... You could use an iterative method like the power method which converges to the dominant eigenvalue/eigenvector pair: start with x=(1,0,...,0), iterate x :=...
Gunter Winkler
guwi17@...
Dec 1, 2003 8:46 am
1302
Hi, ... This seems to be correct only until version 4.0: UMFPACK 4.1 has added an algorithm to handle symmetric matrices. ... Best regards, Julius...
Hi I don't know if this is a bug or a feature, and I have no opinion about what if anything you should do about it, but for me it occurred in a real world...
Salut, ... I would call this the correct extension to matrices with zero columns or zero rows. If you don't sum you get nothing. The only problem is, that...
Dear Derek, The IETL also works with sparse matrices and we have a new version (not yet released because the documentation is not fully updated yet) that...
Hi, I found what I believe is a bug with matrix_range/matrix_row. A matrix_row can be constructed from a matrix_range, but the constructed object contains...
Hi Desmond, ... A feature ;-) ... Hm, unsure about the 'spuriously'. ... [...] The idea when allowing zero-sized vectors/matrices was to eliminate code for ...
Hi Desmond, ... 1.30.2 is the same as 1.30.0 w.r.t. ublas. [...] ... Oops. My local copy compiled with GCC 3.3.2 gives ... INFORM: : create a matrix with...
I was bitten once again by a zero-insertion problem similar to that discussed before in http://groups.yahoo.com/group/ublas-dev/message/1193?threaded=1 The...
Hi, ... I already tried to fix that as described in http://groups.yahoo.com/group/ublas-dev/message/1225 ... With a slightly modified test program ... #include...
I recently profiled an application which uses ublas on FreeBSD 4.5, with g++ 2.95.3 and ublas from boost 1.29.0. While doing Kalman filtering with some fairly...
... I have to apologize. As I noticed before and said in the old thread the fix doesn't work for VC7.0. I somehow failed to remember that and draw the right...
Hi Ben, ... Oh that is a old compiler and old version of ublas! ... The issue of next products is something that causes a great deal of confusion. It even came...
Hi, ... Of course it would be good if there were no nasty surprises with VC7.0 and VC7.1! From reading your initial posting it seems there were two related ...
When profiling my program, I noticed that much time was spent in an unexpected part of the code. It turned out that some ublas sparse matrix operations are...
... I support the request. In my code, I have the following: //some function modifying a vector void do_something(VECTOR &v) { ... } ... MATRIX dmat; ... //now...
... Did you try A = A + B? It is usually much faster to create a new matrix from both operands than to insert elements into A because you can merge the ...
Gunter Winkler
guwi17@...
Dec 3, 2003 2:58 pm
1320
... Why don't you use a templated function template <typename V> do_something(V& v) { ... } Karl...
Karl Meerbergen
Karl.Meerbergen@...
Dec 3, 2003 3:08 pm
1321
Hi Guys, ... I think you are cunfusing object oriented type compatibility with generic code. matrix_row and vector are two completely different types. They...
... Sounds good. I think it is a quite natural wish to handle the proxy types (row, column, range...) in the same way you handle the solid types (matrix,...
... Thanks, that is good to know. I was already wondering if there is a function to show me the amount of memory used by a matrix. Maybe "non_zeros" is an...
I tried to switch from VC 7.0 to VC 7.1 and ran into some problems with the element proxies. It seems that I can't get a reference to a sparse matrix element....
I helped myself by adding a function get() to sparse_matrix_element in matrix_sparse.hpp: BOOST_UBLAS_INLINE reference get() { dirty_ = true; return d_; } So I...
... Right. I would like to stress that the invalidation problem is not fixed. I modified my initial code sample to demonstrate that. The code creates a vector...
Hi, this is a short notice to let you know that we have our own sourceforge project on http://www.sf.net/projects/ublas now. I'll stop distributing snapshots...
Hi, ... Yes. ... This is due to the bug fix documented here: http://groups.yahoo.com/group/ublas-dev/message/1137 ... That's dangerous. ... Interesting idea....
Hi, ... Yes. ... [snip] ... Here viter may or may not be invalidated depending on the storage type. ... This behaviour probably should be documented. Thanks, ...
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.
