| 5da5aae0 | 25-Nov-2010 |
Jed Brown <jed@59A2.org> |
support empty MatType
Hg-commit: afbbf429dfb69190f21a9d125681a36dfcb72eed |
| 89715bbe | 25-Nov-2010 |
Jed Brown <jed@59A2.org> |
The function body is already #if 0'd out and I'm tired of seeing unused variables warnings
Hg-commit: b3430528bb36243daa0d6e3295a26f064a7e141e |
| 1411c6ee | 23-Nov-2010 |
Jed Brown <jed@59A2.org> |
Promote LocalToGlobalMapping and BlockSize from DA to DM
Hg-commit: adc5020243929ce40bd18f2655d78fe82dc79cc4 |
| b17742ca | 20-Nov-2010 |
Sean Farley <sean@mcs.anl.gov> |
Merged fix from petsc-3.1 in fdda.c
Hg-commit: 6cc7450d72f73509df836d2079bee90c8e00fc55 |
| 784ac674 | 19-Nov-2010 |
Jed Brown <jed@59A2.org> |
Distinguish row and column mappings for MatSetValuesLocal
Hg-commit: 71968a52236f04a9bab5e1d4a884aa59ef0c7cfd |
| 11afa6f6 | 14-Nov-2010 |
Barry Smith <bsmith@mcs.anl.gov> |
commit after merge
Hg-commit: cad48a04353cb40f118ecb9f477b2133c90a5fb4 |
| 454e267f | 12-Nov-2010 |
Lisandro Dalcin <dalcinl@gmail.com> |
DMGetElements for DA now supports P1/Q1 in 1D/2D/3D
Hg-commit: 9ce2ad348aa3a26019a8f5bb283ffeb866aa9875 |
| 2f4ee27d | 12-Nov-2010 |
Barry Smith <bsmith@mcs.anl.gov> |
fix for --with-single-library=0
unneeded double precision requirment for characteristic removed
Hg-commit: 65669c03ea518ca0756e6833510df9ccb117486c |
| c47c1910 | 09-Nov-2010 |
Barry Smith <bsmith@mcs.anl.gov> |
fixed variable formating problem introduced when DM was attached to DA types
Hg-commit: 4eddb55e7997bda6108b1e2674f97189697f0382 |
| b1757049 | 07-Nov-2010 |
Jed Brown <jed@59A2.org> |
Add DMGetInjection_DA_3D and test from Dave May
Hg-commit: 61b9996763ec9977a6f017f74e058a735548fe9d |
| 7242296b | 07-Nov-2010 |
Jed Brown <jed@59A2.org> |
Re-enable -da_view, this is what the comment said, not sure if it's the way Barry intended
Hg-commit: 40304ed5103b7487f837a45f569a7b99a7608aaf |
| 616157d6 | 07-Nov-2010 |
Jed Brown <jed@59A2.org> |
Enable 3D VTK viewing with DMView
Hg-commit: 95e87bc9da0599f239ca1e0028e57e93dbf7f209 |
| 2be375d4 | 06-Nov-2010 |
Jed Brown <jed@59A2.org> |
Use DMDACoarsenOwnershipRanges to make DMCoarsen_DA more robust.
Hg-commit: c9469d8ac25d61fcaba2cc41c8bfc6ee2e12162f |
| 7aca7175 | 06-Nov-2010 |
Jed Brown <jed@59A2.org> |
Use more robust/understandable algorithm for DMDARefineOwnershipRanges.
Easier algorithm to reuse for the corresponding coarsening algorithm.
