ersion (MPI), running on40 processors
running the same inputs with 7.1 does not cause this issue for me. I am
using intel compilers 21.
Best,
Chris
Christoph Wolf
IBS Center for Quantum Nanoscience, Seoul, Korea
___
The Quantum ESPRESSO community stan
Many thanks guys!
Chris
On Wed, Jun 21, 2023 at 9:14 PM Paolo Giannozzi
wrote:
> It is fixed in v.7.2
>
> Paolo
>
> On 6/18/23 04:53, Christoph Wolf wrote:
> > Hi guys,
> >
> > I have a slight problem with projwfc and the new Hubbard U formalism.
> >
Dear Iurii,
Thanks for getting back to me so quickly. I briefly "debugged" and two
manifolds don't seem to be a problem when only one atom species is present,
e.g.
nat=1
ntyp = 1
...
ATOMIC_SPECIES
C 12.011 C.pbe-n-kjpaw_psl.1.0.0.UPF
Dy 1.008 Dy.GGA-PBE-paw-v1.0.UPF
...
ATOMIC_POSITIONS (angst
Hi guys,
I have a slight problem with projwfc and the new Hubbard U formalism.
The scf and nscf finish without any issue using the following Hubbard U
input:
HUBBARD ortho-atomic
U C-2p 1D-10
U Dy-5d 1
U Dy-6s 1
(C and Dy are there only species, for C I use the PSL PP and for Dy the PAW
dataset
Hi Mahmoud,
I believe the "output U" does depend on the "input U" - but only to a small
degree. You can study it by performing a loop updating U_in and checking
the value of U_out:
#i U_in U_out
1 1D-10 7.5154
2 7.5154 7.3759
3 7.3759 7.2442
4 7.2442 7.3925
5 7.3925 7.3820
HTH!
Chris
--
IBS Cen
Dear Iurii,
Thanks a lot for the explanation, that makes a lot of sense. I did use the
newest version for AtomPAW (atompaw-4.2.0.1.tar.gz (5.7mb) 5/26/2022) - I
guess the changes you mentioned will be available from the next version.
I will check out the new version on gitlab, thanks for being so
Dear Iurii,
Thank you very much for your fix. I am still facing an issue here, probably
specific to my PPs - I wonder if you can help me out:
These PPs have two "partial waves" per wave function, e.g. two 5d and two
4f waves. It seems that this somehow leads to an issue with the occupations
as fo
Dear Matteo,
I just tested with 22.04 and intel one API and the gitlab version with a
simple ./configure --prefix=... worked without issues. On an older machine
I had some obscure error related to git, which was simply solved by
updating git.
check if your LD_LIBRARY_PATH is properly set; I had
Oct 2022 at 05:57, Paolo Giannozzi
wrote:
> On 06/10/2022 07:07, Christoph Wolf wrote:
>
> > virtual_v2.o: In function `MAIN__':
> > */usr/local/qe-7.1-intel21/q-e-qe-7.1/upflib/virtual_v2.f90:102:
> > undefined reference to `for_stop_core_quiet'*
>
> &
Dear All,
I can compile QE 7.1 without any issue with the older intel 2019 compilers
but with the newer intel one API version from this year I am having some
strange issue related to some routine:
mpiifort -o virtual_v2.x virtual_v2.o atom.o atomic_number.o dqvan2.o
dylmr2.o gth.o interp_atwfc.o
Dear Iurii,
just a short follow-up: the error disappears (but I can't say if the
resulting calculation will be any good) if I overwrite the 0.00 occupation
with a small number, say 0.01 in the pseudo file directly. I tested this
for this pseudo: Mg.pbe-spn-kjpaw_psl.1.0.0.UPF (3S^2 3P0 configurati
Dear Iurii,
Thank you very much for your quick response - that did the trick and
removed the error but now I am facing another problem:
My cards are as follows:
ATOMIC_SPECIES
Gd 157.25 Gd.GGA-PBE-paw-v1.0-labels.UPF
O 15.999 O.pbe-n-rrkjus_psl.1.0.0.UPF
Mg 25.305 Mg.pbe-n-kjpaw_psl.0
Dear All,
I am trying the new (QE 7.1) Hubbard U implementation and I am struggling
with a pseudopotential that seems to have a format mismatch because it
complains:
Error in routine offset_atom_wfc (1):
The pseudo for Gd does not contain labels for atomic orbitals! Please add
them by hand in th
Dear Iurii,
Thanks a lot, I just checked out the new version and the new syntax is a
massive improvement - thanks to all the developers!
