[Wien] :WARN : WARNING: RKmax reduced due to NMATMAX

Tue Nov 7 16:54:51 CET 2017

Gavin and Professor Marks are right in saying that you will need more
computing resources (MPI a must) if you finally decide to use a small RMT
for O (in my experience only an RMT = 1.1 or smaller works well for O2
dimer's relaxation). That being said, one can use some workarounds for this
problem. The following procedures come handy for oxides, especially when
you are interested in computing cohesive energies and thermodynamic
stability but don't want to use small RMT of O for subsequent calculations
involving oxides and their supercells.

(1) Compute your own cohesive energy (will need MPI and small cluster)

(a) first relax an O2 dimer and compute energy of a single O atom using RMT
= 1.1 and RKMAX = 5.5.

(b) work out the cohesive energy E[cohesive_O_rmt1.1] = E[O2_rmt1.1] – 2 *
E [O_atom_rmt1.1]

(c) compute energy of an O atom with RMT and RKMAX of your choice (e.g. RMT
= 1.6 and RKMAX = 8.0); E [O_atom_rmt1.6]

(d) To get the energy of O2 dimer with a (imaginary) RMT of 1.6 add the
above computed cohesive energy into E [O_atom_rmt1.6]: i.e. E[O2_rmt1.6] =
2* E [O_atom_rmt1.6] + E[cohesive_O_rmt1.1]

(2) Use an already reported value of cohesive energy (works fine for small
computers)

(a) Since the cohesive energies of O2 for frequently used XC functionals
within WIEN2k are already available in literature (e.g.
https://doi.org/10.1103/PhysRevB.75.115131), the energy of O2 dimer with a
(imaginary) RMT of 1.6 can be easily had from

E[O2_rmt1.6] = 2* E [O_atom_rmt1.6] + E[cohesive_O_computed_using_wien]

The above procedures are logical but do not allow one to play much with the
RMT values of O2 dimer within the muffin-tin model here. I did however
perform some tests with a P2 dimer some time ago and found both the methods
to be correct.

Best,

Alay

On Tue, Nov 7, 2017 at 7:00 PM, Laurence Marks <L-marks at northwestern.edu>
wrote:

