Fix ratios when combining -da_refine_hierarchy_[xyz] with -da_refineHg-commit: 4dc457ce830894ca4b8a4bc350094f4f669e1a8f
Split DMCreateDecomposition() into Field and Domain decompositions. Reenabled use of DM in ASM.Hg-commit: 2e2ee6a01d1a543670ea1a565999371852551596
Splitting up DM decomposition into field and subdomain decompositions.Hg-commit: 25ae1465cd37db36ff0f764d65ffd3f1bf1fc0cc
Make default coarsening match refinementHg-commit: 971971e3008bdad75b79c007cfc118a06308cc17
removed outdated SNESDMDACompute functions except in 1 example Peter will fixHg-commit: aabf823a9e38cebb8b847b8d026d77b3a6a44a5b
Bug fix in DMDAGetReducedDAHg-commit: 6240a18647e1132da52a0bd94a9b8a72d39eafcb
Added new FEM functions for DA- Remove requirement that coordinates for DA have a given blocksize- DMDASetUniformCoordinates() now works with PetscSection version- Added dageometry.c which has DMD
Added new FEM functions for DA- Remove requirement that coordinates for DA have a given blocksize- DMDASetUniformCoordinates() now works with PetscSection version- Added dageometry.c which has DMDAVecGet/SetClosure()Hg-commit: c4a59f6551ea5ba42e66e722270f8b4772c42349
show more ...
dmrefine and coarsen for DA must not change the N and P for lower dimensional problems since the other code depends on them always being 1curse whomever is assuming their changes are breaking thing
dmrefine and coarsen for DA must not change the N and P for lower dimensional problems since the other code depends on them always being 1curse whomever is assuming their changes are breaking things and don't run make alltests before pushesHg-commit: c968c7d453ecc319668d5ac88c17fd75f0269375
Fix DMRefine_DA and DMCoarsen_DA: DMSetUp() must be called before DMDAGetCoordinateDA()Hg-commit: 551cf33f734cbe22fbe6fe1cfaab3e1295998608
fix unused variable warningsHg-commit: e1d588309ed693445a1a0abe7433fe7bda885ec4
Initialize dmda->coarsen_{x,y,z}Hg-commit: fe5d4f67cdc3d9d87b6818f863f6510c8428c00c
Fix gfortran version in man pageHg-commit: c46fd41f629077de789519c77473879672f17916
Distinguish da->coarsen_{x,y,z} from da->refine_{x,y,z} and mostly make -da_refine_hierarchy_{x,y,z} work without DMMGHg-commit: 9943253a5ac277e1008b2c957cb962101a7c4f64
Port field names in DMCoarsen_DA, as is done by DMRefine_DAHg-commit: 4496ca57d9ac5d09c6a1f280573e42b11204e477
DM CHANGE: DMCreateFieldIS() and DMCreateLocalToGlobalMapping() now use the default PetscSection- DMComplex uses this- DMDA now uses section for DMCreateMatrix(), but preallocation does not workH
DM CHANGE: DMCreateFieldIS() and DMCreateLocalToGlobalMapping() now use the default PetscSection- DMComplex uses this- DMDA now uses section for DMCreateMatrix(), but preallocation does not workHg-commit: 440c9423bbc30f4bdf8dd3ca01ebe76023f766ac
Fixed for flexible DA in parallelHg-commit: 6c2e217ad8b431f9c4ef53f9649d0bfd2776de21
DM CHANGE: Now DMGlobalToLocalBegin/End() and DMLocalToGlobalBegin/End() use the default PetscSF if present- DMComplex now uses theseHg-commit: 6fdfec99b8fc8426d2b385cbe72264449a63e681
DM CHANGE: Moved PetscSections describing layout and PetscSFs describing overlap int DM- This allows unstructured layout on any DM, in particular DA- Creation flows goes localSection --> pointSF --
DM CHANGE: Moved PetscSections describing layout and PetscSFs describing overlap int DM- This allows unstructured layout on any DM, in particular DA- Creation flows goes localSection --> pointSF --> globalSection --> dofSF- DMGetGlobalVector() and DMGetLocalVector() use these layouts if they exist (DMComplex uses this now)- Put in DA code for creating the Section and SF- Cleaned up some memory problemsHg-commit: 612ec4062f8f069f58c4aa2b66de5ea016ab8a63
PetscTypeCompare -> PetscObjectTypeCompareHg-commit: 3fc8f665421ddb862bf4d1ba95505378246340be
Added ex67 where a different layout is used with a DAHg-commit: 51435c26afcf8c9b770276864ab7aff7619e8770
Small fixHg-commit: 3706fa2ae8171492c7ba72f71570ff49d564932d
Added PetscSection into DMDA so that we can start to do better discretizationsHg-commit: ac71d0bd83c68110b23424a72d8727d4ca619f4a
MergeHg-commit: 7fb9acec46a6e81e7f3158a015209bcec3298b62
Added a field type to the VTK viewer- This distinction is present in the VTKHg-commit: 31f6548f16931414c7468bef0b15792a835ccac9
removed more darn {SEATERR}Hg-commit: 67fe7512e8fcb191ac49d6fe48793bfdfcb31b4b
1...<<31323334353637383940>>...48