[Getdp] Out of memory error message

John_V jvillar.john at gmail.com
Sun Aug 22 23:36:33 CEST 2010


  Hi Lars,

Thanks for the diagnosis and suggestion. Unfortunately it did not work 
for me.

    getdp Lines -pre EleSta_v -cal -petsc_prealloc 50 -pos Map

was only a little slower (52 s), but gave the same out-of-memory error.

I went as low as

    getdp Lines -pre EleSta_v -cal -petsc_prealloc 15 -pos Map

This was very slow (~3 h), but the end point was the same.

John

On 8/20/2010 2:06 PM, Lars Rindorf wrote:
>
> Hi John
>
> What goes wrong is that petsc tries to preallocate to much memory. Before your sparse matrix is constructet petsc 'pre'allocates memory to speed up the assembly of the matrix. You can reduce the preallocated memory by the command line option:
>
> -petsc_prealloc 300
>
> I would try with 50. Assembly will probably be slow. Then I would 
> increase it.
>
> KR Lars
>
> *Fra:* John_V [mailto:jvillar.john at gmail.com]
> *Sendt:* 20. august 2010 15:50
> *Til:* Lars Rindorf
> *Cc:* getdp at geuz.org
> *Emne:* Re: [Getdp] Out of memory error message
>
> Lars,
>
> The mesh file is large (148 MB), as you might expect since it produces 
> an out-of-memory error. I'm not sure I should post something that 
> large to the list, but you could regenerate it by 3-D mesh (using 
> gmsh) of the attached Lines.geo file. That's how I did it. The 
> remaining inputs are also attached.
>
> Output follows:
>
> Info    : Running 'getdp Lines -pre EleSta_v -cal -pos Map'
> Info    : Started on Fri Aug 20 09:29:56 2010
> Info    : Loading problem definition 'Lines.pro'
> Info    : Loading problem definition 'parameters.pro 
> <http://parameters.pro>'
> Info    : Loading problem definition 'Jacobian_Lib.pro'
> Info    : Loading problem definition 'Integration_Lib.pro'
> Info    : Loading problem definition 'EleSta_v.pro'
> Info    : Selected Resolution 'EleSta_v'
> Info    : Loaded Geometric data 'Lines.msh'
> Info    : System 'Sys_Ele' : Real
> P r e - P r o c e s s i n g . . .
> Info    : Treatment Formulation 'Electrostatics_v'
> (0 %)
> Info    :   Generate ExtendedGroup '_CO_Entity_13' (NodesOf)
> (10 %)
> (20 %)
> (30 %)
> (40 %)
> (50 %)
> (50 %)
> (60 %)
> (70 %)
> (80 %)
> (90 %)
> (100 %)
> Done!
> Info    : (CPU = 21.8594s)
> E n d   P r e - P r o c e s s i n g
> P r o c e s s i n g . . .
> Info    : Generate[Sys_Ele]
> (0 %)
> (10 %)
> (20 %)
> (30 %)
> (40 %)
> (50 %)
> (50 %)
> (60 %)
> (70 %)
> (80 %)
> (90 %)
> (100 %)
> Done!
> Info    : (CPU = 51.4063s)
> Info    : Solve[Sys_Ele]
> Info    : N: 390857
>
> [0]PETSC ERROR: --------------------- Error Message 
> ------------------------------------
> [0]PETSC ERROR: Out of memory. This could be due to allocating
> [0]PETSC ERROR: too large an object or bleeding by not properly
> [0]PETSC ERROR: destroying unneeded objects.
> [0]PETSC ERROR: Memory allocated 0 Memory used by process 0
> [0]PETSC ERROR: Try running with -malloc_dump or -malloc_log for info.
> [0]PETSC ERROR: Memory requested 943957912!
> [0]PETSC ERROR: 
> ------------------------------------------------------------------------
> [0]PETSC ERROR: Petsc Release Version 3.0.0, Patch 7, Mon Jul  6 
> 11:33:34 CDT 2009
> [0]PETSC ERROR: See docs/changes/index.html for recent updates.
> [0]PETSC ERROR: See docs/faq.html for hints about trouble shooting.
