Skip to content
[Jenkins]Jenkins
log in
Started 3 mo 27 days ago
Took 2 hr 0 min
Success

Build #2 (Jul 19, 2024, 9:56:03 AM)

Changes
  1. Reenable TestHIP_Memory_Requirements (details / githubweb)
  2. SYCL: Cleanup device selection (details / githubweb)
  3. Redeine deep_copy for UnorderMap (details / githubweb)
  4. Remove unused typedef (details / githubweb)
  5. Adding SECURITY.md file (details / githubweb)
  6. Only call deep_copy_view() from deep_copy(), add deprecation warning (details / githubweb)
  7. Cleanup KOKKOS_CONFIGURE_CORE (details / githubweb)
  8. [4.2.01]: changelog update  (#6656) (details / githubweb)
  9. Clean up shift_{right, left}_team_impl (#6821) (details / githubweb)
  10. Avoid detecting unwanted fences in the parallel_scan_no_fence test (details / githubweb)
  11. Add test for current CTAD support with RangePolicy (#6803) (details / githubweb)
  12. multi-GPU support: Add test for all policies (#6782) (details / githubweb)
  13. Implement sort_by_key (#6801) (details / githubweb)
  14. Fix scorecard workflow (#6831) (details / githubweb)
  15. Remove redundant RangePolicy constructor (#6841) (details / githubweb)
  16. Bump Google Benchmark version v1.{6.2 -> 7.1} in CMake FetchContent (details / githubweb)
  17. Remove variadic range policy constructor (#6845) (details / githubweb)
  18. Avoid unused variable warning in TestRangePolicyCTAD.cpp (details / githubweb)
  19. bytes_and_flops with CMake (details / githubweb)
  20. bytes_and_flops fix a small bug in command line argument (details / githubweb)
  21. add gather benchmark to CMake (details / githubweb)
  22. add atomic benchmark to CMake (details / githubweb)
  23. add policy_performance benchmark to CMake (details / githubweb)
  24. fix some warnings in policy_performance benchmark (details / githubweb)
  25. policy_benchmark: fix indentation (details / githubweb)
  26. Benchmarks: disable 2 benchmarks for OpenMPTarget (details / githubweb)
  27. bytes_and_flops: fix a counter name (details / githubweb)
  28. Fix C-style cast (details / githubweb)
  29. Fix amdclang++ compilation (#6857) (details / githubweb)
  30. Fix fallback implementation for sort_by_key (#6856) (details / githubweb)
  31. Update the unsafe implicit conversion error message in MDRangePolicy (#6855) (details / githubweb)
  32. Split some classes from Kokkos_ViewMapping (#6859) (details / githubweb)
  33. update mdspan tpl (details / githubweb)
  34. CTAD (deduction guides) for RangePolicy (#6850) (details / githubweb)
  35. CTAD (deduction guides) for MDRangePolicy (#5516) (details / githubweb)
  36. Add support for rocThrust in sort when using HIP (#6793) (details / githubweb)
  37. Add HIP specialization for sort-by-key (details / githubweb)
  38. Fix unused variable warning in TestMDRangePolicyCTAD.cpp (details / githubweb)
  39. Improve message on view out of bounds access and always abort (#6861) (details / githubweb)
  40. Fix linking with rothrust in downstream applications (details / githubweb)
  41. SYCL: Make sure to call find_dependency for oneDPL if necessary  (#6870) (details / githubweb)
  42. Don't use Fedora development version in GitHub CI (details / githubweb)
  43. kokkos_tpls.cmake: update default option to enable rocthrust (details / githubweb)
  44. Scratch space fix for MultiGPU (#6866) (details / githubweb)
  45. [ci skip] Bump version number to 4.3.99 (details / githubweb)
  46. Fix deep copy when filling Rank-7 views (details / githubweb)
  47. TestViewCopy_c.hpp: add new unit test for deep copy (ViewFill) (details / githubweb)
  48. TestViewCopy_c.hpp: better handling for OpenMPTarget (details / githubweb)
  49. Use builtin for atomic_fetch in the HIP backend (details / githubweb)
  50. Fix Makefile.kokkos for Threads (details / githubweb)
  51. Fix a bug in Makefile when using AMD GPU architectures (#6892) (details / githubweb)
  52. Cuda: Fix configuring with CMake 3.28.4 (#6898) (details / githubweb)
  53. Update Intel GPU architectures in Makefile (#6895) (details / githubweb)
  54. Drop Experimental::LayoutTiled class template (details / githubweb)
  55. Get rid of now unnecessary use of is_layouttiled trait (details / githubweb)
  56. Deprecate is_layouttiled trait (details / githubweb)
  57. Move `Kokkos::Array` tests to a more suitable place (#6905) (details / githubweb)
  58. Do not return a copy of the input functor for Kokkos::Experimental::for_each (details / githubweb)
  59. Drop specialization of ViewMapping for Kokkos::Array (details / githubweb)
