Trilinos merge requestshttps://gitlab.osti.gov/jmwille/Trilinos/-/merge_requests2018-07-10T21:36:48Zhttps://gitlab.osti.gov/jmwille/Trilinos/-/merge_requests/3086MueLu: Modifying Avatar Interface2018-07-10T21:36:48ZJames WillenbringMueLu: Modifying Avatar Interface*Created by: csiefer2*
Auto-PR for SHA 615f11d*Created by: csiefer2*
Auto-PR for SHA 615f11dhttps://gitlab.osti.gov/jmwille/Trilinos/-/merge_requests/2902MueLu TentativePFactory_kokkos: Add no-QR option for NSDim>12018-06-07T22:36:47ZJames WillenbringMueLu TentativePFactory_kokkos: Add no-QR option for NSDim>1*Created by: cgcgcg*
@trilinos/muelu
## Description
Add no-QR option for nullspaces of dimension > 1 to `MueLu_TentativePFactory_kokkos`.
## Motivation and Context
The current QR implementation allocates too much scratch memory...*Created by: cgcgcg*
@trilinos/muelu
## Description
Add no-QR option for nullspaces of dimension > 1 to `MueLu_TentativePFactory_kokkos`.
## Motivation and Context
The current QR implementation allocates too much scratch memory for RefMaxwell runs on Cuda which leads to an error. Disabling the QR altogether can circumvent this issue until we have a better Kokkos QR. The default behavior has not been changed.https://gitlab.osti.gov/jmwille/Trilinos/-/merge_requests/3024MueLu repartitioning: Allow use of graph partitioners2018-07-19T21:09:56ZJames WillenbringMueLu repartitioning: Allow use of graph partitioners*Created by: cgcgcg*
@trilinos/muelu
## Description
Repartitioning using Zoltan2 only allowed the use of coordinate based repartitioning. This PR adds the option to use other, graph-based partitioners such as Scotch or ParMetis thr...*Created by: cgcgcg*
@trilinos/muelu
## Description
Repartitioning using Zoltan2 only allowed the use of coordinate based repartitioning. This PR adds the option to use other, graph-based partitioners such as Scotch or ParMetis through Zoltan2. The default behavior is not changed.
## Motivation and Context
This allows repartitioning without having coordinates. I have not done any timing runs, but graph partitioning might be more expensive, so the setup cost might go up. On the other hand, graph partitioning minimizes the number of edge cuts, so we could potentially see faster solves.https://gitlab.osti.gov/jmwille/Trilinos/-/merge_requests/3379MueLu: fix #33782018-09-01T04:45:56ZJames WillenbringMueLu: fix #3378*Created by: lucbv*
@trilinos/muelu
## Description
Fixing a bug by adding the appropriate typedef.
## Motivation and Context
This fixes a build issue in MueLu when kokkos is enabled
## Related Issues
* Closes #3378
##...*Created by: lucbv*
@trilinos/muelu
## Description
Fixing a bug by adding the appropriate typedef.
## Motivation and Context
This fixes a build issue in MueLu when kokkos is enabled
## Related Issues
* Closes #3378
## How Has This Been Tested?
Local build is now successful with `KOKKOS_REFACTOR=ON`
## Checklist
- [x] My commit messages mention the appropriate GitHub issue numbers.
- [x] My code follows the code style of the affected package(s).
- [] My change requires a change to the documentation.
- [] I have updated the documentation accordingly.
- [x] I have read the [code contribution guidelines](../blob/master/CONTRIBUTING.md) for this project.
- [x] I have added tests to cover my changes.
- [x] All new and existing tests passed.
- [x] No new compiler warnings were introduced.
- [] These changes break backwards compatibility.https://gitlab.osti.gov/jmwille/Trilinos/-/merge_requests/3419MueLu: Missing output when proc 0 drops out2018-09-27T03:01:53ZJames WillenbringMueLu: Missing output when proc 0 drops out*Created by: cgcgcg*
@trilinos/muelu
## Description
MueLu is handling which messages get printed on which rank itself, in `MueLu_VerboseObject`. There, we set the default output stream of the underlying Teuchos object to a staticly...*Created by: cgcgcg*
@trilinos/muelu
## Description
MueLu is handling which messages get printed on which rank itself, in `MueLu_VerboseObject`. There, we set the default output stream of the underlying Teuchos object to a staticly defined stream, and set it to output on all ranks. However, in `MueLu_HierarchyManager` we take that stream and set it to output on rank 0 only. Teuchos states that the behavior of calling `setOutputToRootOnly` more than once is undefined. What happens is that all but rank 0 irreversibly replace their streams with blackhole streams. This means that all MueLu objects everywhere except on rank 0 do not print anymore. This can also happen when `setOutputToRootOnly` is called from somewhere else on the Teuchos default stream.
