MathGroup Archive 2009

[Date Index] [Thread Index] [Author Index]

Search the Archive

Re: Eigenvalue bug in Mathematica 7.0.0

  • To: mathgroup at smc.vnet.net
  • Subject: [mg101233] Re: Eigenvalue bug in Mathematica 7.0.0
  • From: ADL <alberto.dilullo at tiscali.it>
  • Date: Sat, 27 Jun 2009 06:05:08 -0400 (EDT)
  • References: <h22974$hqb$1@smc.vnet.net>

Mark, with version 7.0 for Microsoft Windows (32-bit) (February 18,
2009), I get this:

{eivals, eivecs} = Eigensystem[H];

In[]:= And @@ Table[
  Chop[Norm[H.eivecs[[i]] - eivals[[i]] eivecs[[i]]]] == 0,
  {i, Length[eivals]}
  ]

Out[]= True

So I think the result is correct.

ADL

On Jun 26, 12:51 pm, Mark Saffman <msaff... at wisc.edu> wrote:
> I find that Mathematica V7.0.0 finds incorrect eigenvalues and eigenvecto=
rs of the 100 x 100 real symmetric matrix listed below. You can easily chec=
k the first eigenvector produced is not a true eigenvector.
>
> I get incorrect results running under
> Windows Vista 64 bit, Windows Server 2003 32 bit, Windows Server 2003 64 =
bit. I get correct results running under Windows Vista 32 bit. All of the a=
bove systems produce correct results if I evaluate with Mathematica V6.0.1.=
0.
>
> I consider this a very serious error. Here is the
> 100 x 100 matrix:
> ....



  • Prev by Date: Re: display of graphics primitives
  • Next by Date: Re: Simple question about named variables.
  • Previous by thread: Re: Eigenvalue bug in Mathematica 7.0.0
  • Next by thread: Re: Eigenvalue bug in Mathematica 7.0.0