Dear all:
I'm trying to do a bulk relaxation for a 3x3x3 and 4x4x4 supercell for WO3. The ultimate goal is to be able to do this relaxation in HSE, but I was thinking of doing it first in GGA and have something more or less closer to the equilibrium structure before doing HSE. The 3x3x3 supercell in GGA appears to have relaxed sucessfully, but the 4x4x4 supercell in GGA has been hanging on the first electronic step of the first ionic step for many hours (~20) and is very positive. I've attempted a similar thing in HSE, and it hangs on the fourth electronic step of the first ionic step; it is similarly very large and positive, but is so far decreasing with each step. I'm not sure if there is something wrong with my KPOINTS file or something else I'm not aware of, but it is quite confusing.
Any light on what is happening would be helpful!
Thank you!
Copied below are the input files used for the 4x4x4 GGA calculations:
==============================================
INCAR
---------------------------------------------------------------------------
SYSTEM = WO3bulk-cubic-supercell !Header
!Precision parameters
ALGO = Normal
PREC = Normal
ENCUT = 500
!Convergence parameters
NSW = 50
EDIFFG = -0.001
EDIFF = -0.0001
NELMIN = 5
NBANDS = 1537 ! few empty bands included, needed?
ISMEAR = 0
SIGMA = 0.1
IBRION = 2
ISIF = 3
---------------------------------------------------------------------
KPOINTS
--------------------------------------------------------------------
automatic generation - MP scheme
0
Monkhorst-pack
4 4 4
0 0 0
------------------------------------------------------------------
POTCAR
------------------------------------------------------------------
PAW_PBE W
PAW_PBE O
---------------------------------------------------------------------
POSCAR
------------------------------------------------------------------
supercell cubic WO3- 4x4x4
3.83200000000000
4.0000000000000000 0.0000000000000000 0.0000000000000000
0.0000000000000000 4.0000000000000000 0.0000000000000000
0.0000000000000000 0.0000000000000000 4.0000000000000000
W O
64 192
Direct
0.125000000 0.125000000 0.125000000 !W first layer
0.375000000 0.125000000 0.125000000
0.625000000 0.125000000 0.125000000
0.875000000 0.125000000 0.125000000
0.125000000 0.375000000 0.125000000
0.375000000 0.375000000 0.125000000
0.625000000 0.375000000 0.125000000
0.875000000 0.375000000 0.125000000
0.125000000 0.625000000 0.125000000
0.375000000 0.625000000 0.125000000
0.625000000 0.625000000 0.125000000
0.875000000 0.625000000 0.125000000
0.125000000 0.875000000 0.125000000
0.375000000 0.875000000 0.125000000
0.625000000 0.875000000 0.125000000
0.875000000 0.875000000 0.125000000
0.125000000 0.125000000 0.375000000 !W second layer
0.375000000 0.125000000 0.375000000
0.625000000 0.125000000 0.375000000
0.875000000 0.125000000 0.375000000
0.125000000 0.375000000 0.375000000
0.375000000 0.375000000 0.375000000
0.625000000 0.375000000 0.375000000
0.875000000 0.375000000 0.375000000
0.125000000 0.625000000 0.375000000
0.375000000 0.625000000 0.375000000
0.625000000 0.625000000 0.375000000
0.875000000 0.625000000 0.375000000
0.125000000 0.875000000 0.375000000
0.375000000 0.875000000 0.375000000
0.625000000 0.875000000 0.375000000
0.875000000 0.875000000 0.375000000
0.125000000 0.125000000 0.625000000 ! W third layer
0.375000000 0.125000000 0.625000000
0.625000000 0.125000000 0.625000000
0.875000000 0.125000000 0.625000000
0.125000000 0.375000000 0.625000000
0.375000000 0.375000000 0.625000000
0.625000000 0.375000000 0.625000000
0.875000000 0.375000000 0.625000000
0.125000000 0.625000000 0.625000000
0.375000000 0.625000000 0.625000000
0.625000000 0.625000000 0.625000000
0.875000000 0.625000000 0.625000000
0.125000000 0.875000000 0.625000000
0.375000000 0.875000000 0.625000000
0.625000000 0.875000000 0.625000000
0.875000000 0.875000000 0.625000000
0.125000000 0.125000000 0.875000000 !W fourth layer
0.375000000 0.125000000 0.875000000
