[molpro-user] Print the CI vector in RS2 and RS2C

José Cortés zolidus at gmail.com
Wed Feb 3 08:33:47 CET 2016


Dear Molpro Users and Developers

I have a doubt about the printing of the CI vector coefficients  (CSF and
contracted bielectronic functions) using Multireference Rayleigh
Schrödinger perturbation theory. In the 2010 version, the above can only be
performed with the "RS2" module.

Reviewing the source code, I found that the CI vector is printed on the
output file using the "ciprin" subroutine present in
Molpro/src/mrci/ciexit.F, and the amount of information is controlled with
"Kref, intern , ksing, kpair" variables.

In the RS2C module, "ciexit" it's not called, but it can be done by
changing a few lines, for example by modifying

    1202|   c.....calculate properties and print results
    1203|   c
    1204|         if(iusecs.ne.0.or.iuseci.ne.0) then
    1205|           call cirs2exit
    1206|           goto 9000
    1207|         else
    1208|           call gptran(0)
    1209|           call cigpi(-1)
    1210|           call cigsi(-1)
    1211|           call ciexit(nstatt)
    1202|         end if

in the cipt2.F (cipro.F) routine, to

    1202|   c.....calculate properties and print results
    1203|   c
    1204|           call ciexit(nstatt)
    1205|

the CI vector is printed for systems where "core,0" and "closed,0" in the
definition of the reference configuration space. If core or closed are not
set to zero, an error occurs and the calculation ends with an error without
print the CI vector.

This is fine for small systems, but it is impractical for medium or large
systems.
My questions are:
         there is an error when the CI vector is printed in this way?
         Why the calculations end abruptly when core and closed are not
equal to 0?
         Or in their respective case, Is there any other way to print the
CI vector with the RS2C module?

I would greatly appreciate any information you could provide.

Best regards
José
Universidad Nacional
Autónoma de México
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.molpro.net/pipermail/molpro-user/attachments/20160203/8e7e6b55/attachment.html>


More information about the Molpro-user mailing list