ccsd(t) C1 structre geometry optimization error

[Elena Jakubikova]

Olga,

2GB file limitation is not a problem for SGI workstations. It's an issue
only for linux. I believe you simply didn't have enough free space on your
disk to run the calculation. Just to make sure I ran the input deck on
our SGI workstation using 2000.1 version of molpro and up to the
ccsdt step everything went fine. As for the error message you got, one
cannot use closed shell ccsd(t) method on an open shell system (61
electrons). In this case the use of spin unrestricted (uccsdt) or
partially spin restricted (rccsdt) open shell cluster theories is
necessary. 

Regards,

Elena

On Thu, 30 Jan 2003, Olga Dmitrenko wrote:

> Elena,
> 
> Both servers are SGIs with IRIX. For sure, it looks like it's a  disk limit
> problem. But why I've been able to go through first step (rhf) using old
> version of molpro? I've got new error msg at the step of ccsd(t):
> 
> 
> 1PROGRAM * CCSD (Closed-shell coupled cluster)     Authors: C. Hampel, H.-J.
> Wer
> ner, 1991, M. Deegan, P.J. Knowles, 1992
> 
> 
>  NUMBER OF OCCUPIED ORBITALS NOT CONSISTENT WITH NUMBER OF ELECTRONS FOR
> CCSD:
>   61     8
> 
>  ERROR EXIT
> 
> 
>  ***********************************************************************
> ***************************************************
>  DATASETS  *  I ENTRIES      LENGTH      RECORD NAMES
>               1      17   151710133     500      600      700      900
> 950
>     1000     1050     1100     1400     1410
>                                         VAR    BASINP    GEOM    SYMINP
> ZMAT
>     DATA     BASIS      S        T        V
>                                        1200     1210     1080     1600
> 1650
>     1300     1700
>                                         H0       H01     AOSYM     SMH
> MOLCAS
>     ERIS     OPER
> 
>               2       4      219328     500     1000     1050     2100
>                                         VAR     DATA     BASIS     RHF
> 
>  PROGRAMS   *      CCSD(T)   RHF-SCF       INT
>  CPU TIMES  *         0.00   1429.08    554.52
>  REAL TIME  *    76 MIN,14.87 SEC     CPU TIME *    33 MIN, 3.71 SEC     I/O
> TIM
> E *    43 MIN,33.90 SEC
>  ***************************************************************************
> ****
> ***************************************************
> 
> ----------
> >From: Elena Jakubikova <elena at lamar.ColoState.EDU>
> >To: Olga Dmitrenko <odmitr at UDel.Edu>
> >Cc: Aleksey Kuznetsov <kuznetsov at cc.usu.edu>, molpro-user at molpro.net
> >Subject: Re: ccsd(t) C1 structre geometry optimization error
> >Date: Thu, Jan 30, 2003, 4:56 PM
> >
> 
> > Olga,
> >
> > is it possible that the first server you used to run Aleksey's job runs
> > Red Hat Linux and the other one is using some different operating
> > system? There is an issue with the support for large files (2GB and
> > larger) under RH Linux which seems to be Aleksey's problem. I think the
> > temporary file that stores the integrals just reached 2GB limit in his
> > calculation - that's why the error message.
> >
> > Elena
> >
> > On Thu, 30 Jan 2003, Olga Dmitrenko wrote:
> >
> >> Dear Aleksey,
> >>
> >> I wrote you that when I used molpro version 2000.1, I've got the
> >> same error msg for your job.
> >>
> >> It's not problem of disk.
> >>
> >> Then I went to another server which still has version 98.1, your job is
> >> running. At the momen t it does RHF-SCF (OPEN SHELL) iterations.
> >>
> >> Best regards,
> >> Olga Dmitrenko
> >>
> >> *******************************************
> >> * Dr. Olga Dmitrenko
> >> * Department of Chemistry and Biochemistry
> >> * University of Delaware, Newark
> >> * Delaware 19716, USA
> >> *******************************************
> >>
> >> ----------
> >> >From: Aleksey Kuznetsov <kuznetsov at cc.usu.edu>
> >> >To: <molpro-user at molpro.net>
> >> >Subject: ccsd(t) C1 structre geometry optimization error
> >> >Date: Wed, Jan 29, 2003, 10:43 PM
> >> >
> >>
> >> > Dear Molpro users,
> >> >
> >> > I'm trying to perform ccsd(t) geometry optimization and frequency
> >> > calculations for li3al4- structure with C1 symmetry. Input file looks like
> >> > this:
> >> >
> >> >  ***,li3al4-Cs_butterfly,
> >> >  memory,80,m
> >> >  gprint,basis,orbital
> >> >  geomtyp=xyz
> >> >  geometry={nosym;
> >> >  7
> >> >  This Li3Al4-butterfly structure
> >> >  Al,0.3596668819,-0.427279092,1.2918672482
> >> >  Al,0.3855193239,-0.5221602827,-1.3691583215
> >> >  Al,-1.3326860078,-1.9691921155,0.0089143085
> >> >  Al,0.7524128566,1.850531256,-0.0455627355
> >> >  Li,-1.8588311025,0.9288599899,0.633823052
> >> >  Li,0.3117413463,1.9448898006,2.543549272
> >> >  Li,0.8324665195,1.7546845577,-2.6836344894
> >> >  }
> >> >  basis=6-311+G*
> >> >  rhf;
> >> >  ccsd(t);
> >> >  optg;
> >> >  freq,
> >> >  print,low,imag
> >> >  put,molden,li3al4-butterfly_ccsdt_optGF.molden
> >> >
> >> >
> >> > However, the jobs terminated and quit with this error message:
> >> >
> >> >
> >> >  Eigenvalues of metric
> >> >
> >> >          1 0.574E-04 0.769E-04 0.159E-03 0.171E-03 0.242E-03 0.339E-03
> >> > 0.409E-03 0.462E-03
> >> >
> >> > wrabsf: Error in writing to file T1100023226.TMP (unit 11), 8192 words at
> >> > word offset 268431360
> >> >
> >> >  ERROR EXIT
> >> >
> >> >
> >> >
> ***************************************************************************
> >> > *******************************************************
> >> > wrabsf: Error in writing to file T1100023226.TMP (unit 11), 8192 words at
> >> > word offset 268431360
> >> >
> >> >  ERROR EXIT
> >> >
> >> >  RECURSIVE CALL OF ERROR HANDLER, FILE INFORMATION MAY BE LOST
> >> >
> >> >  EMERGENCY STOP
> >> >
> >> >
> >> > Could anybody please help me with this calculation? I cannot understand
> >> > where error could be hidden.
> >> >
> >> > Thank you very much in advance.
> >> >
> >> > Sincerely,
> >> > Aleksey Kuznetsov.
> >> 
>