Molcas Forum
Support and discussions for Molcas and OpenMolcas users and developers
You are not logged in.
- Topics: Active | Unanswered
Announcement
Welcome to the Molcas forum.
You can choose an avatar and change the default style by going to "Profile" → "Personality" or "Display".
#1 2024-11-12 21:17:40
- mjr0463
- Member
- Registered: 2024-08-09
- Posts: 5
Numerical gradient failing after RASSCF/CASPT2
Hello,
I have ran the following input file:
&GATEWAY
coord= /something.xyz
basis=ANO-RCC-VQZP
group=nosym
&SEWARD
&SCF
iterations=40
&RASSCF
spin = 5
charge=0
nactel=4
inactive= 36
RAS2= 4
symmetry=1
ALTER
1
1 36 38
ITER= 80,25
CIMX= 25
&CASPT2
prop
maxiter=30
&MCKINLEY
perturbation
hessianI ran everything successfully up to CASPT2 in a separate file. When I add McKinley, it fails:
()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()
&MCKINLEY
only a single process is used, running in SERIAL mode
available to each process: 2.0 GB of memory, 1 thread?
pid: 459626
()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()
--- Stop Module: mckinley at Mon Nov 11 12:36:00 2024 /rc=_RC_INVOKED_OTHER_MODULE_ ---
*** files: xmldump
saved to directory /tmp/
*** symbolic link created: INPORB -> something.RasOrb
--- Start Module: slapaf at Mon Nov 11 12:36:00 2024 ---
()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()
&SLAPAF
only a single process is used, running in SERIAL mode
available to each process: 2.0 GB of memory, 1 thread?
pid: 459629
()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()
++ Slapaf input parameters:
------------------------
Maximum number of iterations: 0
Convergence test a la Schlegel.
Convergence criterion on gradient/para.<=: 3.0E-04
Convergence criterion on step/parameter<=: 3.0E-04
Convergence criterion on energy change <=: 0.0E+00
Parameters for step-restricted optimization
Maximum step length (initial seed): 3.00E-01
-RFO activated with parameters:
Maximum number of data points used in RFO: 5
Line search is performed
-Optimization for transition state.
Optimization method: RS-P-RFO
No mode to follow is specified!
Optimization will follow mode with the lowest eigenvalue.
-Initial Hessian guess is estimated with finite differences.
Two point symmetric formula, Delta= 1.00E-02
-Hessian update method: Broyden-Fletcher-Goldfarb-Shanno
Maximum number of points in Hessian update: 5
-Relaxation will be done on non-redundant internal coordinates, based on
force constant weighted redundant internal coordinates.
--
******************************************
* Statistics of the internal coordinates *
******************************************
Translations and Rotations: 0
Bonds : 3
Angles : 3
Torsions : 0
Out-of-plane angles : 3
Accumulate the gradient for selected numerical differentiation.
1 of 13
--- Stop Module: slapaf at Mon Nov 11 12:36:00 2024 /rc=_RC_CONTINUE_LOOP_ ---
..
--- Start Module: seward at Mon Nov 11 12:36:00 2024 ---
..
--- Module seward spent 4 minutes 1 second ---
..
--- Start Module: rasscf at Mon Nov 11 12:40:02 2024 ---
..
--- Stop Module: rasscf at Mon Nov 11 12:45:15 2024 /rc=_RC_ALL_IS_WELL_ ---
..
--- Start Module: caspt2 at Mon Nov 11 12:45:15 2024 ---
..
--- Stop Module: caspt2 at Mon Nov 11 12:46:05 2024 /rc=_RC_ALL_IS_WELL_ ---
..
--- Start Module: slapaf at Mon Nov 11 12:46:05 2024 ---
--- Stop Module: slapaf at Mon Nov 11 12:46:05 2024 /rc=_RC_INVOKED_OTHER_MODULE_ ---
..
--- Start Module: alaska at Mon Nov 11 12:46:05 2024 ---
--- Stop Module: alaska at Mon Nov 11 12:46:05 2024 /rc=_RC_INVOKED_OTHER_MODULE_ ---
..
