[QE-users] Problem about requirement on computing powers

[Pietro Davide Delugas]

Hi

- to facilitate the convergence with SrTiO3 one has usually to either 
increase the number of bands used or use   randomly generated starting 
wave functions. Try to use occupations = smearing it could help  the scf 
cycle in the first iterations.
- you are using noncollinear + spinorbit coupling, to make these 
calculation converge usually one needs to specify a starting 
magnetization in your case, if SrTiO_3 is insulating, you should try to 
initialize Ti with starting magnetization ( put it as first species and 
then put starting_magnetization = 0.1)  even if the Ti d states should 
be empty so no final magnetization should be there. Again spin orbit 
should only affect very weakly Ti d states which are empty why you do 
you need to use spin orbit to compute phono dispersion ?
- try to use occup

-   last but just because I spotted it only now the pseudopotential for 
Ti isTl.rel-pbe-dn-rrkjus_psl.1.0.0.UPF check that you are not using a 
pseudo fot Tallium instead o pseudo for Titanium

hope it helps
regards
Pietro

On 7/29/19 9:07 AM, Che Min Wu wrote:
> Dear users,
> I am trying to calculate phonon dispersion on SrTiO3 supercell and 
> SrTiO3 slab with Quantum Espresso and thermo_pw package. However, it 
> took hours to converge in each iteration of Self-consistent 
> calculation, some times it even aborted because it used up 100% of CPU.
>
>  This is my fist time calculating structure larger than an unit cell, 
> I am running on a server with 16 intel I5 cpu core and 320GB of RAM. I 
> ran it with command
> mpirun -np 4 $BIN_DIR/thermo_pw.x < srtio3_slab.in 
> <http://srtio3_slab.in> > srtio3_slab.out
> I wonder if it is me doing things wrong, or is it simply impossible to 
> do these calculation with my computing power.
> I have attached my input file , thanks for any reply or suggestion.
>
> Thanks and Regards
> Che Min Wu
>
> _______________________________________________
> Quantum ESPRESSO is supported by MaX (www.max-centre.eu/quantum-espresso)
> users mailing list users at lists.quantum-espresso.org
> https://lists.quantum-espresso.org/mailman/listinfo/users


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.quantum-espresso.org/pipermail/users/attachments/20190729/2a0035ef/attachment.html>