| Re: [eigen] Alternative SVD bidiagonalization |
[ Thread Index |
Date Index
| More lists.tuxfamily.org/eigen Archives
]
- To: eigen@xxxxxxxxxxxxxxxxxxx
- Subject: Re: [eigen] Alternative SVD bidiagonalization
- From: Benoit Jacob <jacob.benoit.1@xxxxxxxxx>
- Date: Thu, 5 Feb 2009 19:17:07 +0100
- Dkim-signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=gamma; h=domainkey-signature:mime-version:received:in-reply-to:references :date:message-id:subject:from:to:content-type :content-transfer-encoding; bh=xx+tbFW2KZBVSkrogjzjfs77f4nLMfM3BuwB3vIT2gA=; b=BlRWGI30bdlHSBHy7lQKuqW84W93dUkm3SXN1x63DTF6F9C5PLfAycmJBXChBqA7BT jR+xLTzdiUP+gl6EALUgV2+g+1Z3KOmi3HVNMI6tWs85nTLU9/Sld23jKwAI7E+z4289 i5TT97mibr4L/U+TT8kn8CidFnVEtlG9D44Qk=
- Domainkey-signature: a=rsa-sha1; c=nofws; d=gmail.com; s=gamma; h=mime-version:in-reply-to:references:date:message-id:subject:from:to :content-type:content-transfer-encoding; b=DpYnnJdvKrWgDe9QdXt/WxqTUVbNp5EjoBhN26FmS1to6xWF8Qozjiy9i5zTQ3/4SA FwqYwOW8gyo4PsPh/Fsrc5FaMXStHSmZqw+5dqqqAZymioJDKk5IThUK8OH6JgQ+QrI2 Cq/FeHl7+ro5fjRDvXc5HtDMySzyqExdo/Qac=
Interesting. Just FYI, see §8.5.4 in Golub & van Loan, it's not
exactly the same since it's diagonalizing a selfadjoint tridiagonal
matrix, but it's very similar to the problem of SVD'ing a bidiagonal
matrix. They also show a divide and conquer approach.
Cheers,
Benoit
2009/2/5 Keir Mierle <mierle@xxxxxxxxx>:
> Apparently this is DGESDD or DGESSD that is in LAPACK and is now the
> preferred method for computing a SVD. It uses some sort of divide and
> conquer for computing the SVD of a bidagonal matrix instead of QR
> iterations.
>
> http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.41.5864
>