commit after merge

Hg-commit: f913cb65dfa3c5b3ff4f42d6f799542c1b5d257f

MPIAIJ forwards block size to SeqAIJ parts

Hg-commit: 0bb703470caafade5efbe04a93d90a41f6d33ed2

commit after merge

Hg-commit: 385772f7a560d67dae5108e5cd455273b99579fc

fixed inheritance of block size on coarse grids in GAMG. Fixed bugs in VecScatterCreate wrt block sizes. Added some debug code.

Hg-commit: fb7e4343253b73d3fb40857991a8c6d5d0c7c56b

Lots of changes. Also addign new files for cusparse matrix type.

Hg-commit: d2f118b395b24fdc836f857e3f221f6c0c9eca58

Added block size inheritence in Mat methods.

Dissable a block IS optimization in VecScatterCreate that was wrong with blocks from Mat mult methods.

Added some debug code for a user.

Refactored GAM

Added block size inheritence in Mat methods.

Dissable a block IS optimization in VecScatterCreate that was wrong with blocks from Mat mult methods.

Added some debug code for a user.

Refactored GAMG for KKT solves.

N.B., there are a lot of changes here, sorry to dump this just before a release ...

Hg-commit: a85625d104df12863d7d1735f9fdbe9b281ef7d2

show more ...

PetscTypeCompare -> PetscObjectTypeCompare

Hg-commit: 3fc8f665421ddb862bf4d1ba95505378246340be

commit after merge

Hg-commit: 719627a6cee54b4367ab263a30985c06b3d9399d

Fix MatAXPY and MatAYPX preallocation for MPIAIJ; off-diagonal indices were not being translated

Hg-commit: 7ac7040f19d7b1dd20c78e714bba73ee3265ca9f

added use of BLASDotu() where appropriate, fixes for complex where original author didn't handle it. Fixed wrong use of bblas in maxpy for threads

Hg-commit: 9cf34ccedfe2cc536bab398b0abcbaa633e5a615

Commit after merge

Hg-commit: dd9de3c9b6c26331f5d70e21b5ec7304b66251d1

Removed unneeded setting of Mat Block size, should respect already set value, removed unneeded checking of negative blocksize this is handled in PetscLayoutSetUp().

Hg-commit: 2d1c2b03ef3a0925025555

Removed unneeded setting of Mat Block size, should respect already set value, removed unneeded checking of negative blocksize this is handled in PetscLayoutSetUp().

Hg-commit: 2d1c2b03ef3a0925025555e660e6834a4bd0a50c

show more ...

MatSetSize() called after MatSetUp() is now forbidden even for Dense matrices
MatSetBlockSize() must be called beofre MatSetUp()/ MatXXXSetPreallocation() for all matrices
Fixed vec nest example that

MatSetSize() called after MatSetUp() is now forbidden even for Dense matrices
MatSetBlockSize() must be called beofre MatSetUp()/ MatXXXSetPreallocation() for all matrices
Fixed vec nest example that pushed format that prevented values from being printed

Hg-commit: 80ec81eb9db92735994e950eda2d2db6657bb6cb

show more ...

namespaced some Mat utilities that lacked the Mat

Hg-commit: 6f712ebe4bb5ddbcbf32ebd864d408e3a35085b0

properly namespaced a couple more things

Hg-commit: 71672d5f24e597f91ca620cb4b297ebd71880c5a

commit after merge

Hg-commit: 8bbde4c6a68210736fec1b806d9ca592257e9fc9

Renamed MatMerge_SeqsToMPI() to MatCreateMPIAIJSumSeqAIJ(); Renamed MatMerge() to MatCreateMPIAIJConcatenateSeqAIJ()

Hg-commit: 24f46f7b98f555b1e552573a5bd3ac3737ff4f2b

Merge

Hg-commit: f997b2abc12025b16065184c2bb279689bbd06ac

Avoid PETSC_COMM_WORLD

Hg-commit: 48c7d720fe4c345dfc0ff96316c713b92508bf36

Rename include/private/ to include/petsc-private/

Hg-commit: 9a1730720b0d9756437b40462617551504641ade

Undid this change since it breaks stuff. Note that the entire MatAssembly Infrastructure depends on these flags being set correctly. I just wasted hours tracking down a bug due to this change. Also n

Undid this change since it breaks stuff. Note that the entire MatAssembly Infrastructure depends on these flags being set correctly. I just wasted hours tracking down a bug due to this change. Also note that our test system sucks since the bug was introduced a month ago.

Note also that assembled flag is MUCH more than just "values need to be communicated". The business of garray and colmap depend on this flag being set correctly.

Note also that all the other MatSetXXX still changed the flags correctly.

If you want some functionality like below that allows changing numerical values (but not nonzero locations) then introduce MatGetRowWrite() and MatRestoreRowWrite() and change MatGetRow() to MatGetRowRead(). But please don't don't underestimate side affects when you make such dramatic changes.

author Matthew G Knepley <knepley@gmail.com>
Tue, 14 Feb 2012 00:45:05 -0600 (4 weeks ago)
changeset 22179 1fc53ac2b8ab
parent 22178 c1d3e1b5603f
child 22180 708aa4809612
Now MatSetValues() does not automatically mark the matrix as unassembled
- Individual types set the assembly flag if values need to be communicated
- This is done to allow programming in the following way:
for(r = rStart; r < rEnd; ++r) {
MatGetRow()
<transform values>
MatRestoreRow()
MatSetValues()
}

Hg-commit: 7ae47bfbd6e9aa96e390d0fe01d428bae6c9eac4

show more ...

commit after merge

Hg-commit: 4f48dde03429d0318c7e23bc7de128a9e6ddc8fd

merge SNES/DM refactor

Hg-commit: 31d66a4917111e3b9e87bf19170c2c9421b4e8be

bugfix for MatMerge_SeqsToMPI(): use MatDuplicate/Copy() when np=1; call MatSetOption(mat,MAT_NEW_NONZERO_ALLOCATION_ERR,PETSC_FALSE).

Hg-commit: 302d2ee020f4889bc6f6235e3950b18af318832c

changed MatCreateMPIXXX() to MatCreateXXX() since that is what they are

Hg-commit: 88fb4066e05405d3c444fb00dd00ab15e805b3e3