[Pw_forum] PW.x homogeneous electric field berry phase calculation in trigonal cell

[Louis Fry-Bouriaux]

Hi Lorenzo,


   Thank you for speedy reply!


    Interesting, I will try that, I will add that I tried the calculation with lelfield=.true. and all efield_cart values to zero and it gives me the same error. I also reduced the number of auto k points to speed up testing (auto: 6 2 2 0 0 0/ nppstr=6, which takes ~160s). I may take a look at the code maybe there is something that can be done if I identify what is going on :/


Thanks!

Louis

________________________________
From: pw_forum-bounces at pwscf.org <pw_forum-bounces at pwscf.org> on behalf of Lorenzo Paulatto <lorenzo.paulatto at impmc.upmc.fr>
Sent: 07 February 2017 18:08:35
To: PWSCF Forum
Subject: Re: [Pw_forum] PW.x homogeneous electric field berry phase calculation in trigonal cell


To be honest I do not remember, what I rmember is that if the field is not

along z (or the third axis) then it will have (had?) a huge bottleneck in

parallel which will make it unusable with more than a couple of cpus.

hth



On Tuesday, February 7, 2017 5:09:03 PM CET Louis Fry-Bouriaux wrote:

> Hi there,

>

> I am trying to perform a berry phase calculation with a homogeneous electric field (lelfield=.true.) however I receive the "wrong k-strings?" in c_phase error while running in parallel with OpenMPI on 4 processors:

> >From the documentation, if K_POINTS is automatic, then field strength

> >should be specified in cartesian coordinates (efield_cart(i)) and 'efield'

> >is not used, however reading the berry phase documentation says that the

> >direction of the efield is taken along gdir when a homogeneous electric

> >field is applied.

> Must I still specify efield_cart values when gdir is specified (it would

> seem so)? The code doesn't complain until after bandstructure computation.

> I have tried efield calculation with the following specification (see below

> for input file used):

>

>

> efield_cart(1) = 0.0001,

>

> efield_cart(2) = 0,

>

> efield_cart(3) = 0,

>

>

> which results in the electric field existing in two directions in crystal

> reference system, but the error is the same whether I do or do not include

> these fields and leave efield=0.0001

>

>

> My process is as follows:

>

> - scf calculation of trigonal alpha-Al2O3 cell (input crystal is correct)

>

> - band structure calculation

>

> - Berry phase calculation *without* electric field (this works and output

> is correct) with input:

>

>

> &CONTROL

> title = 'alpha-Alumina' ,

> calculation = 'nscf' ,

> outdir = '${PWD}' ,

> pseudo_dir = '${PWD}/pseudo/' ,

> verbosity = 'high' ,

> prefix = 'run_${RUN_NUMBER}',

> lberry = .true.,

> gdir = 1,

> nppstr = 10,

> lelfield = .false.,

> nberrycyc = 1

> /

> &SYSTEM

> ibrav = 5,

> A = 5.136,

> cosAB = 0.56956647115,

> nat = 10,

> ntyp = 2,

> occupations = 'fixed',

> ecutwfc = 50,

> ecutrho = 350,

> nbnd = 48

> /

> &ELECTRONS

> conv_thr = 1d-8,

> diagonalization = 'cg',

> mixing_beta = 0.7,

> electron_maxstep = 200,

> efield = 0,

> efield_phase = 'write'

> /

> ATOMIC_SPECIES

> O 15.99900 O.pbe-van_ak.UPF

> Al 26.98200 Al.pbe-n-van.UPF

> ATOMIC_POSITIONS crystal

> O 0.556000000 0.944000000 0.250000000

> O 0.444000000 0.056000000 0.750000000

> O 0.250000000 0.556000000 0.944000000

> O 0.750000000 0.444000000 0.056000000

> O 0.944000000 0.250000000 0.556000000

> O 0.056000000 0.750000000 0.444000000

> Al 0.352000000 0.352000000 0.352000000

> Al 0.648000000 0.648000000 0.648000000

> Al 0.148000000 0.148000000 0.148000000

> Al 0.852000000 0.852000000 0.852000000

> K_POINTS automatic

> 10 4 4 0 0 0

>

> - Then Berry phase calculation *with* electric field:

>

>

> &CONTROL

> title = 'alpha-Alumina' ,

> calculation = 'nscf' ,

> outdir = '${PWD}' ,

> pseudo_dir = '${PWD}/pseudo/' ,

> verbosity = 'high' ,

> prefix = 'run_${RUN_NUMBER}',

> lberry = .true.,

> gdir = 1,

> nppstr = 10,

> lelfield = .true.,

> nberrycyc = 1

> /

> &SYSTEM

> ibrav = 5,

> A = 5.136,

> cosAB = 0.56956647115,

> nat = 10,

> ntyp = 2,

> occupations = 'fixed',

> ecutwfc = 50,

> ecutrho = 350,

> nbnd = 48,

> nosym = .false.

> /

> &ELECTRONS

> conv_thr = 1d-8,

> diagonalization = 'cg',

> mixing_beta = 0.7,

> electron_maxstep = 200,

> efield = 0.0001,

> efield_phase = 'read'

> /

> ATOMIC_SPECIES

> O 15.99900 O.pbe-van_ak.UPF

> Al 26.98200 Al.pbe-n-van.UPF

> ATOMIC_POSITIONS crystal

> O 0.556000000 0.944000000 0.250000000

> O 0.444000000 0.056000000 0.750000000

> O 0.250000000 0.556000000 0.944000000

> O 0.750000000 0.444000000 0.056000000

> O 0.944000000 0.250000000 0.556000000

> O 0.056000000 0.750000000 0.444000000

> Al 0.352000000 0.352000000 0.352000000

> Al 0.648000000 0.648000000 0.648000000

> Al 0.148000000 0.148000000 0.148000000

> Al 0.852000000 0.852000000 0.852000000

> K_POINTS automatic

> 10 4 4 0 0 0

>

> My last thought is that when the electric field is specified in cartesian

> coordinates as above, the fact that the field in the crystal reference

> system is along two directions causes the problem, I have struggled to get

> the field to exist along one direction in the crystal reference system

> (does anyone know a reference I could use to do this properly?). Maybe the

> sensible thing to do is not use automatic K_POINTS?

>

> Kindest regards,

> Louis





--

Dr. Lorenzo Paulatto

IdR @ IMPMC -- CNRS & Université Paris 6

+33 (0)1 44 275 084 / skype: paulatz

http://www.impmc.upmc.fr/~paulatto/

23-24/4é16 Boîte courrier 115,

4 place Jussieu 75252 Paris Cédex 05
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.quantum-espresso.org/pipermail/users/attachments/20170207/2ccd4f9c/attachment.html>