Re: [petsc-users] Best way to save entire solution history of a TS

[Jed Brown]

Miguel Angel Salazar de Troya <salazardetr...@gmail.com> writes:

> Hello
>
> I'm performing an adjoint sensitivity analysis in a transient problem and I
> need to save the entire solution history. I don't have many degrees of
> freedom or time steps so I should be able to avoid memory problems. I've
> thought of creating a std::vector<Vec> and save all the PETSc Vec that I
> get from TSGetSolution there. Would this be a problem if I'm running the
> simulation in parallel? Is there a better way to save the entire solution
> history in memory?

Your suggestion is fine.  I would do it from a monitor (TSMonitorSet)
and VecCopy the solution vector into your array of solutions.

pgpVLSS3LNgMY.pgp
Description: PGP signature