Firefly and PC GAMESS-related discussion club

Learn how to ask questions correctly  
We are NATO-free zone

Re: hessian_asymmetry

Alex Granovsky

Dear Richard,

When computed semi-numerically (i.e. using finite differences of
analytical gradients) or fully numerically, resulting hessian matrix
is not exactly symmetric as it should. SYMFCM checks for this
asymmetry and prints pairs of Cartesian coordinates having largest
asymmetry. This asymmetry is the measure of a numerical noise and
anharmonicity. One may check for largest absolute asymmetry and
ensure it is not too high (say, less than 3d-3 or so). If asymmetry
is high, the quality of frequencies is questionable.  

Kind regards,

On Tue Sep 2 '14 3:48pm, Richard wrote
>Lines such as

>SYMFCM: LARGEST HESSIAN ASYMMETRY: PAIR #    1: I =  15 J =   12 REL. ASYM = 0.2210484995D-02 ABS. ASYM = 0.2920159034D-04 etc.

>are printed after a hessian calculation.

>Is there any useful information here, maybe about the reliability of the frequencies?


[ Previous ] [ Next ] [ Index ]           Tue Sep 2 '14 6:40pm
[ Reply ] [ Edit ] [ Delete ]           This message read 430 times