Equivalent of "spparms" to track algorithm choice for dense matrices ?

Hello,

I would like to solve a linear system A.x = b using the mldivide function. Since I don't completely know the characteristics of the A matrix (modified in a for loop) the scheme given on the help of mldivide is not sufficient.

Thus, I would like to track which algorithm MATLAB is using to solve the inversion by plotting it on my command window. I found a function that allows this algorithm's tracking for sparse matrices ssparms and I am looking for the exact same thing but for dense matrices.

Does anyone have an idea ?

Answer 1

[Steven Lord]

Why do you need information about what the mldivide function (or the backslash operator \) is doing? How do you intend to use that information?

The decomposition object may be of interest to you, but it does not give as much information as spparms does.

[ELSTER NATACHA]

According to the mldivide algorithm's scheme the matrix I constructed is supposed to be solved using the LDL solver.

Indeed, I know for sure that the matrix A is dense, square and hermitian (checked with the function ishermitian). However on its diagonal its values can be negative and positive, so it checks the conditions for the LDL solver.

Using the decomposition object I imposed the solver as being LDL but I did not find the same results as with the regular mldivide function. After that I set the LU solver and it gave me the same results than the mldivide so I'm wondering why my hermitian condition is not taken into account.