<p>Thanks for the report. I believe Peter is on travel, otherwise he would have responded.</p>
<p>---------------------------<br>
Professor Laurence Marks<br>
Department of Materials Science and Engineering<br>
Northwestern University<br>
<a href="http://www.numis.northwestern.edu">www.numis.northwestern.edu</a> 1-847-491-3996<br>
"Research is to see what everybody else has seen, and to think what nobody else has thought"<br>
Albert Szent-Gyorgi<br>
</p>
<div class="gmail_quote">On Feb 14, 2013 2:50 AM, "Georg Eickerling" <<a href="mailto:georg.eickerling@physik.uni-augsburg.de">georg.eickerling@physik.uni-augsburg.de</a>> wrote:<br type="attribution"><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
Dear WIEN2k users,<br>
<br>
I finally want to report the bug which is responsible for the<br>
resolution dependency of the calculated HKLs mentioned in this thread.<br>
<br>
We have tracked down the problem to the file fourir.frc and in<br>
particular to the lines 201 and 203 were the PW and the MT part of<br>
the structure factors are finally added:<br>
<br>
!_REAL STRF(K)=STRF(K)+F<br>
<br>
!_COMPLEX STRF(K)=STRF(K)+conjg(F)<br>
<br>
It turns out, that in some cases the order of the HKLs is not the<br>
same for the two components, i.e. the array KZZ (used to obtain F in<br>
the above lines) contains the HKLs in a different order than STRF(K)<br>
assumes.<br>
<br>
This bug only affects HKLs with the same sin theta/labmda, which<br>
probably indicates a rounding error problem when sorting the<br>
reflexes according to their st/l.<br>
<br>
Example:<br>
<br>
For st/l 1.1 the order of the reflexes in the MT AND the PW part is<br>
<br>
5 -3 -5<br>
3 -1 -7<br>
<br>
so the result is correct.<br>
<br>
In case of st/l = 1.2 this is not true: Here, the order of the HKLs<br>
for the MT is<br>
<br>
3 -1 -7<br>
5 -3 -5<br>
<br>
while in the PW part it is still<br>
<br>
-3 -5 -5<br>
-1 -3 -7<br>
<br>
BUT in fourir.frc lapw3 simply adds up the reflexes in the order<br>
they appear without checking for consistent HKLs - causing the<br>
errors described below.<br>
<br>
We did not track down the problem to the part were the "re-ordering"<br>
is taking place, but we fixed it in a local version by using a new<br>
array as intermediate storage for the "correct" order of HKLs. I do<br>
not append this new version here as I see this more like a<br>
workaround than a fix.<br>
<br>
regards<br>
<br>
Georg Eickerling<br>
<br>
<br>
<br>
Am 23.08.2012 08:50, schrieb Georg Eickerling:<br>
> Thank you very much for the replies!<br>
><br>
> Here are some more details:<br>
><br>
> Sphere part:<br>
><br>
> st/l = 1.1:<br>
> 5 -3 -5 1.0767 -0.7523 -0.7411 0.00244 0.00000 0.00107<br>
> 0.00000 -0.01465<br>
> 3 -1 -7 1.0767 -0.7386 -0.7411 -0.00127 0.00000 -0.00030<br>
> 0.00000 0.00410<br>
><br>
> st/l = 1.2:<br>
