Qchem input files from edft and idft

[Michele Turelli]

Hello,

so far I have been running simulations using the approximate ZINDO method for the evaluation of transfer integrals.

I want to compare these calculations with others based on DIPRO ( noCP+noSCF).

I have VOTCA-ctp on my workstation and I have to rely on a cluster for QChem calculations (Gaussian09).

Getting to grips with both the edft and idft calculators has not proven so straightforward unfortunately.
In particular, would it be possible to produce the Gaussian input file independently without running them and transfer them on the cluster to run the calculations there?

One way around it I have tried so far, has been to perform a "mock run" to force the edft calculator to create the monomer input files which I have run on the cluster.
Of these jobs I have saved the fort.7 file (obtained via the command Punch=MO), named it "mol_N.orb" and put it back where idft would look for it,
but reading the manual I suspect it does not contain all the data necessary to idft.

Infact the same tentative procedure with idft has produced inputfiles with problems (small interatomic distances between atoms) and this

I've chalked up to not having all the relevant data about the monomers.

Here's the commands I've run:
        ctp_parallel -o options.xml -f state_${file}.sql -e edft -j "write run"
        ctp_parallel -o options.xml -f state_${file}.sql -e idft -j "write run"

Could you give me an example of what a proper mol_N.orb should look like?

I am also a bit confused by the commands to give Gaussian calculations for DIPRO (noCP + no SCF).

I attach an example of a monomer calculation and a pair calculation together with the orb for the monomer [with only first and last MO so that the file is not too big].

Thanks,
Michele

[Denis Andrienko]

Hi Michele,

orb file is a binary file where MO coefficients and energies are stored.

I think you are not talking about nwchem, but Gaussian?

If first run Gaussian, you need to keep both fort.7 files and log files.

After this you run edft, extract information from fort.7 and log, save it to the orb file.

The same with idft, for the dimer.

If you want to script everything yourself, it is simpler not to use ctp_parallel, but run ctp_tools -e coupling, have a look here:
https://groups.google.com/d/msg/votca-ctp/nOWktKbkuyk/1q3SHdPqBQAJ

If you want I can send you a folder with an example for both calculators.

Best,

Denis

--
You received this message because you are subscribed to the Google Groups "votca-ctp" group.
To unsubscribe from this group and stop receiving emails from it, send an email to votca-ctp+...@googlegroups.com.
To post to this group, send email to votc...@googlegroups.com.
Visit this group at https://groups.google.com/group/votca-ctp.
For more options, visit https://groups.google.com/d/optout.

[Michele Turelli]

Thank you very much for the quick reply.
Yes I am talking about G09 calculations (I wrote QChem for quantum chemistry, sorry it was misleading).

Yes it would be very helpful if you could attach those files.
Thank you very much!

In order to run a noCP+noSCF job it is sufficient to not specify the "Counterpoise=..." in the G09 input for monomers (so basically running them as I am doing now) while the dimer noSCF calculation in what specific G09 command would it translate? 

Thanks again!

[Denis Andrienko]

Here is the example for the coupling test without using ctp_parallel

http://www.mpip-mainz.mpg.de/~andrienk/public_html/coupling_test.tar.gz

if you run

ctp_tools -e coupling -o coupling.xml

you will get

coupling.out.xml

with coupling elements.

There are also gaussian input files (the dimer input has a full SCF loop, but you can also stop it after the first iteration by chaging scf option to scf=(maxcycle=1,conver=1).

Best,

Denis

--

[Denis Andrienko]

To run the entire pair list using ctp_parallel, you need

1. Create list of all monomers to run:

   ctp_parallel -e edft -j write -f state.sql -o options.xml -s 0

   this is saved into edft.jobs

2. Run gaussian for all monomers

   ctp_parallel -e edft -f state.sql -o options.xml

   this will create folder OR_FILES with sub-folders  gaussian  and molecules

   make sure that g09 executable can be found by the shell before you run this

3. Create a list of all pairs to run:

   ctp_parallel -e idft -j write -f state.sql -o options.xml -s 0

   it is saved into idft.jobs

4. Run gaussian for all pairs

   ctp_parallel -e idft -f state.sql -o options.xml

   again, make sure that the gaussian executable is available

5. Import the integrals into the job file

You can find a working example here:

www.mpip-mainz.mpg.de/~andrienk/hub/coupling_parallel_test.tar.gz

Best,

Denis

[Denis Andrienko]

sorry, the link was wrong:

http://www.mpip-mainz.mpg.de/~andrienk/hub/coupling_test.tar.gz

On Wed, Mar 20, 2019 at 11:41 AM Denis Andrienko <andr...@googlemail.com> wrote: