| 54037d71 | 24-Feb-2012 |
Barry Smith <bsmith@mcs.anl.gov> |
removed comment symbol inside comment in manual page since some compilers cannot handle it
Hg-commit: ee22ba586a49874be15c04dc43a63ced8b9fdfe5 |
| 7da23a43 | 24-Feb-2012 |
Jose Roman <jroman@dsic.upv.es> |
Fix nested comment
Hg-commit: c121a16ba358d67fa2eaa66e85942e27006b5553 |
| c848bfd8 | 24-Feb-2012 |
Barry Smith <bsmith@mcs.anl.gov> |
added clearer docs to VecGhostGetLocalForm() indicating the calling of VecGhostUpdateBegin/End()
Hg-commit: 20fe1186fa1b3c1097e77d8461b3fde17f7d0f53 |
| bf4db9bb | 22-Feb-2012 |
Shri Abhyankar <abhyshr@mcs.anl.gov> |
Clean up pthread code. i) Added header pthreadimpl.h ii) Removed error checking for now, will add this later.
Hg-commit: 0d5bfd86c86f938f731b5d37bbc35e2475ad173e |
| 682efee0 | 20-Feb-2012 |
Shri Abhyankar <abhyshr@mcs.anl.gov> |
Renamed MainJob to PetscThreadsRunKernel and MainWait to PetscThreadsWait
Hg-commit: 44b125778f815ce49647dd42cf739fc68ee59169 |
| b5c23020 | 14-Feb-2012 |
Jed Brown <jed@59A2.org> |
Ensure that obj->ops->destroy is zeroed in XXSetType() after the old implementation has been destroyed.
There are circumstances where VecSetType() is called more than once to change the type without
Ensure that obj->ops->destroy is zeroed in XXSetType() after the old implementation has been destroyed.
There are circumstances where VecSetType() is called more than once to change the type without an intervening XXCreate_YY(). An example is changing to VECSEQCUSP via VECCUSP.
Hg-commit: 1262149b3e206062ad0b002925300684fd3420e8
show more ...
|
| 43cbfc5b | 12-Feb-2012 |
Matthew G Knepley <knepley@gmail.com> |
Added PetscSectionGetOffsetRange()
Hg-commit: 39172cd792b79ff312011f8c29b341b468ba44ce |
| fcb5fbc9 | 11-Feb-2012 |
Matthew G Knepley <knepley@gmail.com> |
Fixed PetscSFCreateSectionSF()
Hg-commit: cd6f6ee99340d131ffe9f5fdac306a4ef43ddcf6 |
| 7c46c1cb | 09-Feb-2012 |
Paul Mullowney <paulm@txcorp.com> |
Slight tweaks give better strong scaling of multiple GPU SpMV. For cfd.2.10, 4 GPUs is 2.7x faster than 1 GPU. 2 GPUs is 1.8x faster than 1 GPU.
Hg-commit: 4340e22148acbdbf50a369c52bb7b938679abf01 |
| ecbd414d | 08-Feb-2012 |
Matthew G Knepley <knepley@gmail.com> |
Fix for global section
Hg-commit: 3c048e99ea3b238681ac0df3d129c05b2940d14b |
| bff241c3 | 07-Feb-2012 |
Matthew G Knepley <knepley@gmail.com> |
Fix determination of SF root space
Hg-commit: 4326e9b6486ee0020524d2f7738d7da1e2d9b1db |
| c9029c4c | 07-Feb-2012 |
Matthew G Knepley <knepley@gmail.com> |
Protecting against empty SF
Hg-commit: f82b2f7b61434715d1087a34258988ad15f9f400 |
| bf1d350e | 07-Feb-2012 |
Paul Mullowney <paulm@txcorp.com> |
Better handling of data ownership between Petsc and txpetscgpu package. The code in VecCUSPAllocateCheck (veccusp.cu) is much cleaner. Also, correct and efficient parallel-GPU SpMV is handled in the
Better handling of data ownership between Petsc and txpetscgpu package. The code in VecCUSPAllocateCheck (veccusp.cu) is much cleaner. Also, correct and efficient parallel-GPU SpMV is handled in the latest txpetscgpu tarball for ELL, COO, and DIA formats. Lastly, we also have initial usage of cusparse SpMV for CSR format. This works well for some matrices. Later, we'll have CUSPARSE ELL, COO, and HYB formats.
