Re: Non repeatability issue

[Barry Smith <bsmith@xxxxxxxxxxx>]

  The 2d results, as noted, are reasonable and expected behavior.

  The 3d results are not acceptable, when sometimes it converges
and sometimes it does not. I would use a 1e-12 KSP rtol and a similar
1e-12 SNES rtol; does it now always converge? If not then I suspect there
may be some error in the residual (function) computation.


   Barry


On Wed, 17 Oct 2007, Aldo Bonfiglioli wrote:

> Dear users,
> in the notes attached here we address an issue concerning
> nonrepeatability.
> We know this is a known issue in parallel floating point programs,
> but we would like to be sure we are interpreting
> things correctly a not mis-using PETSc.
> 
> Comments are welcome.
> 
> Regards,
> Aldo
> 
>