  60. Accommodate users that depend on a code that define silly macros (#6909) (details / githubweb)
  61. SYCL: Fix multi-GPU support and add test (#6887) (details / githubweb)
  62. SYCL: Unroll shuffle loops for top-level parallel_reduce and parallel_scan (#6750) (details / githubweb)
  63. Preparing readme for develop as the default branch (#6796) (details / githubweb)
  64. CHANGELOG.md: 4.3.00 update (details / githubweb)
  65. Refactor: Uniformize `create_mirror*` parameter name for views (#6917) (details / githubweb)
  66. SYCL: Improve team_reduce implementation (#6562) (details / githubweb)
  67. Use recommended/max team size functions in Cuda ParallelFor and Reduce constructors (#6891) (details / githubweb)
  68. SYCL: Fix range in subgroup scan for workgroup_scan (details / githubweb)
  69. Add ROCm 6.0 in the nightly CI (details / githubweb)
  70. MI300 support unified memory support (#6877) (details / githubweb)
  71. cuda: reduction with `RangePolicy`: fix grid dimensions to work for large values and avoid overflow (#6578) (details / githubweb)
  72. Fix nightly CI (details / githubweb)
  73. Deprecate trailing Proxy template argument in Kokkos::Array (details / githubweb)
  74. CUDA: Update nvcc_wrapper (details / githubweb)
  75. bring back --fmad option to nvcc_wrapper (#6931) (details / githubweb)
  76. Remove unnecessary header include (details / githubweb)
  77. Adding converting constructor in Kokkos::RandomAccessIterator (#6929) (details / githubweb)
  78. Prefer standard C++ feature testing to guard the C++20 requires expression (details / githubweb)
  79. Add `kokkos_swap(Array<T, N>)` sepcialization (details / githubweb)
  80. Deprecate specialization of Kokkos::pair for a single element (details / githubweb)
  81. Fix deprecated warning from `Kokkos::Array` specialization (#6945) (details / githubweb)
  82. [ci skip] Enable deprecated code and deprecated warnings in nightly CI (details / githubweb)
  83. Serial: Use the provided execution space instance in TeamPolicy (details / githubweb)
  84. Fixed the link for P6601 (Threads backend change) (details / githubweb)
  85. Add test (details / githubweb)
  86. Use Kokkos::AUTO for OpenMPTarget (details / githubweb)
  87. kokkos_swap(Array) member friend should not be templated on some other type U (details / githubweb)
  88. Fix noexcept specification for kokkos_swap on zero-sized arrays (details / githubweb)
  89. Replace deprecated sycl::device_ptr/sycl::host_ptr (details / githubweb)
  90. Introduce alias based on feature macro (details / githubweb)
  91. Kokkos::Impl::SYCLTypes:: -> Kokkos::Impl::sycl_ (details / githubweb)
  92. Fix wrong macro guards for deprecated Kokkos::pair<T1,void> specialization (details / githubweb)
  93. Fix support for Kokkos::Array of const-qualified element type (details / githubweb)
  94. Try to fix the CUDA 11.0 build (details / githubweb)
  95. Revert "Try to fix the CUDA 11.0 build" (details / githubweb)
  96. Alternate definition of Impl::is_nothrow_swappable_v for NVCC version less than 11.4 (details / githubweb)
  97. Avoid introducing a 2nd definition of the Impl::swappable trait (details / githubweb)
  98. Do not require OpenMP support for languages other than CXX (details / githubweb)
  99. Update version (details / githubweb)
  100. Add Linux Foundation notice and fix C++ standard (details / githubweb)
  101. SYCL: Avoid deprecated floating-point number abs overloads (#6959) (details / githubweb)
  102. OpenMPTarget: Update loop order in MDRange (#6925) (details / githubweb)
  103. Threads: Don't silently allow m_instance to be a nullptr (#6969) (details / githubweb)
  104. Fix enabling OpenMP with HIP and "compile as CMake language" (details / githubweb)
  105. Fix OpenMP+CUDA when `Kokkos_ENABLE_COMPILE_AS_CMAKE_LANGUAGE` is `ON` (details / githubweb)
  106. Fix fedora CI builds with flang-new (details / githubweb)
  107. Fix Copyright file (details / githubweb)
  108. Also use is_nothrow_swappable workaround for Intel Classic Compilers (#6983) (details / githubweb)
  109. Add thread-safety tests (#6938) (details / githubweb)
  110. Fix deprecation warnings with GCC for pair<T1,void> comparison operators (details / githubweb)
  111. Fix TPL_LIBRARY_SUFFIXES for 32-bit build (details / githubweb)
  112. Avoid duplicated definition of KOKKOS_IMPL_32BIT (details / githubweb)
  113. Disable failing parallel_scan_with_reducers test (details / githubweb)
  114. Fix a bug when using realloc on views of non-default constructible element types (#6993) (details / githubweb)
  115. SYCL: Prepare Parallel* for Graphs (#6988) (details / githubweb)
