News for hackers. Stuff that matters.

A comment on Outperforming LAPACK with C++ metaprogramming. Parent comment

Re:Outperforming LAPACK with C++ metaprogramming by sevensor on Monday January 29, 2018 @09:53AM (#16257795)
Personally, my biggest reason to prefer LAPACK in general is that its authors have already put a great deal of effort into correctness and numerical stability, so I don't have to. Even basic LAPACK is pretty fast, let alone the optimized libraries. Hand-optimizing my own special case is an absolute last resort.
[ Parent | on HN ]
Re:Outperforming LAPACK with C++ metaprogramming by pletnes on Monday January 29, 2018 @11:01AM (#16258212)
Yes, Cramer’s rule (referred to in the article) does not give numerically stable results AFAIK - which is yet another can of worms. Good catch.
[ Parent | on HN ]
Re:Outperforming LAPACK with C++ metaprogramming by thanatropism on Monday January 29, 2018 @12:06PM (#16258765)
I don't think I have heard Cramer's rule mentioned without the caveat "but don't use this for real problems".
[ Parent | on HN ]
Re:Outperforming LAPACK with C++ metaprogramming by pletnes on Tuesday January 30, 2018 @10:50AM (#16266204)
It’s useful for certain theoretical proofs in analytic math - or so I’ve heard. Not all math is well suited to be coded into reliable software.
[ Parent | on HN ]