Thank you for your help!
Best,
Chris
--
Christoph Wolf
IBS Center for Quantum Nanoscience
Seoul, South Korea
Dear Iurii,
Thank you for your swift response! I uploaded the files here. For V and O I
inserted main and secondary manifolds to the necessary files (tabdf.f90 in
PW/src, set_hubbard_l and n in Modules).
https://drive.google.com/drive/folders/1aewyPfVasZLdGGAmtWA5XP7tWyUChupv?usp=sharing
I set
Dear all,
I am running into some problem when using the same species twice in a
calculation (eg. O1, O2 for two inequivalent oxygen atoms) and then
applying "standard U" on both species, e.g.
ATOMIC_POSITIONS (angstrom)
V 0.0 0.0 0.0
O10.0 0.0 2.0
O20.0 0.0 -2.
Dear Fabrizio,
Thank you for your reply and help. I tried to compile QE 6.8 with libxc
5.1.7 but this leads to a core dump during execution with
'rvv10-scan'/'MGGA_X_SCAN MGGA_C_SCAN'
Program PWSCF v.6.8 starts on 21Dec2021 at 13:49: 8
...
Max number of k-points (npk) = 4
Hi all,
I am having a strange problem when using the rvv10-scan on QE6.8 compiled
with libxc-6.0.0. The calculation works without problem on QE 6.6 compiled
with the same libxc but in QE 6.8 I get the following error when using
input_dft='rvv10-scan'
Error in routine tau_xc (5):
This ca
Dear all,
This is probably a silly question but I could have sworn that a previous
version of QE (projwfc.x to be precise) outputted a separate "lowdin.txt"
with the lowdin charges that are usually printed at the end of the
projwfc.x output - but I can't seem to find the switch for that.
Has it e
Dear Paolo,
Thanks a lot, that worked. Only needed to modify force_hub_gpu.f90, the
other gpu.f90 didn't throw any error!
Thanks again!
Chris
--
Group Leader "Theory of Quantum Systems at Surfaces"
IBS Center for Quantum Nanoscience
Seoul, South Korea
___
Dear all,
I tried to compile the newest version of QE using intel compilers but I am
running into a problem when executing make pw:
mpiifort -O2 -assume byterecl -g -traceback -nomodule -fpp -D__DFTI -D__MPI
-I/usr/local/qe-6.8/q-e//external/devxlib/src
-I/usr/local/qe-6.8/q-e//include -I/usr/lo
Dear all,
I am converting a wave-function from k to R space using wfck2r.x. There is
only 1 k-point (particle in a box of 15*15*15 Ang) and the wfc looks like:
e( 8) =-4.84501 eV
psi = 0.998*[# 2]
|psi|^2 = 0.999
state # 2: atom 1 (Fe ), wfc 2 (l=0 m= 1)
So this
Dear all,
I hope you are well and the summer is treating you well, despite external
circumstances... I have a probably silly question but:
When using fully-relativistic PPs and a calculation that includes
spin-orbit and non-collinear magnetism, is there a way to print the
magnitude of the spin-or
Dear all,
Thanks to the help of Giuseppe and Paolo (here:
https://www.mail-archive.com/users@lists.quantum-espresso.org/msg40236.html)
I managed to converge a cp calculation, but when attempting to restart cp.x
from pw.x the cp code complaints that it cannot find the wave function .dat
file;
As f
Dear all,
I am trying to run an SCF calculation in cp.x (because I want to use the
spin contamination analysis of cp.x), however reaching convergence with
cp.x seems painfully slow - probably my fault as I am not familiar with
cp.x.
System is a spin polarized calculation for a molecule in a vacuu
Thanks for the update, much appreciated!
