Multiprecision algorithms for computing the matrix logarithm

Computation of the matrix logarithm in multiprecision.

This code computes the logarithm and the Frechét derivative of a square matrix with non-negative eigenvalues in multiple precision. Both functions require the Advanpix Multiprecision Computing Toolbox for MATLAB (see www.advanpix.com) to work correctly.
The submission contains two functions:
* LOGM_MP computes the matrix logarithm and its Frechét derivative by using the inverse scaling and squaring algorithm and exploiting the complex Schur decomposition of the matrix.
* LOGM_MP_FULL computes the matrix logarithm without transforming the input matrix to upper triangular form.
A detailed description of the algorithms can be found in:
M. Fasi and N. J. Higham, Multiprecision Algorithms for Computing the Matrix Logarithm. Technical Report 2017.16, Manchester Institute for Mathematical Sciences, The University of Manchester, UK, May 2017.