> You do not need want RKMAX 6.5 for O2, because you will need RMTs of about
> 1.2. 6.0 or 5.75 will be fine. BUT you have to use the same O RMT & RKMAX
> for the oxide.
>
> You need more computing resources.
>
>
>
>
> On Nov 7, 2017 13:46, "chin Sabsu" <chinsabsu at yahoo.in> wrote:
>
> Thank you Sir,
>
> runsp_lapw -it gave me an error:  foreach. So I am running with runsp_lapw
> ....  script.
>
> If I run it with Rkmax=6 then it does not show me any warning. At the FAQ
> of Wienk page standard RMT for O is 6.5 but it also gave me RKmax reduced
> warning.
>
> If someone advises me that I can take RKMAX=6 for O2 system then only I
> can proceed.
>
> I do not have any good facility.
>
>
>
>
>
> On Tuesday, 7 November 2017 6:59 PM, Gavin Abo <gsabo at crimson.ua.edu>
> wrote:
>
>
> To make that calculation feasible, it looks like you need a better
> computing system like a small cluster and mpi.
> If ~8 GB is your total RAM, keep in mind that the Linux operating might
> use around 1 GB.  Using top [ https://superuser.com/question
> s/282867/why-does-the-memory-usage-in-top-not-add-up
> <https://urldefense.proofpoint.com/v2/url?u=https-3A__superuser.com_questions_282867_why-2Ddoes-2Dthe-2Dmemory-2Dusage-2Din-2Dtop-2Dnot-2Dadd-2Dup&d=DwMFaQ&c=yHlS04HhBraes5BQ9ueu5zKhE7rtNXt_d012z2PA6ws&r=U_T4PL6jwANfAy4rnxTj8IUxm818jnvqKFdqWLwmqg0&m=7P5UhBS9GKV9ZVnjwc4fSYw4NNdswsNYfUvNn4rZiZI&s=2O7Uggfs21S8VM2B67fINP9etTQoh-Me6-8d-WyrJSE&e=>
> ] or a system monitor [ https://askubuntu.com/question
> s/29757/what-can-replace-system-monitoring-in-the-top-gnome-panel-in-unity
> <https://urldefense.proofpoint.com/v2/url?u=https-3A__askubuntu.com_questions_29757_what-2Dcan-2Dreplace-2Dsystem-2Dmonitoring-2Din-2Dthe-2Dtop-2Dgnome-2Dpanel-2Din-2Dunity&d=DwMFaQ&c=yHlS04HhBraes5BQ9ueu5zKhE7rtNXt_d012z2PA6ws&r=U_T4PL6jwANfAy4rnxTj8IUxm818jnvqKFdqWLwmqg0&m=7P5UhBS9GKV9ZVnjwc4fSYw4NNdswsNYfUvNn4rZiZI&s=tlxLaeinGiOXOpeuOOsygjJTvXx1krPYykP85S2l1rI&e=>
> ], you can check for any unnecessary processes and kill them to try to gain
> back more "free" memory.  Though, while this might help a little with
> memory, it likely won't help with the speed issue.
> For speed, you likely need mpi or iterative diagonalization.  Are using
> iterative diagonalization (runsp -it) as was mentioned before:
> http://zeus.theochem.tuwien.ac.at/pipermail/wien/2017-November/027303.html
> <https://urldefense.proofpoint.com/v2/url?u=http-3A__zeus.theochem.tuwien.ac.at_pipermail_wien_2017-2DNovember_027303.html&d=DwMFaQ&c=yHlS04HhBraes5BQ9ueu5zKhE7rtNXt_d012z2PA6ws&r=U_T4PL6jwANfAy4rnxTj8IUxm818jnvqKFdqWLwmqg0&m=7P5UhBS9GKV9ZVnjwc4fSYw4NNdswsNYfUvNn4rZiZI&s=bYFJI4_QBufzea0m3bDcoKFstJcTfD06rpV4Ou0f7-U&e=>
> If none of that helps, it looks like your stuck with doing less accurate
> calculations using small RKmax values.
> On 11/6/2017 9:24 AM, chin Sabsu wrote:
>
> Thank you Sir,
>
>
> As per memory test, I got the figure: 46743  it means ~4.5 GB RAM is
> enough to run the case. But my system is having ~8 GB RAM:
>
> 7.6 GiB
> Intel® Core™ i5-3570K CPU @ 3.40GHz × 4
> Intel® Ivybridge Desktop
>
>
> Dear Lyudmila Dobysheva
>
> I increase NMATMAX  upto45000 and NUME at 10000. At this value, it gave
> me lapw1 error while at  NMATMAX (NUME)=35000(8000) it taking too much
> time and even after three hrs lapw1 is running.
>
> I want to keep consistent RKMAX with the literature. Otherwise, my job is
> to calculate the thermodynamic stability and for the same, I need the
> cohessive energy of O also.
>
>
>
> Can I use RKmax 6.0 for O2 and for others I am tacking 7. If not then what
> should I do to run this O2 system?
>
>
> Chin
>
> _______________________________________________
> Wien mailing list
> Wien at zeus.theochem.tuwien.ac.at
> http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien
> <https://urldefense.proofpoint.com/v2/url?u=http-3A__zeus.theochem.tuwien.ac.at_mailman_listinfo_wien&d=DwMFaQ&c=yHlS04HhBraes5BQ9ueu5zKhE7rtNXt_d012z2PA6ws&r=U_T4PL6jwANfAy4rnxTj8IUxm818jnvqKFdqWLwmqg0&m=7P5UhBS9GKV9ZVnjwc4fSYw4NNdswsNYfUvNn4rZiZI&s=_abSDRV3mleDHcpzpini35XEwrqchTeGKib0LE-XjnY&e=>
> SEARCH the MAILING-LIST at:  http://www.mail-archive.com/wi
> en at zeus.theochem.tuwien.ac.at/index.html
> <https://urldefense.proofpoint.com/v2/url?u=http-3A__www.mail-2Darchive.com_wien-40zeus.theochem.tuwien.ac.at_index.html&d=DwMFaQ&c=yHlS04HhBraes5BQ9ueu5zKhE7rtNXt_d012z2PA6ws&r=U_T4PL6jwANfAy4rnxTj8IUxm818jnvqKFdqWLwmqg0&m=7P5UhBS9GKV9ZVnjwc4fSYw4NNdswsNYfUvNn4rZiZI&s=lHwBjp42cwfD5FjQsF0xeC1jjSncHbR9wyle1GJnOXw&e=>
>
>
>
>
> _______________________________________________
> Wien mailing list
> Wien at zeus.theochem.tuwien.ac.at
> http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien
> SEARCH the MAILING-LIST at:  http://www.mail-archive.com/
> wien at zeus.theochem.tuwien.ac.at/index.html
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://zeus.theochem.tuwien.ac.at/pipermail/wien/attachments/20171107/8238890b/attachment.html>