Re: [ublas] lu_factorize failing

From: Paul C. Leopardi (paul.leopardi_at_[hidden])
Date: 2007-09-06 06:46:04

Hi Rick,
Did you compile with -DNDEBUG ?
Best, Paul
On Thu, 6 Sep 2007, Rick Harding wrote:
> Hi,
>
> Using lu_factorize() on the following matrix will cause an assertion
> to be thrown:
>
> A=
> [8,8](
> (138.625,14.2945,.00134662,-.00731894,-.00731894,-.00731894,2.717e-018,-.01
>5625),
> (14.2945,5.24033,-0.0193141,0.00646566,0.00646566,0.00646566,0,0.125),
> (0.00134662,-0.0193141,47.8208,0.0470879,0.0470879,0.0470879,0,-0.0078125),
> (-0.00731894,0.00646566,0.0470879,0.477962,0,0,0,0),
> (-0.00731894,0.00646566,0.0470879,0,0.477962,0,0,0),
> (-0.00731894,0.00646566,0.0470879,0,0,0.477962,0,0),
> (2.717e-018,0,0,0,0,0,0.477962,0),
> (-0.015625,0.125,-0.0078125,0,0,0,0,0)
> )
>
> Assertion failed in file C:\boost\numeric\ublas\lu.hpp at line 148:
> pm (i) == i_norm_inf
>
> The matrix is not singular. Have I done something wrong or is there a
> problem in lu_factorize()?
>
> I'm using uBLAS found in Boost 1.33.1.
>
> Thanks!
>
> _______________________________________________
> ublas mailing list
> ublas_at_[hidden]
> http://lists.boost.org/mailman/listinfo.cgi/ublas