Molecular orbitals exchanged:
In symmetry 1 : 36 38
Variance of |WF0>: 6.9959730497
Molecular orbitals exchanged:
In symmetry 1 : 36 38
Variance of |WF0>: 6.9959730499
Molecular orbitals exchanged:
In symmetry 1 : 36 38
Variance of |WF0>: 7.2211654871
Molecular orbitals exchanged:
In symmetry 1 : 36 38
Variance of |WF0>: 7.2211654871
Molecular orbitals exchanged:
In symmetry 1 : 36 38
Variance of |WF0>: 7.2213593298
Molecular orbitals exchanged:
In symmetry 1 : 36 38
Variance of |WF0>: 7.2209713836
Molecular orbitals exchanged:
In symmetry 1 : 36 38
Variance of |WF0>: 7.2175085807
Molecular orbitals exchanged:
In symmetry 1 : 36 38
Variance of |WF0>: 7.2248551760
Molecular orbitals exchanged:
In symmetry 1 : 36 38
Variance of |WF0>: 7.2268011683
Molecular orbitals exchanged:
In symmetry 1 : 36 38
Variance of |WF0>: 7.2268011728
Molecular orbitals exchanged:
In symmetry 1 : 36 38
Variance of |WF0>: 7.2209810660
Molecular orbitals exchanged:
In symmetry 1 : 36 38
Variance of |WF0>: 7.2214081097
--- Stop Module: numerical_gradient at Mon Nov 11 15:02:59 2024 /rc=_RC_ALL_IS_WELL_ ---
*** files: xmldump
saved to directory /tmp/
--- Module numerical_gradient spent 2 hours 16 minutes 53 seconds ---
*** files: xmldump
saved to directory /tmp/
--- Module numerical_gradient spent 2 hours 16 minutes 53 seconds ---
>>> EXPORT MOLCAS_TRAP = ON
*** symbolic link created: INPORB -> something.RasOrb
--- Start Module: slapaf at Mon Nov 11 15:02:59 2024 ---
Accumulate the gradient for selected numerical differentiation.
2 of 13
--- Stop Module: slapaf at Mon Nov 11 15:02:59 2024 /rc=_RC_CONTINUE_LOOP_ ---
*** files: something.slapaf.h5 xmldump
saved to directory /tmp/
>>> END DO
>>> IF (ITER != 1)
*** symbolic link created: INPORB -> something.RasOrb
--- Start Module: seward at Mon Nov 11 15:02:59 2024 ---
--- Stop Module: seward at Mon Nov 11 15:07:03 2024 /rc=_RC_ALL_IS_WELL_ ---
*** files: xmldump
saved to directory /tmp/
--- Module seward spent 4 minutes 3 seconds ---
*** symbolic link created: INPORB -> something.RasOrb
--- Start Module: rasscf at Mon Nov 11 15:07:03 2024 ---
Molecular orbitals exchanged:
In symmetry 1 : 36 38
--- Stop Module: rasscf at Mon Nov 11 15:10:26 2024 /rc=_RC_ALL_IS_WELL_ ---
*** files: something.RasOrb something.RasOrb.1 something.rasscf.h5 something.SpdOrb.1 xmldump
saved to directory /tmp/
--- Module rasscf spent 3 minutes 22 seconds ---
*** symbolic link created: INPORB -> something.RasOrb
--- Start Module: caspt2 at Mon Nov 11 15:10:26 2024 ---
Variance of |WF0>: 7.2213012289
--- Stop Module: caspt2 at Mon Nov 11 15:11:18 2024 /rc=_RC_ALL_IS_WELL_ ---
*** files: something.Pt2Orb xmldump
saved to directory /tmp/
--- Module caspt2 spent 52 seconds ---
*** symbolic link created: INPORB -> something.RasOrb
--- Start Module: slapaf at Mon Nov 11 15:11:18 2024 ---
--- Stop Module: slapaf at Mon Nov 11 15:11:19 2024 /rc=_RC_INVOKED_OTHER_MODULE_ ---
*** files: something.slapaf.h5 xmldump
saved to directory /tmp/
*** symbolic link created: INPORB -> something.RasOrb
--- Start Module: alaska at Mon Nov 11 15:11:19 2024 ---
--- Stop Module: alaska at Mon Nov 11 15:11:19 2024 /rc=_RC_INVOKED_OTHER_MODULE_ ---
*** files: xmldump
saved to directory /tmp/
*** symbolic link created: INPORB -> something.RasOrb
--- Start Module: numerical_gradient at Mon Nov 11 15:11:19 2024 ---
Molecular orbitals exchanged:
In symmetry 1 : 36 38
Numerical_Gradient failed ...
RASSCF returned with return code, rc = 96
for the perturbation iDisp = 1
--- Stop Module: numerical_gradient at Mon Nov 11 15:41:41 2024 /rc=-6 ---
*** files: xmldump
saved to directory /tmp/
--- Module numerical_gradient spent 30 minutes 22 seconds ---
.########################.
.# Non-zero return code #.
.########################.
Timing: Wall=11827.08 User=10569.17 System=1206.65I appreciate any and all insight.
Offline
#2 2024-11-13 12:00:01
- Ignacio
- Administrator
- From: Uppsala
- Registered: 2015-11-03
- Posts: 1,203
Re: Numerical gradient failing after RASSCF/CASPT2
You probably need to have two RASSCF inputs: one to prepare the orbitals and a "clean" one, without ALTER:
&RASSCF
spin = 5
charge=0
nactel=4
inactive= 36
RAS2= 4
symmetry=1
ALTER
1
1 36 38
ITER= 80,25
CIMX= 25
&RASSCF
spin = 5
charge=0
nactel=4
inactive= 36
RAS2= 4
symmetry=1
ITER= 80,25
CIMX= 25Offline
Last refresh: Today 12:49:21