PC GAMESS/Firefly-related discussion club

Learn how to ask questions correctly

Thomas Patko
tpatko@gmail.com

I am performing RHF optimizations for various conformations of some biological systems that I wish to model with force field methods.  In order to obtain the necessary charges with R.E.D. (http://q4md-forcefieldtools.org/RED/http://q4md-forcefieldtools.org/RED/) I need to get good geometry optimization that result in biological relevant conformations.  When I used the DLC optimization engine, I got the lowest energy conformation, but as it turned out not a biologically relevant one (see options below).

$CONTRL SCFTYP=RHF RUNTYP=OPTIMIZE
      ICHARG=0 MULT=1 COORD=ZMTMPC nzvar=348 maxit=100
$END
$system mwords=100 timlim=1000000 kdiag=0 nojac=1 $end
$p2p p2p=1 dlb=1 $end
$scf dirscf=.t. $end
$statpt nstep=1024 method=gdiis flags=0 noreg=10 regtol=1.5 $end
$zmat dlc=1 auto=1 $end
$BASIS GBASIS=N31 NGAUSS=6 NDFUNC=1 NPFUNC=1 $END

When I simply removed the nzvar option (this should set it to the default of zero) then the optimizations yielded optimized structures that were biologically relevant conformations.

My question is what optimization engine is then actually used then nzvar is zero?

I suspect that the $zmat dlc=1 auto=1 $end is then completely ignored (and thus could be removed on the options), correct?  Are any other options also then ignored?  Is this search still conducted in internal coordinates (initial structure is provided as MOPAC style z-matrix) but just not using the DLC engine?  Should this still be faster than conducting such a search in Cartesian coordinates, correct?

Regards,

Thomas