LEPSILON with gamma point VASP

Message

tholme · #1 Post by **tholme** » Mon Nov 30, 2009 11:34 pm

Is the gamma point only version of VASP supposed to be able to calculate the dielectric properties with LEPSILON = .TRUE. ? I get an error

My output looks like this:

Code: Select all

...
DAV:  47    -0.456858097619E+03   -0.11918E-05    0.64768E-07   520   0.243E-04    0.872E-05
DAV:  48    -0.456858098563E+03   -0.94424E-06    0.65805E-07   520   0.183E-04
   1 F= -.45685810E+03 E0= -.45532004E+03  d E =-.307611E+01
 Linear response reoptimize wavefunctions to high precision
DAV:   1    -0.456858098944E+03   -0.38155E-06    0.66497E-07   584   0.139E-04
DAV:   2    -0.456858098946E+03   -0.12224E-08    0.66872E-07   496   0.897E-05
DAV:   3    -0.456858098945E+03    0.64028E-09    0.67072E-07   496   0.652E-05
 Linear response G [H, r] |phi>, progress :
  Direction:   1
       N       E                     dE             d eps       ncg     rms
 internal error in LRF_COMMUTATOR: for the gamma-point code, LI must be set to .FALSE.

#2 Post by **admin** » Tue Dec 01, 2009 1:05 pm

as written in the error statement: LI=.True. causes the gamma-only
(real) version to stop immediately.
please either set LI to false explicitely, or use the standard version of vasp

tholme · #3 Post by **tholme** » Thu Dec 03, 2009 5:32 pm

If I set LI=.FALSE. in my INCAR file, I get the same error. Apparently the code does not read the value LI from INCAR ? So must we use the standard version of VASP? This can be inconvenient since LEPSILON calculations are quite time consuming.

Thanks for any input you have.

#4 Post by **admin** » Mon Dec 07, 2009 8:27 am

LI is not read from INCAR, you have to change this statement in the
source-code and re-compile