| Re: [eigen] symmetric matrix |
[ Thread Index |
Date Index
| More lists.tuxfamily.org/eigen Archives
]
- To: eigen@xxxxxxxxxxxxxxxxxxx
- Subject: Re: [eigen] symmetric matrix
- From: Gael Guennebaud <gael.guennebaud@xxxxxxxxx>
- Date: Thu, 9 Feb 2012 22:20:42 +0100
- Dkim-signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=gamma; h=mime-version:in-reply-to:references:from:date:message-id:subject:to :content-type:content-transfer-encoding; bh=4VOOJNXEDuBNG/7vbTzg0g/ETDviUMLEgqXLumaGZGw=; b=uPywFn3zM13xnB4le1WdqUxWu7wNeJSZmHkJJ8rzmO8ISBFFVMboOyR9xbeGuVfG5f 2EJTC/ciLdShOkccdJqVlBX566T7VQc0O5m0G+CI3VLo/S2rfLJ2QmM829K/klusSzZW 55vwyNyIrOSY3NSgibQLKm3O4rnqWdrbQcsq4=
Hi,
there is no compact symmetric storage if that's what you were looking
for. On the other hand the selfadjoint-view allows you to do many
optimized symmetric operations by working on an half dense matrix
(products, rank updates, cholesky factorization, etc.). You can also
pack two symmetric matrices into a single dense one to save some
bytes.... If you don't want to write .selfadjointView<Lower>() every
time you can name a view:
MatrixXd mat(n,n);
SelfadjointView<MatrixXd,Lower> sym(mat);
sym.rankUpdate(v),
....
cheers,
gael
On Thu, Feb 9, 2012 at 8:42 PM, Christopher Petrovitch
<cpetrovi@xxxxxxxxxx> wrote:
> Hello,
>
> I noticed that there isn't a symmetric matrix class in any of the documentation (or doxygen docs) and was wondering if I was just missing it or if there is no symmetric class? I saw there was a self-adjoint view however... What is the difference between a view and the actual matrix type class?
>
>
> Thanks,
> Chris
>