And I now get the following error when I try to run this in parallel. Note, Blaise, I do not have coordinates, so perhaps you could check that you still get divergence w/o coordinates. (I don't need anything else form you) Mark [1]PETSC ERROR: --------------------- Error Message ------------------------------------ [1]PETSC ERROR: Nonconforming object sizes! [1]PETSC ERROR: Sum of local lengths 29784 does not equal global length 29791, my local length 2482 likely a call to VecSetSizes() or MatSetSizes() is wrong. See http://www.mcs.anl.gov/petsc/documentation/faq.html#split! [1]PETSC ERROR: ------------------------------------------------------------------------ [1]PETSC ERROR: Petsc Development HG revision: 73637941c41b51aa3e638c914e6cb05248ece8b0 HG Date: Mon Jun 11 14:55:24 2012 -0400 [1]PETSC ERROR: See docs/changes/index.html for recent updates. [1]PETSC ERROR: See docs/faq.html for hints about trouble shooting. [1]PETSC ERROR: See docs/index.html for manual pages. [1]PETSC ERROR: ------------------------------------------------------------------------ [1]PETSC ERROR: ./ex10 on a arch-maco named madams-macbk-3.local by markadams Tue Jun 12 17:31:55 2012 [1]PETSC ERROR: Libraries linked from /Users/markadams/Codes/petsc-dev/arch-macosx-gnu-g/lib [1]PETSC ERROR: Configure run at Mon Jun 11 17:03:36 2012 [1]PETSC ERROR: Configure options CXX=/sw/lib/gcc4.4/bin/g++-4 CC=/sw/lib/gcc4.4/bin/gcc-4 CFLAGS=-malign-double CXXFLAGS=-malign-double FC=/sw/lib/gcc4.4/bin/gfortran --download-hypre=1 --download-parmetis=1 --download-metis=1 --download-triangle=1 --with-x=0 --with-debugging=1 --download-mpich=1 PETSC_ARCH=arch-macosx-gnu-g [1]PETSC ERROR: ------------------------------------------------------------------------ [1]PETSC ERROR: PetscSplitOwnership() line 93 in /Users/markadams/Codes/petsc-dev/src/sys/utils/psplit.c [1]PETSC ERROR: PetscLayoutSetUp() line 147 in /Users/markadams/Codes/petsc-dev/src/vec/vec/impls/mpi/pmap.c [1]PETSC ERROR: MatMPIAIJSetPreallocation_MPIAIJ() line 3194 in /Users/markadams/Codes/petsc-dev/src/mat/impls/aij/mpi/mpiaij.c On Jun 12, 2012, at 5:19 PM, Blaise Bourdin wrote:

Hi,

Sorry, I should have mentioned that the .h5 and xml files are not for public eyes. I am uploading the binary files created with -U_ksp_view_binary, but my internet connection sucks.

Jed, I put the file on your account on schur. can you access them? The datafiles.tgz archive contains the the result files in hdf5 format. ksp_view_binary.tgz contains only the binary output of ksp_view_binary

If you give me until tonight, I can try to set a prefix one the vecs.

Thanks a lot for your help.

Blaise

On Jun 12, 2012, at 4:06 PM, Mark F. Adams wrote:

...

On Jun 12, 2012, at 4:56 PM, Jed Brown wrote:

...

On Tue, Jun 12, 2012 at 3:54 PM, Mark F. Adams <[email protected]> wrote: How is what? The ex10 code or "I set the block size for matrices before SetFromOptions".

This how I create matrices with blocks sizes:

ierr = MatCreate( wcomm, &Prol ); CHKERRQ(ierr); ierr = MatSetSizes(Prol,nloc*bs,nLocalSelected*col_bs,PETSC_DETERMINE,PETSC_DETERMINE); CHKERRQ(ierr); ierr = MatSetBlockSizes( Prol, bs, col_bs ); CHKERRQ(ierr);

Sorry, I thought you were talking about vectors because the error is coming from a vector load.

There is a vector load, at least ex10 loads a vec and its seems to work (so I assume Blaise put one in there).

...

Where is this input file? I'll grab it sort out the issue.

You can get an archive containing the KSP and the data files at https://filestogeaux.lsu.edu/public/download.php?FILE=bourdin/52338Q0at2Y

These are just standard "binaryoutput" files I think,

Mark

-- Department of Mathematics and Center for Computation & Technology Louisiana State University, Baton Rouge, LA 70803, USA Tel. +1 (225) 578 1612, Fax +1 (225) 578 4276 http://www.math.lsu.edu/~bourdin