> 3 -1 -7 1.0767 -0.7386 -0.7411 -0.00127 0.00000 -0.00030<br>
> 0.00000 0.00410<br>
> 5 -3 -5 1.0767 -0.7523 -0.7411 0.00244 0.00000 0.00107<br>
> 0.00000 -0.01465<br>
><br>
> st/l = 1.3:<br>
> -3 -5 -5 1.0767 -0.7523 -0.7411 0.00244 0.00000 0.00107<br>
> 0.00000 -0.01465<br>
> -1 -3 -7 1.0767 -0.7386 -0.7411 -0.00127 0.00000 -0.00030<br>
> 0.00000 0.00410<br>
><br>
> st/l = 1.8<br>
> -3 -5 -5 1.0767 -0.7523 -0.7411 0.00244 0.00000 0.00107<br>
> 0.00000 -0.01465<br>
> -1 -3 -7 1.0767 -0.7386 -0.7411 -0.00127 0.00000 -0.00030<br>
> 0.00000 0.00410<br>
><br>
> PW part:<br>
><br>
> st/l = 1.1:<br>
> -3 -5 -5 0.0665201058405473 1.0766653378172495<br>
> -1 -3 -7 0.0345681054384858 1.0766653378172495<br>
><br>
> st/l = 1.2:<br>
> -3 -5 -5 0.0665201058405473 1.0766653378172495<br>
> -1 -3 -7 0.0345681054384858 1.0766653378172495<br>
><br>
> st/l = 1.3:<br>
> -3 -5 -5 0.0665201058405473 1.0766653378172495<br>
> -1 -3 -7 0.0345681054384858 1.0766653378172495<br>
><br>
><br>
> stl/l = 1.8:<br>
> -3 -5 -5 0.0665201058405473 1.0766653378172495<br>
> -1 -3 -7 0.0345681054384858 1.0766653378172495<br>
><br>
><br>
> But then:<br>
><br>
><br>
> "Sum":<br>
><br>
> st/l = 1.1:<br>
> -3 -5 -5 1.0766653 -1.4379800081<br>
> -1 -3 -7 1.0766653 -1.4425910379<br>
><br>
> st/l = 1.2:<br>
> -3 -5 -5 1.0766653 -1.4106390375<br>
> -1 -3 -7 1.0766653 -1.4699320085<br>
><br>
> st/l = 1.3:<br>
> -3 -5 -5 1.0766653 -1.4379800081<br>
> -1 -3 -7 1.0766653 -1.4425910379<br>
><br>
> stl/l = 1.8:<br>
> -3 -5 -5 1.0766653 -1.4379800081<br>
> -1 -3 -7 1.0766653 -1.4425910379<br>
><br>
><br>
><br>
> Second example at higher res (data in the order st/l =1.2, 1.3, 1.8):<br>
><br>
> Sphere:<br>
> 6 0 -6 1.1894 0.8887 0.8902 -0.00212 0.00000 0.00068<br>
> 0.00000 0.00000<br>
> 6 0 -6 1.1894 0.8887 0.8902 -0.00212 0.00000 0.00068<br>
> 0.00000 0.00000<br>
> 0 -6 -6 1.1894 -0.8887 -0.8902 0.00212 0.00000 -0.00068<br>
> 0.00000 0.00000<br>
><br>
> PW:<br>
> 0 -6 -6 -0.0362599474060042 1.1893809383585805<br>
> 0 -6 -6 -0.0362599474060042 1.1893809383585805<br>
> 0 -6 -6 -0.0362599474060042 1.1893809383585805<br>
><br>
> Sum:<br>
> 0 -6 -6 1.1893809 1.7411783356<br>
> 0 -6 -6 1.1893809 -1.8246688899<br>
> 0 -6 -6 1.1893809 -1.8136982305<br>
><br>
><br>
> This is a "default" run by the way, so I am using a GMAX of 12 which<br>
> gives st/l = 1.8. What I see in addition in output3 is this:<br>
><br>
> KVEC 0 -8 -8 IN DENSITY<br>
> LIST NOT FOUND IN GENERATED K LIST<br>
> KVEC -1 -7 -9 IN DENSITY<br>
> LIST NOT FOUND IN GENERATED K LIST<br>
> KVEC -2 -8 -8 IN DENSITY<br>
> LIST NOT FOUND IN GENERATED K LIST<br>
> KVEC 0 -6 -10 IN DENSITY<br>
> LIST NOT FOUND IN GENERATED K LIST<br>
> KVEC -3 -7 -9 IN DENSITY<br>
> LIST NOT FOUND IN GENERATED K LIST<br>
> KVEC -2 -6 -10 IN DENSITY<br>
> LIST NOT FOUND IN GENERATED K LIST<br>
> KVEC -4 -8 -8 IN DENSITY<br>
> LIST NOT FOUND IN GENERATED K LIST<br>