With the changes in this PR, we do not set the output stream used by MueLu to be the default Teuchos output stream in order to that somewhere else `setOutputToRootOnly` is called on it.
The change was tested by running
```
mpirun -np 2 ./MueLu_Driver.exe --matrixType=Laplace3D --xml=scaling.xml --nx=34 --ny=34 --nz=34 --rebuild=1 --solver=none --notimings
```
## Related Issues
* Addresses #1471 and #1489https://gitlab.osti.gov/jmwille/Trilinos/-/merge_requests/3352MueLu RefMaxwell updates2018-08-27T15:02:48ZJames WillenbringMueLu RefMaxwell updates*Created by: cgcgcg*
@trilinos/muelu
## Description
- Allow to use RefMaxwell as preconditioner or as standalone solver.
- Correctly interpret 'use kokkos refactor' parameter.
- Show hierarchy label in multigrid summary. (Helpful...*Created by: cgcgcg*
@trilinos/muelu
## Description
- Allow to use RefMaxwell as preconditioner or as standalone solver.
- Correctly interpret 'use kokkos refactor' parameter.
- Show hierarchy label in multigrid summary. (Helpful, when there are multiple of them.)
- Allow different pre- and postsmoothers on finest level of RefMaxwell. (E.g. forward/backward GS)https://gitlab.osti.gov/jmwille/Trilinos/-/merge_requests/3226Write region data files into subdirectory (#3061)2018-08-14T14:26:51ZJames WillenbringWrite region data files into subdirectory (#3061)*Created by: mayrmt*
@trilinos/muelu
## Description
All region-related data files are now written into a user-specified subdirectory to keep the directory structure clean.
* Specify path directory for region data files in Matlab...*Created by: mayrmt*
@trilinos/muelu
## Description
All region-related data files are now written into a user-specified subdirectory to keep the directory structure clean.
* Specify path directory for region data files in Matlab script `createInput.m`
as well as on the command line (command line argument --regDataDir)
* Save/read all region-related input data from this directory
* Rename file `readRegionalFile.m` to `readRegionFile.m` to avoid Matlab warning
## Motivation and Context
Keeping the directory structure clean and moving problem specific output to a subdirectory.
## Related Issues
* Closes #3061
* Related to #2709
## How Has This Been Tested?
Results of toy problems did not change. However, there is no actual test yet.
## Checklist
- [x] My commit messages mention the appropriate GitHub issue numbers.
- [x] My code follows the code style of the affected package(s).
- [x] My change requires a change to the documentation.
- [x] I have updated the documentation accordingly.
- [x] I have read the [code contribution guidelines](../blob/master/CONTRIBUTING.md) for this project.
- [ ] I have added tests to cover my changes.
- [x] All new and existing tests passed.
- [x] No new compiler warnings were introduced.
- [ ] These changes break backwards compatibility.https://gitlab.osti.gov/jmwille/Trilinos/-/merge_requests/3722MueLu: kokkos openmp refactor tests fix2018-10-24T02:53:18ZJames WillenbringMueLu: kokkos openmp refactor tests fix*Created by: lucbv*
@trilinos/muelu
## Description
Fixing a couple of tests as per issue #3717 simply changing logic in xml files to make them consistent with new kokkos refactor and removing a Chebyshev smoother call to avoid issu...*Created by: lucbv*
@trilinos/muelu
## Description
Fixing a couple of tests as per issue #3717 simply changing logic in xml files to make them consistent with new kokkos refactor and removing a Chebyshev smoother call to avoid issue with complex type.
## Motivation and Context
This fixes issue #3717
## Related Issues
* Part of #3717
## How Has This Been Tested?
This has been tested locally using a build with `MUELU_ENABLE_KOKKOS_REFACTOR=ON` and `Trilinos_ENABLE_COMPLEX_DOUBLE=ON`
## Checklist
- [x] My commit messages mention the appropriate GitHub issue numbers.
- [x] My code follows the code style of the affected package(s).
- [ ] My change requires a change to the documentation.
- [ ] I have updated the documentation accordingly.
- [x] I have read the [code contribution guidelines](../blob/master/CONTRIBUTING.md) for this project.
- [x] I have added tests to cover my changes.
- [x] All new and existing tests passed.
- [x] No new compiler warnings were introduced.
