#2446: Cactus: New cGH fields tile_min, tile_max
| Reporter: | Erik Schnetter |
| Status: | open |
| Milestone: | |
| Version: | |
| Type: | proposal |
| Priority: | major |
| Component: | Cactus |
Comment (by Roland Haas):
I am afraid this may require more work. Sorry for the piecemeal review (though some of the comments were present before):
omp singleclause by removing the copyprivate(cctki3_bndsize, cctki3_is_physbnd) which will leave cctki3_bndsize uninitialized in the threads that do not execute the single clause it seems since those arrays are automatic, private variables. They could be made static though it seems, which hopefully won’t make them threadprivate.GetBoundarySizesAndTypes should not be duplicated. Thorns using CCTK_LOOPX_INT must themselves USE or REQUIRE the aliased function at which point a prototype will be available when the macro is expandedthe
```c
``
* does not seem to be required since bothcctk_Loop.handcctk_GroupsOnGH.h` are in the same repository so will be kept in sync by git
* no support for Fortran is provided (F90 is still in use, see http://lists.einsteintoolkit.org/pipermail/users/2020-February/007298.html)
--
Ticket URL: https://bitbucket.org/einsteintoolkit/tickets/issues/2446/cactus-new-cgh-fields-tile_min-tile_max