Chris
--
IBS Center for Quantum Nanoscience
Seoul, South Korea
___
Quantum ESPRESSO is supported by MaX (www.max-centre.eu)
users mailing list users@lists.quantum-espresso.org
https://lists.quantum-espresso.org/
Dear all,
As far as I can see GIPAW 6.7, which make attempts to download from
https://github.com/dceresoli/qe-gipaw/archive/6.7MaX.tar.gz does not exist
(yet) - is it OK to compile QE 6.7 with GIPAW for QE 6.6 considering some
of the changes with IO?
Best,
Chris
--
IBS Center for Quantum Nanosc
Dear all,
I tested k-point parallelization and I wonder if the following results can
be normal or if my cluster has some serious problems...
the system has 74 atoms and a 2x2x1 k-point grid resulting in 4 k-points
number of k points= 4 Fermi-Dirac smearing, width (Ry)= 0.0050
Dear all,
I am facing a weird issue when calculating the phonon DOS using matdyn.x.
>From the phonon calculation and the dynmat mode analysis I know that I have
a stretching mode at 1529 /cm (mode 102)
freq (100 -100) =619.3 [cm-1] --> A_2 D_2 S_2 R
freq (101 -101) =
Dear all,
I am having some troubles compiling QE with libxc. I compiled libxc and it
ran a make check without issues.
Then I run a ./configure without flags in QE and change the make.inc as
follows:
DFLAGS = -D__DFTI -D__MPI -D__SCALAPACK -Duse_beef* -D__LIBXC*
IFLAGS = -I$(TOPD
Dear Pietro,
thank you very much for your reply, I guess I just haven't tried enough PPs
to encounter this behavior before! Good to know it is by design and not by
"my mistake"!
Best,
Chris
--
Group Leader "Theory of Quantum Systems at Surfaces"
IBS Center for Quantum Nanoscience
Seoul, South K
Dear all,
I saw some strange behavior when calculating the WF using GBRV (1.4, 1.5)
pseudos; basically what happens is that the potential inside the slab is
always above Ef instead of below (i.e. the potential seems "mirrored"
compared to any other PP I used). Everything else (EOS, ionic distances
There is a "known problem" (known at least to Don Hamann, the creator of
the ONCV), I will paste part of a conversation I had with him a few years
ago pertaining to the same problem:
" The Fe SG15 input data must have been on the "hairy edge" of ghost behavior.
A small change of the local potentia
Dear all,
We are trying to map the g-vectors to the coefficients of a wave function
(in order to extrapolate to its asymptotic value in vacuum for some STM
image simulation). After looking into some of the related codes in PP I am
still struggling and I am hoping someone with better knowledge can
tamination": CPV/src/spinsq.f90. Hope this helps (but I am not sure it
> will!)
>
> Paolo
>
> On Sun, Jul 26, 2020 at 8:36 AM Christoph Wolf
> wrote:
>
>> Dear all,
>>
>> I am navigating in circles here, I hope someone can "kick" me in t
Dear all,
I am navigating in circles here, I hope someone can "kick" me in the right
direction ;)
When estimating the "spin-contamination" literature (e.g. JP Malrieu, Chem
Rev 2014, 114, 429 ff) mentions that the degree of contamination is
estimated by the deviation of and from their ideal
(Is
Dear all,
Does anyone know where to find the orbital order for the occupation matrix
in PWSCF's Hubbard U for l=3? for l=2 I can see the orders on the manual
for projwfc.x but I was unable to find it for the f-electrons!
Best,
Chris
--
Postdoctoral Researcher
Center for Quantum Nanoscience, Ins
Dear all,
I was wondering if there is a particular reason that a hybrid calculation
converges at the same threshold as the SCF as
tr2_final = tr2
Since my dexx converges painfully slowly after the first few iterations I
was wondering if it could be "safe" to terminate the calculation earlier?
Dear Paolo,
thank you for the hint about bsplines, it works like a charm!
Best,
Chris
--
Postdoctoral Researcher
Center for Quantum Nanoscience, Institute for Basic Science
Ewha Womans University, Seoul, South Korea
___
Quantum ESPRESSO is supported b
Dear all,
I am trying to get smaller xsf files from post-proc but I can't figure out
how to get pp to accept my nx,ny,nz.
