Hi,everyone.
I want to perform Scf calculation using quantum espresso. In the Input
describtion of PW , naming an atom as X_* is allowed and
max total length cannot exceed 3 characters (Card: ATOMIC_SPECIES). But
, as I test, if X has three letters (for example Ni1),
an error will occure as f
On 2021-09-24 14:22, Zahra Mosleh wrote:
Hi,everyone.
I want to perform Scf calculation using quantum espresso. In the Input
describtion of PW , naming an atom as X_* is allowed and
max total length cannot exceed 3 characters (Card: ATOMIC_SPECIES).
However , as I test, if X has four letter
It is not obvious to me if -nd parallelization works with dfpt in ph.x. Is ELPA
used? What is the best parallelization strategy for effective charges in large
systems? phonons? Thanks for any advice.
Sincerely,
Ron
---
Ronald Cohen
Extreme Materials Initiative
Carnegie Institution for Science
I think -nd is used only for the initial calculation of the needed
Kohn-Sham \psi_{k+q} states. There is little dense-matrix linear algebra in
the phonon code.
Paolo
On Fri, Sep 24, 2021 at 8:26 PM Ronald Cohen
wrote:
> It is not obvious to me if -nd parallelization works with dfpt in ph.x. Is
Dear all,
I have a question related to the wave functions in the phonon code. I think
the wave function at the finite q vector (xq variable) \Psi_n(k+q) is
contained in the evq variable.
This wave function appears different from the wave function at an
equivalent q point, for instance if I look at
Dear QE Users,
I have doped a 3X2X3 supercell of the compound CsPbI3 with one Cu atom. My
input file looks like this:
&CONTROL
calculation = 'vc-relax'
etot_conv_thr = 9.00d-04
forc_conv_thr = 1.00d-04
outdir = './outdir'
prefix = 'basic'
pseudo_dir = './'
nstep