[gpaw-users] Problems with calculating all electron wave functions in parallel mode
Kuisma, Mikael
mikael.j.kuisma at jyu.fi
Fri Sep 15 11:04:53 CEST 2017
Hi!
This is the closest thing I can find to match your needs:
def calculate_matrix_elements(self, psit1_nG, P1_ani, A, dA,
psit2_nG=None, P2_ani=None):
So you supply two sets of wave functions, two sets of projector overlaps P_ani = <p^a_i| psit_nG >.
https://gitlab.com/gpaw/gpaw/blob/master/gpaw/hs_operators.py#L251
with A and dA defined like this to get the overlap:
https://gitlab.com/gpaw/gpaw/blob/master/gpaw/overlap.py#L68
BR,
Mikael
________________________________
Lähettäjä: Varadharajan Srinivasan [varadharajan.srinivasan at gmail.com]
Lähetetty: 15. syyskuuta 2017 11:51
Vastaanottaja: Kuisma, Mikael
Kopio: gpaw-users
Aihe: Re: [gpaw-users] Problems with calculating all electron wave functions in parallel mode
Dear Mikael,
Thank you very much. I think this is very helpful. If I understand correctly, the overlap operators can be recovered from paw.wfs.setups() as given here https://wiki.fysik.dtu.dk/gpaw/devel/developersguide.html.
I will try this out.
Just one more question. The projectors are atom-centred. So would it be correct to take overlaps between wave functions belonging to different atomic positions (for instance, along a trajectory) the same way as
above?
Best,
Vardha.
On Fri, Sep 15, 2017 at 2:09 PM, Kuisma, Mikael <mikael.j.kuisma at jyu.fi<mailto:mikael.j.kuisma at jyu.fi>> wrote:
Hi!
You want to project real time propagation grid wave functions to ground state wave functions? What you need to do is to calculate the inner product
B_nn' = <psit_nG(0) | O | psit_n'G(t)>,
where O is the overlap operator
O = I + \sum_aii' |p^a_i> O_ii' <p^a_i'|
For sanity check, you can make sure that the ground state orbitals are orthogonal B_nn' = delta_nn' for t=0. Perhaps somebody in this list can show you how to do it.
You don't want to calculate all electron wave functions, because it is time consuming and not accurate.
BR,
Mikael Kuisma
Academy of Finland Post Doc
Nanoscience Center
Department of Chemistry
University of Jyväskylä
P.O. Box 35
FIN-40014 University of Jyväskylä
Finland
E-mail: mikael.j.kuisma (at) jyu.fi<http://jyu.fi>
Phone: +358 40 077 4114
https://scholar.google.se/citations?user=w0eP9esAAAAJ
Visiting address: NSC, 3rd floor, YN320
________________________________
Lähettäjä: Varadharajan Srinivasan [varadharajan.srinivasan at gmail.com<mailto:varadharajan.srinivasan at gmail.com>]
Lähetetty: 15. syyskuuta 2017 10:05
Vastaanottaja: Kuisma, Mikael
Kopio: gpaw-users
Aihe: Re: [gpaw-users] Problems with calculating all electron wave functions in parallel mode
Thanks, Mikael. I saw this paper just this morning and was wondering if I could find the implementation and adapt to FD as well.
Best,
Vardha.
On Fri, Sep 15, 2017 at 12:27 PM, Kuisma, Mikael <mikael.j.kuisma at jyu.fi<mailto:mikael.j.kuisma at jyu.fi>> wrote:
Hi!
>I am actually interested in computing projections of time-propagated Kohn-Sham states with the initial KS states. Does a tool for this already exist?
Yes! In case of LCAO-TDDFT, it does! Check out our recent development just accepted JCTC. In practice, we project the time dependent density matrix to initial KS states and analyse the content of excitations within linear response calculations in real time. However, the method itself is not restricted to linear response calculations.
http://pubs.acs.org/doi/abs/10.1021/acs.jctc.7b00589
It is not yet added to git, but soon will be.
