On Tue, Jun 3, 2014 at 1:13 PM, Chris Foster <chris...@gmail.com> wrote: > So you can compute the real part L of the Cholesky decomposition exactly as > usual. Given that you now have L, you want the lower triangular matrix M. > Because L and M are lower triangular that's actually quite easy: matrix > elements of M can be computed from the top left to bottom right in a process > that basically looks like forward substitution.
I implemented this for the heck of it, and it seems to work. Here's the code: https://gist.github.com/c42f/40356c153e90ff5071bc Unfortunately it seems pretty slow as written but perhaps you could test it on your system for comparison? It would be nice to do those dot products without allocating any temporary arrays but brief fiddling with Base.BLAS.dot only got me as far as a segfault so far. ~Chris
