History log of /petsc/src/mat/impls/baij/mpi/baijov.c (Results 226 – 250 of 460)
Revision Date Author Comments
# 8382e67f 26-Jun-2012 Paul Mullowney <paulm@txcorp.com>

committing after previous merge.

Hg-commit: 62ea3b3249ac52e1d599b51313301ac9e25ed491


# f1301baf 25-Jun-2012 Matthew G Knepley <knepley@gmail.com>

Merge

Hg-commit: 5fd7344ebb01d47fc51348189c58908ebeb94702


# c7e6e2c7 04-Jun-2012 Jed Brown <jed@59A2.org>

Fix bad cast syntax

Hg-commit: b5972847114825b1d6f566981883d5ed67bbb148


# 29dcf524 03-Jun-2012 Dmitry Karpeev <karpeev@mcs.anl.gov>

Make MatGetSubMatrices_MPIXAIJ() error out on unsorted col indices. Disable broken MatPermute_MPIXAIJ(). This indicates a problem with MatGetSubMatrix_MPIXAIJ() with unsorted indices().

Hg-commit:

Make MatGetSubMatrices_MPIXAIJ() error out on unsorted col indices. Disable broken MatPermute_MPIXAIJ(). This indicates a problem with MatGetSubMatrix_MPIXAIJ() with unsorted indices().

Hg-commit: 6f7bc96353fea1dfc5aad18ee74e4f8a12fd99e8

show more ...


# 158b5388 04-Jun-2012 Jed Brown <jed@59A2.org>

merge release-3.3

Hg-commit: 5fedf93c78c2bdcaabc4b8ee7380a68199d19b3c


# 94bacf5d 01-Apr-2012 Barry Smith <bsmith@mcs.anl.gov>

Phase 2 of update of PetscBT: PetscBTDestroy() takes pointer

Hg-commit: aca1f1363559db957297829187a5e034ace731de


# 53b8de81 01-Apr-2012 Barry Smith <bsmith@mcs.anl.gov>

phase 1 of PetscBT conversion: changed from Macros to static functions

Hg-commit: 7b64fe46b771f1e27250f571813ebf7716745437


# 51322221 28-Feb-2012 Jungho Lee <julee@mcs.anl.gov>

commit after merge

Hg-commit: e8d19a834d936a4816c3abf3af0b68733c00be11


# 983ee111 28-Oct-2011 Matthew G Knepley <knepley@gmail.com>

Merge

Hg-commit: b584b5d00755e10f2f8b8f1e271250ea3a7b046c


# c5287449 28-Oct-2011 Matthew G Knepley <knepley@gmail.com>

Merge

Hg-commit: 2a372a7e1e4585876ca01c369fbe01ef7cdc023e


# d673de6f 28-Oct-2011 Jed Brown <jed@59A2.org>

merge ASM fix from 3.2

Hg-commit: 104686ef66c5037d2a00e9be84f8547c026a7909


# ca31476a 28-Oct-2011 Jed Brown <jed@59A2.org>

Fix MatIncreaseOverlap (-pc_type asm) with MPIAIJ and bs>1 by reverting PetscLayoutFindOwner to original API.

In 46250962cc92: eliminate non-scalable "rtable" from MatIncreaseOverlap_MPIBAIJ_Once, u

Fix MatIncreaseOverlap (-pc_type asm) with MPIAIJ and bs>1 by reverting PetscLayoutFindOwner to original API.

In 46250962cc92: eliminate non-scalable "rtable" from MatIncreaseOverlap_MPIBAIJ_Once, using updated PetscLayoutFindOwner(),

the API for PetscLayoutFindOwner() was changed to index by blocks, but
the calls from mpiov.c were not updated, so this function was returning
erroneous data which broke -pc_type asm for MPIAIJ if bs>1. This
semantic change was not accompanied by a documentation update. Since
PetscLayoutFindOwner() does not include the word "block", I have
reverted it to operate on scalar indices instead of block indices, so
now the original calls from mpiov.c are correct once again. This patch
updates the calls from baijov.c to use scalar indices.

Hg-commit: 7782710e3f96d2562f32a7b21cb4bef00b1dbc9d

show more ...