> KVEC -1 -5 -11 IN DENSITY<br>
> LIST NOT FOUND IN GENERATED K LIST<br>
> KVEC -4 -6 -10 IN DENSITY<br>
> LIST NOT FOUND IN GENERATED K LIST<br>
> KVEC -5 -7 -9 IN DENSITY<br>
> LIST NOT FOUND IN GENERATED K LIST<br>
> KVEC -3 -5 -11 IN DENSITY<br>
> LIST NOT FOUND IN GENERATED K LIST<br>
> KVEC 0 -4 -12 IN DENSITY<br>
> LIST NOT FOUND IN GENERATED K LIST<br>
> KVEC -6 -8 -8 IN DENSITY<br>
> LIST NOT FOUND IN GENERATED K LIST<br>
> KVEC -2 -4 -12 IN DENSITY<br>
> LIST NOT FOUND IN GENERATED K LIST<br>
><br>
><br>
><br>
><br>
> Am 23.08.2012 07:37, schrieb Peter Blaha:<br>
>> Thank's for the report. I'll check that.<br>
>><br>
>> But in the meantime you could do some more analysis and compare the<br>
>> case.output3 files of two different runs.<br>
>> Are the differences coming from inside the spheres or from the<br>
>> interstital<br>
>> region ? (I expect the latter !)<br>
>><br>
>> Am 22.08.2012 18:32, schrieb Georg Eickerling:<br>
>>> Dear WIEN users,<br>
>>><br>
>>> I noticed a strange behavior of lapw3 which I do not understand:<br>
>>><br>
>>> Take for example a simple diamond case and calculate structure<br>
>>> factors from clmsum, lets say up to sin theta/lambda = 1.0:<br>
>>><br>
>>> 0 0 0 0.0000000 12.0000251726<br>
>>> -1 -1 -1 0.2427814 -4.6536716863<br>
>>> 0 0 -2 0.2803398 0.0000000000<br>
>>> 0 -2 -2 0.3964603 -3.9459734623<br>
>>> -1 -1 -3 0.4648909 -2.3988162763<br>
>>> -2 -2 -2 0.4855627 0.2266152403<br>
>>> 0 0 -4 0.5606796 -3.1297582248<br>
>>> -1 -3 -3 0.6109864 2.2069300710<br>
>>> 0 -2 -4 0.6268588 0.0000000000<br>
>>> -2 -2 -4 0.6866894 2.8777607788<br>
>>> -3 -3 -3 0.7283441 1.9393120414<br>
>>> -1 -1 -5 0.7283441 1.9663566686<br>
>>> 0 -4 -4 0.7929206 2.6458269883<br>
>>> -1 -3 -5 0.8292562 1.8056937118<br>
>>> -2 -4 -4 0.8410193 -0.0171687161<br>
>>> 0 0 -6 0.8410193 0.0000000000<br>
>>> 0 -2 -6 0.8865122 2.4363270068<br>
>>> -3 -3 -5 0.9191554 -1.6841664183<br>
>>> -2 -2 -6 0.9297818 -0.0087306004<br>
>>> -4 -4 -4 0.9711255 -2.2589221048<br>
>>><br>
>>> Repeating this calculation with sin theta/lambda = 1.1 and a diff<br>
>>> with the old hkl will just show the additional reflections as<br>
>>> expected:<br>
>>><br>
>>> diff hkl.10 hkl.11<br>
>>> 20a21,26<br>
>>>> -1 -5 -5 1.0010132 1.6970975057<br>
>>>> -1 -1 -7 1.0010132 1.5391902602<br>
>>>> 0 -4 -6 1.0107794 0.0000000000<br>
>>>> -2 -4 -6 1.0489354 -2.0944311378<br>
>>>> -3 -5 -5 1.0766653 -1.4379800081<br>
>>>> -1 -3 -7 1.0766653 -1.4425910379<br>
>>><br>
>>> Now again repeat the calculation with sin theta/lambda = 1.2:<br>
>>><br>
>>> diff hkl.11 hkl.12<br>
>>> 25,26c25,32<br>
>>> < -3 -5 -5 1.0766653 -1.4379800081<br>
>>> < -1 -3 -7 1.0766653 -1.4425910379<br>
>>> ---<br>
>>>> -3 -5 -5 1.0766653 -1.4106390375<br>
>>>> -1 -3 -7 1.0766653 -1.4699320085<br>
>>>> 0 0 -8 1.1213591 1.9443586030<br>
