Re: [sage-devel] Re: Matrices over QQbar

Thank you for the Flint reference, I will check it out, and thanks again for your hard work debugging! Best, Håkan On Mon, May 6, 2024 at 9:57 AM Vincent Delecroix <20100.delecr...@gmail.com> wrote: > The fix was easy. Much harder was to find the root of the problem! > > If you are interested i

Re: [sage-devel] Re: Matrices over QQbar

The fix was easy. Much harder was to find the root of the problem! If you are interested in intensive computations over QQbar I strongly advise you to look towards https://flintlib.org/ maintained by Fredrik Johansson. This library has *two* implementations of QQbar which are way more performant (

Re: [sage-devel] Re: Matrices over QQbar

Great, thanks for working out and fixing the (quite non-trivial, it seems) root cause of this issue! Just a curious question: would it be useful to add the functionality of your commit 8de6ba5 to potentially simplif

Re: [sage-devel] Re: Matrices over QQbar

Turns out to be a serious bug in complex interval fields. The problem is hopefully fixed by https://github.com/sagemath/sage/pull/37941 which should make its way to the next sage release. Thanks again for your report. Vincent On Fri, 3 May 2024 at 18:13, vdelecroix <20100.delecr...@gmail.com> wro

[sage-devel] Re: Matrices over QQbar

Thanks for your report! I simplified a bit your example and posted an issue https://github.com/sagemath/sage/issues/37927. Vincent Le vendredi 3 mai 2024 à 15:05:03 UTC+2, Hakan Granath a écrit : > Hi, > > I think sometimes matrices over QQbar give erroneous results (sorry for > the messy exam