[molpro-user] CCSD(T) job crash

[Peter Knowles]

The job has crashed on receipt of a SIGINT signal. The first thing I  
would think of is that this signal came from outside the program - for  
example from running out of time in a batch environment - but it may  
be something else.

Peter

On 12 May 2009, at 11:07, Andrew Mukhanov wrote:

> Dear members of Molpro community,
>
> I have a problem with the large (5Z) basis set CCSD(T) job for a low- 
> symmetry (C_s) ScF3 molecular geometry. There are no problems with  
> the TZ and QZ basis set jobs for the same geometry .
>
> The input is:
> ******************************************************
>  ***,ScF3 CCSD(T), F aug-cc-pV5Z-DK Sc cc-pwCV5Z-DK
>  memory,800,M
>  GTHRESH,ENERGY=1.0D-10
>  GTHRESH,ONEINT=1.0e-18
>  GTHRESH,TWOINT=1.0e-18
>  GTHRESH,prefac=1.0d-15
>  GTHRESH,orbital=1.0d-10
>  dkroll=1
>  geomtyp=xyz
>  geometry={
>  4
>  Sc     0.0000000000000     0.0000000000000     0.0170823080240
>  F      1.8420121778718     0.0000000000000    -0.0134739529400
>  F     -0.9210060889359    -1.5952293401170    -0.0134739529400
>  F     -0.9210060889359     1.5952293401170    -0.0134739529400
>  }
>  basis={
>  spdfghi,SC,WCV5Z-DK;c;
>  spdfgh,F,V5Z-DK;c;
>  s,F,0.0806000;
>  p,F,0.0550000;
>  d,F,0.1720000;
>  f,F,0.3310000;
>  g,F,0.6630000;
>  h,F,1.3260000;
>  }
>  {rhf,MAXIT=250;
>  WF,48,1,0
>  }
>  {ccsd(t),MAXIT=150;
>  DIIS,1,1,5,2
>  CORE,6,2
>  }
> ****************************************************************
> The job crashes in the CCSD routine:
>
> ***********************************************************
> 1PROGRAM * CCSD (Closed-shell coupled cluster)     Authors: C.  
> Hampel, H.-J. Werner, 1991, M. Deegan, P.J. Knowles, 1992
>
>  Convergence thresholds:  THRVAR = 1.00D-14  THRDEN = 1.00D-10
>  CCSD(T)     terms to be evaluated (factor= 1.000)
>
>  Number of core orbitals:           8 (   6   2 )
>  Number of closed-shell orbitals:  16 (  10   6 )
>  Number of external orbitals:     568 ( 315 253 )
>
>  Molecular orbitals read from record     2100.2  Type=RHF/CANONICAL  
> (state 1.1)
>  Number of N-1 electron functions:              16
>  Number of N-2 electron functions:             136
>  Number of singly external CSFs:              4668
>  Number of doubly external CSFs:          20667856
>  Total number of CSFs:                    20672525
>
>  Length of J-op  integral file:               0.00 MB
>  Length of K-op  integral file:              12.20 MB
>  Length of 3-ext integral record:             0.00 MB
>
>  For full I/O caching in triples, increase memory by944233811 words  
> to1744.5 Mword
> ***************************************************************
> with the following message:
> ***************************************************************
> /usr/local/bin/molprop_2008_1_Linux_x86_64_i8.exe, len=49
>    -master, len=7
>         s2, len=2
>      43342, len=5
>          2, len=1
>          2, len=1
>          0, len=1
>          0, len=1
> /usr/local/bin/molprop_2008_1_Linux_x86_64_i8.exe, len=49
>    -master, len=7
>         s2, len=2
>      53998, len=5
>          2, len=1
>          2, len=1
>          1, len=1
>          1, len=1
> tmp = /home/sol/pdir//usr/local/bin/ 
> molprop_2008_1_Linux_x86_64_i8.exe.p
>  Creating: host=s2, user=sol,
>            file=/usr/local/bin/molprop_2008_1_Linux_x86_64_i8.exe,  
> port=43342
>  Creating: host=s2, user=sol,
>            file=/usr/local/bin/molprop_2008_1_Linux_x86_64_i8.exe,  
> port=53998
>   2: interrupt(1)
> 1:SigIntHandler: interrupt signal was caught: 2
> 1:SigIntHandler: interrupt signal was caught: 2
> Last System Error Message from Task 1:: Bad file descriptor
>   1: ARMCI aborting 2 (0x2).
>   1: ARMCI aborting 2 (0x2).
> system error message: Bad file descriptor
> WaitAll: Child (21964) finished, status=0x100 (exited with code 1).
> WaitAll: No children or error in wait?
> *********************************************************************
>
> I would highly appreciate your advice.
> Thank you
>
> Andrew Mukhanov
> PhD Student
> Laboratory of Quantum Chemistry
> Ivanovo State University of Chemistry and Technology
> Russia
>
> _______________________________________________
> 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
Telephone: +44 29208 79182 Fax: +44 2920874030 Email KnowlesPJ at Cardiff.ac.uk
WWW http://www.cardiff.ac.uk/chemy/contactsandpeople/academicstaff/knowles.html