thartmann15

svd_r_raw is not meant to be called directly. use svd_r instead: >>> from mpmath import mp >>> A = mp.matrix([[2, -2, -1], [3, 4, -2], [-2, -2, 0]]) >>> S...

Sadly the matrix implementation of mpmath is using dictionaries which makes it very slow. But you can hack mpmath to use numpy arrays with mpmath objects as entries. Examples: >...

Regarding numpy+scipy datatypes: You cannot take a numpy array with mpmath entries and feed it to the numpy/scipy eigen (or other matrix manipulation) routines. It does not work this way...

I started implementing the generalized eigenvalue problem but stopped halfway because I had to take care of other things. If there is a strong need for this, I could try...

I can try to code a preliminary version of the generalized Schur decomposition, but this will take one to three weeks as the details are crazy complicated. The reason I...

In my testing, I got similar results: Numpy array with mpmath content give ca. a 10x speed improvement for matrix operations. The main question is if one is willing to...

Here is a status update on my attempts to implement the generalized eigenvalue problem in mpmath. I have read the relevant passages in some books, looked at my old attempts...

Concerning flamp: On second thought it would be nice to have something like flamp (numpy arrays filled with mpmath or gmpy2 objects) directly inside mpmath. There is certainly a need...

Hi. I have cherry-picked the fix. Edit: the routines which failed pytest-3 are test_quad.py::test_expmath_fail test_matrices.py::test_interval_matrix_matrix_mult

more code coverage tests added