Zoltan2 metric refactor #1
Closed
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
added 5 commits
April 8, 2016 15:52
…es classes to have a derived base class.
…rs to not reference virtual functions.
…fies the use of those classes to use the general base class MetricBase
… removed the test code in test driver.
MicheldeMessieres
pushed a commit
that referenced
this pull request
Aug 16, 2016
This is mostly for extra information. For instance, running 2D laplacian produces Setup #1: no reuse 0.02187 (1) Setup #2: reuse none 0.01018 (1) Setup trilinos#3: reuse smoothers 0.01025 (1) Setup trilinos#4: reuse tentative P 0.006917 (1) Setup trilinos#5: reuse smoothed P and R 0.002708 (1) This means that about 50% of no reuse time is actually spent in the interpreter, and the real construction time take 50%. This also shows that there is no reuse happening in smoothers which could not be seen if one only compare #1 and trilinos#3.
MicheldeMessieres
added a commit
that referenced
this pull request
Jun 21, 2017
This reworks the tests so that we run 4 modes:
Original - uses the original zoltan code
Relic - similar to original but includes some C++ memory allocation
Tpetra - runs using a Tpetra map
Kokkos - runs using a Kokkos map
Performance can now be evaluated directly since they run as a batch.
This is the current output:
Start 1: Zoltan2_directoryTest_Original_MPI_4
1/4 Test #1: Zoltan2_directoryTest_Original_MPI_4 ...***Failed 0.92 sec
Start 2: Zoltan2_directoryTest_Relic_MPI_4
2/4 Test #2: Zoltan2_directoryTest_Relic_MPI_4 ...... Passed 1.42 sec
Start 3: Zoltan2_directoryTest_Tpetra_MPI_4
3/4 Test trilinos#3: Zoltan2_directoryTest_Tpetra_MPI_4 ..... Passed 2.46 sec
Start 4: Zoltan2_directoryTest_Kokkos_MPI_4
4/4 Test trilinos#4: Zoltan2_directoryTest_Kokkos_MPI_4 ..... Passed 1.94 sec
Note that the first fail is expected since the original zoltan
doesn't handle Add which this test checks for success of, but it should
be running a similar code sequence so times are still relevant.
Initial times indicate there is some work to be done to get this
running with similar performance but even Tpetra is only about 2x
worse so it's not terrible.
MicheldeMessieres
added a commit
that referenced
this pull request
Jun 29, 2017
This reworks the tests so that we run 4 modes:
Original - uses the original zoltan code
Relic - similar to original but includes some C++ memory allocation
Tpetra - runs using a Tpetra map
Kokkos - runs using a Kokkos map
Performance can now be evaluated directly since they run as a batch.
This is the current output:
Start 1: Zoltan2_directoryTest_Original_MPI_4
1/4 Test #1: Zoltan2_directoryTest_Original_MPI_4 ...***Failed 0.92 sec
Start 2: Zoltan2_directoryTest_Relic_MPI_4
2/4 Test #2: Zoltan2_directoryTest_Relic_MPI_4 ...... Passed 1.42 sec
Start 3: Zoltan2_directoryTest_Tpetra_MPI_4
3/4 Test trilinos#3: Zoltan2_directoryTest_Tpetra_MPI_4 ..... Passed 2.46 sec
Start 4: Zoltan2_directoryTest_Kokkos_MPI_4
4/4 Test trilinos#4: Zoltan2_directoryTest_Kokkos_MPI_4 ..... Passed 1.94 sec
Note that the first fail is expected since the original zoltan
doesn't handle Add which this test checks for success of, but it should
be running a similar code sequence so times are still relevant.
Initial times indicate there is some work to be done to get this
running with similar performance but even Tpetra is only about 2x
worse so it's not terrible.
MicheldeMessieres
pushed a commit
that referenced
this pull request
Nov 13, 2018
… into FROSch_Thyra * 'thyra_FROSch_sort' of https://github.com/roeverf/Trilinos: Clean Up Epetra Version Eliminate Epetra Version->Files located at src/adapters/Epetra Thyra_FROSCh_Xpetra Use Amesos Klu as subdomain solver fix #1 thyra test New initialize() ExtractFromParameterList TwoLevelPreconditioner+ExtractCoordinates(not tested ) no message no message example stratimikos ParameterList Issue ZwischenStand Thyra Thyra_FROSch_Epetra Thyra -> New start Test Thyra Adapters-> still not working -> wrong Map Test Thyra_FROSch —not working yet! interface start
kddevin
pushed a commit
that referenced
this pull request
Jan 18, 2019
MicheldeMessieres
pushed a commit
that referenced
this pull request
May 8, 2019
MicheldeMessieres
pushed a commit
that referenced
this pull request
Jun 4, 2019
MicheldeMessieres
pushed a commit
that referenced
this pull request
Jun 4, 2019
…kCrsMatrix-v001 Xpetra: ETI for Xpetra_TpetraBlockCrsMatrix (attempt #1)
MicheldeMessieres
pushed a commit
that referenced
this pull request
Jun 19, 2019
MicheldeMessieres
pushed a commit
that referenced
this pull request
Aug 7, 2019
Tpetra assembly Updates
MicheldeMessieres
pushed a commit
that referenced
this pull request
Jun 9, 2020
ATDM: ats1: Final fixups (ATDV-357)
jrobcary
pushed a commit
that referenced
this pull request
Dec 7, 2021
Add super robust gather_build_stats.py and other stuff for PR trilinos#8638
jrobcary
pushed a commit
that referenced
this pull request
May 10, 2022
atdm/sems-rhel7: Add drivers and update supported builds
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
No description provided.