0.625000000 0.125000000 0.875000000
0.875000000 0.125000000 0.875000000
0.125000000 0.375000000 0.875000000
0.375000000 0.375000000 0.875000000
0.625000000 0.375000000 0.875000000
0.875000000 0.375000000 0.875000000
0.125000000 0.625000000 0.875000000
0.375000000 0.625000000 0.875000000
0.625000000 0.625000000 0.875000000
0.875000000 0.625000000 0.875000000
0.125000000 0.875000000 0.875000000
0.375000000 0.875000000 0.875000000
0.625000000 0.875000000 0.875000000
0.875000000 0.875000000 0.875000000 !end W
0.125000000 0.125000000 0.000000000 !O first layer
0.375000000 0.125000000 0.000000000
0.625000000 0.125000000 0.000000000
0.875000000 0.125000000 0.000000000
0.125000000 0.375000000 0.000000000
0.375000000 0.375000000 0.000000000
0.625000000 0.375000000 0.000000000
0.875000000 0.375000000 0.000000000
0.125000000 0.625000000 0.000000000
0.375000000 0.625000000 0.000000000
0.625000000 0.625000000 0.000000000
0.875000000 0.625000000 0.000000000
0.125000000 0.875000000 0.000000000
0.375000000 0.875000000 0.000000000
0.625000000 0.875000000 0.000000000
0.875000000 0.875000000 0.000000000
0.125000000 0.125000000 0.250000000 !O second layer
0.375000000 0.125000000 0.250000000
0.625000000 0.125000000 0.250000000
0.875000000 0.125000000 0.250000000
0.125000000 0.375000000 0.250000000
0.375000000 0.375000000 0.250000000
0.625000000 0.375000000 0.250000000
0.875000000 0.375000000 0.250000000
0.125000000 0.625000000 0.250000000
0.375000000 0.625000000 0.250000000
0.625000000 0.625000000 0.250000000
0.875000000 0.625000000 0.250000000
0.125000000 0.875000000 0.250000000
0.375000000 0.875000000 0.250000000
0.625000000 0.875000000 0.250000000
0.875000000 0.875000000 0.250000000 !
0.125000000 0.125000000 0.500000000 !O third layer
0.375000000 0.125000000 0.500000000
0.625000000 0.125000000 0.500000000
0.875000000 0.125000000 0.500000000
0.125000000 0.375000000 0.500000000
0.375000000 0.375000000 0.500000000
0.625000000 0.375000000 0.500000000
0.875000000 0.375000000 0.500000000
0.125000000 0.625000000 0.500000000
0.375000000 0.625000000 0.500000000
0.625000000 0.625000000 0.500000000
0.875000000 0.625000000 0.500000000
0.125000000 0.875000000 0.500000000
0.375000000 0.875000000 0.500000000
0.625000000 0.875000000 0.500000000
0.875000000 0.875000000 0.500000000
0.125000000 0.125000000 0.750000000 !O fourth
0.375000000 0.125000000 0.750000000
0.625000000 0.125000000 0.750000000
0.875000000 0.125000000 0.750000000
0.125000000 0.375000000 0.750000000
0.375000000 0.375000000 0.750000000
0.625000000 0.375000000 0.750000000
0.875000000 0.375000000 0.750000000
0.125000000 0.625000000 0.750000000
0.375000000 0.625000000 0.750000000
0.625000000 0.625000000 0.750000000
0.875000000 0.625000000 0.750000000
0.125000000 0.875000000 0.750000000
0.375000000 0.875000000 0.750000000
0.625000000 0.875000000 0.750000000
0.875000000 0.875000000 0.750000000
0.125000000 0.000000000 0.125000000 ! O face
0.125000000 0.000000000 0.375000000
0.125000000 0.000000000 0.625000000
0.125000000 0.000000000 0.875000000
0.375000000 0.000000000 0.125000000
0.375000000 0.000000000 0.375000000
0.375000000 0.000000000 0.625000000
0.375000000 0.000000000 0.875000000
0.625000000 0.000000000 0.125000000
0.625000000 0.000000000 0.375000000
0.625000000 0.000000000 0.625000000
0.625000000 0.000000000 0.875000000
0.875000000 0.000000000 0.125000000
0.875000000 0.000000000 0.375000000
0.875000000 0.000000000 0.625000000
0.875000000 0.000000000 0.875000000
0.000000000 0.125000000 0.125000000 ! O face
0.000000000 0.125000000 0.375000000
0.000000000 0.125000000 0.625000000
0.000000000 0.125000000 0.875000000
0.000000000 0.375000000 0.125000000
0.000000000 0.375000000 0.375000000
0.000000000 0.375000000 0.625000000
0.000000000 0.375000000 0.875000000
0.000000000 0.625000000 0.125000000
0.000000000 0.625000000 0.375000000
0.000000000 0.625000000 0.625000000
0.000000000 0.625000000 0.875000000
0.000000000 0.875000000 0.125000000
