I've compiled NWChem 6.6-rev27746 and applied all 19 patches on an CentOS 7.2 system using the Intel 2016.3.210 compilers. The compile was configured with 64_to_32, MKL, SCALAPACK, ELPA.
I've been running some initial calculations of small molecules as a test. The B3LYP, M11, M06-2X simlulations all complete. Furthermore, the B3LYP-D3 calculations for the same molecules also work.
As for B3LYP-D3(BJ), most of the small molecules run; however, it fails for the following input. Here is the input file:
echo
start ch3-dat
geometry noautosym
H 0.32788591 0.15780840 1.03021043
C 0.13494670 0.00676764 -0.02429037
H -0.02354775 0.95705220 -0.67247537
H 0.21289023 -0.93440752 -0.43099729
X -1.13583065 -0.05709849 0.20418594
end
basis
* library 6-311++g**
end
dft
MULT 2
XC B3LYP
DIRECT
NOIO
MAXITER 150
DISP vdw 4
end
task dft optimize
task dft freq
And here are the first iterations of the SCF:
Memory utilization after 1st SCF pass:
Heap Space remaining (MW): 8.54 8541542
Stack Space remaining (MW): 15.20 15203396
convergence iter energy DeltaE RMS-Dens Diis-err time
---------------- ----- ----------------- --------- --------- --------- ------
d= 0,ls=0.0,diis 1 NaN NaN 2.09D-03 4.03D-02 1.0
1.98D-03 4.07D-02
d= 0,ls=0.0,diis 2 NaN NaN 8.50D-04 1.84D-03 1.5
9.59D-04 2.39D-03
...
d= 0,ls=0.0,diis 11 NaN NaN 2.51D-09 6.93D-15 5.3
1.11D-09 3.00D-15
d= 0,ls=0.0,diis 12 NaN NaN 2.63D-10 8.38D-17 5.7
2.47D-10 1.44D-16
Singularity in Pulay matrix. Error and Fock matrices removed.
Singularity in Pulay matrix. Error and Fock matrices removed.
Singularity in Pulay matrix. Error and Fock matrices removed.
Singularity in Pulay matrix. Error and Fock matrices removed.
The main difference between this molecule and others is the use of the 'X' atom. I'm wondering if that is causing the error.
Thanks.
Joseph
|