My FFT grid is 451 451 226 and the grid in the xsf file is always identical
with that grid, irrespective of my choice of nx, ny, nz. I use the
following input file:
&inputpp
Dear all,
I think pp.x does not follow the rule printed in the manual in the latest
version 6.5. The manual says that if emax is not specified the LDOS will
only be integrated for emin, however in my test it still calculates it for
other energies as well:
&inputpp
prefix= "agmgofe",
outdi
Dear all,
a very convenient way is PWSCF -> xcrysden -> export xsf -> ovito (
https://www.ovito.org/ ) it is free and gives you a lot of control over
visualization!
HTH!
Chris
--
Postdoctoral Researcher
Center for Quantum Nanoscience, Institute for Basic Science
Ewha Womans University, Seoul, S
Dear Sydney,
the following tools are very convenient to build surface slabs with
adsorbates as you described:
http://hjkgrp.mit.edu/content/molsimplify-tutorial-2-slab-builder
https://wiki.fysik.dtu.dk/ase/
If you want to do it more "manually" you can of course also use Avogadro,
VESTA, or simp
Dear all,
I am seeing something funny after running a calculation with pw.x conv_thr
6.5. I have set
conv_thr=1.d-9,
but in my calculation I see that estimated scf accuracy has been less than
1E-10 for over 30 SCF cycles - is this a glitch of did something change in
the new version of PWSCF?
O
to plot the local density of states. OR integrate the LDOS
> over a certain region at the specified bias - "simulating" an
> experimental broadening... Option 10.
>
> Hope that helps! Kind regards
>
> Thomas
>
>
> P.S.: Numerical derivative of the STM picture
Dear all,
I was wondering if there is a tool that is able to calculate the dI/dV for
output from PWSCF? I guess the way it is currently implemented would be to
calculate a set of STM images for different biases and then take the
numerical derivative but for larger systems this is actually really t
Dear all,
I am looking for suggestions how to converge a noncolin+spin-orbit
calculation in a direction other than the preferred one (in my system the
preferred direction is out-of-plane and I need to converge the calculation
in plane to get the energy differences).
The error is "too many bands n
Dear all,
I am having some issues with the energy convergence of a noncollinear
calculation to determine energy difference for different spin alignments.
The expected energies are ~1...5 meV so energy convergence should be rather
tight.
However I find that with high degauss values I cannot reach
Dear colleague,
unfortunately your email attachment was barely readable, not sure what
happened.
1) you have to converge your calculations wrt. cutoff and ecutrho, k-point
sampling, smearing (!), and also the convergence threshold as the MAE might
be tiny
2) you have another option called "force
Dear Sitangshu,
I am not sure if the "force theorem" works well for insulators, but here is
a nice discussion on different methods and their accuracy for determining
the MAE: https://arxiv.org/abs/1811.12100
HTH!
Chris
--
Postdoctoral Researcher
Center for Quantum Nanoscience, Institute for Bas
Dear all,
I am trying to calculate the magnetic anisotropy energy of a molecule
(VOPc). But I am facing a weird behavior that makes the total magnetization
huge and strongly oscillating; I already tried to use "lambda" with "atomic
direction" to align the magnetic moment but nothing seems to help.
Hi Brendan,
did you check the LDOS for spin up and down; try seeing if you introduce
some spurious magnetization in the case that looks wrong.
I remember a discussion a while back leading to the conclusion that LDA/GGA
are probably not well suited to describe a localized spin. You might have
bett
Getting the correct band gap from DFT is usually a coincidence due to the
approximations made in exchange and correlation potential.
You can try GW calculations on top of the DFT ground state, in QE you can
use YAMBO. In addition to the above you also have to consider if your
experimental value is
Dear Gaoxue,
I cannot see the attached output but I was wondering what range of alpha
you were using? In my experience small alpha ranges give very reliable
linear responses and they do not change the resulting U by more than the
uncertainty from inversion.
HTH!
Chris
--
Postdoctoral Researcher
Dear Jay,
for graphene you will need a much finer k-grid, for phonons people often
use 72x72 or 96x96 grids to get close to the Dirac point.
DFPT on graphene is also a bit tricky due to its semi-metallic character;
try with smearing instead of treating it as an insulator, of course metals
have in
Oh wow I haven't seen that one before! Thanks!