# deeea0db 07-Oct-2011 Barry Smith <bsmith@mcs.anl.gov>

commit after merge

Hg-commit: 2e38230693469702fd02f8db8b5417ede06c7aed


# 4da72fa9 06-Oct-2011 Hong Zhang <hzhang@mcs.anl.gov>

add 'allrows' to MatGetSubMatrices_MPIBAIJ_local() to avoid calling nonscalable ISGetIndices()

Hg-commit: 4f6c9b1ff2e5d7832f9e1a1716751b1cb7e0a64e


# 9066e3d5 04-Oct-2011 Hong Zhang <hzhang@mcs.anl.gov>

initialize cmap_i to fix nightly build complains

Hg-commit: 92a5f4c7a2fc8ee99878a3b5a7fee2415ed86445


# 19fa5cdb 04-Oct-2011 Barry Smith <bsmith@mcs.anl.gov>

commit after merge

Hg-commit: 920144eab3d902372b0a9e6dfd80b2f44b55e4ec


# 3c13f0ed 03-Oct-2011 Peter Brune <brune@mcs.anl.gov>

commit after merge

Hg-commit: 3afde820c7daf5b48f91c94669fe8be22a1e3d40


# 8fa52d88 03-Oct-2011 Hong Zhang <hzhang@mcs.anl.gov>

skip non-scalable ISGetIndices(iscol[i],&icol[i]) when allcolumns are wanted
for MatGetSubMatrices_MPIBAIJ_local() in the case !PETSC_USE_CTABLE

Hg-commit: 549d79c855e3ad9cdca633761d1bb206ca0c02d9


# c17803e7 01-Oct-2011 Jed Brown <jed@59A2.org>

Merge TS adaptivity work, some conflicts with Emil's explicit stages for Rosenbrock

In merging Emil's work, I added a flag for an explicit stage instead of
redirecting through a different function.

Merge TS adaptivity work, some conflicts with Emil's explicit stages for Rosenbrock

In merging Emil's work, I added a flag for an explicit stage instead of
redirecting through a different function. It is less code and more
clear, I think.

Hg-commit: 1aa32ce28475592f35bc148baeecb41ed3995a50

show more ...


# bb47af7d 01-Oct-2011 Barry Smith <bsmith@mcs.anl.gov>

commiit after merge

Hg-commit: 2f9449e8e0b477318b1cea8d6a46b8fe0836bd5c


# 307b7a18 30-Sep-2011 Hong Zhang <hzhang@mcs.anl.gov>

add special case "allcolumns" to MatGetSubMatrices_MPIBAIJ_Local()
for scalability. Remove non-scalable colmap. See changeset 20197:1093ea5868ec on MPIAIJ

Hg-commit: abe18fb0e4d548c12fe822dc2a9d43a1

add special case "allcolumns" to MatGetSubMatrices_MPIBAIJ_Local()
for scalability. Remove non-scalable colmap. See changeset 20197:1093ea5868ec on MPIAIJ

Hg-commit: abe18fb0e4d548c12fe822dc2a9d43a1bcc36153

show more ...


# a6963c91 27-Sep-2011 Victor Minden <victorminden@gmail.com>

commit after merge

Hg-commit: 5f0cb0de0a8c7783ecc1afa8ea3838c3bf8dcac2


# 3b3c4b7e 27-Sep-2011 Hong Zhang <hzhang@mcs.anl.gov>

merge from release with minor edit on isblock.c

Hg-commit: 70fe517ef28a01974b3bd58d45be5b1ed8f022e8


# 05d8c843 27-Sep-2011 Hong Zhang <hzhang@mcs.anl.gov>

add 'nkeys' to ISCompressIndicesGeneral() to make it scalable when PETSC_USE_CTABLE
Ref: http://petsc.cs.iit.edu/petsc/petsc-dev/rev/39de4196260f (a bad change made in 2004)

Hg-commit: 8cc43958d6493

add 'nkeys' to ISCompressIndicesGeneral() to make it scalable when PETSC_USE_CTABLE
Ref: http://petsc.cs.iit.edu/petsc/petsc-dev/rev/39de4196260f (a bad change made in 2004)

Hg-commit: 8cc43958d6493e4eeb199b935ab256ab96f3dc6d

show more ...


# 472b90b3 26-Sep-2011 adams <adams@jaguarpf-login1.ccs.ornl.gov>

fix for emtpty procs on input grid (2nd commit).

Hg-commit: 416bc569082d603eba1a683707921f8a1b880bdb


12345678910>>...19