[molpro-user] CCS(T) calculation restart

[Peter Knowles]

It looks to me as if this job has died during the calculation of the perturbative triples correction. Without knowing the number of electrons and basis functions, it is difficult to be sure, but looking at the time taken for CCSD iterations, this looks like a very big job for which the time needed to do (T) may be very large indeed. Unfortunately there is no restart facility for the triples.

Perhaps this is a case that could be treated reasonably using local correlation methods?

Peter

On 21 May 2012, at 13:37, Vera Cathrine wrote:

> Dear All,
> 
> Sorry molpro users and developers! I have sent this message before but I didn't receive any plausible answer yet. 
> I would like to perform a UCCS(T) calculation using molpro. The job has been killed as following because of lack of wall time after a long time.
> 
>  Starting UCCSD calculation
> 
>  ITER.      SQ.NORM     CORR.ENERGY   TOTAL ENERGY   ENERGY CHANGE        DEN1      VAR(S)    VAR(P)  DIIS     TIME
>    1      1.73650269    -2.25301369 -1833.81547384    -2.25301369    -0.06306107  0.14D-01  0.13D-01  1  1 49728.85
>    2      1.80563606    -2.31734026 -1833.87980041    -0.06432657    -0.00582553  0.62D-03  0.20D-02  2  2 96666.08
>    3      1.83097906    -2.32771986 -1833.89018001    -0.01037960    -0.00094578  0.53D-03  0.19D-03  3  3146418.61
>    4      1.84369652    -2.33172029 -1833.89418044    -0.00400043    -0.00022539  0.12D-03  0.56D-04  4  4192003.98
>    5      1.85090109    -2.33320576 -1833.89566591    -0.00148547    -0.00006876  0.48D-04  0.14D-04  5  5238330.05
>    6      1.85475494    -2.33359052 -1833.89605067    -0.00038476    -0.00002376  0.17D-04  0.52D-05  6  6287357.85
>    7      1.85739583    -2.33382197 -1833.89628212    -0.00023145    -0.00000762  0.51D-05  0.19D-05  6  1335993.18
>    8      1.85856688    -2.33394486 -1833.89640502    -0.00012290    -0.00000282  0.20D-05  0.66D-06  6  2387682.78
>    9      1.85905003    -2.33392064 -1833.89638079     0.00002423    -0.00000090  0.60D-06  0.21D-06  6  3434022.59
>   10      1.85943124    -2.33397126 -1833.89643141    -0.00005062    -0.00000029  0.20D-06  0.64D-07  6  4483288.55
>   11      1.85944210    -2.33396099 -1833.89642115     0.00001027    -0.00000010  0.70D-07  0.24D-07  6  5532495.01
>   12      1.85955574    -2.33397179 -1833.89643194    -0.00001080    -0.00000003  0.24D-07  0.81D-08  6  6581230.20
>   13      1.85955326    -2.33397219 -1833.89643234    -0.00000039    -0.00000001  0.83D-08  0.32D-08  6  1629169.58
>   14      1.85957562    -2.33397253 -1833.89643268    -0.00000034     0.00000000  0.30D-08  0.11D-08  6  2677344.94
> 
>  Norm of t1 vector:      0.22309390      S-energy:    -0.00234446      T1 diagnostic:  0.01616919
>                                                                        D1 diagnostic:  0.06950833
>  Norm of t2 vector:      0.89989151      P-energy:    -2.33162806
>                                          Alpha-Beta:  -1.73604116
>                                          Alpha-Alpha: -0.33690408
>                                          Beta-Beta:   -0.25868281
> 
> 
> the point here is that actually I don't know what are the files and directive which I have to define to make sure I could restart my job efficiently.
> or how can I restart this job to save the time for long jobs? I have enclosed my input file below.
> 
> Thank you so much for your kind help.
> Vera
> 
> 
> ***, ab;
> memory,500,m;                         
> file,1,ab.int
> file,2,ab.wfn
> punch,ab.pun
> RESTART;
> gprint,basis,orbitals;          
> NOSYM
> NOORIENT
> set, charge=2,spin=5
> geometry={
> 28 ! number of atoms 
> GeomXYZ
> XXXXXXXXX
> }
> proc setbasis
>    basis={
>    default,6-31g(d,p)
>      }
> endproc
> 
> setbasis
> {hf; orbital,2100.2
> shift, -1.0, -0.5;
> 
> }
> 
> cartesian
> setbasis
> int
> {hf; start,2100.2
> shift, -1.0, -0.5;
> 
> }
> put,molden,ab.molden;
> uccsd(t)
> ---
> _______________________________________________
> Molpro-user mailing list
> Molpro-user at molpro.net
> http://www.molpro.net/mailman/listinfo/molpro-user

--
Prof. Peter J. Knowles             
School of Chemistry, Cardiff University, Main Building, Park Place, Cardiff CF10 3AT, UK
Email KnowlesPJ at Cardiff.ac.uk 
WWW http://www.cardiff.ac.uk/chemy/contactsandpeople/academicstaff/knowles.html
Secretary: Alison Thompson, ThompsonAF at Cardiff.ac.uk Telephone +44 29 208 74805