[Columbus] ciudg to do single reference CI (CIS)

[Hans Lischka]

Dear Spiridoula and Felix,

Felix, thank you very much for your solution. I have only one question 
for my understanding. The occminr is cumulative, this means that the 
values should always increase or remain the same. You suggest the values 
1 0 2. What is the explanation for the 0?

Spiridoula:
The explanation for the problem with the SR-CIS is that Columbus does 
not do a simple diagonalization of a given Hamiltonian matrix, but uses 
a two-step procedure: first comes the generation of reference 
configurations, one set for each state, and corresponding starting 
vectors, followed by the actual Davidson diagonalization. You want three 
states, but have only one reference, and the program will break, 
depending how the starting vectors are created. To solve the problem, 
one has to find fake references which do not distort the CIS. I hope the 
procedure of Felix helped.

Best regards, Hans

On 11/8/2017 6:19 PM, Felix Plasser wrote:
> Hi Spiridoula,
> 
> I just played around with it to see how it works. One way to do this is 
> you first setup an MR-CIS with a CAS(2,3) active space and then you 
> change in cidrtin:
> 
>    1 0 2  /occminr
> 
>   1 1 1  /occmin
> 
> Then it works with the default ciudgin file that colinp creates. I am 
> attaching the files that are different in my job when compared to yours. 
> I am also sending the standard output "runls".
> 
> -Felix
> 
> 
> On 11/06/2017 06:42 PM, SPIRIDOULA MATSIKA wrote:
>> Dear Hans and everyone else who sent some suggestions,
>>
>> Thank you all! My systems is very simple: closed shell H2 +CIS.  So I 
>> only have 1 reference (the closed shell HF) and I add singles at the 
>> CI level. I have been able to create the list using cidrt (I had to 
>> play some tricks and manually edit cidrtin but I think I got that 
>> part). But ciudg fails. I am asking for 3 states. In the beginning it 
>> was failing because I had less references than the number of states I 
>> was asking for, but I think I have solved that part using IVMODE=0 and 
>> ISTRT=1. Now it is failing but there is no clear error message, so I 
>> am not sure why. I suspect that there may still be some 
>> inconsistencies between the parameters I manually changed (IVMODE and 
>> ISTRT) and other default values ( I do not fully understand all the 
>> parameters in ciudg). I have included my compressed directory here. I 
>> copied the ciudgls in the main directory since that is the program 
>> that fails. Any hint would be greatly appreciated.
>>
>> Thanks!
>> Spiridoula
>>
>>
>>
>>> On Nov 4, 2017, at 4:20 AM, Hans Lischka <hans.lischka at univie.ac.at 
>>> <mailto:hans.lischka at univie.ac.at>> wrote:
>>>
>>> Dear Spiridoula,
>>>
>>> Could you please provide a bit more information about the 
>>> calculation. Was it an closed shell or open shell? Could you please 
>>> send us the a tar file with the inputs and outputs, this would make 
>>> things easier.
>>>
>>> Thank you, Hans
>>>
>>> On 11/3/2017 11:09 PM, SPIRIDOULA MATSIKA wrote:
>>>> Dear Columbus colleagues,
>>>> I have tried to make Columbus do a CIS calculation using a single 
>>>> reference. It seems to me that it cannot be done. I have tried to use
>>>>  IVMODE = 0
>>>> ISTRT = 1
>>>> to generate initial vectors from both z and y walks but this has not 
>>>> worked. Is there a way to do such calculation?
>>>> Thank you,
>>>> Spiridoula
>>>> Spiridoula Matsika
>>>> Professor of Chemistry
>>>> Department of  Chemistry
>>>> Temple University
>>>> 1901 N.13th Street
>>>> 242 Beury Hall
>>>> Philadelphia, PA 19122
>>>> Phone: 215 204 7703
>>>> Fax: 215 204 1532
>>>> smatsika at temple.edu 
>>>> <mailto:smatsika at temple.edu><mailto:smatsika at temple.edu>
>>>> _______________________________________________
>>>> Columbus mailing list
>>>> Columbus at lists.osc.edu <mailto:Columbus at lists.osc.edu>
>>>> https://lists.osu.edu/mailman/listinfo/columbus
>>>
>>> _______________________________________________
>>> Columbus mailing list
>>> Columbus at lists.osc.edu <mailto:Columbus at lists.osc.edu>
>>> https://lists.osu.edu/mailman/listinfo/columbus
>>
>>
>>
>> _______________________________________________
>> Columbus mailing list
>> Columbus at lists.osc.edu
>> https://lists.osu.edu/mailman/listinfo/columbus
> 
> -- 
> Felix Plasser
> Institute for Theoretical Chemistry
> University of Vienna
> Währingerstr. 17/404
> 1090 Wien
> felix.plasser at univie.ac.at
> http://homepage.univie.ac.at/felix.plasser/
> Tel.: +43-1-4277-52757
> 
> 
> 
> _______________________________________________
> Columbus mailing list
> Columbus at lists.osc.edu
> https://lists.osu.edu/mailman/listinfo/columbus
>