[CPMD-list] Scratch array too small in CPMD 3.7

[Veronika Brazdova]

Dear all,

I ran a calculation (input file attached) which used to work fine with 
CPMD 3.5 & 3.4, but with CPMD 3.7 it crashed at this point (the platform 
was a Cray T3E):

 ***    RWFOPT|   CURRENT HEAP LENGTH             3989 KWords ***
LSCR=   178337 2*NHGB+IBRESET*(2*NHGB+2*IBRES=*********


 PROGRAM STOPS IN SUBROUTINE UPDRHO| SCRATCH ARRAY TOO SMALL [PROC=   0]
LSCR=   178479 2*NHGB+IBRESET*(2*NHGB+2*IBRES=*********


 PROGRAM STOPS IN SUBROUTINE UPDRHO| SCRATCH ARRAY TOO SMALL [PROC=   8]
LSCR=   178487 2*NHGB+IBRESET*(2*NHGB+2*IBRES=*********


 PROGRAM STOPS IN SUBROUTINE UPDRHO| SCRATCH ARRAY TOO SMALL [PROC=  10]
LSCR=   178467 2*NHGB+IBRESET*(2*NHGB+2*IBRES=*********


 PROGRAM STOPS IN SUBROUTINE UPDRHO| SCRATCH ARRAY TOO SMALL [PROC=   1]
LSCR=   178467 2*NHGB+IBRESET*(2*NHGB+2*IBRES=*********


 PROGRAM STOPS IN SUBROUTINE UPDRHO| SCRATCH ARRAY TOO SMALL [PROC=   5]
LSCR=   178491 2*NHGB+IBRESET*(2*NHGB+2*IBRES=*********

(plus more lines of the same message)

I found that it is the keyword 'LANCZOS DIAGONALIZATION' 
(or 'LANCZOS PARAMETERS') which causes the trouble. The problem
has also occured with other systems when I tried to use the
Lanczos diagonalisation.

Here is the input file: 
&CPMD
    OPTIMIZE WAVEFUNCTION
    BIG MEMORY
    HESSIAN UNIT
    FREE ENERGY FUNCTIONAL
    TEMPERATURE ELECTRON
      300.
    LANCZOS PARAMETERS
      1   6   20  1.D-18
    BROYDEN MIXING
      0.15  200 0.01 0 8
    CONVERGENCE
     1.D-4 5.D-4
&END

&SYSTEM
   SCALE
   ANGSTROM
   SYMMETRY
    8
   CELL ABSOLUTE
    11.618000 3.586000 4.443000  0.0 0.0 0.0
   CUTOFF
    120.
   TESR
    8
   KPOINTS MONKHORST-PACK MESH
    1 3 3
   POINT GROUP DELTA=3.D-5
    AUTO
   STATES
    80
&END

&DFT
   NEWCODE
   FUNCTIONAL GGA
   NO EXCHANGE CORRELATION TABLE
&END

&ATOMS
*MT/O_MT_GGA_115 KLEINMAN-BYLANDER RAGGIO=1.2
  LMAX=P
  10
   0.85390   -0.00000    0.46336
   0.14610   -0.00000    0.46336
   0.64611    0.50000    0.53664
   0.35389    0.50000    0.53664
   0.68173    0.00000   -0.00738
   0.31827    0.00000   -0.00738
   0.81827    0.50000    0.00738
   0.18173    0.50000    0.00738
   0.00000   -0.00000    0.00360
   0.50000    0.50000   -0.00360
*MT/V_MT_GGA_semi KLEINMAN-BYLANDER RAGGIO=1.2
  LMAX=D LOC=S
  4
   0.85140   -0.00000    0.10664
   0.14860   -0.00000    0.10664
   0.64860    0.50000    0.89334
   0.35140    0.50000    0.89334
&END

Any fixes for this problem? 

Thanks for help.

	Veronika Brazdova

-- 
_______________________________________________________________________________

                             Veronika Brazdova
                Humboldt University, Institute of Chemistry, 
                   Brook-Taylor-Strasse 2, 12489, Berlin
                          vb at chemie.hu-berlin.de 
                 http://atrey.karlin.mff.cuni.cz/~verunka
               phone: +49 30 2093 7140 fax: +49 03 2093 7136
_______________________________________________________________________________