[molpro-user] What goes into optconv?

[Dr. Seth Olsen]

Thank you, Professor Werner, for this reply and also to the reply to my 
question about improving CASSCF convergence.  Regarding the step vector 
and optconv, I have noticed that optconv is dependent upon the 
specifications for 'step' that are used as additional options for 'opt' 
or 'optg' (particularly for conical intersection optimizations).  At the 
root of my original question was a desire to double-check that I was not 
artificially forcing convergence of a minimal energy intersection search 
via the specification of very small initial and maximum step sizes.  I 
would therefore ask the list: is it possible to artificially force 
convergence in this manner?

Happy New Year to all!

Cheers,
Seth

H. -J. Werner wrote:

>It is the length of the step vector. If z-matrix
>coordinates are used, bond lengths are in bohr and angles in radian.
>If cartesian coordinates are used, it is simply the length
>of the stepvector in 3N cartesian coordinates (in bohr).
>H.-J. Werner
>
>On Di, 10 Jan 2006, Dr Seth Olsen wrote:
>
>  
>
>>Hi Molpro-users,
>>
>>I've got a question about MolPro that I think is pretty fundamental. 
>>What exactly is optconv?  What goes into the quantity?  I'm particularly
>>interested with respect to it's use as a measure of convergence in
>>minimal energy conical intersection searches.  I've looked at the code
>>but I confess that I'm not the best at extracting a 'big picture' from
>>large scale FORTRAN code.  
>>
>>Cheers,
>>
>>Seth
>>
>>
>>
>>ccmsccmsccmsccmsccmsccmsccmsccmsccmsccmsccmsccmsccmsccmsccmsccms
>>
>>Dr Seth Olsen, PhD
>>Postdoctoral Fellow, Biomolecular Modeling Group
>>Centre for Computational Molecular Science
>>Chemistry Building,
>>The University of Queensland
>>Qld 4072, Brisbane, Australia
>>
>>tel (617) 33653732
>>fax (617) 33654623
>>email: s.olsen1 at uq.edu.au
>>Web: www.ccms.uq.edu.au 
>>
>>ccmsccmsccmsccmsccmsccmsccmsccmsccmsccmsccmsccmsccmsccmsccmsccms
>>
>>
>>    
>>
>
>  
>