[petsc-dev] Any substitute for -ts_theta_adapt

[Jed Brown]

Lulu Liu <lulu.liu at kaust.edu.sa> writes:

> I want to register a new adaptive time-step  scheme for the backward Euler
> method.  I used   options:
>
>  -ts_adapt_type myscheme   -ts_theta_adapt
>
> The error message is show as:
>
> No time-step adaptivity implemented for 1st order theta method;

Please show the full error trace.  Presumably you are calling
TSEvaluateStep, but if you are trying to step with a first-order method,
you can't use "zeroth order" as an estimate of the error.

> Is there any substitute for -ts_theta_adapt such that I could have my own
> time-step adaptivity for backward Euler?

Why not use '-ts_type arkimex -ts_arkimex_type 1bee', which uses an
extrapolation-based error estimator?  If you want something else, just
write it as a Runge-Kutta method like '1bee'.