[molpro-user] Restart CASSCF Calculation

[Dr. Seth Olsen]

Hi Angelo,

I recommend that you use MATROP to output the orbitals to an ascii file, 
then read in that file for the new geometry. In the context of your 
example...

For the first job, append the following MATROP deck after the CASSCF 
(assuming the CASSCF orbitals are saved in record 2141.2)

{matrop
load,oldorbitals,orb,2141.2
write,oldorbitals,orbitalfile.orb}

This will write the orbitals as a matrix of (human-readable) numbers to 
a file called 'orbitalfile.orb'.  Note that 'oldorbitals' is the name 
that you have assigned to the matrix in MATROP - another name is just as 
good. 

For the second job, place the following MATROP deck in before your 
CASSCF deck:

{matrop
read,oldorbitals,file=oldorbitals.orb
save,orbitals,2140.2,orbitals}

This reads in the ascii file and saves it as an orbital record on 
2140.2.  No other changes to your input are necessary.

I recommend, as a molpro user of about 10 years, that you always save 
your orbitals to ascii files.  In the end, they are easier to use and 
manipulate than the binary file and (even better) it is not 
system-dependent so if you need to regenerate your data on another 
machine later you won't have a problem.  It is a generally more 
transparent technique, so it is easier to tell sometimes if there is a 
problem, whereas with the binary file you need to rely on Molpro's 
report of the error - which as you notice may not be transparent.

Cheers,

Seth




rossi at york.cuny.edu wrote:

>Hello:
>
>Gosh, I am frustrated.
>
>Can someone point me to an example input (or give me pointers) on how to
>perform a CASSCF calculation, save the orbitals in one job to a permanent
>file and,then in a separate job, to use those orbitals as the starting
>guess for a new CASSCF calculation with the same basis set but a NEW
>geometry.
>
>1) The first CASSCF job works perfectly.
>
>2) The second job (input shown below) crashes.
>
>***,New Molecule - Excited state
>memory,128,m
>
>print, basis,orbitals
>
>FILE,1,./int/second.int,new
>FILE,2,./wfu/second.wfu,new
>FILE,3,./wfu/first.aux,old
>
>BASIS=6-31G*
>
>geomtyp=xyz
>geometry={
>...
>...
>...
>}
>
>{casscf
>start,2140.3
> closed,5,5
>    occ,11,10
>     wf,30,1,0
> state,2; weight,0,1;
> orbprint,20
>}
>optg
>---
>
>And this is the error I get
>
>IREC   NAME  TYPE        OFFSET    LENGTH   IMPLEMENTATION   EXT   PREV  
>PARENT  MPP_STATE
>   1     500  VAR          4096.    54801.         df          0      0     
>0      1
>   2     700  GEOM        58897.    31749.         df          0      0     
>0      1
>   3     610  BASINP      90646.     8192.         df          0      0     
>0      1
>   4    1000  BASIS       98838.     1879.         df          0      0     
>0      1
>   5    2140  MCSCF      100717.    13056.         df          7      0     
>0      1
>   6    2141  MCSCF      113773.    16128.         df          0      0     
>0      1
>   7    2140  MCSCF      129901.     3072.         df          0      5     
>5      1
>
> ? Error
> ? I/O error
> ? The problem occurs in readw
>Read error in iow_direct_read; fd=11, l=16128, p=113773; read returns 0
>
> GA ERROR fehler on processor   0
>0:0:fehler on processor   0:: -1079207560
>  0: ARMCI aborting -1079207560 (0xbfac9978).
>1:SigIntHandler: interrupt signal was caught: 2
>
>Can someone please help me?
>
>Regards,
>
>Angelo
>  
>


-- 
ccmsccmsccmsccmsccmsccmsccmsccmsccmsccmsccmsccmsccmsccmsccmsccms

Dr Seth Olsen, PhD
Postdoctoral Fellow, Biomolecular Modeling Group
Centre for Computational Molecular Science
Australian Institute  For Bioengineering and  Nanotechnology  (75)
The University of Queensland
Brisbane Qld 4072 Australia 

Tel: +61 7 3346 3946
Fax: +61 7 3346  3992
email: s.olsen1 at uq.edu.au
Web: www.ccms.uq.edu.au 

ccmsccmsccmsccmsccmsccmsccmsccmsccmsccmsccmsccmsccmsccmsccmsccms

Unless stated otherwise, this e-mail represents only the views of the Sender and not the views of The University of Queensland