[molpro-user] READM: TRYING TO READ BEYOND RECORD BOUNDARY: for CASSCF-gradient on "many" point charges

[keyarash sadeghian]

Hi Everyone!

I am having a trouble doing a QM/MM (QM = CASSCF) calculation with 
MolPro. Problem arises at the stage where the QM-gradient for point 
charges are being evaluated. Here is the interesting part of my output:


----------------------------------------------------------------
Primary working directories    : /scr/sak/locscratch
  Secondary working directories  : /scr/sak/locscratch
  Wavefunction directory         : /home/sak/wfu/
  Main file repository           : /scr/sak/locscratch/

  ARCHNAME  : Linux/x86_64
  FC        : /opt/intel/compilerpro-12.0.2.137/bin/intel64/ifort
  FCVERSION : 12.0.2
  BLASLIB   :
  id        : cupuni

  Nodes     nprocs
  b15          6
  ga_uses_ma=false, calling ma_init with nominal heap.
  GA-space will be limited to   8.0 MW (determined by -G option)

  Using customized tuning parameters: mindgm=4; mindgv=4; mindgc=1; 
mindgr=1; noblas=0; mincuda=1000; minvec=7
  default implementation of scratch files=sf

  ***,verd_p1n_nc1x_attack_cas_fr1_blypsvpD_qmset2_3.4
  memory,200,m
  file,2,verd_p1n_nc1x_attack_cas_fr1_blypsvpD_qmset2_3.4,new
  gdirect
  print,civectors,orbitals
  orient,noorient
  bohr
  geomtype=xyz
  geometry={
  45
  QM coords
  N      -10.5329462465  2.4762487760  1.2122001001

.....

                                          ***  PROGRAM SYSTEM MOLPRO  ***
                          Copyright, University College Cardiff 
Consultants Limited, 2008

                                     Version 2010.1 linked 15 Sep 2011 
12:01:52


  **********************************************************************************************************************************
  LABEL *   verd_p1n_nc1x_attack_cas_fr1_blypsvpD_qmset2_3.4
  Linux-2.6.32.27-0.2-default/b15(x86_64) 64 bit mpp 
version                              DATE: 11-Dec-11          TIME: 
15:32:58
  **********************************************************************************************************************************

  SHA1:             c82fc9886d3ef081defdb91dc08d0bb97a8ad3e0
  **********************************************************************************************************************************

  Variable memory set to  200000000 words,  buffer space   230000 words

.........

<<< HF is finished ....>>>


.................


  SOLVING CPF EQUATIONS
  =====================

  Number of parameters: 15996
  Dimension of Hessian:     0

  ITER.    NORM       VARIANCE    ACCURACY   ALPHA(IT)

    0   0.2137D-01  0.2137D-01  0.0000D+00  0.1000D+01
    1   0.1686D-01  0.7895D-02  0.4682D+00  0.1686D-01
    2   0.2234D-01  0.2808D-02  0.1257D+00  0.8531D-02
    3   0.2340D-01  0.1315D-02  0.5617D-01  0.2584D-02
    4   0.2388D-01  0.4207D-03  0.1762D-01  0.1515D-02
    5   0.2399D-01  0.1899D-03  0.7916D-02  0.5097D-03
    6   0.2402D-01  0.6640D-04  0.2764D-02  0.2042D-03
    7   0.2403D-01  0.2550D-04  0.1061D-02  0.6666D-04
    8   0.2403D-01  0.8534D-05  0.3552D-03  0.2213D-04
    9   0.2403D-01  0.2876D-05  0.1197D-03  0.9347D-05
   10   0.2403D-01  0.1155D-05  0.4807D-04  0.3192D-05
   11   0.2403D-01  0.5122D-06  0.2132D-04  0.1327D-05
   12   0.2403D-01  0.2483D-06  0.1033D-04  0.6336D-06
   13   0.2403D-01  0.1230D-06  0.5121D-05  0.2851D-06
   14   0.2403D-01  0.4740D-07  0.1973D-05  0.1304D-06
   15   0.2403D-01  0.1702D-07  0.7084D-06  0.5399D-07

  CPF equations solved. CPU time:                141.19
  CPF finished. Total CPU time:                  155.51
  READM: TRYING TO READ BEYOND RECORD BOUNDARY:
  RECORD      300.4  EXTENSION= 1  OFFSET=          0.  ADDRESS=   
140505.  LEN=************  LENREST=************

  Records on file 4

  IREC   NAME  TYPE        OFFSET    LENGTH   IMPLEMENTATION   EXT   
PREV   PARENT  MPP_STATE
    1     800                  0.    70225.         sf          0      
0      0      0
    2     500              70225.       55.         sf          0      
0      0      0
    3     100              70280.    70225.         sf          0      
0      0      0
    4     300             140505.    35245.         sf          0      
0      0      0
    5     400             175750.    35245.         sf          0      
0      0      0
    6    8000             210995.    35245.         sf          0      
0      0      0
    7    8001             246240.    35245.         sf          0      
0      0      0
    8     600             281485.    35245.         sf          0      
0      0      0
    9     200             316730.    35245.         sf          0      
0      0      0
   10     700             351975.    35245.         sf          0      
0      0      0


  GLOBAL ERROR fehler on processor   0

----------------------------------------------------------------

There are some 50000 point charges for which the gradient should be 
evaluated. If I reduce the number of point charges down to say 10000, 
then everything is fine and NO problems occur. So I guess there must be 
some limit for the number of point charges which is HARD-CODED somewhere 
in the code. Are there any patches to avoid this problem (I am using 
Molpro2010, patch level 24).
best regards

Keyarash Sadeghian
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.molpro.net/pipermail/molpro-user/attachments/20111212/3ca24667/attachment.html>