Chris
On Tue, Aug 27, 2019, 7:34 PM Nicola Marzari wrote:
>
>
> Very cool work ~10 years ago by Wei Cai and Giulia Galli (PRL) - but I
> think it's not available anywhere.
>
> nicola
>
>
As far as I know magnetic fields are incompatible with periodic boundary
conditions and plane-wave DFT. I did find however several implementations
based on LAPW where the magnetic field is applied in the muffin tin only -
check out the ELK code for example, it is free.
HTH!
Chris
--
Postdoctoral
Dear Aijun,
I am not a developer but most of the support for EPW is done via the
messaging board (forum) at http://epwforum.uk/index.php
there is no integrated plot utility for the spectral functions but you can
plot it via matlab, python or even gnuplot once you understand the format
HTH!
Chris
Dear all,
I am curious if some of you have experience with the following: in normal
LDA/GGA calculations my graphene phonon dispersions always end up with a
small problem at the "M" point of the BZ; the highest optical mode is split
from the lower mode by maybe 50/cm (almost all published papers s
Dear all,
I was trying to find the implementation of spin-spin interaction in GIPAW
(for total spin S>1/2). This is mentioned in Phys. Rev. B 97, 115135 (2018)
but I guess it has not been publicly released as the GIPAW distribution
does not mention it?
Anyone got any leads on this?
Thanks in adv
Dear Paolo,
grazie mille!
Chris
--
Postdoctoral Researcher
Center for Quantum Nanoscience, Institute for Basic Science
Ewha Womans University, Seoul, South Korea
___
Quantum ESPRESSO is supported by MaX (www.max-centre.eu/quantum-espresso)
users maili
Dear all,
This might be a silly question but since the codebase of PWSCF is all but
small I wanted to make sure that the wave function binary IO is handled by
io_base.f90 lines 101 ff in the current version.
OPEN ( UNIT = iuni, FILE = TRIM(filename)//'.dat', &
FORM='unforma
Dear all,
I have been using XCrySDen for years now and I find it very convenient for
most cases but sometimes I would like to just render a bunch of xsf files
(molecular orbitals) by orienting the cell in a way, setting an Isovalue
and then writing out raster images. Ideally this can be done remot
Dear colleagues,
thank you for your replies, they were very helpful!
Best,
Chris
--
Postdoctoral Researcher
Center for Quantum Nanoscience, Institute for Basic Science
Ewha Womans University, Seoul, South Korea
___
Quantum ESPRESSO is supported by MaX
Dear all,
after reading the works of Prof. Cococcioni regarding the "on-site" U and
"inter-site" V correction to DFT I was wondering if this was actually
implemented in PWSCF and is undocumented or if this was never implemented
in the public version?
Any pointers in the right direction are very w
Dear all,
I am having a rather obscure problem with dipole correction in QE 6.4.1
(compiled with intel compilers).
I am simulating a slab with atoms between z=0 and z=0.8 (in crystal units)
and a field appliad at z=0.90 with a dipole size of 0.05 (the slab lies in
the physical region).
The estim
I can recommend, as a first step, to use the older 0.3.1 for Fe; it is also
the recommended pseudo here
https://www.materialscloud.org/discover/sssp/plot/precision/Fe
I am not sure why, but I also had some problems with the 1.0.0 PP;
HTH!
Chris
--
Postdoctoral Researcher
Center for Quantum Nano
Dear all,
I noticed something interesting: when testing the convergence of total
energy wrt. ecutwfc (keeping ecutrho at 4*ecutwfc and using NC PPs) the
convergence without vdW is fairly straightforward and monotonous, however
with grimme-d3 VdW correction applied the behavior of total energy beco
Dear all,
when putting a molecule in a "vacuum box", why does the absolute position
of the molecule influence the outcome of some calculations? Should the
molecule be actually centered around (0,0,0) or in the middle of the box
(0.5,0.5,0.5)? I already converged the box size and I don't think this
Dear all,
I am giving the linear response method for Hubbard U and Hund J another go.
A recent publication has obtained quite resonable values for U and J based
on a hexahydrated transition metal model and I thought it might be
worthwhile trying (
https://journals.aps.org/prb/abstract/10.1103/Phys
s can happen, due to the fact that mv in these cases
> can have two solutions for the fermi energy, and mp can have three).
