last day (490 days later) » 

Feeds
Feeds
3:38 PM
 
hBy2Py
hBy2Py
Ok. First one is a ZMAT opt of isosorbide mononitrate
This didn't actually converge -- I attributed it to the low-barrier internal rotor of the nitro-group, but based on how well a different molecule converged with a Cartesian opt, it makes me think that ZMAT OPT in GAMESS is indeed a bad idea
Input: 
! Isosorbide mononitrate
! Run 01 - RHF/PCseg-1 SP

 $CONTRL SCFTYP=RHF MAXIT=200 RUNTYP=OPTIMIZE $END
 $CONTRL UNITS=ANGS COORD=UNIQUE NZVAR=60 ISPHER=1 $END
 $SCF DIRSCF=.TRUE. $END
 $SYSTEM TIMLIM=100 MWORDS=150 $END
 $STATPT METHOD=QA NSTEP=200 $END
 $BASIS  GBASIS=PCseg-1 $END
 $DATA
Isosorbide nitrate at the cleaned-up Avogadro SMILES-generated geometry
C1
OXYGEN     8.0     0.75906       -2.54863       -2.43610
NITROGEN   7.0     0.62814       -1.54688       -2.39741
OXYGEN     8.0    -0.21003       -1.13071       -2.51432
Snip from middle of output: 
BEGINNING GEOMETRY SEARCH POINT NSERCH=   0 ...

 COORDINATES OF ALL ATOMS ARE (ANGS)
   ATOM   CHARGE       X              Y              Z
 ------------------------------------------------------------
 OXYGEN      8.0   0.7590600000  -2.5486300000  -2.4361000000
 NITROGEN    7.0   0.6281400000  -1.5468800000  -2.3974100000
 OXYGEN      8.0  -0.2100300000  -1.1307100000  -2.5143200000
 OXYGEN      8.0   1.5768100000  -1.1514600000  -2.1235700000
 CARBON      6.0   2.4158900000  -1.6500300000  -1.7038100000
Other one is a Cartesian $STATPT of the non-nitrated isosorbide 
! Run 01 - RHF/PCseg-0 OPT
!

 $CONTRL SCFTYP=RHF MAXIT=200 RUNTYP=OPTIMIZE $END
 $CONTRL UNITS=ANGS COORD=UNIQUE ISPHER=1 $END
 $SCF DIRSCF=.TRUE. $END
 $SYSTEM TIMLIM=100 MWORDS=150 $END
 $STATPT METHOD=QA NSTEP=200 $END
 $BASIS  GBASIS=PCseg-0 $END
 $DATA
Isosorbide at the cleaned-up Avogadro SMILES-generated geometry
C1
OXYGEN     8.0     1.78215       -1.29939       -1.93203
CARBON     6.0     2.41719       -2.01202       -1.54993
HYDROGEN   1.0     2.84716       -1.55764       -0.99470
CARBON     6.0     2.57263       -3.15237       -0.99646
Note the absence of NZVAR in $CONTRL, and the entire removal of the $ZMAT card.
Output snip: 
BEGINNING GEOMETRY SEARCH POINT NSERCH=  10 ...

 COORDINATES OF ALL ATOMS ARE (ANGS)
   ATOM   CHARGE       X              Y              Z
 ------------------------------------------------------------
 OXYGEN      8.0   1.3242730126  -0.8891606646  -2.0611573107
 CARBON      6.0   2.4474079563  -1.6993863650  -1.6820071962
 HYDROGEN    1.0   3.0894764327  -1.2042910054  -0.9638698532
 CARBON      6.0   2.0976333097  -2.9835620731  -0.9918383892
 HYDROGEN    1.0   1.1109383167  -3.4315224927  -1.0604076714
No mention of internals at all.
So, either it's converting under the hood, silently, or just optimizing directly in Cartesians.
Regardless, I first attempted this same isosorbide $STATPT using $ZMAT, and it didn't converge w/in 200 steps. This Cartesian opt converged in ~55 steps.
 
Martin - マーチン
Martin - マーチン
4:10 PM
@hBy2Py I'll look at it tomorrow, I'm off to bed now :D
 
hBy2Py
hBy2Py
Heheh, g'night
 
 
1 hour later…
orthocresol
orthocresol
5:21 PM
room topic changed to Optimising optimising: Discussion on optimisations &c. in quantum chem packages (no tags)
*ahem*
I have corrected an, err, critical mistake in the room title.
;)
room topic changed to Optimizing optimizing: Discussion on optimizations &c. in quantum chem packages (no tags)
 
 
4 hours later…
hBy2Py
hBy2Py
9:15 PM
@orthocresol :-P
MURICA
GAMESS seems to want to take great leaping bounds around the equilibrium geometry of this molecule. I'm having to manually tighten the leash by reducing the trust radius settings. Annoying!
 

  last day (490 days later) »