Homepage Books Releases Docs Discourse

Higher-order time (and space) integration

Hi everybody

First, what’s the current support for higher-order time solution (possibly 3rd order)?

Reading the documentation I found the linear_multi_step tag, but there are currently no infos on how to set it up. Referring to example
Tests/Data/Parabolic/ComponentTransport/ReactiveTransport/calcite.prj
where should the tag be put? I guess at the end amongst the solver variables, but then in the time_loop section (lines 174-355) all the variables involved have BackwardEuler as time_discretization parameter.
Should I (re-)compile with some additional Cmake directive? Such as -DOGS_USE_CVODE=1.

As for solving in space, does integration_order equal to 2 to enable higher-order solution in space, as in line 13 of the same calcite.prj above? Did I get it correctly?

Thank you in advance