Hello, after a few weeks of absence I feel now confused regarding location of uBLAS files. I used to get the latest version from ...
Zdenek Hurak
z.hurak@...
Oct 1, 2003 7:56 am
1091
... Hash: SHA1 ... [(stiffness) Matrix assembly] ... No. But have a look at my sample program. Joerg and I spent lots of hours to optimize matrix assembly. I...
Gunter Winkler
guwi17@...
Oct 1, 2003 9:02 am
1092
Hi Joerg, ... All a lot of work!! I think the crucuial point is that there are no useful definitions of matrices and their operators that form a field. ...
Hello, I finished the first part of an ublas howto. Please have a look at http://www-user.tu-chemnitz.de/~wgu/ublas and feel free to send comments and more...
Gunter Winkler
guwi17@...
Oct 1, 2003 1:00 pm
1094
Ok, I 'll try with sparse vector. Thx, jean marc mercier ... From: "jean_marc_mercier" <jean_marc_mercier@...> To: <ublas-dev@yahoogroups.com> Sent:...
Hi Joerg, Beman is looking at getting a Boost 1.31 release out. This seem like a good time to ask if you need any help in getting the current uBLAS into a new...
Hi all, You may recall that a while ago, I was asking about something similar: http://groups.yahoo.com/group/ublas-dev/message/395 The GluCat multivector...
... Hash: SHA1 ... I think requiring the underlying data type of ublas::matrix and ublas::vector to form a field is to much for most operations. For most ...
Gunter Winkler
guwi17@...
Oct 1, 2003 3:08 pm
1098
... I took a look. THanks! ... Your matrix sparsity pattern is much more random than mine (you actually use a random number generator to choose indii). My...
This is a very exciting and challenging topic for UBLAS. This library may become a very powerful tool and a brick toward linear algebra if such features were...
Dear Joerg, Do you think there is an easy way to handle sparse_vector with other key map than std::size_t ? jean marc mercier <jean_marc_mercier@...>...
Dear all, It seems that my last message was not relevant. I don't understand why, but the following codes compile and does give results :) The only bad point...
Oops, I just understood that my last request is not so easy to achieve. I made some change in the UBLAS sparse_vector class to handle other key type. I began...
I have thought about a beam finite element with 12 degree of freedom: I't s just the classic one. I have also thought about a an element with 16 degree of...
Have you ever thought about providing diadic product through prod (vector, vector) ? or is there a simple way to compute it that I am missing? Bye Antonio...
Did you look at outer_prod? ... Julius Muschaweck _____________________________________________________ OEC AG Paul-Gerhardt-Allee 42 81245 Muenchen, Germany ...
Hi, I find that prod( m, cm ) where m is a small (3x3) dense matrix and cm is a (3xn) compressed matrix with about 30 nonzeroes takes O(n) time whereas I ...
Please take a look at the attached code snippet, where a class with a compressed_vector member returns a reference to an element of that vector, to be modified...
Hi, We are now using ublas extensively to replace our old matrix classes but encountered a severe performance problem when scaling to large sparse matrices,...
... Yes. You probably use sparse_matrix<>. ... Please have a look at http://www-user.tu-chemnitz.de/~wgu/ublas/matrix_sparse_usage.html (and report your...
Gunter Winkler
guwi17@...
Oct 13, 2003 9:01 am
1111
Thanks! That will hopefully help and get the performance-critical parts of our codes back into the matrix element calculation. I will play with the various...
Hi, I tried to use these faster versions but the Boost version of ublas is missing the file boost/numeric/ublas/vector_of_vector.hpp Can somebody send me this...
... until this happens you can replace the ublas files by the files of ublas_2003_09_23.zip provided on my page. (This is the current (quite stable) and...
Gunter Winkler
guwi17@...
Oct 13, 2003 11:03 am
1114
Hi Gunter, Thanks, I got that and tried. While coordinate_matrix is by far the slowest and sparse_matrix not much faster, compressed_matrix seems the best so...
... This should not happen - looks like a (new) bug. This program failed: generalized_vector_of_vector< double, row_major, vector<compressed_vector<double> > >...
Gunter Winkler
guwi17@...
Oct 13, 2003 2:50 pm
1116
... From: "Gunter Winkler" <guwi17@...> To: <ublas-dev@yahoogroups.com> Sent: Monday, October 13, 2003 4:50 PM Subject: Re: [ublas-dev] Performance problems...
Having problems with message search?
Fill out this form to ensure your group is one of the first to be migrated to the new message search system.