- [ ] These changes break backwards compatibility.https://gitlab.osti.gov/jmwille/Trilinos/-/merge_requests/3802MueLu: Fixes from rstumin2018-11-05T21:34:37ZJames WillenbringMueLu: Fixes from rstumin*Created by: csiefer2*
Auto-PR for SHA 534aab8*Created by: csiefer2*
Auto-PR for SHA 534aab8https://gitlab.osti.gov/jmwille/Trilinos/-/merge_requests/2803Muelu analysis update2018-05-23T16:31:40ZJames WillenbringMuelu analysis update*Created by: aprokop*
@trilinos/muelu
## Description
Introduces command line filtering in `setup_timers`. And fixes the update in `transitions` python package.
No testing required.
## Checklist
- [ ] My commit messages men...*Created by: aprokop*
@trilinos/muelu
## Description
Introduces command line filtering in `setup_timers`. And fixes the update in `transitions` python package.
No testing required.
## Checklist
- [ ] My commit messages mention the appropriate GitHub issue numbers.
- [x] My code follows the code style of the affected package(s).
- [ ] My change requires a change to the documentation.
- [ ] I have updated the documentation accordingly.
- [ ] I have read the [code contribution guidelines](../blob/master/CONTRIBUTING.md) for this project.
- [ ] I have added tests to cover my changes.
- [x] All new and existing tests passed.
- [x] No new compiler warnings were introduced.
- [ ] These changes break backwards compatibility.https://gitlab.osti.gov/jmwille/Trilinos/-/merge_requests/2619MueLu: Missing header fixes2018-04-24T13:54:53ZJames WillenbringMueLu: Missing header fixes*Created by: csiefer2*
Auto-PR for SHA 1a98b5b*Created by: csiefer2*
Auto-PR for SHA 1a98b5bhttps://gitlab.osti.gov/jmwille/Trilinos/-/merge_requests/2643MueLu interface tests: Add gold files for kokkos and kokkos+complex2018-04-26T00:04:34ZJames WillenbringMueLu interface tests: Add gold files for kokkos and kokkos+complex*Created by: cgcgcg*
@trilinos/muelu
## Description
Adds gold files for interface tests using Kokkos. Simplifies rebasing by adding a Makefile target: "make rebase" now picks the correct set of gold files to rebase.*Created by: cgcgcg*
@trilinos/muelu
## Description
Adds gold files for interface tests using Kokkos. Simplifies rebasing by adding a Makefile target: "make rebase" now picks the correct set of gold files to rebase.https://gitlab.osti.gov/jmwille/Trilinos/-/merge_requests/4185MueLu: update rocketman crontab2019-02-23T02:01:13ZJames WillenbringMueLu: update rocketman crontab*Created by: jhux2*
Auto-PR for SHA c5faedc
<!-- Reviewable:start -->
---
This change is [<img src="https://reviewable.io/review_button.svg" height="34" align="absmiddle" alt="Reviewable"/>](https://reviewable.io/reviews/trilinos/trili...*Created by: jhux2*
Auto-PR for SHA c5faedc
<!-- Reviewable:start -->
---
This change is [<img src="https://reviewable.io/review_button.svg" height="34" align="absmiddle" alt="Reviewable"/>](https://reviewable.io/reviews/trilinos/trilinos/4185)
<!-- Reviewable:end -->
https://gitlab.osti.gov/jmwille/Trilinos/-/merge_requests/2297Sacado: Modifications for deep_copy2019-03-15T04:24:01ZJames WillenbringSacado: Modifications for deep_copy*Created by: ndellingwood*
Changes to the deep_copy implementation in Kokkos are incompatible with
the existing implementation of deep_copy in Sacado. This is due in part
to custom layouts in Sacado that are not recognized by Kokkos.
...*Created by: ndellingwood*
Changes to the deep_copy implementation in Kokkos are incompatible with
the existing implementation of deep_copy in Sacado. This is due in part
to custom layouts in Sacado that are not recognized by Kokkos.
The changes in this commit handle the layout issue and revise the assign
method of Sacado's ViewMapping to properly convert View's of (nested)
Fad types into Kokkos Views (with additional dynamic ranks) to use the
new deep_copy implementation in Kokkos.
<!---
Be sure to select `develop` as the `base` branch against which to create this
pull request. Only pull requests against `develop` will undergo Trilinos'
automated testing. Pull requests against `master` will be ignored.
-->
<!---
Provide a general summary of your changes in the Title above. If this pull
request pertains to a particular package in Trilinos, it's worthwhile to start
the title with "PackageName: ".
-->
<!---
Note that anything between these delimiters is a comment that will not appear
in the pull request description once created.
-->
<!---
Replace <teamName> below with the appropriate Trilinos package/team name.
