Molcas Forum

Support and discussions for Molcas and OpenMolcas users and developers

You are not logged in.

Announcement

Welcome to the Molcas forum. You can choose an avatar and change the default style by going to "Profile" → "Personality" or "Display".

#1 2018-08-24 10:31:28

Sarah
Member
Registered: 2016-08-01
Posts: 40

[SOLVED] MEP-restart

Hello,

I have been running an MEP calculation for quite some time with the following input:

> export MOLCAS_MEM = 10000
&gateway
 XBAS = ANO-L-VTZP
 ZMAT
...
End  of  z-matrix
>>> EXPORT MOLCAS_MAXITER=800
>>> Do while
&SEWARD
>>> If ( Iter = 1 ) <<<
&SCF
>>> EndIf <<<
&RASSCF
  Title
  charge   = 0
  spin     = 1
  nactel   = 8 0 0
  inactive = 16
  ras2     = 11
  ciroot   = 6 6 1
  MDRlxRoot = 6
  &RASSI
    HOP
&SLAPAF
  MEP-search
  MEPStep=0.1
>>> EndDo
&GRID_IT
 All

After several successful iterations, I got the following error, is there a way to restart the calculations?

 ###############################################################################
 ###############################################################################
 ###                                                                         ###
 ###                                                                         ###
 ###    Location: rdVec_                                                     ###
 ###    File: TMPORB                                                         ###
 ###                                                                         ###
 ###                                                                         ###
 ###    Error during reading INPORB                                          ###
 ###      *******  2.1507  0.1138  0.0172 *******  1.5587                    ###
 ###    0.0056 -0.0000 -0.0000 -0.0000                                       ###
 ###                                                                         ###
 ###                                                                         ###
 ###############################################################################
 ###############################################################################
--- Stop Module:  mclr at Tue Aug 21 00:59:43 2018 /rc= _INTERNAL_ERROR_ ---
--- Module mclr spent 1 hour and 11 minutes
Non-zero return code - check program input/output
--- Module auto spent 1 hour and 11 minutes
Non-zero return code - check program input/output
--- Module auto spent 1 hour and 11 minutes
Non-zero return code - check program input/output
--- Stop Module:  auto at Tue Aug 21 00:59:44 2018 /rc= _INTERNAL_ERROR_ ---
--- Module auto spent 1 hour and 37 minutes
--- Stop Module:  auto at Tue Aug 21 00:59:44 2018 /rc= _INTERNAL_ERROR_ ---
--- Module auto spent 1 hour and 37 minutes

Offline

#2 2018-08-24 10:40:49

Ignacio
Administrator
From: Uppsala
Registered: 2015-11-03
Posts: 1,203

Re: [SOLVED] MEP-restart

If you have the old RunFile in the WorkDir, and do not run GATEWAY again, the calculation should continue at the last iteration.

But, in the worst case, a MEP optimization is a set of mostly independent optimizations. You can restart from the last optimized point in the path (check the $Project.mep.molden file, or the convergence data in the log file), with possibly a change in MDRlxRoot.

Offline

#3 2018-08-24 15:19:47

Sarah
Member
Registered: 2016-08-01
Posts: 40

Re: [SOLVED] MEP-restart

Thanks, Ignacio,

The worst case scenario applies here, so I did previously restart the MEP using the last cartesian coordinate printed on *.mep.molden file starting from the last root according to the output. However, the resulting MEP looked quite different, i.e, I had a sudden change in the geometry from the restart point. I realized that the last converged geometry in the output looks a little bit different than *.mep.molden last geometry. Now I am running a new MEP from the outputs' last geometry. let's see ...

Offline

#4 2018-08-27 17:16:37

Sarah
Member
Registered: 2016-08-01
Posts: 40

Re: [SOLVED] MEP-restart

Just to complete the discussion, the converged coordinates taken directly from output file gave more reasonable MEP, i.e, no sudden change in the geometry. The restarted MEP converged to a minimum after 2-3 runs.

Offline

Board footer

Powered by FluxBB 1.5.11

Last refresh: Today 18:11:14