VMD-L Mailing List

From: Francesco Pietra (chiendarret_at_gmail.com)
Date: Sun Aug 16 2015 - 05:54:24 CDT

forgot to specify the input

Gaussian 09: EM64L-G09RevD.01 24-Apr-2013
                15-Aug-2015
 ******************************************
 %chk=fmnh-opt.chk
 %nproc=16
 Will use up to 16 processors via shared memory.
 %mem=120GB
 -------------------------------------------------------
 # MP2/6-31G* Stable=Opt SCF=Tight Geom=PrintInputOrient

......
 Symbolic Z-matrix:
 Charge = -3 Multiplicity = 1

---------- Forwarded message ----------
From: Francesco Pietra <chiendarret_at_gmail.com>
Date: Sun, Aug 16, 2015 at 12:43 PM
Subject: ffTK problems with Gaussian OPT
To: VMD Mailing List <vmd-l_at_ks.uiuc.edu>

Hello:

I generated the Gaussian .gau input with ffTK (vmd 1.9.2 on GNU Linux amd64
jessy) and I am getting problems in optimizing the structure (heavy atoms
for HR X-ray diffraction, H-atoms added with CHIMERA, 52 total atoms).

First, the Gaussian optimization crashed because of "The wavefunction has
an RHF -> UHF instability".

On adding "Stable=Opt" to the to the route section, segmentation violation
occurred.

On replacing "Opt=(Redundant)" with "Stable=Opt" in the route section, the
log file tells

Eigenvectors of the stability matrix:
>
> Eigenvector 1: Triplet-A Eigenvalue=-0.0264435 <S**2>=2.000
> 106 ->121 -0.18719
> 110 ->121 0.18052
> 110 ->122 0.34361
> 113 ->121 0.23785
> 114 ->121 0.13098
> 117 ->121 -0.18294
> 117 ->122 0.23046
> SavETr: write IOETrn= 770 NScale= 10 NData= 16 NLR=1 NState= 1
> LETran= 28.
> The wavefunction has an RHF -> UHF instability.
> Operation on file out of range.
> FileIO: IOper= 2 IFilNo(1)= -554 Len= 46680 IPos= 0 Q=
> 47912003969888
>
>
> dumping /fiocom/, unit = 1 NFiles = 82 SizExt = 33554432 WInBlk =
> 524288
> defal = F LstWrd = 104333312 FType=2 FMxFil=10000
>

............................
> dumping /fiocom/, unit = 3 NFiles = 1 SizExt = 524288 WInBlk =
> 524288
> defal = T LstWrd = 1048576 FType=2 FMxFil=10000
>
> Number 0
> Base 524288
> End 1048576
> End1 1048576
> Wr Pntr 524288
> Rd Pntr 524288
> Length 524288
> Error termination in NtrErr:
> NtrErr Called from FileIO.
>

I know what Gaussian says about the last message above. However, the
wavefunction instability is not mentioned by Gaussian in that context.

--------------------------

I have been working on a full node, 16 cores, 120GB, and there is no
problem of disk space, either for final files or scratch.

I suspect that the RHF->UHF message originated from misinterpreting the
singlet (conjugated, charge -3) as a triplet.

thanks for advice

francesco pietra