[ Thread Index |
Date Index
| More lists.tuxfamily.org/eigen Archives
]
- To: eigen@xxxxxxxxxxxxxxxxxxx
- Subject: Re: [eigen] SVD Bug
- From: Benoit Jacob <jacob.benoit.1@xxxxxxxxx>
- Date: Thu, 23 Sep 2010 07:43:58 -0400
- Dkim-signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=gamma; h=domainkey-signature:mime-version:received:received:in-reply-to :references:date:message-id:subject:from:to:content-type; bh=8d9irLhalbrxJeKuWAFcp8ElHseoHHb2meEjUKmK4Y0=; b=KLmY4vrH1h+35In96UyQ3IqnfYVtopL5neVL4yJQEIumZ7KOBmEbOx2QSavIpyVPRQ rtm/IIoYH5tVCbCoKfsLvuW5LEcNWXHlSZ/2TS87JGQIi4Qi+g5RkbJzauinjKs3XYx2 8wfmT/tqHGDnK+d1qUhu89Fm+plO0b6hlC2W0=
- 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; b=ZP750QeKaSmpwWKAF6g0uULmw4P7VUbASNMgFmjjbPqamIPBoFxY51u0P5UauvytaE LePBNqhnmy7ac9IYcxdiWe+jAQlFfGPH1oUQka6V/HX3nWUc0GJRhUidwrE3U5/mn62f v7dDO72g0rGGKTwfgAa/4H/oCk6k2RzrCPoBA=
Hi,
I confirm that before your patch, valgrind finds a lot of errors in
his testcase, and that they are all gone with your patch.
Now looking at the unit test...
Of course we must maintain the current SVD implementation until we
replace it with something better.
Benoit
2010/9/23 Hauke Heibel <hauke.heibel@xxxxxxxxxxxxxx>:
> Hi,
>
> the user chef_seppel has found a bug in the SVD. Here is the post
> http://forum.kde.org/viewtopic.php?f=74&t=90425. I can confirm it is a
> bug.
>
> In principle, the attached patch should fix it. I also changed the
> unit test and according to our docs
> (http://eigen.tuxfamily.org/dox/classEigen_1_1SVD.html) I would expect
> any m-by-n matrix A to be working (when m>n). But the unit test fails
> even for the most simple example where A is 3-by-2, x is 2-by-1 and b
> is 3-by-1. The failure is caused by the solve part.
>
> I tried to look into it but failed. I can only say that already the
> matrix U is strange since its last row is zero and this should not be
> the case.
>
> I know that Benoit is still working on rewriting the SVD but at least
> the bad memory access should be fixed.
>
> Cheers,
> Hauke
>