The ABINIT Users Mailing List ( CLOSED )

[BOTTIN Francois <francois.bottin@cea.fr>]

Dear Paul,

I feel it's coming from iprcel 45.
Your job seems to crash within the susceptibility matrix calculation.
If this part is allowed using paral_kgb=1, at this time it is very time 
consuming and ressource demanding.
So it is only allowed for testing purpose.

If not, please sent your whole ouptput and log files.

Regards,
Francois

Paul Fons a écrit :
>   I have been attempting to study the use of parallization keywords 
> for abinit and have run into a situation where abinip crashes with a 
> segmentation fault (brief snip of the log below).  I have also 
> included the input file for reference.  I must admit I don't have a 
> very good grasp of the parallization options at this point, but 
> whatever I choose, I would guess that I should not end up with a 
> segmentation fault.  Any suggestions?
> The code was compiled with ifort (specific options listed below) on an 
> eight core Mac Pro.
>
>
>     Thanks
>
>             Paul Fons
> ############Log File Tail with error ####################
>
> -P-0000  leave_test : synchronization done...
>  mkrho: loop on k-points and spins done in parallel
>  Total charge density [el/Bohr^3]
> ,     Maximum=    8.8868E-02  at reduced coord.    0.9083    0.7778    
> 0.9667
> ,     Minimum=    1.2180E-03  at reduced coord.    0.5250    0.0000    
> 0.5250
>  Total charge density [el/Bohr^3]
> ,     Maximum=    8.8868E-02  at reduced coord.    0.9083    0.7778    
> 0.9667
> ,     Minimum=    1.2180E-03  at reduced coord.    0.5250    0.0000    
> 0.5250
> -P-0000  leave_test : synchronization done...
>   suscep_stat : loop on k-points and spins done in parallel
>   shape (susmat) =            2        1785           1        
> 1785           1
>   shape (susmat) =            2        1785           1        
> 1785           1
>   shape (susmat) =            2        1785           1        
> 1785           1
>   shape (susmat) =            2        1785           1        
> 1785           1
>   shape (susmat) =            2        1785           1        
> 1785           1
>   shape (susmat) =            2        1785           1        
> 1785           1
>   shape (susmat) =            2        1785           1        
> 1785           1
>   shape (susmat) =            2        1785           1        
> 1785           1
> -------------------------------------------------------------------------- 
>
> mpiexec noticed that process rank 5 with PID 68966 on node 
> neutrino.a04.aist.go.jp exited on signal 11 (Segmentation fault).
>
> ##############Compilation Options ####################
>
> === Build Information ===
>   Version       : 5.8.4
>   Build target  : i386_darwin9.8.0_intel11.0
>   Build date    : 20090919
>
>  === Compiler Suite ===
>   C compiler       : gnu
>   CFLAGS           : -O2 -m64
>   C++ compiler     : gnu4.0
>   CXXFLAGS         : -O2 -m64
>   Fortran compiler : intel11.0
>   FCFLAGS          : -O3 -g
>   FC_LDFLAGS       :
>
>  === Optimizations ===
>   Debug level        : symbols
>   Optimization level : standard
>   Architecture       : _
>
>  === MPI ===
>   Parallel build : yes
>   Parallel I/O   : yes
>   MPI CPPFLAGS   : -DMPI=1 -DMPI2=1      -DMPI_IO=1
>
>  === Linear algebra ===
>   Library type  : external
>   Use ScaLAPACK : no
>
>  === Plug-ins ===
>   BigDFT    : yes
>   ETSF I/O  : yes
>   LibXC     : yes
>   FoX       : no
>   NetCDF    : yes
>   Wannier90 : yes
>   XMLF90    : no
>
>  === Experimental features ===
>   Bindings            : no
>   Error handlers      : no
>   Exports             : no
>   GW double-precision : no
>   Macroave build      : yes
>
>
> ####################Input File#####################################
>
>
> The input file is attached below for reference
>
> #    Lattice parameter relaxation (including re-optimization of
> #         internal coordinates)
>
>  dilatmx   1.2    # Maximum scaling allowed for lattice parameters
>  ionmov   3       # Use BFGS algorithm
>  ntime   120      # Maximum number of optimization steps
>  optcell   2       # Fully optimize unit cell geometry, keeping symmetry
>  tolmxf   1.0e-6  # Convergence limit for forces as above
>  strfact   100     # Test convergence of stresses (Hartree/bohr^3) by
>                     # multiplying by this factor and applying force
>                     # convergence test
> prtgeo   6
> ecutsm 0.5
> #=================================================
>
> # parallization variables
>  paral_kgb   1
>  wfoptalg    14
>  nloalg      4
>  fftalg      401
>  iprcch      4
>  intxc       0
>  istwfk      1
>  fft_opt_lob 2
>  npfft       2
>  npband      16
>
>
> #=================================================
> # Calculation
> ixc 7   #   LDA or LSD, Teter Pade parametrization
> ecut 20
> #nband 200
> occopt 3
> tsmear 0.001 Ha
> diemix 1
> diemac 12
> iscf 17
> npulayit 20