BR,
Mikael Kuisma
Academy of Finland Post Doc
Nanoscience Center
Department of Chemistry
University of Jyväskylä
P.O. Box 35
FIN-40014 University of Jyväskylä
Finland
E-mail: mikael.j.kuisma (at) jyu.fi<http://jyu.fi>
Phone: +358 40 077 4114
https://scholar.google.se/citations?user=w0eP9esAAAAJ
Visiting address: NSC, 3rd floor, YN320
________________________________
Lähettäjä: gpaw-users-bounces at listserv.fysik.dtu.dk<mailto:gpaw-users-bounces at listserv.fysik.dtu.dk> [gpaw-users-bounces at listserv.fysik.dtu.dk<mailto:gpaw-users-bounces at listserv.fysik.dtu.dk>] käyttäjän Varadharajan Srinivasan via gpaw-users [gpaw-users at listserv.fysik.dtu.dk<mailto:gpaw-users at listserv.fysik.dtu.dk>] puolesta
Lähetetty: 15. syyskuuta 2017 9:33
Vastaanottaja: Jens Jørgen Mortensen
Kopio: gpaw-users
Aihe: Re: [gpaw-users] Problems with calculating all electron wave functions in parallel mode
Not yet. But hopefully eventually.
Vardha,
On Fri, Sep 15, 2017 at 11:01 AM, Jens Jørgen Mortensen <jjmo at dtu.dk<mailto:jjmo at dtu.dk>> wrote:
On 09/14/2017 09:08 PM, Varadharajan Srinivasan wrote:
Dear Jens,
I am actually interested in computing projections of time-propagated Kohn-Sham states with the initial KS states. Does a tool for this already exist? I wasn't aware so I wanted to code one. However, I was not sure if pseudo wave functions in PAW would be useful for this. So I am trying to get all-electron wave functions along my time-propagated simulation. Since that is parallel I also wanted the PS2AE to work in the same framework.
This might be a very round about way of getting the projections and would be unnecessary if such a tool is already in existence. If not, I guess my trials might eventually contribute to making one. I would really appreciate any help in this regard.
Do your atoms move?
Jens Jørgen
Vardha.
On Thu, Sep 14, 2017 at 1:46 PM, Jens Jørgen Mortensen <jjmo at dtu.dk<mailto:jjmo at dtu.dk> <mailto:jjmo at dtu.dk<mailto:jjmo at dtu.dk>>> wrote:
On 09/13/2017 08:16 PM, Varadharajan Srinivasan wrote:
Hi Jens,
I installed and tried out the development version and am
getting it to work in parallel mode as well.
Thanks for checking. Note that the current implementation will
not run any faster in parallel!
Jens Jørgen
Thanks,
Vardha.
On Tue, Sep 12, 2017 at 7:32 PM, Varadharajan Srinivasan
<varadharajan.srinivasan at gmail.com<mailto:varadharajan.srinivasan at gmail.com>
<mailto:varadharajan.srinivasan at gmail.com<mailto:varadharajan.srinivasan at gmail.com>>
<mailto:varadharajan.srinivasan at gmail.com<mailto:varadharajan.srinivasan at gmail.com>
<mailto:varadharajan.srinivasan at gmail.com<mailto:varadharajan.srinivasan at gmail.com>>>> wrote:
Ok I will try and report.
Vardha.
On Tue, Sep 12, 2017 at 3:02 PM, Jens Jørgen Mortensen
<jjmo at dtu.dk<mailto:jjmo at dtu.dk> <mailto:jjmo at dtu.dk<mailto:jjmo at dtu.dk>> <mailto:jjmo at dtu.dk<mailto:jjmo at dtu.dk>
<mailto:jjmo at dtu.dk<mailto:jjmo at dtu.dk>>>> wrote:
On 09/11/2017 07:44 PM, Varadharajan Srinivasan wrote:
Yes it does. I am sorry I had forgotten to add
that. It
only fails when run in parallel.
It should work Ok in parallel also. Can you try the
latest
development version of ASE+GPAW?
Jens Jørgen
Vardha.
On Mon, Sep 11, 2017 at 3:02 PM, Jens Jørgen Mortensen
<jjmo at dtu.dk<mailto:jjmo at dtu.dk> <mailto:jjmo at dtu.dk<mailto:jjmo at dtu.dk>>
<mailto:jjmo at dtu.dk<mailto:jjmo at dtu.dk> <mailto:jjmo at dtu.dk<mailto:jjmo at dtu.dk>>> <mailto:jjmo at dtu.dk<mailto:jjmo at dtu.dk>
<mailto:jjmo at dtu.dk<mailto:jjmo at dtu.dk>>
<mailto:jjmo at dtu.dk<mailto:jjmo at dtu.dk> <mailto:jjmo at dtu.dk<mailto:jjmo at dtu.dk>>>>> wrote:
On 09/10/2017 02:26 PM, Varadharajan
Srinivasan via
gpaw-users wrote:
Dear all,
I am trying to obtain all electron wave
functions
from a
ground-state calculation on guanine. Having
performed the GS
calculation in domain parallel mode, I wrote a
script to get
the all electron wave function using the ps2ae
utility. The
code used is attached below. However, I
get the
following
error. I have seen some similar posts in
the forum
but there
was no response. I would greatly
appreciate some
advice on
this issue.
