[molpro-user] IRC restart
Gabriel Freitas
gabrielnfreitas at gmail.com
Wed Nov 17 15:07:33 GMT 2010
Dear Molpro users,
Does anybody know how to restart an IRC job? In other words, if I get some
convergence problem (like '? Error ? No convergence ? The problem occurs
in rhfpro ') in an optimization step of IRC, can I restart my IRC from where
it stuck?
Here is an example:
*** INPUT (part of it)
basis=aug-cc-pV(T+d)Z
rhf,maxdis=50,maxit=100;
rccsd(t),maxit=70,thrden=1.d-6,thrvar=1.d-10;
{optg,root=2,method=qsd,maxit=100,saveact=TS3,rewind;print,history}
{optg,method=qsdpath,dir=+1,numhess=1,hesscentral,maxit=100,saveact=TS3_IRC;print,history}
****
*** OUTPUT (part of it)
PROGRAM * OPT (Geometry optimization) Authors: F. Eckert and H.-J.
Werner
Geometry optimization using default procedure for command RCCSD(T)
Geometry written to block 1 of record 700
Optimized active variables along the reaction path will be saved in file
ts3_irc.act
Numerically approximating hessian using central energy differences
Task list generated. Total number of displacements: 42
(...)
Numerical RCCSD(T) hessian completed. CPU-time: 58207.17 sec, Elapsed:
59983.65 sec
RCCSD(T) hessian saved to record 5301.2
Combined Powell-Murtagh-Sargent Update of Hessian
Quadratic Steepest Descent - Reaction Path Following using exact Hessian
Hessian eigenvalues: -0.720210 0.045186 0.086122 0.220480 0.252733
0.333235
QSDPATH1: Starting at a critical point. Gradient norm= 0.44D-04 Stepping
along the transition vector, idir = 1
Optimization point 1
Variable Last Current
Next Gradient Hessian
E(RCCSD(T)) / Hartree 0.00000000 -548.40580762 0.00000000
R1 / ANG 0.00000000 1.68000597
1.67535167 -0.00008214 -0.06944810
R2 / ANG 0.00000000 1.58945503
1.59220702 -0.00000047 0.46843406
A / DEGREE 0.00000000 66.05724619
66.95062670 -0.00000006 -0.00013657
R3 / ANG 0.00000000 1.35229995
1.35212903 0.00000623 0.88658757
B / DEGREE 0.00000000 102.58308447
102.51347761 0.00000002 0.00006953
G / DEGREE 0.00000000 90.11384708
90.00488570 0.00000005 0.00002365
Convergence: 1.00000000 (line search)
0.01878095 0.00004385 (total)
Numerically approximating hessian using central energy differences
Task list generated. Total number of displacements: 42
(...)
Numerical RCCSD(T) hessian completed. CPU-time: 58106.44 sec, Elapsed:
59815.04 sec
RCCSD(T) hessian saved to record 5301.2
Hessian eigenvalues: -0.603756 0.044188 0.077717 0.215977 0.252942
0.332157
Performing a regular QSD step. Stepsize = 0.135181 Curvature =
.14652D+01
Optimization point 2
Variable Last Current
Next Gradient Hessian
E(RCCSD(T)) / Hartree -548.40580762 -548.40592241 0.00000000
R1 / ANG 1.68000597 1.67535167
1.63672546 0.01118126 -0.06876999
R2 / ANG 1.58945503 1.59220702
1.61238313 -0.00547148 0.53891992
A / DEGREE 66.05724619 66.95062670
73.54907404 -0.00015855 -0.00010933
R3 / ANG 1.35229995 1.35212903
1.35084905 0.00036726 0.88529322
B / DEGREE 102.58308447 102.51347761
101.91805140 0.00001545 0.00006969
G / DEGREE 90.11384708 90.00488570
89.25817619 0.00001837 0.00002126
Convergence: 0.00000000 (line search)
0.14257717 0.01130670 (total)
Numerically approximating hessian using central energy differences
Task list generated. Total number of displacements: 42
(...)
Numerical RCCSD(T) hessian completed. CPU-time: 50556.55 sec, Elapsed:
52471.89 sec
RCCSD(T) hessian saved to record 5301.2
Hessian eigenvalues: -0.132078 0.038530 0.146748 0.203447 0.251917
0.334504
Performing a regular QSD step. Stepsize = 0.108723 Curvature =
.35494D+01
Optimization point 3
Variable Last Current Next
Gradient Hessian
E(RCCSD(T)) / Hartree -548.40592241 -548.40943457 0.00000000
R1 / ANG 1.67535167 1.63672546
1.58902162 0.04235023 0.24875090
R2 / ANG 1.59220702 1.61238313
1.61361588 0.00201373 0.84044207
A / DEGREE 66.95062670 73.54907404
78.25132174 -0.00033263 0.00000493
R3 / ANG 1.35212903 1.35084905
1.35035885 0.00052066 0.88601265
B / DEGREE 102.51347761 101.91805140
100.67110349 0.00011472 0.00007094
G / DEGREE 90.00488570 89.25817619
88.29840320 0.00006811 0.00001246
Convergence: 0.00000000 (line search)
0.12498993 0.03041555 (total)
? Error
? No convergence
? The problem occurs in rhfpro
****
I will appreciate very much any help.
Best regards,
Gabriel
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.molpro.net/pipermail/molpro-user/attachments/20101117/d3199105/attachment.html>
More information about the Molpro-user
mailing list