Dear all,
I have been trying to optimize a [Zn(NH3)4]2+ complex with NWChem, and COSMO, using a rather tight SCF convergence criteria (the criteria recommended by Frank Neese). I could not converge it with the regular procedure so I decided to use cgmin. Cgmin converges very nicely, but it crashes soon after. In order to pinpoint the problem, I switched to the default convergence criteria and ran a few tests.
Below I attach the input and the error I get. I also tried a few modifications: taking out the charge density fitting, wich avoided the problem but revealed a new one, this time with COSMO, among others. After the input and the error, I detail all modifications I used and the errors I got (and the combination that avoided errors). Everything preceded by a # is my annotation. It appears that using charge density fitting or COSMO cause cgmin to fail.
All the tests were ran in nwchem-6.3.revision2-src.2013-10-17, in a x86-64 machine.
Any help is very much appreciated.
Cheers!
Raul
INPUT:
start ZnComplex
echo
charge 2
memory total 4000 mb
geometry units angstroms noautosym
Zn -1.262 0.914 -0.069
N -1.516 2.882 0.229
H -1.195 3.199 1.153
H -1.012 3.466 -0.451
H -2.501 3.173 0.167
N -2.366 -0.099 1.267
H -2.904 -0.869 0.847
H -1.807 -0.531 2.014
H -3.059 0.487 1.749
N 0.671 0.424 0.155
H 0.834 -0.592 0.138
H 1.279 0.810 -0.578
H 1.071 0.752 1.044
N -1.860 0.441 -1.925
H -2.173 1.254 -2.471
H -1.123 0.002 -2.491
H -2.648 -0.220 -1.942
end
basis "ao basis"
Zn library def2-tzvp
N library def2-svp
H library def2-svp
end
basis "cd basis"
* library "Ahlrichs Coulomb Fitting"
end
cosmo
dielec 80.0
do_gasphase False
end
dft
cgmin
vectors output ZnComplex.movecs
iterations 100
grid fine
xc xtpss03 ctpss03
disp vdw 3
print convergence information parameters
mult 1
end
driver
maxiter 200
trust 0.3
gmax 0.0500
grms 0.0300
xmax 0.1800
xrms 0.1200
xyz ZnComplex_prev
end
task dft optimize
driver
maxiter 200
trust 0.1
gmax 0.003
grms 0.0001
xmax 0.004
xrms 0.002
xyz ZnComplex
end
task dft optimize
- Error 1, after converging the firs SCF.
NWChem DFT Gradient Module
--------------------------
charge = 2.00
wavefunction = closed shell
------------------------------------------------------------------------
dftg_cdfit_gen: file cdfit does not exist 0
------------------------------------------------------------------------
An error occured while trying to read or write to disk space
------------------------------------------------------------------------
For more information see the NWChem manual at http://nwchemgit.github.io/index.php/NWChem_Documentation
For further details see manual section:
0:0:dftg_cdfit_gen: file cdfit does not exist:: -1
(rank:0 hostname:c15 pid:6529):ARMCI DASSERT fail. ../../ga-5-2/armci/src/common/armci.c:ARMCI_Error():208 cond:0
- Variant 2: Without charge density fitting:
- Error 2, in the second step of the first optimization.
Screening Tolerance Information
-------------------------------
Density screening/tol_rho: 1.00D-10
AO Gaussian exp screening on grid/accAOfunc: 14
CD Gaussian exp screening on grid/accCDfunc: 20
XC Gaussian exp screening on grid/accXCfunc: 20
Schwarz screening/accCoul: 1.00D-08
cosmo_initialize: bq_create failed 0
------------------------------------------------------------------------
There is an error related to the specified geometry
------------------------------------------------------------------------
For more information see the NWChem manual at http://nwchemgit.github.io/index.php/NWChem_Documentation
- This message repeates 3 more times, and then:
For further details see manual section:
For further details see manual section:
3:3:cosmo_initialize: bq_create failed:: -1
(rank:3 hostname:d15 pid:1497):ARMCI DASSERT fail. ../../ga-5-2/armci/src/common/armci.c:ARMCI_Error():208 cond:0
- Variant 3: With spherical atomic basis set:
- Error 3: Same as Error 1
- Variant 4: With spherical atomic basis set and not charge density fitting
- Error 4: Same as Error 2
- Variant 5: Cosmo but do_gasphase True:
- Error 5: Same as Error 1
- Variant 6: No Cosmo:
- Error 6: Same as Error 1
- Variant 7: No Cosmo, no charge density fitting.
- Error 7: No error, all SCF converge easily, the first optimization converges after 3 SCF, and the second after 6. Even if I added back the tight SCF convergence criteria, the SCF and the optimizations converge easily.
|