>
> But your case seem to show a problem with the tetrahedron method, and
> not sure that what is the problem.
>
> nicola
>
&
Dear all,
I am facing something rather interesting. My system is an insulator
supercell with a metallic impurity. I tried to converge the Fermi energy
wrt. to k-points and smearing in particular for "low temperatures" (~20 K).
I observed the following behavior for a chosen number of k-points (e.g.
runs fairly fast in about 3 min
General routines
calbec : 14.21s CPU 14.28s WALL ( 144 calls)
fft : 23.74s CPU 23.91s WALL ( 206 calls)
ffts : 1.04s CPU 1.04s WALL ( 36 calls)
fftw : 0.88s CPU
Dear all,
I apologize for the previous email; apparently I loaded, then unloaded and
forgot, the parallel environment; the build went smoothly after loading it
and the performance is quite good (ausurf benchmark (I removed 28 atoms
because of memory limitations)
--
Postdoctoral Researcher
Cente
Dear all,
I am trying to build the most recent stable version of QE-GPU with CUDA/PGI
community edition. My environment is set up as follows
echo $LD_LIBRARY_PATH
/usr/local/cuda-9.2/lib64:/usr/local/pgi/18.10/linux86-64/2018/cuda/9.2/lib64:/usr/local/pgi/18.10/linux86-64/18.10/lib
echo $CUDA
/
Dear all,
please forgive this "beginner" question but I am facing a weird problem.
When compiling qe-6.4 (intel compiler, intel MPI+OpenMP) with or without
intel's fftw libs I find that in openMP with 2 threads per core the intel
fftw version is roughly "twice as slow" as the internal one
"intern
Dear all,
I saw "basin hopping" referenced in several presentations about pwscf but I
cannot find the corresponding implementation - can anyone point me in
the right direction?
Thanks in advance!
Chris
--
Postdoctoral Researcher
Center for Quantum Nanoscience, Institute for Basic Science
Ewha W
Hallo Lukas,
you can also give the "open_grid.x" example a shot; just do a scf
calculation with empty bands included at an arbitrarily fine q-grid and go
though wannier using open_grid.x; I wrote a more detailed tutorial about it
somewhere, it seems to be working really well if you know how to wan
Dear all,
Thank you for elaborating; I was not aware of that limitation. That being
said, I managed to calculate MgO with HSE EXX and got quite good results
with ONCV. They just required a relatively dense k-grid to converge (but I
guess that is a totally different issue).
I will keep that in mi
short comment: Don Hamann recommended me to use the newer input files from
abinit for QE pseudos as well as the SG15 inputs are not recommended for
the new oncvpsp versions
https://github.com/pipidog/ONCVPSP
HTH!
Chris
--
Postdoctoral Researcher
Center for Quantum Nanoscience, Institute for Bas
I think there is a typo
lelfield = .TURE. ,
for your convergence issue: run a calc without electric field first, then
restart by reading the charge density from file(startingpot='file') with
field turn on at a small value (you can play that game several times and
increase the value gradually)!
H
Dear all,
I know how to calculate Hubbard U from the linear response mechanism but I
found that this is apparently not feasible for J;
"However, this calculation has not been actually attempted yet and it is
impossible to comment on its reliability" (
https://www.cond-mat.de/events/correl12/manus
Dear all,
in case this happens to someone else, here for reference how I solved the
problem:
a 1x1 in plane slab does not work, however a 2x2xN slab (with a bit of
vacuum) seems to be working in all cases, even at lower k-point density. I
am not quite clear yet why that is but I found it works fo
Dear Prof. Mazari and colleagues,
thank you again for your input. I have now run the calcs for a series of
smearings (degauss=0.2, 0.1, 0.01, 0.001) and alpha_mix(1...4) from 0.7
down to 0.0001) but there does not seem to be a trend emerging. If you take
a look at the attached plot it seems that f
Dear all,
thank you for your help so far. Unfortunately I am still not able to reach
self consistency during the ph.x calculation. I have tested all smearing
methods (and tetrahedra_opt), several smearing widths between 0.001 and
0.02, in-plane k-point grids up to 21x21x1 (121 k points) and relaxe
Dear all,
after reading your comments I have tried the following:
- three different Pseudos ("just to make sure..") that all work in the bulk
- increasing the layer numbers 6,7,8,9 ML
- test if the vacuum distance has any difference (but apparently it does
not in the range of 10-20A).
but for Cu
Dear Stefano,
Thank you for the prompt response; I will give the recommended method a
try. The reason why I built relatively thick slabs was that some of the
metals have slowly decaying surface states and I wanted to make sure they
are sufficiently decayed in the bulk region.
