Page 4 of 7
Re: Open Shell Systems
Posted: Mon Mar 24, 2014 9:52 am
by Daniele Varsano
Dear Vito,
1) Yes, but you have to find a system with a stable triplet initial state.
4) For my experience, it is mandatory for finite system (molecules), usually unnecessary for bulk semiconductors. Of course a testing is always useful.
Best,
Daniele
Re: Open Shell Systems
Posted: Mon Mar 24, 2014 6:18 pm
by vitoversace
Dear Daniele,
From whatever reason this command ./ypp -e s just does not work for me.
I am using 4.4.1 version. Attached you find my files.
Bests
Re: Open Shell Systems
Posted: Mon Mar 24, 2014 6:40 pm
by Daniele Varsano
Dear Vito,
please post the report and standard output of your BSE calculation, and the results of
>ls ./SAVE
as well.
Best,
Daniele
Re: Open Shell Systems
Posted: Mon Mar 24, 2014 6:51 pm
by vitoversace
Dear Daniele,
The problem is solved.
Since i did the whole procedure exactly according to tutorial, db.BS_diago_Q01 and db.BS_Q1 are saved in 06_BSE file.
So ./ypp -e s simply can not find db.BS_diago_Q01 and db.BS_Q1.
The solution was just copy & paste these two files into SAVE file. And now it finds the necessary data.
But a question : what are all these fragments data ? which information do they contain exactly ?
With best wishes
Re: Open Shell Systems
Posted: Mon Mar 24, 2014 9:22 pm
by Daniele Varsano
Dear Vito,
it is just a matter of I/O. In case of cumbersome calculations, better to have fragmented databases of reasonable size than a single database of many Gb.
Best,
Daniele
Re: Open Shell Systems
Posted: Tue Mar 25, 2014 2:37 am
by vitoversace
Dear Daniele,
1) what is exactly the formula for the determination of dimension of bse matrix ? I know that depends on number of k-points, number of occupied and empty states. But for instance multiplication of these parameters do not give me
the dimension reported in the o.exc_E_sorted file.
2) For maximally how big bse matrix is exact diagonalization effective and time reasonable ?
Bests
Re: Open Shell Systems
Posted: Tue Mar 25, 2014 6:10 am
by Daniele Varsano
Dear Vito,
1) Hdim=Nc x Nv x NK_points(BZ)
if you are considering the coupling there is a factor 2.
2) Some thousands.
Best,
Daniele
Re: Open Shell Systems
Posted: Tue Mar 25, 2014 4:27 pm
by vitoversace
Dear Daniele,
when typing ypp after ypp -e w i get the following attached error !!!
Bests
Re: Open Shell Systems
Posted: Tue Mar 25, 2014 4:53 pm
by Daniele Varsano
No idea,
it could be whatever,
anyway if you do not post any input/report and revision of the code it is absolutely impossible to help you.
Are you using the last revision of the code?
Please post all the useful information can help to spot the problem.
Best,
Daniele
Re: Open Shell Systems
Posted: Tue Mar 25, 2014 5:12 pm
by vitoversace
Attached all my files. Every thing works very well except ypp -e w after running bse calculation !!!
If you need more information, let me please know and by which commands.
I am using 3.3.0 version.
Bests