[petsc-users] DIVERGED_LS_FAILURE

[Rebecca Xuefei Yuan]

Dear Barry,

Thanks for your explanation!

R



Quoting Barry Smith <bsmith at mcs.anl.gov>:

>
>   The linear solver did not fail to converge. It says it converged.
>
>    The nonlinear solver failed because the line search failed, not   
> the linear solver. This is common when the changes are tiny, tiny   
> like they are here. You cannot expect that corrections that are the   
> same size as the round error to improve the nonlinear system. That   
> is your initial guess is so small you cannot do any better.
>
>    I will change the LS in the message to line_search for clarity in  
>  petsc-dev
>
>
>
>    Barry
>
> On Sep 10, 2010, at 9:54 AM, Rebecca Xuefei Yuan wrote:
>
>> Dear all,
>>
>> I have the convergence history as
>>
>>
>>
>>  0 SNES Function norm 1.930639214656e-14
>>    0 KSP Residual norm 1.930639214656e-14
>>    1 KSP Residual norm 1.897969082784e-14
>>    2 KSP Residual norm 1.890943705731e-14
>>    3 KSP Residual norm 1.890499250642e-14
>>    4 KSP Residual norm 1.882719263807e-14
>>    5 KSP Residual norm 1.881639616800e-14
>>    6 KSP Residual norm 1.881624492545e-14
>>    7 KSP Residual norm 1.876497311488e-14
>>    8 KSP Residual norm 1.864714377471e-14
>>    9 KSP Residual norm 1.861061216430e-14
>>   10 KSP Residual norm 1.856993815540e-14
>>   11 KSP Residual norm 1.856245191617e-14
>>   12 KSP Residual norm 1.855993722295e-14
>>   13 KSP Residual norm 1.855756795426e-14
>>   14 KSP Residual norm 1.855542358114e-14
>>   15 KSP Residual norm 1.855188622531e-14
>>   16 KSP Residual norm 1.854906019037e-14
>>   17 KSP Residual norm 1.851397858397e-14
>>   18 KSP Residual norm 1.851152033377e-14
>>   19 KSP Residual norm 1.845662066987e-14
>>   20 KSP Residual norm 1.810881990601e-14
>>   21 KSP Residual norm 9.715701270773e-15
>>   22 KSP Residual norm 2.134473951854e-15
>> Linear solve converged due to CONVERGED_RTOL iterations 22
>> Nonlinear solve did not converge due to DIVERGED_LS_FAILURE
>>
>> Why the linear solver is failed?
>>
>> Thanks a lot!
>>
>>
>>
>>
>>
>> Rebecca Xuefei YUAN
>> Department of Applied Physics and Applied Mathematics
>> Columbia University
>> Tel:917-399-8032
>> www.columbia.edu/~xy2102
>>
>
>
>



Rebecca Xuefei YUAN
Department of Applied Physics and Applied Mathematics
Columbia University
Tel:917-399-8032
www.columbia.edu/~xy2102