Hg-commit: d52e04f97453e9e2790982a5f6208d3f353cab08
show more ...
|
| 3686e829 | 07-Feb-2012 |
Satish Balay <balay@mcs.anl.gov> |
merge from 32
Hg-commit: 3673b1a67ed63e75f273e277339e9ab78d942979 |
| 9c95f1f0 | 07-Feb-2012 |
Matthew G Knepley <knepley@gmail.com> |
Fixed HDF5 VecLoad
Hg-commit: 75349537bfb65c3634fce11d1e6dc994fcb5da16 |
| 58c52dc8 | 03-Feb-2012 |
Matthew G Knepley <knepley@gmail.com> |
Fixed global section creation
Hg-commit: 4e74fb70da1f334e354ba608caa686ebd11b14dc |
| ff965184 | 01-Feb-2012 |
Jed Brown <jed@59A2.org> |
Doc fixes for VecDestroyVecsF90()
Hg-commit: 195b801a7caf0154c321379b3de6a3d74018bc3e |
| 2a3259e9 | 01-Feb-2012 |
Matthew G Knepley <knepley@gmail.com> |
DMComplex now works for the first ex62 test - Added PetscSection argument to VecGet/SetClosure() and MatSetClosure() - Fixed global section to handle constraints - Added PetscSectionGetConstrainedSto
DMComplex now works for the first ex62 test - Added PetscSection argument to VecGet/SetClosure() and MatSetClosure() - Fixed global section to handle constraints - Added PetscSectionGetConstrainedStorageSize()
Hg-commit: 97aa74253c6f88e4bbb1aee586498fa96aeb97bb
show more ...
|
| 941020cf | 31-Jan-2012 |
Matthew G Knepley <knepley@gmail.com> |
PetscSFCreateSectionSF() tolerates an empty SF
Hg-commit: 67107ac344bb2a3713c0d5cca8de13aa950f29d2 |
| 2e6f6002 | 31-Jan-2012 |
Matthew G Knepley <knepley@gmail.com> |
Check for negative size to ISCreateStride()
Hg-commit: 2e8b02156917693dd4b70ab6d0d832412b676bd7 |
| 32fd6b96 | 30-Jan-2012 |
Barry Smith <bsmith@mcs.anl.gov> |
added P to report number for ANL reports
Hg-commit: b235f9ce30533f49d7b19e2c7612ef8dbefd53ce |
| ac710365 | 30-Jan-2012 |
Shri Abhyankar <abhyshr@mcs.anl.gov> |
Fixed stupid bug in VecMDot_SeqPThread. Trying to figure out the bug in pthread matrix class causing snes to have more iterations.
Hg-commit: 2ee5b725da5e34e60a1dbd9f275c8b17e67d8198 |
| 7da65231 | 29-Jan-2012 |
Matthew G Knepley <knepley@gmail.com> |
Lots of work on DMComplex - Added PetscSectionCreateGlobalSection() - This makes a Section which has global offsets, negative for ghosts - Now negative sizes are not counted for PetscSectionGetSt
Lots of work on DMComplex - Added PetscSectionCreateGlobalSection() - This makes a Section which has global offsets, negative for ghosts - Now negative sizes are not counted for PetscSectionGetStorageSize() - Added DMPrintCellVector() and DMPrintCellMatrix() - Added PetscQuadrature structure - Added SNESDMComplexComputeFunction/Jacobian()
Hg-commit: 93e35c5fb2848c5fd712d4bd760214daf5e64247
show more ...
|
| e5000d1f | 28-Jan-2012 |
Lisandro Dalcin <dalcinl@gmail.com> |
merge release-3.2
Hg-commit: 92ecc738af2e54343bf34f9c473146d1479f768d |
| 4de1be5d | 28-Jan-2012 |
Lisandro Dalcin <dalcinl@gmail.com> |
Fix a few format strings
Hg-commit: 71f429175a71a799d01a1cefdb2806fbcdd7ae81 |