
> >> Evanston, IL 60201, USA
> >> Tel: (847) 491-3996 Fax: (847) 491-7820
> >> email: L-marks at northwestern dot edu
> >> http://www.numis.northwestern.edu
> >> -----------------------------------------------
> >>
> >> _______________________________________________
> >> Wien mailing list
> >> Wien at zeus.theochem.tuwien.ac.at
> >> http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien
> >>
> >
> > _______________________________________________
> > Wien mailing list
> > Wien at zeus.theochem.tuwien.ac.at
> > http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien
> >
> 
> Note: if you have an old email address for me, please note that "nwu" has
> been changed to "northwestern".
> -----------------------------------------------
> Laurence Marks
> Department of Materials Science and Engineering
> MSE Rm 2036 Cook Hall
> 2220 N Campus Drive
> Northwestern University
> Evanston, IL 60201, USA
> Tel: (847) 491-3996 Fax: (847) 491-7820
> email: L-marks at northwestern dot edu
> http://www.numis.northwestern.edu
> -----------------------------------------------
> 
> _______________________________________________
> Wien mailing list
> Wien at zeus.theochem.tuwien.ac.at
> http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien
> 



More information about the Wien mailing list