-
Best way to fix this would be to use iterator base classes in all of uBLAS so this kind of thing can be fixed in one spot.
2006-12-01 15:46:46 UTC by mistevens
-
Always check the return values of factorisations!!
In your case lu_factorize returns non zero because the result is singular. Latter on the lu_subsitute correctly fails as no inverse is possible.
Your matrix can however be easily factorised with partial pivoting. You simply need to pass a permutation_matrix to the lu_ functions.
Probably best to simply use the complete function at...
2006-12-01 15:36:44 UTC by mistevens
-
1. IB seems reasonable as we expect base to be small and size_type to support +/- size_t for IB correction.
compressed and coordinate types need a change to template parameters so difference_type can be passed. IB could go and be replaced by row_major_1 etc
2. Fixed
3. Expression functors have been fixed. No more size_t in functional.hpp.
2006-12-01 10:23:12 UTC by mistevens
-
I got singular error while inversing a single matrix
Assertion failed in file /usr/include/boost/numeric/ublas/
triangular.hpp at line 2507:
e1 () (n, n) != value_type (0)
terminate called after throwing an instance of 'boost::numeric::
ublas::singular'
what(): singular
Aborted
The matrix is
[1 1 1
1 1 2
1 2 2]
the code I used to inverse is
matrix A(input);
// perform...
2006-02-03 19:49:46 UTC by windmaomao
-
Logged In: YES
user_id=556024
As of Boost 1.33.0 all documentation has corret function
signatures.
2005-08-22 15:57:54 UTC by mistevens
-
Logged In: YES
user_id=556024
FIX potential fix as outlined commited to Boost post 1.33.0
required testing.
2005-08-18 13:23:21 UTC by mistevens
-
Logged In: YES
user_id=556024
FIX as outlined above.
2005-08-18 13:08:27 UTC by mistevens
-
Logged In: YES
user_id=556024
FIX potential fix as outlined commited to Boost post 1.33.0
required testing.
2005-08-18 13:07:01 UTC by mistevens
-
// this produces a wrong result
diagonal_matrix z(sz), x(sz);
matrix y(sz, sz);
matrix B = prod(z, x - y);
In matrix_binary the increment and decrement operators
are dispatched depending on the underlying iterator types,
but the += and -= have the old behavior.
Quick fix would be to call ++ or -- as often as requested.
Full fix...
2005-07-11 12:01:02 UTC by mistevens
-
Logged In: YES
user_id=556024
Data accessors are documented in Vector/Matrix concepts.
2005-07-11 11:55:19 UTC by mistevens
