| #
c1acdb04
|
| 28-Sep-2019 |
Satish Balay <balay@mcs.anl.gov> |
Merge remote-tracking branch 'origin/jczhang/feature-sf-on-gpu'
Add GPU-aware VecScatter/PetscSF
See merge request petsc/petsc!1995
|
| #
b23bfdef
|
| 13-Aug-2019 |
Junchao Zhang <jczhang@mcs.anl.gov> |
Update pack/unpack routines to do packing/unpacking for all neighbors in at most two routines
One is used to pack data in self to self communication; The second is used for remote communication.
So
Update pack/unpack routines to do packing/unpacking for all neighbors in at most two routines
One is used to pack data in self to self communication; The second is used for remote communication.
So that on GPU, we can use at most two kernels to do packing/packing for all neighbors instead of multiple kernels
show more ...
|
| #
05393080
|
| 25-Jul-2019 |
Karl Rupp <me@karlrupp.net> |
Merge branch 'jczhang/sf-more-opts' [PR #1567]
* jczhang/sf-more-opts:
Add more optimizations in SF and use it as the default for VecScatter.
|
| #
f12d10cc
|
| 18-Jun-2019 |
Junchao Zhang <jczhang@mcs.anl.gov> |
Add message logging for MPI collectives and also update code style
In Petsc, the output arguments are usually put at the end
|
| #
1c575b32
|
| 07-Jul-2019 |
Barry Smith <bsmith@mcs.anl.gov> |
Merge branch 'maint'
|
| #
c9b973be
|
| 07-Jul-2019 |
Barry Smith <bsmith@mcs.anl.gov> |
Fixed incomplete set of changes, did not remove __ from most use of __MPIUNI in the source
Commit-type: bug-fix
Reported-by: nightly tests
|
| #
26bd1501
|
| 05-Jul-2019 |
Barry Smith <bsmith@mcs.anl.gov> |
Remove use of _ and __ in front of PETSc include guards. Reason: C99 Reserved Identifiers
Commit-type: portability-fix
|
| #
7a71495b
|
| 04-Jul-2019 |
Karl Rupp <me@karlrupp.net> |
Merge branch 'hannah/gpu-computation-logging' [PR #1843]
* hannah/gpu-computation-logging:
Adding GPU flop rate and GPU time.
|
| #
7a052e47
|
| 03-Jul-2019 |
hannah_mairs <hannah.mairs@gmail.com> |
PetscLogGpuTimeStart -> Begin
|
| #
958c4211
|
| 01-Jul-2019 |
hannah_mairs <hannah.mairs@gmail.com> |
Adding Gpu flop rate and GPU time
|
| #
04f8278e
|
| 05-Jun-2019 |
Karl Rupp <me@karlrupp.net> |
Merge branch 'hannah/gpu-communication-logging' [PR #1746]
* hannah/gpu-communication-logging:
This branch logs the number and size of CPU to GPU and GPU to CPU vector copies when PETSc is built wit
Merge branch 'hannah/gpu-communication-logging' [PR #1746]
* hannah/gpu-communication-logging:
This branch logs the number and size of CPU to GPU and GPU to CPU vector copies when PETSc is built with ViennaCL or CUDA and adds four columns to log_view reporting this communication.
show more ...
|
| #
bec0b493
|
| 30-May-2019 |
hannah_mairs <hannah.mairs@gmail.com> |
Adding logging code
|
| #
24ac310b
|
| 18-May-2019 |
BarryFSmith <bsmith@mcs.anl.gov> |
Merged in barry/log-memory (pull request #1677)
Introduce logging of memory usage to the PETSc event logging infrastructure.
|
| #
e3ed9ee7
|
| 29-Apr-2019 |
Barry Smith <bsmith@mcs.anl.gov> |
Introduce logging of memory usage to the PETSc event logging infrastructure.
The logging can be turned on and viewed with the options -log_view -log_view_memory
The logging requires the use of the
Introduce logging of memory usage to the PETSc event logging infrastructure.
