More XXXDestory() fixes.Hg-commit: bc145f9b1bfad0d7f0314178ac4fa0c1b25f3a6e
Better error message for ML with non-AIJ matricesHg-commit: cb21765ba515e199ee3266d864f7b54f9eae95e3
Bad rebase of my PCReset_ML fixHg-commit: 0d0d4be6ee004c7cf30d8cc46959b66903e4917c
Fixed PCReset_ML()Hg-commit: 7eb805e6536e8cec28ea9f113282802dc036581e
Fixed PCReset_MLHg-commit: 93b1ce1b60a8268ceff90ab5d5b75796378a2467
got gfortran quad precision also working for 4.6added more pcreset for other PCsHg-commit: 9ecce33931895bae4575b370a4bd88f21e9fd656
Use #include <header.h> instead of #include "header.h" when there is no intent to search the current directory for header.hHg-commit: bd7216e80cc035071a5136364ab9d04bf9f41a07
Rename petscmg.h to petscpcmg.hHg-commit: 23be2aaeb8cd5463a8fa34e98fa9b96d440658e1
removed #define PETSCXXX_DLL from top of all source filesHg-commit: b784aa18ddcf5337d4011d34f0fa008681695ac1
further work on users manual section on block solversHg-commit: 73448a3773ada8770a10da680d636d8fe9359035
added ML_free() in MatWrapML_MPIAIJ() on gordering to prevent memory leakHg-commit: 5cfc9f44a5c1c84db14136f0df0edcf48046f93e
moved definition of PETSC_STDCALL to private/fortranimpl.h where it belongsremoved all DLL export/import stuff since it was never used and never will beremoved MatPermuteSparsify() since it sucked.
moved definition of PETSC_STDCALL to private/fortranimpl.h where it belongsremoved all DLL export/import stuff since it was never used and never will beremoved MatPermuteSparsify() since it sucked.Hg-commit: 72f7a5e60333f6c9780d56fcce45657940d16b13
show more ...
PetscOptionsTruth -> PetscOptionsBool and relatedHg-commit: dfea0d194d94180043ddc1c2d9d074a0b8717c77
Changed PetscTruth to PetscBool and PETSC_TRUTH to PETSC_BOOLnote this requires a new sowing versionHg-commit: 8b4c4277f05e3fb874633288193dab1105bf142a
Finally fixed MLHg-commit: a9fb6da56033f52b8dc6886316032032f3432557
Fixed MatMultAdd() in MLHg-commit: 37cc00fbb62c7bb5a8798e59b722e067d4d90be4
ML interface: error needs communicator from something that is in scope.Hg-commit: 69a9083af90374bb043901e5af2b379bacd0d976
the last work I will do on reformatting use of SETERRQX() and passing in the appropriate comm; I am bored to tearsHg-commit: 9e0ef782022dbdc055a74473b60f135ee06122ac
merge from releaseHg-commit: 0a243b58bcc67382f0974df4db9343f8c7c5e794
commit after mergeHg-commit: a943693fc1dc75fc6929b1e3c8d2179e228e3223
changed many calls to SETERRQ() to use correct (non-PETSC_COMM_SELF) communicatorreformated many badly formatted SETERRQ(); there was nonsense like {} around them or they were wrongly on a seperate
changed many calls to SETERRQ() to use correct (non-PETSC_COMM_SELF) communicatorreformated many badly formatted SETERRQ(); there was nonsense like {} around them or they were wrongly on a seperate lineHg-commit: 7550f3779f18eaf7ce68bae52b964532aa97c1d5
Two SETERRQs not given communicator, use PC communicator since these are options database compatibility checksHg-commit: 03b0678a4b1913290be970855566dd31cbcaef51
add patch contributed by lawrence.mitchell@ed.ac.uk (see [petsc-maint #45870] PCSetUp_ML needs to set communicator )Hg-commit: 00a52a8194d242ac0ff263979c726708bd2a7865
added MPI_Comm as first argument to PetscError() and SETERRQX()Hg-commit: fe5a2ff050abc00ca2c2979c0af22b117b62874e
Update ML interface to support energy minimization and other options.This uses many features that are not documented in the reference manual.In particular, the C interface does not seem to be supp
Update ML interface to support energy minimization and other options.This uses many features that are not documented in the reference manual.In particular, the C interface does not seem to be supported any more,the current one is based on strings and is only available fromC++/Epetra with associated baggage and certain limitations relative tothe C interface. Fortunately, the C++ interface has a nearly one-to-onecorrespondence with fields in the ML_Aggregate structure, so the presentcode sets options in that structure directly (when not available fromC++).The biggest problem with exposing this extra functionality is that MLoptions are rather non-orthogonal so it's easy to produce options setsthat are not actually functional. Rather than providing useful(user-understandable) information about what is wrong, ML calls exit(1).So if ML crashes, you have to attach a debugger with a breakpoint on'exit' just to get a stack trace (for example, to know which ML instancein your code is causing trouble). I am less than thrilled with thiserror handling strategy.Hg-commit: e721d2184f87ad8008519982dabd8955327e31f2
12345678910