Hankel method is rather slow

[martin-ueding]

I am currently waiting for all correlators of two ensembles to be processed through the fixed Hankel method. It will take over an hour. I would think that perhaps this could benefit from being implemented in Rcpp at some point.

[urbach]

I know why it is slow for fixed t0: the matrix is inverted over and over again. This could be improved, but for the cost of some generality.

Using Rcpp would help, of course. How urgent is this?

[martin-ueding]

It's not at all urgent. It is more of an idea with the importance of “would be nice”.