> iprcel 45  #SCF preconditioning, compute the RPA dielectric matrix at 
> the first step, and recompute each step
> nstep 200
> tolvrs 1.0d-15
> #=================================================
>
>
> #=================================================
> # KptGrid
> kptopt 1
> ngkpt 2 2 2
> #=================================================
>
>
>
> #=================================================
> # Structure
>   acell   2.3472990156E+01  2.1416202665E+01  2.2829820755E+01
>   rprim   9.9999040546E-01 -4.3278800423E-03 -6.7708842004E-04
>          -8.4365915123E-03  9.9995267729E-01 -4.8442866378E-03
>          -9.2839564304E-04 -8.2630793413E-03  9.9996542920E-01
>
> natom 56
> ntypat 3
> znucl 32 51 52
> typat 8*1 16*2 32*3
> xcart
>   7.64903050727687E+00  3.70786933047274E+00  2.03854248709848E+01
>   1.53777783328731E+01  1.75129722033248E+01  8.91144538019716E+00
>   1.61043791880839E+01  8.42811621834851E+00  8.30391356166042E+00
>   1.92648347440380E+01  4.48585851017361E+00  1.41075418556874E+01
>   7.65528402890655E+00  1.32373536776582E+01  2.13621839955297E+01
>   3.53536790751728E+00  1.77037554287189E+01  7.83953650844295E+00
>   2.01669253437016E+01  1.43678663967696E+01  1.99921563932553E+01
>   3.41610243671658E+00  3.27583538347910E+00  2.18451293492937E+00
>   1.77658032309385E+01  1.44381527352672E+01  1.21808728166705E+01
>   3.99726773805650E+00  1.55301797871776E+01  1.44367783628060E+00
>   3.24381188162565E-02  9.90763595116681E+00  2.31616068517336E+01
>   1.15634952674770E+01  1.63469867258440E+01  1.67858050920974E+01
>   1.17234135518916E+01  5.15206944025922E+00  5.57997270756973E+00
>   3.38545937719249E-01  5.27152903896988E+00  1.64909066477600E+01
>  -1.25956396013325E+00  1.51941337814835E+01  5.66856539691514E+00
>   1.67427745587491E+01  1.01677363481710E+01  1.70356580625314E+01
>   1.74889112537567E+01  2.14415960427150E+01  5.08012987634538E+00
>   5.20436950708091E+00  2.04510938227703E+01  1.69328729383300E+01
>   5.83248173079816E+00  1.00668269543405E+01  5.82511338366536E+00
>   1.69058930916342E+01  5.44158534726656E+00 -2.16875982251231E-01
>   1.14478213641201E+01  2.14067680887845E+01 -7.37133630516936E-02
>  -4.56781545787971E-03  2.15082711873171E+01  1.11914229522603E+01
>   6.89325977321322E+00  5.19085086051822E+00  1.20732280739244E+01
>   1.18260878315009E+01  1.10106693193841E+01  1.09554286623317E+01
>   5.55293700386207E+00 -9.08379756199719E-01  3.66465273998066E-03
>   1.79985411043133E+01  6.24218498329430E-01  1.12402686810674E+01
>   1.71942001152461E+01  1.13101891917151E+01 -4.84440266489979E-01
>   1.26878781059539E+01  5.13251738639552E+00  1.09978256028475E+01
>  -6.50193761804745E-01  1.57802036826258E+01  1.11596387346547E+01
>  -1.02604457424129E+00  4.96634865210211E+00 -6.73097999699995E-01
>   2.33537680056396E+01  9.88707912220360E+00  6.10826699096729E+00
>   2.29852463452618E+01  2.12114518251038E+01  1.66599708565945E+01
>   1.22232885036521E+01  2.08727445643273E+01  5.53194801348795E+00
>   1.14667297727525E+01  1.08883768941973E+01  1.68133908463315E+01
>   1.20152100923160E+01  1.61669109268310E+01 -5.25089342964899E-01
>   6.50738025520855E+00  4.48682210053020E+00  6.33107290863955E+00
>   6.09706234778241E+00  1.51808280126207E+01  1.68378955464732E+01
>   1.69665142853349E+01  1.36189954044245E+01  5.85004443906740E+00
>   1.54155719512525E+01  5.02533339195595E+00  1.72012760815797E+01
>   1.97226735071461E+01  8.87990788178025E+00  1.17767449660632E+01
>   1.68905182361589E+01  2.12690677624622E+01  2.23479489549080E+01
>   5.75725578561531E+00  2.07078547580210E+01  1.13763969715566E+01
>   5.21245919927846E+00  8.44225866564750E+00  2.31958914257793E+01
>   1.08818580526721E+01  1.63624727256228E+01  1.08918294189764E+01
>   1.13124635080077E+01  6.30393394077432E+00  2.26878820078345E+01
>   2.49257212897598E+01  5.50572088462659E+00  1.11122967668993E+01
>   2.18464733892219E+01  1.75141649461871E+01  2.31127995799147E+01
>  -9.97470518267632E-02  1.15891619518773E+01  1.76507937284874E+01
>   9.50044364651022E-01 -1.64226282922227E-01  5.40194333962429E+00
>   1.07022490316033E+01  7.88369968438184E-01  1.72497450152962E+01
>   1.12766649184463E+01  1.07058794844567E+01  5.32218315850765E+00
>   1.74924886915205E+01  1.70921011246861E+01  1.67711014044835E+01
>   1.88674326277119E+01  5.26568571353990E+00  5.20322800151578E+00
>   5.95117393151507E+00  7.09327889287457E+00  1.71421788222188E+01
>   7.54305999769611E+00  1.58557245081277E+01  5.40369066115511E+00
>   6.52780647971927E+00  1.05300127764927E+01  1.13754713069850E+01


--
##############################################################
Francois Bottin                    tel: 01 69 26 41 73
CEA/DIF                            fax: 01 69 26 70 77
BP 12 Bruyeres-le-Chatel         email: Francois.Bottin@cea.fr
##############################################################