-->
@trilinos/sacado
<!---
Reviewers: If you know someone who is knowledgeable about what you are
changing, or perhaps someone who should be, and you would like them to review
your changes before they are accepted, select them from the Reviewers drop-down
on the right.
-->
<!---
Assignees: If you know anyone who should likely handle bringing this pull
request to completion, select them from the Assignees drop-down on the right.
If you have write-access to Trilinos, this should likely be you.
-->
<!---
Lables: Choose any applicable package names from the Labels drop-down on the
right. Additionally, choose a label to indicate the type of issue, for
instance, bug, build, documentation, enhancement, etc.
-->
## Motivation and Context
<!--- Why is this change required? What problem does it solve? -->
This PR is needed due to changes in Kokkos deep_copy (currently on Kokkos' develop branch) that are incompatible with the existing implementation in Sacado
## Related Issues
<!---
If applicable, let us know how this merge request is related to any other open
issues or pull requests:
-->
* Closes
* Blocks
* Is blocked by
* Follows
* Precedes
* Related to Kokkos integration testing for the next Kokkos snapshot into Trilinos
* Part of
* Composed of
## How Has This Been Tested?
<!---
Please describe in detail how you tested your changes. Include details of your
testing environment and the tests you ran to see how your change affects other
areas of the code. Consider including configure, build, and test log files.
-->
Locally with Kokkos' develop branch and within the kokkos-develop branch of Trilinos with gcc/5.4 on Serial and OpenMP backends.
## Screenshots
<!--- Not obligatory, but is there anything pertinent that we should see? -->
## Additional Information
<!--- Anything else we need to know in evaluating this merge request? -->
This PR is labeled WIP for review purposes - may be more appropriate to merge into the kokkos-develop branch pending review and further testing.
Will undergo more extensive testing as part of the Trilinos integration test process with Kokkos.
https://gitlab.osti.gov/jmwille/Trilinos/-/merge_requests/2701Rol trilinos develop as of afff363 "Adding cantilever beam tests."2018-05-09T14:55:21ZJames WillenbringRol trilinos develop as of afff363 "Adding cantilever beam tests."*Created by: bartlettroscoe*
CC: @trilinos/rol
## Description
This is a PR that merged the state of rol-trilinos/develop as of the commit:
```
afff363 "Adding cantilever beam tests."
Author: Drew P. Kouri <dpkouri@sandia.gov...*Created by: bartlettroscoe*
CC: @trilinos/rol
## Description
This is a PR that merged the state of rol-trilinos/develop as of the commit:
```
afff363 "Adding cantilever beam tests."
Author: Drew P. Kouri <dpkouri@sandia.gov>
Date: Mon May 7 12:53:50 2018 -0600 (31 hours ago)
M packages/rol/test/step/CMakeLists.txt
A packages/rol/test/step/test_16.cpp
A packages/rol/test/step/test_17.cpp
```
https://gitlab.osti.gov/jmwille/Trilinos/-/merge_requests/4218TriBITS Snapshot: Fix bad SOVERSION when Version.cmake is missing , change 'W...2019-01-20T00:45:45ZJames WillenbringTriBITS Snapshot: Fix bad SOVERSION when Version.cmake is missing , change 'WARNING' to 'NOTE'*Created by: bartlettroscoe*
CC: @jmgate
Bring in TriBITS updates to address TriBITSPub/TriBITS#270 (bad SOVERSION for shared libs) TriBITSPub/TriBITS#271 (replace 'WARNING:' with 'NOTE:' to avoid CDash warnings).
*Created by: bartlettroscoe*
CC: @jmgate
Bring in TriBITS updates to address TriBITSPub/TriBITS#270 (bad SOVERSION for shared libs) TriBITSPub/TriBITS#271 (replace 'WARNING:' with 'NOTE:' to avoid CDash warnings).
https://gitlab.osti.gov/jmwille/Trilinos/-/merge_requests/2261MueLu: print data in advanced xml mode (#2260)2018-02-20T15:56:00ZJames WillenbringMueLu: print data in advanced xml mode (#2260)*Created by: mayrmt*
Allow to print "Nullspace" and "Coordinates" to files from the advanced xml input deck.
## Description
In the 'expert' xml interface (Tobias' style), the parameter list 'DataToWrite' controls output of quantitie...*Created by: mayrmt*
Allow to print "Nullspace" and "Coordinates" to files from the advanced xml input deck.
## Description
In the 'expert' xml interface (Tobias' style), the parameter list 'DataToWrite' controls output of quantities to files. Now, we can also print 'Nullspace' and 'Coordinates'. This has been available in the simple xml-file for quite a while.
