| 9dcd2ab3 | 07-May-2019 |
Stefano Zampini <stefano.zampini@gmail.com> |
Minor code formatting |
| 4edbe3a6 | 12-May-2019 |
Karl Rupp <me@karlrupp.net> |
Merge branch 'barry/feature-pintocpu' [PR #1641]
* barry/feature-pintocpu:
Adding a MatPinToCPU() and VecPinToGPU() capability
For matrices this will prevent copies to the GPU when they will never b
Merge branch 'barry/feature-pintocpu' [PR #1641]
* barry/feature-pintocpu:
Adding a MatPinToCPU() and VecPinToGPU() capability
For matrices this will prevent copies to the GPU when they will never be used there.
For vectors this will prevent vectors from boucing back and forth between the CPU.
show more ...
|
| 7eda4a2b | 11-May-2019 |
Satish Balay <balay@mcs.anl.gov> |
Merge remote-tracking branch 'origin/maint' |
| 2bd48813 | 11-May-2019 |
Junchao Zhang <junchao.zhang@gmail.com> |
Merged in jczhang/feature-add-vscat-example (pull request #1652)
Add an example to show transfer vectors between comms
Approved-by: BarryFSmith <bsmith@mcs.anl.gov> |
| 69aaa56e | 11-May-2019 |
Satish Balay <balay@mcs.anl.gov> |
do not build this example with mpiuni. This fixes warnings on arch-mswin-uni
c:\cygwin64\home\petsc\petsc.next-3\src\vec\vscat\examples\ex9.c(144) : warning C4700: uninitialized local variable 'inte
do not build this example with mpiuni. This fixes warnings on arch-mswin-uni
c:\cygwin64\home\petsc\petsc.next-3\src\vec\vscat\examples\ex9.c(144) : warning C4700: uninitialized local variable 'intercomm' used
show more ...
|
| 7729c4fc | 11-May-2019 |
Satish Balay <balay@mcs.anl.gov> |
fix clang warnings
/sandbox/petsc/petsc.next/src/vec/vscat/examples/ex9.c:54:5: warning: Value stored to 'ierr' is never read |
| cfebd5f6 | 10-May-2019 |
Junchao Zhang <jczhang@mcs.anl.gov> |
And code to do reverse communication that is needed for coupling |
| 458980a0 | 10-May-2019 |
Junchao Zhang <jczhang@mcs.anl.gov> |
Add an example to show transfer vectors between comms |
| ffe67aa5 | 10-May-2019 |
Václav Hapla <vaclav.hapla@erdw.ethz.ch> |
docs: clarify deallocation |
| ea7d9e87 | 09-May-2019 |
Vaclav Hapla <vaclav.hapla@erdw.ethz.ch> |
sf tut ex1: remove redundant degree computation |
| 66dfcd1a | 08-Apr-2019 |
Vaclav Hapla <vaclav.hapla@erdw.ethz.ch> |
PetscSFComputeMultiRootOriginalNumbering: add outarg nMultiRoots |
| f3d5ece7 | 09-May-2019 |
Satish Balay <balay@mcs.anl.gov> |
Merge remote-tracking branch 'origin/pr1605/m_diehl/fortran-tests/master' [PR #1605] |
| 1a05bf98 | 08-May-2019 |
Vaclav Hapla <vaclav.hapla@erdw.ethz.ch> |
comment: removed wrong one above VecLoad_HDF5 |
| 874842be | 08-May-2019 |
Vaclav Hapla <vaclav.hapla@erdw.ethz.ch> |
doc: fix PetscViewerHDF5ReadSizes manpage |
| bc307625 | 08-May-2019 |
Vaclav Hapla <vaclav.hapla@erdw.ethz.ch> |
doc: no need to include petscsys.h because petscviewerhdf5.h does already |
| 4e97f8eb | 08-May-2019 |
Vaclav Hapla <vaclav.hapla@erdw.ethz.ch> |
Move HDF5 array IO stuff from sys to vec.
- Fix issue 288.
- Moved PetscViewerHDF5Load,PetscViewerHDF5ReadSizes and their helpers from
src/sys/classes/viewer/impls/hdf5/hdf5v.c
into the new fi
Move HDF5 array IO stuff from sys to vec.
- Fix issue 288.
- Moved PetscViewerHDF5Load,PetscViewerHDF5ReadSizes and their helpers from
src/sys/classes/viewer/impls/hdf5/hdf5v.c
into the new file
src/vec/is/utils/hdf5io.c
- PetscViewerHDF5Load() header moved from viewerimpl.h to petscis.h
- PetscViewerHDF5ReadSizes() header moved from petscviewerhdf5.h to petscis.h
show more ...
|
| e7e92044 | 07-May-2019 |
Barry Smith <bsmith@mcs.anl.gov> |
Based on discussion with Oana I am adding a MatPinToCPU() and VecPinToGPU() capability. For matrices this
will prevent copies to the GPU when they will never be used there. For vectors this will
prev
Based on discussion with Oana I am adding a MatPinToCPU() and VecPinToGPU() capability. For matrices this
will prevent copies to the GPU when they will never be used there. For vectors this will
prevent vectors from boucing back and forth between the CPU and GPU when most of the work is in the CPU. An
example of the place that needs to avoid bouncing is in MatFDColoringApply_XXXX()
Commit-type: feature, documentation, example
Thanks-to: Oana Marin <oanam@mcs.anl.gov>
show more ...
|
| 324b6635 | 09-May-2019 |
Satish Balay <balay@mcs.anl.gov> |
more arch-linux-cxx-cmplx-pkgs-64idx fixes |
| 90c762ab | 08-May-2019 |
Matthew Knepley <knepley@gmail.com> |
Merged in knepley/feature-plex-viz-hybrid-mesh (pull request #1629)
Knepley/feature plex viz hybrid mesh
Approved-by: BarryFSmith <bsmith@mcs.anl.gov> |
| 4a347d90 | 08-May-2019 |
Martin Diehl <mail@martin-diehl.net> |
compatible with varying integer kinds |
| c63b7f52 | 08-May-2019 |
Martin Diehl <mail@martin-diehl.net> |
some compilers are case sensitive when linking |
| b7fb9943 | 08-May-2019 |
Martin Diehl <mail@martin-diehl.net> |
output only real part in case of complex build |
| 8693ff2a | 06-May-2019 |
Barry Smith <bsmith@mcs.anl.gov> |
shrink length of PetscLogEventRegister strings to acceptable size so it does not mess up the output table from -log_view
Commit-type: bug-fix |
| b267dc21 | 06-May-2019 |
Matthew G. Knepley <knepley@gmail.com> |
IS: Fix for writing 0 size IS to HDF5 |
| 80534ad6 | 06-May-2019 |
BarryFSmith <bsmith@mcs.anl.gov> |
Merged in barry/special-case-mat-shift-0 (pull request #1614)
MatShift() and VecShift() should handle the special case of 0.0 shift and return immediately without the time spent in the shift or the
Merged in barry/special-case-mat-shift-0 (pull request #1614)
MatShift() and VecShift() should handle the special case of 0.0 shift and return immediately without the time spent in the shift or the change the objects state.
show more ...
|