Re: [Libmesh-users] Newton iteration nonlinear elasticity slow

[John P. <jwp...@gm...>]

On Tue, Dec 10, 2013 at 8:58 AM, Lorenzo Zanon
<za...@ai...> wrote:
> Hi,
>
> After re-installing petsc with Hypre should I recompile libmesh as well?

Definitely.

> Otherwise here are the detailed outputs (JFNK, LU, and -ksp_rtol 1e-4):
>
>
> NL step  0, |residual|_2 = 3.464102e-05
>     0 KSP Residual norm 3.212616421517e-03
...
>    27 KSP Residual norm 5.868687294700e-08
>    28 KSP Residual norm 8.654056711463e-09
>   NL step  1, |residual|_2 = 2.417677e-04

Here you can see you're "oversolving" the linear system quite a bit.


>   NL step  2, |residual|_2 = 6.140305e-08
>     0 KSP Residual norm 4.390107230670e-07
...
>    28 KSP Residual norm 3.523137533760e-12

Here, near the root, you possibly could have reached your nl_tol by
using a smaller ksp_rtol and avoided doing another Newton step

-snes_ksp_ew will theoretically help with both of these issues.

-- 
John