[petsc-users] Strange error(?) message
Just built petsc-dev and it did not help. I'm going to look into the code to see if my graph is ill-formed in some sense. Just hope the problem is from my side not a real bug in ParMetis! On Mon, Apr 9, 2012 at 8:46 PM, Mohammad Mirzadeh mirzadeh at gmail.comwrote: ok. Thanks Jed. I'll try petsc-dev to see if it fixes the problem. Thanks everyone. On Mon, Apr 9, 2012 at 8:42 PM, Jed Brown jedbrown at mcs.anl.gov wrote: On Mon, Apr 9, 2012 at 22:37, Mohammad Mirzadeh mirzadeh at gmail.comwrote: Thanks Sean. I'm using Petsc 3.2-p6 along with ParMetis 4.0.2. Since this was not supported with 3.2-p6, and previous versions had bugs, I built parmetis myself and used --with-parmetis-include and --with-parmetis-lib flags to build petsc. Should I switch to petsc-dev? Yes, and use --download-metis --download-parmetis because the version upstream has some bugs for which the patches have not been applied. -- next part -- An HTML attachment was scrubbed... URL: http://lists.mcs.anl.gov/pipermail/petsc-users/attachments/20120410/99e9ebfd/attachment.htm
[petsc-users] petsc binary matrix file for dense matrices
Dear all, For a simple interface to SLEPc, I need to write my sparse or dense matrices in PETSc binary format, at first, I need to solve reduced dense generalized eigenvalue problems which are in dense matrix format. I first had a look at the MATLAB routines to read and write the matrices as examples and MatLoad function manual. I got the idea however I have a question regarding the write operation for dense matrices, in MATLAB code, it is given as [m,n] = size(A); write(fd,[1211214,m*n],'int32'); write(fd,A,'double'); I guess the dense matrix is written as a 1d array, is this right? and there is no need for row and column information for the matrix. I am guessing that MATLAB writes the matrix, A, in column-major orientation, is not this a problem for petsc dense matrices? Most probably this info is somewhere in the documentation, but could not find. Any pointers are appreciated. Best, Umut
[petsc-users] petsc binary matrix file for dense matrices
On Tue, Apr 10, 2012 at 04:27, Umut Tabak u.tabak at tudelft.nl wrote:
I first had a look at the MATLAB routines to read and write the matrices
as examples and MatLoad function manual. I got the idea however I have a
question regarding the write operation for dense matrices, in MATLAB code,
it is given as
[m,n] = size(A);
write(fd,[1211214,m*n],'int32'**);
write(fd,A,'double');
I suggest just using
PetscBinaryWrite('filename',sparse(A))
The scripts are in $PETSC_DIR/bin/matlab.
PETSc binary format does not contain information for storing dense matrices
as dense. Note that you can still read the sparse matrix on disk as a
dense matrix. Also, if you compress the binary (e.g. gzip), it will be
essentially the same size (because the column indices are so regular). For
example, writing a 1000x1000 random matrix (the entries essentially don't
compress for a random matrix), I have
-rw-r--r-- 1 jed users 808 Apr 10 06:52 densematrix
-rw-r--r-- 1 jed users 12004016 Apr 10 06:52 sparsematrix
-rw-r--r-- 1 jed users 7539706 Apr 10 06:52 densematrix.gz
-rw-r--r-- 1 jed users 7568387 Apr 10 06:52 sparsematrix.gz
-- next part --
An HTML attachment was scrubbed...
URL:
http://lists.mcs.anl.gov/pipermail/petsc-users/attachments/20120410/c2bda468/attachment.htm
[petsc-users] petsc binary matrix file for dense matrices
On 04/10/2012 01:55 PM, Jed Brown wrote: On Tue, Apr 10, 2012 at 04:27, Umut Tabak u.tabak at tudelft.nl mailto:u.tabak at tudelft.nl wrote: I first had a look at the MATLAB routines to read and write the matrices as examples and MatLoad function manual. I got the idea however I have a question regarding the write operation for dense matrices, in MATLAB code, it is given as [m,n] = size(A); write(fd,[1211214,m*n],'int32'); write(fd,A,'double'); I suggest just using Dear Jed, I looked at the MATLAB codes as an example, I should interface my matrices which are in the C++ code(they are MTL4 compressed or dense matrices), so they are not in MATLAB. However, I guess I should use the format given on MatLoad page whether the matrices are sparse or dense. I wrote a simple test case which writes the matrix in binary format, which in this case a sparse matrix, now I guess I can use the same routine to write a dense matrix as far as I could understand from MatLoad(There is no separate specification for a dense matrix, right?). Best, U. -- next part -- An HTML attachment was scrubbed... URL: http://lists.mcs.anl.gov/pipermail/petsc-users/attachments/20120410/7d65aa19/attachment-0001.htm
[petsc-users] petsc binary matrix file for dense matrices
On Tue, Apr 10, 2012 at 07:34, Umut Tabak u.tabak at tudelft.nl wrote: However, I guess I should use the format given on MatLoad page whether the matrices are sparse or dense. I wrote a simple test case which writes the matrix in binary format, which in this case a sparse matrix, now I guess I can use the same routine to write a dense matrix as far as I could understand from MatLoad(There is no separate specification for a dense matrix, right?). Yes -- next part -- An HTML attachment was scrubbed... URL: http://lists.mcs.anl.gov/pipermail/petsc-users/attachments/20120410/f94211bc/attachment.htm
