new cGH field cctk_patch

[Roland Haas]

Cactus currently does not contain a easy way to find out which patch (“map”) one is on for multipatch (Llama) simulations, instead requiring the use of the aliased function GetMap provided by Carpet.

[Roland Haas]

Pull request is here: https://bitbucket.org/cactuscode/cactus/pull-requests/135/cactus-new-cgh-field-cctk_patch

[Roland Haas]

Please review.

[Roland Haas]

Please apply.

Requests:

• document cctk_patch in the UsersGuide
• in the UsersGuide replace all use of “patch” meaning “disjoint grid patch” by “component” to avoid using “patch” for 2 different things (at least in the docs. Carpet still uses “patch” for “component” eg in CarpetIOHDF5’s checkpoint file handling code).
• add code to PUGH / the flesh to set cctk_patch to 0 by default

‌

[Roland Haas]

Note that to find out the total number of patches (maps) one still has to call the aliased functon GetMaps. Not sure if there is a legit use for that information though (there is also not official way to find out the total number of components [not MPI ranks]). Also note that the assumption cctk_patch == 0 → Cartesian patch is most likely not true for eg the 6 patch cubed sphere system (which has no Cartesian block in the center). The function to query this would be MultiPatch_MapIsCartesian in the Coordinates thorn (which only requires knowledge of cctk_patch and of course calls it map).

[Roland Haas]

Applied as git hash f767e6fc "Cactus: initialize cctk_patch and cctk_npatches to single-block values" of cactus