Release 4.2.00 #2040
Merged
Release 4.2.00 #2040
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
- Rename uploChar to uplo in fillMode for consistency with usage - Dereference uplo in fillMode usage for comparison to char
Correcting 'syr test' bug causing compilation errors with Trilinos
MKL: support indices properly
Fix for rocblas builds
…kkos-4.0 Backward-compatible fix with [email protected]
docs: Add testing table
- Have common infrastructure detect "-h" and "--help", and print options - In help text, distinguish between enabled and non-enabled backends
This should be good for now but eventually more work is needed for the multivector case which will use the gemm function instead of gemv from onemkl Signed-off-by: Berger-Vergiat <[email protected]>
oneMKL spmv
Use extraction timestamps for fetched Google Benchmark files
in batched dense
KokkosKernels is not even calling the wrapper macro kokkoskernels_exclude_autotools_files() which calls tribits_exclude_autotools_files(). KokkosKernels has no autotools files anymore and support for tribits_exclude_autotools_files() was stripped out of TriBITS because it is not needed anymore. There will be a slight hack in Trilinos 'develop' to avoid needing to make this change on Trilinos 'develop'. But the next time KokkosKernels is snapshotted into Trilinos, we can remove that (minor) hack.
Co-authored-by: Shane W. D. Hart <[email protected]>
cusparse 12 requires that the y (output) vector has 16-byte alignment. So check for this and call the native fallback if y has less alignment. This fixes Trilinos #11926.
…lude_autotools_files Remove calling tribits_exclude_autotools_files()
ode/ in the binary dir doesn't exist as a header directory (because the ODE component has no auto-generated headers for ETI). The unit tests might be there, but contain no headers that need to be found. This fixes errors when building against a Trilinos installation, where CMake looks for this directory that doesn't exist.
Adds an abstraction over the diagonals merge-matrix M of two sorted views A and B, where M[i,] = A[i] > B[j]. A diagonal of such a matrix is the entries running from the bottom-left to the top-right (from large I / small J to small I / large J). MergeMatrixDiagonal presents a 1D-View-Like interface into a specific diagonal of the merge matrix of two views A and B, allowing the caller to retrieve the value of entries along the diagonal.
BSPGEMM: removing cusparse testing for version older than 11.4.0
Common: remove half and bhalf implementations
Update corresponding to kokkos/kokkos#6520
…n-check CMakeLists.txt: Update Kokkos version to 4.2.99 for version check
Remove /etc/profile.d/modules.sh from cm_test_all_sandia for blake
iostream clean-up in benchmarks
Update Kokkos version check to include 4.2.00 (upcoming release)
Revsions suggested by @lucbv - Allow versions greater than or equal to existing release - If Kokkos_VERSION is greater than expected develop branch version, add warning to update the version check
par_ilut: Update documentation for fill_in_limit (cherry picked from commit ab0a32d)
bhalf_t fix for isnan function (cherry picked from commit b9f3d78)
Experimental hip cleanup (cherry picked from commit d9a67b9)
hide native merge-path SpMV behind "native-merge" (cherry picked from commit 89df0f9)
…aliases KokkosKernelsConfig.cmake: add all_libs target and necessary aliases (cherry picked from commit 2c70f24)
Creation of the 'lapack' subdirectory, parallel to 'blas' (cherry picked from commit e7b6c12)
Quick fix on 'lapack' subdirectory, for night compilation with Trilinos (cherry picked from commit 9caa7ca)
SYCL: fix for Trilinos build with MKL (cherry picked from commit 78455b4)
* Adding Changelog for Release 4.2.0 Part of Kokkos C++ Performance Portability Programming EcoSystem 4.2 * Formatting the changelog a bit more Mentioning more clearly LAPACK vs BLAS, grouping PRs by logical work unit, etc... * Remove minor revisions, improve text descriptions * Changelog: add spmv perftest detail --------- Co-authored-by: Luc Berger <[email protected]> Co-authored-by: Carl Pearson <[email protected]> Co-authored-by: brian-kelley <[email protected]>
Part of Kokkos C++ Performance Portability Programming EcoSystem 4.2
|
Snapshot to Trilinos part of trilinos/Trilinos#12505 |
crtrott
approved these changes
Nov 20, 2023
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.