0.000000000 0.875000000 0.375000000
0.000000000 0.875000000 0.625000000
0.000000000 0.875000000 0.875000000
0.125000000 0.250000000 0.125000000 ! O inside
0.125000000 0.250000000 0.375000000
0.125000000 0.250000000 0.625000000
0.125000000 0.250000000 0.875000000
0.375000000 0.250000000 0.125000000
0.375000000 0.250000000 0.375000000
0.375000000 0.250000000 0.625000000
0.375000000 0.250000000 0.875000000
0.625000000 0.250000000 0.125000000
0.625000000 0.250000000 0.375000000
0.625000000 0.250000000 0.625000000
0.625000000 0.250000000 0.875000000
0.875000000 0.250000000 0.125000000
0.875000000 0.250000000 0.375000000
0.875000000 0.250000000 0.625000000
0.875000000 0.250000000 0.875000000
0.125000000 0.500000000 0.125000000 ! O inside
0.125000000 0.500000000 0.375000000
0.125000000 0.500000000 0.625000000
0.125000000 0.500000000 0.875000000
0.375000000 0.500000000 0.125000000
0.375000000 0.500000000 0.375000000
0.375000000 0.500000000 0.625000000
0.375000000 0.500000000 0.875000000
0.625000000 0.500000000 0.125000000
0.625000000 0.500000000 0.375000000
0.625000000 0.500000000 0.625000000
0.625000000 0.500000000 0.875000000
0.875000000 0.500000000 0.125000000
0.875000000 0.500000000 0.375000000
0.875000000 0.500000000 0.625000000
0.875000000 0.500000000 0.875000000
0.125000000 0.750000000 0.125000000 ! O inside
0.125000000 0.750000000 0.375000000
0.125000000 0.750000000 0.625000000
0.125000000 0.750000000 0.875000000
0.375000000 0.750000000 0.125000000
0.375000000 0.750000000 0.375000000
0.375000000 0.750000000 0.625000000
0.375000000 0.750000000 0.875000000
0.625000000 0.750000000 0.125000000
0.625000000 0.750000000 0.375000000
0.625000000 0.750000000 0.625000000
0.625000000 0.750000000 0.875000000
0.875000000 0.750000000 0.125000000
0.875000000 0.750000000 0.375000000
0.875000000 0.750000000 0.625000000
0.875000000 0.750000000 0.875000000
0.250000000 0.125000000 0.125000000 ! O face
0.250000000 0.125000000 0.375000000
0.250000000 0.125000000 0.625000000
0.250000000 0.125000000 0.875000000
0.250000000 0.375000000 0.125000000
0.250000000 0.375000000 0.375000000
0.250000000 0.375000000 0.625000000
0.250000000 0.375000000 0.875000000
0.250000000 0.625000000 0.125000000
0.250000000 0.625000000 0.375000000
0.250000000 0.625000000 0.625000000
0.250000000 0.625000000 0.875000000
0.250000000 0.875000000 0.125000000
0.250000000 0.875000000 0.375000000
0.250000000 0.875000000 0.625000000
0.250000000 0.875000000 0.875000000
0.500000000 0.125000000 0.125000000 ! O face
0.500000000 0.125000000 0.375000000
0.500000000 0.125000000 0.625000000
0.500000000 0.125000000 0.875000000
0.500000000 0.375000000 0.125000000
0.500000000 0.375000000 0.375000000
0.500000000 0.375000000 0.625000000
0.500000000 0.375000000 0.875000000
0.500000000 0.625000000 0.125000000
0.500000000 0.625000000 0.375000000
0.500000000 0.625000000 0.625000000
0.500000000 0.625000000 0.875000000
0.500000000 0.875000000 0.125000000
0.500000000 0.875000000 0.375000000
0.500000000 0.875000000 0.625000000
0.500000000 0.875000000 0.875000000
0.750000000 0.125000000 0.125000000 ! O face
0.750000000 0.125000000 0.375000000
0.750000000 0.125000000 0.625000000
0.750000000 0.125000000 0.875000000
0.750000000 0.375000000 0.125000000
0.750000000 0.375000000 0.375000000
0.750000000 0.375000000 0.625000000
0.750000000 0.375000000 0.875000000
0.750000000 0.625000000 0.125000000
0.750000000 0.625000000 0.375000000
0.750000000 0.625000000 0.625000000
0.750000000 0.625000000 0.875000000
0.750000000 0.875000000 0.125000000
0.750000000 0.875000000 0.375000000
0.750000000 0.875000000 0.625000000
0.750000000 0.875000000 0.875000000
Positive Energy and Hanging on first few electronic steps in supercell
Moderators: Global Moderator, Moderator
-
wwwennie
- Newbie
- Posts: 3
- Joined: Sat Sep 21, 2013 12:55 am
Positive Energy and Hanging on first few electronic steps in supercell
Last edited by wwwennie on Mon Nov 18, 2013 2:20 am, edited 1 time in total.