> [0]PETSC ERROR: See docs/index.html for manual pages.
> [0]PETSC ERROR: 
> ------------------------------------------------------------------------
> [0]PETSC ERROR: Lines on a umfpack-c named GALILEO by jvillar Fri Aug 
> 20 09:29:56 2010
> [0]PETSC ERROR: Libraries linked from 
> /home/Administrator/src/petsc-3.0.0-p7/umfpack-cxx-opt/lib
> [0]PETSC ERROR: Configure run at Wed Mar 24 13:55:40 2010
> [0]PETSC ERROR: Configure options --with-debugging=0 
> --with-scalar-type=complex --with-clanguage=cxx --with-shared=0 
> --with-mpi=0 --with-x=0 --with-fortran=0 --with-cc="gcc -mno-cygwin" 
> --with-cxx="g++ -mno-cygwin" COPTFLAGS=-O2 CXXOPTFLAGS=-O2 
> --with-umfpack=1 --download-umfpack=ifneeded --useThreads=0 --useThreads=0
> [0]PETSC ERROR: 
> ------------------------------------------------------------------------
> [0]PETSC ERROR: PetscMallocAlign() line 61 in src/sys/memory/mal.c
> [0]PETSC ERROR: PetscFreeSpaceGet() line 14 in src/mat/utils/freespace.c
> [0]PETSC ERROR: MatILUFactorSymbolic_SeqAIJ() line 1322 in 
> src/mat/impls/aij/seq/aijfact.c
> [0]PETSC ERROR: MatILUFactorSymbolic() line 5308 in 
> src/mat/interface/matrix.c
> [0]PETSC ERROR: PCSetUp_ILU() line 293 in 
> src/ksp/pc/impls/factor/ilu/ilu.c
> [0]PETSC ERROR: PCSetUp() line 794 in src/ksp/pc/interface/precon.c
> [0]PETSC ERROR: KSPSetUp() line 237 in src/ksp/ksp/interface/itfunc.c
> [0]PETSC ERROR: KSPSolve() line 353 in src/ksp/ksp/interface/itfunc.c
> [0]PETSC ERROR: User provided function() line 976 in 
> unknowndirectory/LinAlg_PETSC.cpp
>
> This application has requested the Runtime to terminate it in an 
> unusual way.
> Please contact the application's support team for more information.
>
> John
>
> On Fri, Aug 20, 2010 at 1:57 AM, Lars Rindorf <lrf at teknologisk.dk 
> <mailto:lrf at teknologisk.dk>> wrote:
>
> Hi John
>
> Could you send the complete output and/or a simple file that 
> reproduces your problem?
>
> KR Lars
>
> -----Oprindelig meddelelse-----
> Fra: getdp-bounces at ace20.montefiore.ulg.ac.be 
> <mailto:getdp-bounces at ace20.montefiore.ulg.ac.be> 
> [mailto:getdp-bounces at ace20.montefiore.ulg.ac.be 
> <mailto:getdp-bounces at ace20.montefiore.ulg.ac.be>] På vegne af John_V
> Sendt: 19. august 2010 19:37
> Til: David Colignon
> Cc: getdp at geuz.org <mailto:getdp at geuz.org>
> Emne: Re: [Getdp] Out of memory error message
>
>
> Thank you Dave. I expected a possible barrier at 2 GB and was
> surprised that the error message indicated only about half that. I have
> also estimated memory usage by just watching the memory used for that
> process in the Task Manager. This method's estimate is consistent with
> the error message--it also indicates a bit over 1 GB usage when the
> program fails. This is what made me wonder if the limit were something
> else (maybe something more easily correctable). --John
>
> On 8/19/2010 3:00 AM, David Colignon wrote:
> > Hi John,
> >
> > if you use the precompiled windows binary, I think it is compiled in
> > 32 bits mode , and then you hit the 2 GB barrier.
> >
> > Regards,
> >
> > Dave
> >
>
> _______________________________________________
> getdp mailing list
> getdp at geuz.org <mailto:getdp at geuz.org>
> http://www.geuz.org/mailman/listinfo/getdp
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.geuz.org/pipermail/getdp/attachments/20100822/3fb6eaf3/attachment.html>