Re: Calling LAPACK's dpotri from Octave ?

[Julien Bect]

Le 19/02/2020 à 11:43, Kai Torben Ohlhus a écrit :
> On 2/19/20 7:02 PM, Julien Bect wrote:
> > Hi all,
> >
> > Suppose I have a real square matrix U, double precision, upper triangular.
> >
> > I need to compute the inverse of A = U' U (in other words, U is the
> > upper-triangular Cholesky factor of A).
> >
> > Is there a way to do so efficiently from Octave, for instance by calling
> > LAPACK's dpotri procedure ?
> >
> > @++
> > Julien
> Some "grep" in the Octave sources [2,3,4] shows, that "chol2inv" [1]
> should almost directly call "dpotri", of course with some dimension and
> type checking.
>
> [1] https://octave.org/doc/v5.2.0/XREFchol2inv.html

Oh, thank you very much !

I don't know why I didn't find chol2inv before, sorry for the noise...