-
alex
- Hero Member
- Posts: 585
- Joined: Tue Nov 16, 2004 2:21 pm
- License Nr.: 5-67
- Location: Germany
Positive Energy and Hanging on first few electronic steps in supercell
Hi wwwennie,
some advice:
a) relax the primitive or the regular unit cell only (you would have to adjust the k-point set). It'll safe plenty(!) of computer time! It just makes absolutely no sense to relax a supercell of that size, if no impurities are present. The quantity you are after is the cell parameter only. The positions are fixed by symmetry anyway.
b) O POTCAR would require about 800eV cut-off for reasonable small Pulay stress.
c) NBANDS is a non-equal number. That's surely wrong and probably the cause of your positive total energy. Maybe it's a remainder of the 3x3x3 calculation.
d) the kpoint-mesh is far too large! (but itself it is correct)
e) You should set NPAR and ALGO
So much for starters. :-)
Cheers,
alex
some advice:
a) relax the primitive or the regular unit cell only (you would have to adjust the k-point set). It'll safe plenty(!) of computer time! It just makes absolutely no sense to relax a supercell of that size, if no impurities are present. The quantity you are after is the cell parameter only. The positions are fixed by symmetry anyway.
b) O POTCAR would require about 800eV cut-off for reasonable small Pulay stress.
c) NBANDS is a non-equal number. That's surely wrong and probably the cause of your positive total energy. Maybe it's a remainder of the 3x3x3 calculation.
d) the kpoint-mesh is far too large! (but itself it is correct)
e) You should set NPAR and ALGO
So much for starters. :-)
Cheers,
alex
Last edited by alex on Mon Nov 18, 2013 10:06 am, edited 1 time in total.
-
wwwennie
- Newbie
- Posts: 3
- Joined: Sat Sep 21, 2013 12:55 am
Positive Energy and Hanging on first few electronic steps in supercell
Thanks for the advice! Do you have general advice for how to best set NPAR? (e.g. anything apart from what is in the manual)
Last edited by wwwennie on Tue Nov 19, 2013 10:17 pm, edited 1 time in total.
-
alex
- Hero Member
- Posts: 585
- Joined: Tue Nov 16, 2004 2:21 pm
- License Nr.: 5-67
- Location: Germany
Positive Energy and Hanging on first few electronic steps in supercell
You are welcome!
General advice: Figure it out by yourself.
For fast networks (like Infiniband or comparable stuff) it's good to start with NPAR = 1. Double it from one calculation to the next and compare the timings. If you see huge drops you'll have to adjust further. You'll go through a minimum of execution time with increasing NPAR (or NPAR = 1 is already the fastest ...).
Cheers,
alex
General advice: Figure it out by yourself.
For fast networks (like Infiniband or comparable stuff) it's good to start with NPAR = 1. Double it from one calculation to the next and compare the timings. If you see huge drops you'll have to adjust further. You'll go through a minimum of execution time with increasing NPAR (or NPAR = 1 is already the fastest ...).
Cheers,
alex
Last edited by alex on Wed Nov 20, 2013 9:53 am, edited 1 time in total.
-
admin
- Administrator
- Posts: 2921
- Joined: Tue Aug 03, 2004 8:18 am
- License Nr.: 458
Positive Energy and Hanging on first few electronic steps in supercell
1. the 4x4x4 GGA-run: please try do use parallelization over k-points in addition to speed up (KPAR = [integer] ) (implemented in vasp-version 5.3.3)
as the first electronic steps starts with a random wavefuction, a positive total energy for the first electronic step does not necessarily mean that sth is wrong.
if the positive energy persists, please re-check
-- the input geometry (in particular the atomic distances, as written in OUTCAR)
-- that you have not continued from an inappropriate WAVECAR by error
2. HSE calculations are MUCH more time consuming than standard DFT ones
and it seems that you runs simply take so much time (without hanging)
for the first HSE-step (after a few default 'startup-electronic steps using DFT only). --> to reduce this need for CPU-time, please
set NKRED in the hybrid (HSE) calculations (unless you have done so anyway.)
as the first electronic steps starts with a random wavefuction, a positive total energy for the first electronic step does not necessarily mean that sth is wrong.
if the positive energy persists, please re-check
-- the input geometry (in particular the atomic distances, as written in OUTCAR)
-- that you have not continued from an inappropriate WAVECAR by error
2. HSE calculations are MUCH more time consuming than standard DFT ones
and it seems that you runs simply take so much time (without hanging)
for the first HSE-step (after a few default 'startup-electronic steps using DFT only). --> to reduce this need for CPU-time, please
set NKRED in the hybrid (HSE) calculations (unless you have done so anyway.)
Last edited by admin on Mon Dec 02, 2013 4:08 pm, edited 1 time in total.