Now, xml parameters for printing level matrices, prolongators, and restrictors are consistently named "A", "P", and "R". A deprecation warning is issued if the advanced xml input uses its old parameter names.
## Related Issues
* Closes #2260
## How Has This Been Tested?
- ctest passes on my machine
- Correct behaviour of new output capability has been verified locally
## Checklist
<!---
Go over all the following points, and put an `x` in all the boxes that apply.
If you are unsure about any of these, please ask—we are here to help.
-->
- [x] My commit messages mention the appropriate GitHub issue numbers.
- [x] My code follows the code style of the affected package(s).
- [x] My change requires a change to the documentation.
- [x] I have updated the documentation accordingly.
- [x] I have read the [code contribution guidelines](../blob/master/CONTRIBUTING.md) for this project.
- [ ] I have added tests to cover my changes.
- [x] All new and existing tests passed.
- [x] No new compiler warnings were introduced.
- [ ] These changes break backwards compatibility.
@trilinos/muelu https://gitlab.osti.gov/jmwille/Trilinos/-/merge_requests/4179MueLu: modifying structured aggregation to compute graph of prolongator2019-01-15T15:20:52ZJames WillenbringMueLu: modifying structured aggregation to compute graph of prolongator*Created by: lucbv*
@trilinos/muelu
## Description
This PR modifies the prolongatorGraph output from structured aggregation.
Instead of getting a node based graph it now returns a dofs based graph.
This means that in the new form...*Created by: lucbv*
@trilinos/muelu
## Description
This PR modifies the prolongatorGraph output from structured aggregation.
Instead of getting a node based graph it now returns a dofs based graph.
This means that in the new format, the prolongator graph can be used to call the constructor of the prolongator matrix directly.
## Motivation and Context
This has the advantage of reducing some overhead and it also simplifies the implementation of the geometric interpolation factory.
## Related Issues
* Closes #4173
## How Has This Been Tested?
All the tests in MueLu have been run and passed
## Checklist
- [x] My commit messages mention the appropriate GitHub issue numbers.
- [x] My code follows the code style of the affected package(s).
- [x] I have read the [code contribution guidelines](../blob/master/CONTRIBUTING.md) for this project.
- [x] I have added tests to cover my changes.
- [x] All new and existing tests passed.
- [x] No new compiler warnings were introduced.https://gitlab.osti.gov/jmwille/Trilinos/-/merge_requests/2738Tempus: List available Steppers if requested one is not found.2018-05-14T17:12:15ZJames WillenbringTempus: List available Steppers if requested one is not found.*Created by: ccober6*
#2736
@trilinos/tempus
## Checklist
- [x] My commit messages mention the appropriate GitHub issue numbers.
- [x] My code follows the code style of the affected package(s).
- [ ] My change requires a ...*Created by: ccober6*
#2736
@trilinos/tempus
## Checklist
- [x] My commit messages mention the appropriate GitHub issue numbers.
- [x] My code follows the code style of the affected package(s).
- [ ] My change requires a change to the documentation.
- [ ] I have updated the documentation accordingly.
- [x] I have read the [code contribution guidelines](../blob/master/CONTRIBUTING.md) for this project.
- [ ] I have added tests to cover my changes.
- [x] All new and existing tests passed.
- [x] No new compiler warnings were introduced.
- [x] These changes break backwards compatibility.
https://gitlab.osti.gov/jmwille/Trilinos/-/merge_requests/2910AztecOO: Remove Copy Constructor Compiler Warning2018-06-08T13:52:01ZJames WillenbringAztecOO: Remove Copy Constructor Compiler Warning*Created by: jmgate*
@trilinos/aztecoo
## Description
Remove compiler warning by explicitly calling base class copy constructors.
## Motivation and Context
Just hoping for a cleaner build. In this case the way things were orig...*Created by: jmgate*
@trilinos/aztecoo
## Description
Remove compiler warning by explicitly calling base class copy constructors.
## Motivation and Context
Just hoping for a cleaner build. In this case the way things were originally written left the meaning unclear, which is why `gcc-7.3.0` is warning about it.
## How Has This Been Tested?
@trilinos-autotester
## Checklist
- [ ] My commit messages mention the appropriate GitHub issue numbers.
- [x] My code follows the code style of the affected package(s).
- [ ] My change requires a change to the documentation.
- [ ] I have updated the documentation accordingly.
- [x] I have read the [code contribution guidelines](../blob/master/CONTRIBUTING.md) for this project.
- [ ] I have added tests to cover my changes.
- [ ] All new and existing tests passed.
- [ ] No new compiler warnings were introduced.
- [ ] These changes break backwards compatibility.