Hg-commit: ed543b71266ed027e04bc06ab7ff12ff70dfa5d8 |
| 0689a922 | 05-Nov-2010 |
Jed Brown <jed@59A2.org> |
da->refine_x == 0 does not make sense, so remove special case for it
Hg-commit: c849172086aebf436e63e1ee1b55a5e442d993cd |
| 3da9ae13 | 04-Nov-2010 |
Jed Brown <jed@59A2.org> |
Call DMView_DA_VTK from DMView_DA_2d, somehow this reference disappeared
Hg-commit: 5de1721540a7e7092c27c2a7d0396073c2709cfa |
| 076202dd | 04-Nov-2010 |
Jed Brown <jed@59A2.org> |
Fix DMGetInjection_DA_2D, was generating mismatched vector sizes for valid configurations
From Dave May:
A couple of cases which crash (with different errors) are
${PETSC_DIR}/${PETSC_ARCH
Fix DMGetInjection_DA_2D, was generating mismatched vector sizes for valid configurations
From Dave May:
A couple of cases which crash (with different errors) are
${PETSC_DIR}/${PETSC_ARCH}/bin/mpiexec -n 6 ./da_interp_test_dev_jb
-mx 10 -my 20 -da_processors_x 2 -da_processors_y 3
${PETSC_DIR}/${PETSC_ARCH}/bin/mpiexec -n 20 ./da_interp_test_dev_jb
-mx 10 -my 30 -da_processors_x 2 -da_processors_y 10
The latter case still does not work, but I think it is related to
DMCoarsen_DA, not the construction of interpolants.
Hg-commit: 3bfbade72e584aedebc636dad30d6d0b5d8f2e1b
show more ...
|
| e40c8662 | 02-Nov-2010 |
Barry Smith <bsmith@mcs.anl.gov> |
commit after merge
Hg-commit: 9ed2cb431947553955f0ce85412d290cc41bb35c |
| 325fc9f4 | 02-Nov-2010 |
Barry Smith <bsmith@mcs.anl.gov> |
fixed some mansec = DA and mansec = DMDA mistakes from previous changes
Hg-commit: 5d8cac658961538ec6e588146350f5f4e566d4b7 |
| 8e181257 | 02-Nov-2010 |
Satish Balay <balay@mcs.anl.gov> |
doc fixes [from builddist]
Hg-commit: d8954d10e0ddff2f5bc8e3dbb1b06b606d80bab4 |
| 30e6f737 | 02-Nov-2010 |
Jed Brown <jed@59A2.org> |
Better PetscStackCallHypre macro, uses "#" for function name, so that they stay consistent
Hg-commit: 85003f8e9e094c6e7e1888cce0aa07e6772586a8 |
| 2dcb2ebc | 31-Oct-2010 |
Barry Smith <bsmith@mcs.anl.gov> |
added VecDuplicate_MPI_DA
Hg-commit: be6d052bd8683d037a1bda7bdc21bb994668a112 |
| 9c7c4993 | 28-Oct-2010 |
Barry Smith <bsmith@mcs.anl.gov> |
VecLoad_HDF5_DA bug in accessing DM_DA pointer from DM
Changed PetscObjectName() to force same name on all processes by broadcasting address of comm from zeroth process
Hg-commit: 91801262b539fd06c5
VecLoad_HDF5_DA bug in accessing DM_DA pointer from DM
Changed PetscObjectName() to force same name on all processes by broadcasting address of comm from zeroth process
Hg-commit: 91801262b539fd06c563086c31a4e2fe8a77db8e
show more ...
|
| 9005068f | 27-Oct-2010 |
Jed Brown <jed@59A2.org> |
Fix Hypre MatZeroRows from bad pasting
Hg-commit: 8a996997f85a72770514d8834d864c1b41d82fff |
| 9c3e2b52 | 26-Oct-2010 |
Barry Smith <bsmith@mcs.anl.gov> |
added error checking for MatZeroRowsLocal_ special cases when x and b are provided
if size of comm is one then MatZeroRowsLocal() just calls (*matzerorows) if matzerorowslocal does not exist
Hg-comm
added error checking for MatZeroRowsLocal_ special cases when x and b are provided
if size of comm is one then MatZeroRowsLocal() just calls (*matzerorows) if matzerorowslocal does not exist
Hg-commit: ae98f442a894e242ae06f464bbd1ec844b6e622a
show more ...
|