Let me give the thin
Dear all,
I am trying to calculate the phonon spectrum for a series of thin metal
layer. However there seem to be a few "pathological" cases such as Cu(111)
where, no matter what, I cannot achieve convergence during the ph.x
calculation.
I have tried, to no avail, change mixing parameter, k point
Dear all,
Is there a trick to converge a calculation with spin-orbit and non-colinear
magnetism?
No matter what the ecutwfc/rho and k-points the "estimated scf accuracy"
does not go below 1e-5 but hovers at the same magnitude whilst slightly
changing the "absolute magnetization" (totl magnetizati
Dear Jibiao ,
the input looks OK (3 types of atoms and 8 atoms in total) but your k-grid
is very coarse. For graphene I usually use a grid of 33 33 1 0 0 0 or
finer. To see if the individual atoms are accounted for you can project the
bands onto the DOS (PDOS) using projwfc.x which will give you
Hi Gui,
a few things that help convergence: mixing_beta, mixing_mode. Is there a
reason you want a certain number of bands? Otherwise smearing will add a
suitable number of empty bands anyway.
Check out thermo_pw; a really great tool to generate and calculate slab
systems:
https://github.com/dal
Dear all,
this is a very general question but after hours of googling I have come up
empty so I was wondering if anyone here is aware of a few worked examples
for the DMFT code AMULET (connected to QE via wannier_ham.x or wannier90)?
I am particularly interested in magnetic impurities embedded in
Dear all,
I was curious why I am unable to calculate a non-colinear scf run on top of
a spin-polarized charge density. e.g on top of
&control
calculation='scf',
prefix='pb',
pseudo_dir = '../',
outdir='./',
wf_collect = .true.
/
&system
ibrav = 2,
celldm(1) = 9.2697
I see you are using the MAX version. If I remember correctly I also had a
minor problem with that one at some point. Did you compile it yourself?
Which compiler and libraries did you use? In my limited experience intels
compiler and mkl works with the least hassle.
What is the required RAM (scf ou
Sometimes a single process can crash and cause others to hang. Are you sure
you have plenty of RAM available (ph.x requires more RAM than pw.x in my
experience).
Try running a smaller job on a single processor first. Also try avoiding
hyperthreading (depending on if you have 30 physical cores or
Dear all,
Some time ago I read about the implementation of a constrained DFT approach
here: http://theossrv1.epfl.ch/Main/OxidationStates
I was naturally curious on how to simply test this and decided to use H-H
dissociation as outlined in the text. However I am quite skeptical if my
approach wor
Dear Jingyang,
If I remember correctly there is one example (B-doped diamond) in the EPW
folder that uses VCA- might help to look at this. There is also a routine
(virtual.x) but I have never used it.
HTH!
Chris
--
Postdoctoral Researcher
Center for Quantum Nanoscience, Institute for Basic Scien
At first glance your system is 2-dimensional but you sample out-of plane
for ph.in (q3=3).
try an appropriate q-mesh such as 3x3x1, 5x5x1, ... You can see that the
convergence problem occurs at a q-point that is out of plane:
q = ( 0.000 0.000 0.1059000 )
HTH,
Ch
Just out of curiosity I ran it over night using the recommended pseudos
from materialscloud.com (accuracy) and grimme-D3 vdw correction
(vdw_corr='grimme-d3', ) and get this
VC-Relax
In:
##
bravais-lattice in
Hi Eric,
your structure is quite complex so the less than perfect agreement can have
several causes.
first I would check if the calculation is converged with respect to cutoff
and k-points. In particular forces require higher cutoffs than energy
convergence (~20%) in my experience.
Are your pseud
1 - 100 of 151 matches
Mail list logo