fixed amsoptions.java to use default localhost
amsoptions code now handles all options and snes ex19 runs to completion
still some glitches, cannot handle changing type.

Hg-commit: 3a2988280069fa75d

fixed amsoptions.java to use default localhost
amsoptions code now handles all options and snes ex19 runs to completion
still some glitches, cannot handle changing type.

Hg-commit: 3a2988280069fa75d8bacfad2ac6478e41eb2d3d

show more ...

PCASMSetTotalSubdomains_ASM: prototype may as well match PCASMSetTotalSubdomains

Hg-commit: 78c71b6654af44f1abf89abd4c1ca6132d20e1e3

asm: VecSet -> VecZeroEntries

Hg-commit: 927e7cb3b6d73bd2a9daace953734152dd328b03

PCASMCreateSubdomains2D: provide both overlapping and nonoverlapping index sets (is and is_local)

* Indicate user-specified overlap with osm->overlap == -1
* Clean up documentation and reporting of

PCASMCreateSubdomains2D: provide both overlapping and nonoverlapping index sets (is and is_local)

* Indicate user-specified overlap with osm->overlap == -1
* Clean up documentation and reporting of overlap for user-specified overlap
* Don't increase overlap for user-specified subdomains

Hg-commit: e331897b281528e947bf569a03879dee1bdcfe5d

show more ...

changed PCILU to automatically switch to MAT_SHIFT_INBLOCKS for block matrices since they do not support nonzero shift.

Hg-commit: 37e323e1d2fdb97f8b6f379ca0d46554880d4efe

clarify why something is a macro MatSeqXAIJReallocateAIJ
make ILU shift default to shift nonzero NOT shift in block

Hg-commit: caed971d3165d970d5d268727e8913640ff687c4

bugfix of shift_pd for inode implementation

Hg-commit: b89092042bcc53148820ab79013a23a5c36160a9

update output files of nightly tests after replacing the default shift_nz to shift_inblocks for baij matrix.

Hg-commit: 38ac1ff390e4eb19acc0ad034d81608cde936242

fixed MATORDERING_xxx I missed when I dropped the _ from various XXXType names

Hg-commit: e51d3588701d8722510e4edb04d52e907153c627

fixed XXXType #define names that had underscores in them such as MAT_SOLVER_SUPERLU MATORDERING_ND or PETSC_VIEWER_ASCII to not have _ also updated style guide to clearly indicate this.

Hg-commit: 9

fixed XXXType #define names that had underscores in them such as MAT_SOLVER_SUPERLU MATORDERING_ND or PETSC_VIEWER_ASCII to not have _ also updated style guide to clearly indicate this.

Hg-commit: 9ee752148eed3df65feb0f7eb0686798c3c0621b

show more ...

fix order of reference/destroy calls in {TS|SNES|KSP|PC}SetDM()

Hg-commit: ff67b5cd67740f40d467711c8533b2f67124eb50

replace ilu default shifttype SHIFT_NONZERO to SHIFT_INBLOCKS because
shift_nonzero is not supported for baij matrices with bs>1.

Hg-commit: 37f8548c89a6fe45f638d5fe8955616b04085d1f

PCFieldSplit: remove command line versus code conflict

* Factor out PCFieldSplitSetRuntimeSplits_Private()

* Splits defined using the options database become permanent, subsequent calls to
PCFiel

PCFieldSplit: remove command line versus code conflict

* Factor out PCFieldSplitSetRuntimeSplits_Private()

* Splits defined using the options database become permanent, subsequent calls to
PCFieldSplitSet(IS|Fields) will not create new splits. This is a problem that
may be slightly confusing, so such cases are logged with PetscInfo.

* Add tests of setting options with both code-defined (string) and options
database (numeric) prefixes.

* Always use PetscSNPrintf() for creating prefixes (note that it now involves
an additional user-defined string)

Hg-commit: 978e0f333fff90b0c3226a36fdc401b53b28895e

show more ...

PCFieldSplitSetFields const correctness

Hg-commit: a5b881ca83b895cb5aaf65364156d8af6aa895d9

PCFieldSplitSetIS() and PCFieldSplitSetFields() now take a split-name parameter.

This is used to define options database keys for controlling the inner solves.

This patch also removes a fragment of

PCFieldSplitSetIS() and PCFieldSplitSetFields() now take a split-name parameter.

This is used to define options database keys for controlling the inner solves.

This patch also removes a fragment of dead code in PCFieldSplitSetDefaults()

Hg-commit: 1e0766c47496eba21827219418d039be3b470b45

show more ...

PCSetOperators() will not generate error if passing in null operators.
added Note to KSPSetOperators() that you cannot change the size of the matrices with future calls

Hg-commit: 7d226c8ba617eca0d3

PCSetOperators() will not generate error if passing in null operators.
added Note to KSPSetOperators() that you cannot change the size of the matrices with future calls

Hg-commit: 7d226c8ba617eca0d310f16a646ad02c07f6af91

show more ...

do not allow setting a new PCSetOperator() with a different size after it has been used
From: Matthew Knepley <knepley@gmail.com>
Subject: [petsc-maint #46134] Re: [petsc-users] Error usin

do not allow setting a new PCSetOperator() with a different size after it has been used
From: Matthew Knepley <knepley@gmail.com>
Subject: [petsc-maint #46134] Re: [petsc-users] Error using the same ksp with different operator sizes
Date: May 13, 2010 9:57:22 AM CDT
To: Aron Ahmadia <aron.ahmadia@kaust.edu.sa>, petsc-maint Maintenance <petsc-maint@mcs.anl.gov>
Reply-To: petsc-maint Maintenance <petsc-maint@mcs.anl.gov>, Matthew Knepley <knepley@gmail.com>
On Thu, May 13, 2010 at 10:44 AM, Aron Ahmadia <aron.ahmadia@kaust.edu.sa>wrote:

In that case, shouldn't you throw the error when they try to set the new
operator?

Hg-commit: afd8a6ab1b20bc643b64afec12cd81f0ec325471

show more ...

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 tears

Hg-commit: 9e0ef782022dbdc055a74473b60f135ee06122ac

removed unneeded { around SETERRQX() calls}

Hg-commit: d55148dbbe8a10f500a7e181b91ba166aec6f665

merge from release

Hg-commit: 0a243b58bcc67382f0974df4db9343f8c7c5e794

commit after merge

Hg-commit: a943693fc1dc75fc6929b1e3c8d2179e228e3223

changed many calls to SETERRQ() to use correct (non-PETSC_COMM_SELF) communicator
reformated 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) communicator
reformated many badly formatted SETERRQ(); there was nonsense like {} around them or they were wrongly on a seperate line

Hg-commit: 7550f3779f18eaf7ce68bae52b964532aa97c1d5

show more ...

Two SETERRQs not given communicator, use PC communicator since these are options database compatibility checks

Hg-commit: 03b0678a4b1913290be970855566dd31cbcaef51

add patch contributed by lawrence.mitchell@ed.ac.uk (see [petsc-maint #45870] PCSetUp_ML needs to set communicator )

Hg-commit: 00a52a8194d242ac0ff263979c726708bd2a7865