Does the script work OK when you run it in serial?
Jens Jørgen
Vardha.
IISER Bhopal
India
Number of Atomic Orbitals: 49
Number of Bands in Calculation: 50
Bands to Converge: 40 Lowest Bands
Number of Valence Electrons: 56
rank=0 L00: Traceback (most recent call last):
rank=0 L01: File "trial.py", line 19, in
<module>
rank=0 L02: wf =
twf_gs.get_wave_function(band, ae=False)
rank=0 L03: File
"/opt/gpaw-intelmpi/lib/python2.7/site-packages/gpaw/utilities/ps2ae.py",
line 97, in get_wave_function
rank=0 L04: pad=True, periodic=True)
rank=0 L05: File
"/opt/gpaw-intelmpi/lib/python2.7/site-packages/gpaw/aseinterface.py",
line 458, in get_pseudo_wave_function
rank=0 L06: periodic=periodic)
rank=0 L07: File
"/opt/gpaw-intelmpi/lib/python2.7/site-packages/gpaw/aseinterface.py",
line 465, in get_pseudo_wave_function
rank=0 L08: periodic=periodic)
rank=0 L09: File
"/opt/gpaw-intelmpi/lib/python2.7/site-packages/gpaw/wavefunctions/base.py",
line 404, in get_wave_function_array
rank=0 L10: psit_G =
self.gd.collect(psit_G)
rank=0 L11: File
"/opt/gpaw-intelmpi/lib/python2.7/site-packages/gpaw/grid_descriptor.py",
line 502, in collect
rank=0 L12: A_xg[..., b0:e0, b1:e1,
b2:e2] = a_xg
rank=0 L13: ValueError: could not
broadcast input
array from
shape (115,107,83) into shape (23,53,83)
GPAW CLEANUP (node 0): <type
'exceptions.ValueError'>
occurred. Calling MPI_Abort!
application called
MPI_Abort(MPI_COMM_WORLD, 42) -
process 0
_______________________________________________
gpaw-users mailing list
gpaw-users at listserv.fysik.dtu.dk<mailto:gpaw-users at listserv.fysik.dtu.dk>
<mailto:gpaw-users at listserv.fysik.dtu.dk<mailto:gpaw-users at listserv.fysik.dtu.dk>>
<mailto:gpaw-users at listserv.fysik.dtu.dk<mailto:gpaw-users at listserv.fysik.dtu.dk>
<mailto:gpaw-users at listserv.fysik.dtu.dk<mailto:gpaw-users at listserv.fysik.dtu.dk>>>
<mailto:gpaw-users at listserv.fysik.dtu.dk<mailto:gpaw-users at listserv.fysik.dtu.dk>
<mailto:gpaw-users at listserv.fysik.dtu.dk<mailto:gpaw-users at listserv.fysik.dtu.dk>>
<mailto:gpaw-users at listserv.fysik.dtu.dk<mailto:gpaw-users at listserv.fysik.dtu.dk>
<mailto:gpaw-users at listserv.fysik.dtu.dk<mailto:gpaw-users at listserv.fysik.dtu.dk>>>>
https://listserv.fysik.dtu.dk/mailman/listinfo/gpaw-users
<https://listserv.fysik.dtu.dk/mailman/listinfo/gpaw-users>
<https://listserv.fysik.dtu.dk/mailman/listinfo/gpaw-users
<https://listserv.fysik.dtu.dk/mailman/listinfo/gpaw-users>>
<https://listserv.fysik.dtu.dk/mailman/listinfo/gpaw-users
<https://listserv.fysik.dtu.dk/mailman/listinfo/gpaw-users>
<https://listserv.fysik.dtu.dk/mailman/listinfo/gpaw-users
<https://listserv.fysik.dtu.dk/mailman/listinfo/gpaw-users>>>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listserv.fysik.dtu.dk/pipermail/gpaw-users/attachments/20170915/41c8a39a/attachment-0001.html>
More information about the gpaw-users
mailing list