NEB calculation seems to be tremendously slow

[Rainbow]

Hi,

I've been trying to run some cNEBs for compression of a methylthiolate species between Cu(100) slabs. However, the calculation seems to be tremendously slow (10+ days with 6 nodes 12 cores/node). Is there anything in the INCAR here that I can change to speed up the process?


SYSTEM = cNEB MeS/Cu(100) compression

PREC = Normal
ENCUT = 350
ALGO = Normal
NELMDL = -2.4
LREAL = Auto
#LWAVE = .FALSE. # do not write WAVECAR file

ISMEAR = 2
SIGMA = 1.4

IMAGES = 12 # number of images, forces NEB method
LCLIMB = .TRUE. # use climbing image method
IOPT = 3 # Q-Min, relaxation method
IBRION = 4 # tell VASP to do MD
POTIM = 1 # tell VASP zero time step
SPRING = -4 # spring constant for images, >0 images don't move
# along path

NELMIN = 6 # do a minimum of four electronic steps
EDIFF = 1E-5 # high accuracy
EDIFFG = -0.4 # accuracy of ions
NSW = 1000 # 1000 ionic steps in ions

LPLANE = .TRUE.
LSCALU = .FALSE.
NSIM = 5
NCORE = 5


Regards!

[graeme]

Probably. If you can post a .tar.gz file of the calculation, I can take a look.

Please remove any CHG* WAV* files. If it is too large, you can post a link to it.

Also, take a look back at the many posts about controlling convergence, and my suggestions.