>>>> -3 -3 -7 1.1473400 1.3618747163<br>
>>>> -4 -4 -6 1.1558705 0.0030399053<br>
>>>> 0 -2 -8 1.1558705 0.0000000000<br>
>>>> 0 -6 -6 1.1893809 1.7411783356<br>
>>>> -2 -2 -8 1.1893809 -1.8133181764<br>
>>><br>
>>> What happens here? Why are reflections which were exactly the same<br>
>>> before suddenly different? Why I worry about this is, that if you go<br>
>>> on increasing the resolution, the differences become more severe<br>
>>> than in the example above, i.e. sin theta/lambda = 1.3:<br>
>>><br>
>>> diff hkl.12 hkl.13<br>
>>> 21,22c21,22<br>
>>> < -1 -5 -5 1.0010132 1.6970975057<br>
>>> < -1 -1 -7 1.0010132 1.5391902602<br>
>>> ---<br>
>>>> -1 -5 -5 1.0010132 -1.5542465484<br>
>>>> -1 -1 -7 1.0010132 1.5480881986<br>
>>><br>
>>> On the other hand, the differences "converge" for a given reflection<br>
>>> but more and more become "affected", i.e. sin theta/lambda = 1.8<br>
>>> vs. 1.2:<br>
>>><br>
>>><br>
>>> diff hkl.12 hkl.18<br>
>>> 21,22c21,22<br>
>>> < -1 -5 -5 1.0010132 1.6970975057<br>
>>> < -1 -1 -7 1.0010132 1.5391902602<br>
>>> ---<br>
>>>> -1 -5 -5 1.0010132 -1.5542465484<br>
>>>> -1 -1 -7 1.0010132 1.5480881986<br>
>>> 25,26c25,26<br>
>>> < -3 -5 -5 1.0766653 -1.4106390375<br>
>>> < -1 -3 -7 1.0766653 -1.4699320085<br>
>>> ---<br>
>>>> -3 -5 -5 1.0766653 -1.4379800081<br>
>>>> -1 -3 -7 1.0766653 -1.4425910379<br>
>>> 28c28<br>
>>> < -3 -3 -7 1.1473400 1.3618747163<br>
>>> ---<br>
>>>> -3 -3 -7 1.1473400 -1.3370357923<br>
>>> 31c31<br>
>>> < 0 -6 -6 1.1893809 1.7411783356<br>
>>> ---<br>
>>>> 0 -6 -6 1.1893809 -1.8136982305<br>
>>><br>
>>><br>
>>> Looking at the result of a refinement of a structural model against<br>
>>> the different HKLs, the "high resolution version" seems to be<br>
>>> "wrong" compared to the low-res one.<br>
>>><br>
>>> Thank you very much in advance for any comments on this.<br>
>>><br>
>>> regards<br>
>>><br>
>>> Georg Eickerling<br>
>>><br>
>>><br>
>><br>
><br>
><br>
<br>
<br>
--<br>
============================<br>
Dr. Georg Eickerling<br>
Universitaet Augsburg<br>
Institut fuer Physik<br>
Lehrstuhl fuer Chemische Physik und Materialwissenschaften<br>
Universitaetsstr. 1<br>
86159 Augsburg<br>
<br>
E-Mail: <a href="mailto:georg.eickerling@physik.uni-augsburg.de">georg.eickerling@physik.uni-augsburg.de</a><br>
Phone: +49-821-598-3362<br>
FAX: +49-821-598-3227<br>
WWW: <a href="http://www.physik.uni-augsburg.de/cpm/" target="_blank">http://www.physik.uni-augsburg.de/cpm/</a><br>
=====================================================<br>
_______________________________________________<br>
Wien mailing list<br>
<a href="mailto:Wien@zeus.theochem.tuwien.ac.at">Wien@zeus.theochem.tuwien.ac.at</a><br>
<a href="http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien" target="_blank">http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien</a><br>
</blockquote></div>