Code for setting matrix data via MatSetOption/MatSetFromOptions is working.

Hg-commit: 32aee254a63d1b8a7e35b24855ebc647d510e965

Code for setting Matrix Storage types (via SetOption, SetFromOptions) appears to be working correctly. Also, I now have MatMultTranspose(Add) working for CUSPARSE type so one can do least squares kry

Code for setting Matrix Storage types (via SetOption, SetFromOptions) appears to be working correctly. Also, I now have MatMultTranspose(Add) working for CUSPARSE type so one can do least squares krylov solvers (lsqr) via PETSc on GPUs.

Hg-commit: db11b8fc0ec0bbe2daeb9fbcbcf6c2bcc2385d4b

show more ...

Fix flop logging in ILU SeqAIJ to count multiplier computation

Hg-commit: c9575be6cbf3e252f9efb1979a29c1469ef43205

Anothr block size fix

Hg-commit: 60e8a7e8c1bea21bcb6a15fa52030957215de23d

updated and merged from petsc-dev

Hg-commit: 4bdbb24181fe1c8ce13e9b82c1d499b152c3d14f

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

Hg-commit: d2f118b395b24fdc836f857e3f221f6c0c9eca58

added matload_block_size to (MPI)AIJView methods.

Hg-commit: 5a6f02d8d055defb1a971274dc0d4632f9289cd5

The heap and btheap versions of MatMatMultSymbolic_SeqAIJ_SeqAIJ can sum into a dense array

Hg-commit: 4ab803940a7fc63cb246c63a553baf0073e1c7cb

Add MatMatMultSymbolic_SeqAIJ_SeqAIJ_BTHeap, competitive with alternatives

Hg-commit: 27307154f71ea5ce9f622656a9aa768fc993becf

Add MatMatMultSymbolic_SeqAIJ_SeqAIJ_Heap, usually slower

Hg-commit: 437bcb349f014615de0ef2e52d039ea785f846ee

Fixes for the interaction of PCMG and TS w.r.t. hacks in MatFDColoring

Hg-commit: ea036b984a55f55ad28d0cc072122128a7d3812c

fixed some divide by zero errors.

Hg-commit: 36cb8ce3a2012f7aa2ef2d17455a2e07517d03d3

Remove unused variables: rbs,cbs

Hg-commit: dc597ff7b29f368b9fa0d9b6c1f5ba28a0cac1ba

Must use block size 1 for SeqSBAIJ factorization with AIJ since the SBAIJ block size affects indexing in the original matrix

Hg-commit: a42cefaf9f496f3c58e1b407631e3ed837e5479c

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

OBJSC is used by both .c and .cu build targets [fix by using OBJSCU for cuda objects]

Hg-commit: 66e850cc6cac3abe40ce8e1a2faabe7560da6fe9

The runtime options for setting the default max. threads and thread affinities now include the main thread too. So,
./ex -nthreads 8 will spawn 7 threads with the main thread being the 8th one.

Hg-c

The runtime options for setting the default max. threads and thread affinities now include the main thread too. So,
./ex -nthreads 8 will spawn 7 threads with the main thread being the 8th one.

Hg-commit: 4b42148a82c9a2971d30baa82b46bd0a614fb6a9

show more ...

Fixed setting of core affinities. Cleaned up lockfree code.

Hg-commit: 3f10440d912e0f7fc251c25444b232f05364f5fb

Made the kernels to return PetscErrorCode instead of void*

Hg-commit: d174ac62d17756536bbb8251eb89f9173757d7ee

not stinking {} around SETERRQ

Hg-commit: f601164082f593da87052a9248bdd8d6ef84f0c8

block size fix from Lisandro
http://lists.mcs.anl.gov/pipermail/petsc-dev/2012-April/008124.html

Hg-commit: d15eff44f7c30727b0eb725f2e2672baa450e7a6

commit after merge

Hg-commit: a7033f44b7153ddbc07db52896b096f5b5ac7e60

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 ...

Removed setting affinity inside the kernels since it is slow on cg.

Hg-commit: e387fd8113ad94926fa7d4cbfc44e3e9dd016270