The logging can be turned on and viewed with the options -log_view -log_view_memory
The logging requires the use of the PETSc malloc so may slow the runs slightly but since it is not
doing error checking of the memory it should not slow down the computations significently.
Four pieces of information about memory are logged and display in four additional columns in the table.
They are documented in the output above the table. Each column provides a particular "view" of the
memory usage in the event and they complement each other.
This should make it much easier for developers focusing on memory usage in their implementations
to understand the usage in the code, where the memory is being utilized, how it scales with problem size and number of processes etc.
Commit-type: feature
show more ...
|
| #
dd364f81
|
| 16-May-2019 |
Lisandro Dalcin <dalcinl@gmail.com> |
Merged in dalcinl/fix-clang-warnings (pull request #1671)
Fix clang warning -Wextra-semi-stmt
Approved-by: BarryFSmith <bsmith@mcs.anl.gov>
|
| #
e83a5d19
|
| 15-May-2019 |
Lisandro Dalcin <dalcinl@gmail.com> |
Fix clang warning -Wextra-semi-stmt
https://clang.llvm.org/docs/DiagnosticsReference.html#wextra-semi-stmt
warning: empty expression statement has no effect;
remove unnecessary ';' to silence this
Fix clang warning -Wextra-semi-stmt
https://clang.llvm.org/docs/DiagnosticsReference.html#wextra-semi-stmt
warning: empty expression statement has no effect;
remove unnecessary ';' to silence this warning [-Wextra-semi-stmt]
ierr = PetscOptionsTail();CHKERRQ(ierr);
^
show more ...
|
| #
10682a7e
|
| 26-Mar-2019 |
Lisandro Dalcin <dalcinl@gmail.com> |
Merged in dalcinl/plex-distribute (pull request #1454)
DMPLEX: Speedup DMPlexDistribute[Overlap]()
Approved-by: Matthew Knepley <knepley@gmail.com>
|
| #
d7fc57b6
|
| 25-Mar-2019 |
Junchao Zhang <junchao.zhang@gmail.com> |
Merged in jczhang/vecscat-add-opts (pull request #1462)
Add two small optimizations in VecScatter and SF before release
Approved-by: BarryFSmith <bsmith@mcs.anl.gov>
Approved-by: Jed Brown <jed@59a
Merged in jczhang/vecscat-add-opts (pull request #1462)
Add two small optimizations in VecScatter and SF before release
Approved-by: BarryFSmith <bsmith@mcs.anl.gov>
Approved-by: Jed Brown <jed@59a2.org>
show more ...
|
| #
ea1c87f7
|
| 20-Mar-2019 |
Junchao Zhang <jczhang@mcs.anl.gov> |
Do not call MPI_Startall(count,reqs) if count=0
MPICH raises an error when count=0 and reqs=NULL. They should not. Before they fix it, we use this extra test.
|
| #
457b14b1
|
| 20-Mar-2019 |
Junchao Zhang <jczhang@mcs.anl.gov> |
Replace MPIU_SCALAR with an MPI datatype arg in MPI_Startall_irecv etc
So that we can log persistent communication for other data types
|
| #
62872c28
|
| 13-Mar-2019 |
Lisandro Dalcin <dalcinl@gmail.com> |
Log: Add PetscLogEventSync()
This routine should be called only if there is not a PetscObject available
to pass to PetscLogEventBegin(), but you rather have a MPI_Comm handle.
|
| #
b0f52d29
|
| 28-Jan-2019 |
Joseph Pusztay <josephpusztay@Josephs-MacBook-Pro.local> |
Merge branch 'master' into jpusztay/feature-swarm-symplectic-example
|
| #
940dfb0a
|
| 28-Jan-2019 |
Joseph Pusztay <josephpu@buffalo.edu> |
Merged petsc/petsc into master
|
| #
30faf514
|
| 27-Jan-2019 |
m_diehl <m.diehl@mpie.de> |
Merged petsc/petsc into master
|
| #
ddad275a
|
| 21-Jan-2019 |
Patrick Farrell <patrick@pefarrell.org> |
Merge branch 'master' into knepley/feature-snes-patch
|