[Wien] 323 atoms, no symmetry ... is it doable ?

[Peter Blaha]

I can only agree with what Laurence said.

Did you look at your structure (xcrysden or VESTA)?

You basically put a Ts atom and an O2 molecule onto a surface. Is this 
what you want ? It looks more than strange.

Your supercell is way too large in x and y (Your Ts have a distance of 
20 ang), but your vacuum is much too small (Your TsO2 is "nearly" in the 
middle of the slab).

I suggest you start with

a) just the surface slab (no supercell at all in x,y) as Laurence 
suggested. Your model contains so far a H-saturated surface + plain Si 
terminated surface. I'd never do that. Either passivate both sides of 
the slab, or none. In addition I'd use 2 "identical" surface-terminations.

b) a 2x2x1 supercell, but put your ad-atoms into "reasonable" positions.
In most cases in WIEN2k it is also better to put the adsorbate on both 
sides AND KEEP inversion symmetry. I don't know if this is possible here

c) You need mpi for this, on many cores (probably 128 - 512 might be a 
good number)

d) init_lapw with the wien2k_23 program and start with -prec 0n ; i.e. 
do not set parameters yourself unless you are an expert. A 300 atom cell 
does not need many k-points, if it is an insulator, probably ONE is 
enough (at least for a first structural optimization with -min).



Am 6/25/23 um 19:38 schrieb Laurence Marks:
> You need to understand more about Wien2k before you try something as 
> ambituous as this. Some comments:
> 
> a) You need to compile and check mpi version of Wien2k first. You are 
> currently not using it.
> b) You have 4x4x5 k-points, which is way too many. The default (23.2) 
> uses 2x2x2 and I would suggest just using 1 without shift at first. This 
> will be 40 times faster.
> c) Your calculation appears to have TsO2 in between an OH terminated 
> silicate on one side, and a (strange) Si terminated silicate on the 
> other. I suggest a far smaller cell using IO2 (similar, probably, to Ts) 
> as a sanity test first. I think your Si termination is probably wrong, 
> and I doubt that TsO2 is even remotely valid.
> d) You appear from the dayfile and job output log to be running 32 
> k-point parallel jobs, each with 4 OMP. I assume that you allocated 128 
> cores for this, although your script appears to show #SBATCH=1 which is 
> one node.
> e) Assuming that you have allocated 128 cores, your machines file should 
> have
> omp_global:4
> 1:localhost:32
> 
> 332 atoms is not that large, although I think it will take forever to 
> reach anything reasonable as I don't think that your structure is 
> realistic. For context, I am currently running a 1212 atom calculation 
> using 3 kpts on a cluster using 224 Gold 6130 cores, with one iteration 
> taking 90 minutes.
> 
> I suggest that you start with just calculating a surface slab of your 
> silicate to get that right and learn, before trying something this 
> ambitious.
> 
> On Sun, Jun 25, 2023 at 11:57 AM Ilias Miroslav, doc. RNDr., PhD. 
> <miroslav.ilias at umb.sk <mailto:miroslav.ilias at umb.sk>> wrote:
> 
>     Greetings,
> 
>     I was trying to kick on at least 1 SCF iteration within  the limit
>     of 8 hours, but now way. The system of 323 atoms was precisely
>     prepared using w2web, I set up 32 threads with omp_global=4, but no way.
> 
>     Calculation remains in parallel lapw1, than got killed.
> 
>     I guess WIen2k can not do anything with such big systems on a good
>     cluster like https://hpc.gsi.de/virgo/preface.html
>     <https://hpc.gsi.de/virgo/preface.html>  ?
> 
> 
>     All files are here:
> 
>     https://github.com/miroi/open-collection/tree/master/theoretical_chemistry/software/wien2k/runs/TsOO_on_quartz_G/TsOOQg <https://github.com/miroi/open-collection/tree/master/theoretical_chemistry/software/wien2k/runs/TsOO_on_quartz_G/TsOOQg>
> 
> 
>     Best,
> 
>     Miro
>     _______________________________________________
>     Wien mailing list
>     Wien at zeus.theochem.tuwien.ac.at <mailto:Wien at zeus.theochem.tuwien.ac.at>
>     http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien
>     <http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien>
>     SEARCH the MAILING-LIST at:
>     http://www.mail-archive.com/wien@zeus.theochem.tuwien.ac.at/index.html <http://www.mail-archive.com/wien@zeus.theochem.tuwien.ac.at/index.html>
> 
> 
> 
> -- 
> Professor Laurence Marks (Laurie)
> Department of Materials Science and Engineering
> Northwestern University
> www.numis.northwestern.edu <http://www.numis.northwestern.edu>
> "Research is to see what everybody else has seen, and to think what 
> nobody else has thought", Albert Szent-Györgyi
> 
> _______________________________________________
> Wien mailing list
> Wien at zeus.theochem.tuwien.ac.at
> http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien
> SEARCH the MAILING-LIST at:  http://www.mail-archive.com/wien@zeus.theochem.tuwien.ac.at/index.html

-- 
Peter Blaha, Inst.f. Materials Chemistry, TU Vienna, A-1060 Vienna
Phone: +43-1-58801-165300          Email: peter.blaha at tuwien.ac.at
WWW: http://www.imc.tuwien.ac      WIEN2k: http://www.wien2k.at