[petsc-dev] MatAXPY with DIFFERENT_NONZERO_PATTERN

[Matthew Knepley]

On Fri, Apr 20, 2012 at 6:07 AM, Alexander Grayver
<agrayver at gfz-potsdam.de>wrote:

> Hello,
>
> Calling MatAYPX with MatStructure == DIFFERENT_NONZERO_PATTERN produces
> following errors:
>
> [30]PETSC ERROR: --------------------- Error Message
> ------------------------------**------
> [30]PETSC ERROR: Argument out of range!
> [30]PETSC ERROR: New nonzero at (0,8486) caused a malloc!
> [30]PETSC ERROR: ------------------------------**
> ------------------------------**------------
> [30]PETSC ERROR: Petsc Development HG revision:
> 7c3c3f8c0d65b793aea33c138419de**755b591ca7  HG Date: Thu Feb 09 14:42:30
> 2012 -0600
> [30]PETSC ERROR: See docs/changes/index.html for recent updates.
> [30]PETSC ERROR: See docs/faq.html for hints about trouble shooting.
> [30]PETSC ERROR: See docs/index.html for manual pages.
> [30]PETSC ERROR: ------------------------------**
> ------------------------------**------------
> [30]PETSC ERROR: /home/prog on a openmpi-i named node229 by agrayver Fri
> Apr 20 11:33:55 2012
> [30]PETSC ERROR: Libraries linked from /home/lib/petsc-dev/openmpi-**
> intel-complex-debug-f/lib
> [30]PETSC ERROR: Configure run at Fri Feb 10 09:28:34 2012
> [30]PETSC ERROR: Configure options --with-petsc-arch=openmpi-**intel-complex-debug-f
> --with-fortran-interfaces=1 --with-scalapack-lib=/opt/**
> intel/Compiler/11.1/072/mkl/**lib/em64t/libmkl_scalapack_**lp64.a
> --with-scalapack-include=/opt/**intel/Compiler/11.1/072/mkl/**include
> --with-blacs-lib=/opt/intel/**Compiler/11.1/072/mkl/lib/**
> em64t/libmkl_blacs_openmpi_**lp64.a --with-blacs-include=/opt/**
> intel/Compiler/11.1/072/mkl/**include --with-mpi-dir=/opt/mpi/intel/**openmpi-1.4.2
> --with-scalar-type=complex --with-blas-lapack-lib="[/opt/**
> intel/Compiler/11.1/072/mkl/**lib/em64t/libmkl_intel_lp64.a,**
> /opt/intel/Compiler/11.1/072/**mkl/lib/em64t/libmkl_intel_**
> thread.a,/opt/intel/Compiler/**11.1/072/mkl/lib/em64t/libmkl_**
> core.a,/opt/intel/Compiler/11.**1/072/lib/intel64/libiomp5.a]"
> --with-precision=double --with-x=0
> [30]PETSC ERROR: ------------------------------**
> ------------------------------**------------
> [30]PETSC ERROR: MatSetValues_MPIAIJ() line 538 in
> /home/lib/petsc-dev/src/mat/**impls/aij/mpi/mpiaij.c
> [30]PETSC ERROR: MatSetValues() line 1129 in /home/lib/petsc-dev/src/mat/*
> *interface/matrix.c
> [30]PETSC ERROR: MatAXPY_**BasicWithPreallocation() line 108 in
> /home/lib/petsc-dev/src/mat/**utils/axpy.c
> [30]PETSC ERROR: MatAXPY_MPIAIJ() line 2235 in /home/lib/petsc-dev/src/mat/
> **impls/aij/mpi/mpiaij.c
> [30]PETSC ERROR: MatAXPY() line 39 in /home/lib/petsc-dev/src/mat/**
> utils/axpy.c
> [30]PETSC ERROR: MatAYPX() line 272 in /home/lib/petsc-dev/src/mat/**
> utils/axpy.c
>
>
> I guess it has something to do with recent changes in mataij
> implementation, doesn't it?


Yes, we started to require that everything be correctly allocated, rather
than have hidden performance problems.

   Matt


> --
> Regards,
> Alexander
>
-- 
What most experimenters take for granted before they begin their
experiments is infinitely more interesting than any results to which their
experiments lead.
-- Norbert Wiener
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mcs.anl.gov/pipermail/petsc-dev/attachments/20120420/11a137ff/attachment.html>