This is the mail archive of the
gsl-discuss@sources.redhat.com
mailing list for the GSL project.
Re: gsl_linalg_SV_decomp goes into infinite loop on reasonable data
- From: Brian Gough <bjg at network-theory dot co dot uk>
- To: tlb at trevorblackwell dot com
- Cc: gsl-discuss at sources dot redhat dot com
- Date: Thu, 13 Jun 2002 18:48:36 +0100 (BST)
- Subject: Re: gsl_linalg_SV_decomp goes into infinite loop on reasonable data
- References: <200206130656.g5D6uB597950@tlb.blackwells.net>
tlb@trevorblackwell.com writes:
>
> On my system, FreeBSD 4.5-RELEASE-p2, gsl_linalg_SV_decomp gets stuck
> in an infinite loop trying to reduce the matrix to diagonal. The test:
>
> <linalg/svd.c:103>
> if (gsl_vector_get (&f.vector, b - 1) == 0.0)
>
> repeatedly fails so the loop it's in never terminates. The value it's
> getting is 1.83429e-07, noticeably more than DBL_EPS. (I'm using
> double throughout.) In the chop_small_elements code, the line which is
> supposed to remove these small values:
>
Hi,
Thanks for the bug report, I see the same. I'll look into it.
regards
Brian Gough