[CPMD-list] Error when trying FIXRHO UPWFN for wfn optimization

[Eduardo J. Lamas]

Hi Dr. Kohlmeyer thanks for replying.

The executables are supposed to be all 64 bits.

This is again the section where CPMD stops, in this case I reduced the cut
off to 50 Ry and the kpoints
to 3 3 1.

 ***  CALC_ALM| THE NEW SIZE OF THE PROGRAM IS   73124/ 317456 kBYTES ***

 DEGREES OF FREEDOM FOR SYSTEM:                                36

 CPU TIME FOR WAVEFUNCTION INITIALIZATION:          63.57 SECONDS
 ***    RWFOPT| THE NEW SIZE OF THE PROGRAM IS  253828/ 317456 kBYTES ***
p4_23699:  p4_error: interrupt SIGSEGV: 11
  EWALD| SUM IN REAL SPACE OVER                     7* 7* 7 CELLS
DDRHO   1.000000E+00  TOL   5.000000E-05
p1_3765:  p4_error: interrupt SIGSEGV: 11
p0_3763:  p4_error: handle_connection_interrupt: read listener: -1
p6_23701:  p4_error: net_recv read:  probable EOF on socket: 1
p6_23701:  p4_error: net_recv read:  probable EOF on socket: 1
p3_3767:  p4_error: interrupt SIGSEGV: 11


This is a successful run without the FIXRHO UPWFN keyword it seems that is
allocating much more
memory without any problem (in both cases the jobs are running alone in the
allocated nodes, and 
the number of nodes and system size are the same in both cases)


 ****************************************************************

 ***  CALC_ALM| THE NEW SIZE OF THE PROGRAM IS  549560/1792032 kBYTES ***

 DEGREES OF FREEDOM FOR SYSTEM:                                36

 CPU TIME FOR WAVEFUNCTION INITIALIZATION:        1513.55 SECONDS
 ***    RWFOPT| THE NEW SIZE OF THE PROGRAM IS 1653952/1792032 kBYTES ***
  EWALD| SUM IN REAL SPACE OVER                     7* 7* 7 CELLS

Thanks,


Eduardo 



-----Original Message-----
From: Axel Kohlmeyer [mailto:axel.kohlmeyer at theochem.ruhr-uni-bochum.de] 
Sent: Wednesday, December 22, 2004 11:12 AM
To: cpmd-list at cpmd.org; Eduardo J. Lamas
Subject: Re: [CPMD-list] Error when trying FIXRHO UPWFN for wfn optimization


On Wed, 22 Dec 2004, Eduardo J. Lamas wrote:

dear eduardo,


EL> Dear List members I am trying to use FIXRHO UPWFN
EL> for wave function optimization. I think that there is a typo in the
manual
EL> where it says :
EL> FIXRHO UPDWF It should say: FIXRHO UPWFN

yep. found that one some time ago. it is already fixed in the soon to be
released version 3.9.2. thanks for the tip anyways.

EL> When trying to use this method I get the following error :
EL> 
EL>  ***  CALC_ALM| THE NEW SIZE OF THE PROGRAM IS  549460/1837104 
EL> kBYTES ***
EL> 
EL>  DEGREES OF FREEDOM FOR SYSTEM:                                36
EL> 
EL>  CPU TIME FOR WAVEFUNCTION INITIALIZATION:        1743.29 SECONDS
EL>  ***    RWFOPT| THE NEW SIZE OF THE PROGRAM IS 1653852/1837104 kBYTES
***
EL>   EWALD| SUM IN REAL SPACE OVER                     7* 7* 7 CELLS
EL> DDRHO   1.000000E+00  TOL   5.000000E-05
EL> p0_26498:  p4_error: interrupt SIGSEGV: 11
EL> p3_26502:  p4_error: interrupt SIGSEGV: 11
EL> p5_29019:  p4_error: net_recv read:  probable EOF on socket: 1
EL> p0_26498: (2234.738281) net_send: could not write to fd=4, errno = 
EL> 32


hmmm, is this a 32-bit or a 64-bit binary? 
it looks as if you are running out of memory (in the MPI library), 
which would be plausible in case of a 32-bit binary. 
does it work after reducing the plane wave cutoff (or using more nodes)?

EL> This error is not present if I use PCG MINIMIZE (I tried different
EL> timesteps)
EL> but PCG MINIMIZE fails to converge.

for the few times, that i have used the free energy functional 
code (successfully), it was always with lanczos diagonalization.

best regards,
	axel.


EL> Does anybody has any suggestion about how to solve this problem ?
EL> 
EL> Thanks in advance,
EL> 
EL> Eduardo
EL> 
EL> &INFO
EL>   Wavefunction optimization bulk platinum
EL> &END
EL> &CPMD
EL>     rESTART WAVEFUNCTIONS DENSITY OCCUPATION KPOINTS COORDINATES LATEST
EL>     oPTIMIZE GEOMETRY
EL>     OPTIMIZE WAVEFUNCTIONS
EL>     FIXRHO UPWFN
EL>     LSD
EL>     FREE ENERGY FUNCTIONAL
EL>     ELECTRON TEMPERATURE
EL>       1000.
EL>     STORE
EL>       5
EL>     STRUCTURE BONDS
EL>     PRINT LSCAL COORDINATES ON
EL>     rEAL SPACE WFN KEEP 1000
EL>     ELECTROSTATIC POTENTIAL
EL>     RHOOUT
EL> &END
EL> &DFT
EL>    FUNCTIONAL BLYP
EL> &END
EL> &SYSTEM
EL>    POINT GROUP
EL>     AUTO
EL>    SYMMETRY
EL>     14
EL>    CELL DEGREE
EL>      5.54846   1   2.2   90   90  120
EL>    CUTOFF
EL>      120.000
EL>    ANGSTROMS
EL>    TESR
EL>      3
EL>    KPOINTS MONKHORST-PACK
EL>      5  5   1
EL> &END
EL> &ATOMS
EL> *Pt_TM_BLYPspd5.psp GAUSS-HERMIT=35 NLCC
EL>  LMAX=D LOC=S
EL>  12
EL> .........
EL> .........
EL> *O_TM_BLYP4.psp GAUSS-HERMIT=35
EL>  LMAX=D LOC=S
EL>  1
EL> .........
EL> .........
EL> &END
EL> &BASIS
EL>  PSEUDO AO 2 OCUPPATION
EL> 0   2
EL> 1   9
EL> SKIP
EL> &END
EL> 
EL> 
EL> _______________________________________________
EL> CPMD-list mailing list
EL> CPMD-list at cpmd.org http://cpmd.org/mailman/listinfo/cpmd-list
EL> 
EL> 

-- 

=======================================================================
Dr. Axel Kohlmeyer   e-mail: axel.kohlmeyer at theochem.ruhr-uni-bochum.de
Lehrstuhl fuer Theoretische Chemie          Phone: ++49 (0)234/32-26673
Ruhr-Universitaet Bochum - NC 03/53         Fax:   ++49 (0)234/32-14045
D-44780 Bochum  http://www.theochem.ruhr-uni-bochum.de/~axel.kohlmeyer/
=======================================================================
If you make something idiot-proof, the universe creates a better idiot.