Troubles with calculation of MEP when using C1/C2-DIIS (bug?)

Posted by Evgeniy on April 18, 2011 at 10:11:15:

Dear Molcas Community,

I have faced problems when calculating minimum energy path (MEP) starting from a transition state (TS). At one side of TS it works fine, but at the other side(which is presumably more steep) the procedure does not converge. I used all default options, i.e. optimization scheme RFO, Hessian update Broyden-Fletcher-Goldfarb-Shanno, Sphere=0.1 etc. When I tried to use the conjugate gradient methods (C1-DIIS and C2-DIIS) the calculation crashed during slapaf with rc=174(unknown). The same happens for the example of MEP search given in the manual when either C1-DIIS or C2-DIIS is used. Is there a bug for C1/C2-DIIS in the case of MEP?

Follow Ups:

Post a Followup

Name:
E-Mail:

Subject:

if B is 1s22s22p1, what is Li?

Passfield:

Comments:
: Dear Molcas Community, : I have faced problems when calculating minimum energy path (MEP) starting from a transition state (TS). At one side of TS it works fine, but at the other side(which is presumably more steep) the procedure does not converge. I used all default options, i.e. optimization scheme RFO, Hessian update Broyden-Fletcher-Goldfarb-Shanno, Sphere=0.1 etc. When I tried to use the conjugate gradient methods (C1-DIIS and C2-DIIS) the calculation crashed during slapaf with rc=174(unknown). The same happens for the example of MEP search given in the manual when either C1-DIIS or C2-DIIS is used. Is there a bug for C1/C2-DIIS in the case of MEP?

[ Follow Ups ] [ Post Followup ] [ Molcas user's WWWBoard ]