  116. SYCL: Fix deprecation in custom parallel_for RangePolicy implementation (details / githubweb)
  117. [ci skip] update changelog for 4.3.1 (#6995) (details / githubweb)
  118. OpenMPTarget: Use mutex lock for parallel scan. (details / githubweb)
  119. Remove cuSPARSE TPL (details / githubweb)
  120. SYCL: Print submission command queue property (#7004) (details / githubweb)
  121. Test DualView resize/realloc for types without default constructor (details / githubweb)
  122. Suppress deprecated warnings via pragma push/pop in the tests (#6999) (details / githubweb)
  123. This PR changes the default execution behavior of the parallel_for(team-policy) constructs in the OpenACC backend. (details / githubweb)
  124. Use parallel_for instead of parallel_reduce for check (details / githubweb)
  125. Fix unique_any_senders nvcc template deduction (details / githubweb)
  126. Properly guard deprecated `Kokkos_Vector.hpp` header self contained test (#7016) (details / githubweb)
  127. mdspan converting constructors (#6830) (details / githubweb)
  128. SYCL: Don't use shuffles for top-level reductions (#7009) (details / githubweb)
  129. Add a CITATION.cff file (#7008) (details / githubweb)
  130. Always run Graph tests (#7011) (details / githubweb)
  131. Fix undefined behavior in is_zero_byte (#7014) (details / githubweb)
  132. Disable OpenMPTarget Kokkos::Graph test (does not compile) (details / githubweb)
  133. Fixup for disabling deprecation warnings with NVC++ (details / githubweb)
  134. Remove unused CudaInternal::cuda_{malloc,free}_async_wrapper (details / githubweb)
  135. Improve `Impl::is_zero_byte()` (#7017) (details / githubweb)
  136. OpenMPTarget: Fix compiling Graph tests (#7020) (details / githubweb)
  137. SYCL: Skip launch_six Graph test (details / githubweb)
  138. HIP: Use builtin atomic for compare_exchange (#7000) (details / githubweb)
  139. Fix Kokkos README's FENL link (details / githubweb)
  140. Copied the deduction guides and test cases over from branch (details / githubweb)
  141. Refactor: Replace SFINAE by `if constexpr` for `create_mirror*` functions (#6955) (details / githubweb)
  142. Workaround icpc warnings (details / githubweb)
  143. Adding `Kokkos::to_array` (#6375) (details / githubweb)
  144. Fix adjacent difference (#6922) (details / githubweb)
  145. TestArray: add intel guard to to_array implicit conversion test (details / githubweb)
  146. Avoid using "#if not defined" (details / githubweb)
  147. (Rebase) Partial fix to compile time issues w/nvcc + Kokkos_ENABLE_DEBUG_BOUNDS_CHECK (#7013) (details / githubweb)
  148. Add dependabot for GitHub Actions (details / githubweb)
  149. Added static_asserts for out of range tuple_element and get (details / githubweb)
  150. Add tests to `ScopeGuard` (#7028) (details / githubweb)
  151. Improve Kokkos Graphs (#7039) (details / githubweb)
  152. Introduce `KOKKOS_DEDUCTION_GUIDE` macro to allow user-defined deduction guide in device code for clang compiler (#6954) (details / githubweb)
  153. Add nightly CI on Frontier (#7048) (details / githubweb)
  154. Update scorecard GitHub workflow (details / githubweb)
  155. Fix unused symbols in CTAD tests (details / githubweb)
  156. Cuda: Fix nvcc warnings (details / githubweb)
  157. Fix quotation marks in CXX flags (details / githubweb)
  158. Fix kokkos_swap (details / githubweb)
  159. Fix array size (details / githubweb)
  160. Fix gtest (details / githubweb)
  161. Fix .jenkins whitespce (details / githubweb)
  162. Only use -Werror all-warnings with explicit nvcc_wrapper (details / githubweb)
  163. Fix dangling reference (details / githubweb)
  164. Suppress 'long double' is treated as 'double' in device code (details / githubweb)
  165. Use -Xcudafe --diag_suppress=20208 for 11.6 build; nothing else seems to help (details / githubweb)
  166. Try moving pragma suppress to tests (details / githubweb)
  167. Use -Xcudafe --diag_suppress=20208 in Makefile build (details / githubweb)
  168. Move dependabot to .github (details / githubweb)
  169. Bump Jimver/cuda-toolkit from 0.2.14 to 0.2.15 (details / githubweb)
  170. Bump DoozyX/clang-format-lint-action from 0.16.2 to 0.17 (details / githubweb)
  171. Bump version in the readme (details / githubweb)
  172. Restrict to_array subtest to NVCC >= 11.4.0 (details / githubweb)
  173. [ci skip] Use ROCM 6.1 in the nightly CI and disable one test (details / githubweb)
  174. Switch to using functors in sort_by_key_via_sort (#7059) (details / githubweb)
  175. Fix mdspan test (details / githubweb)
  176. Update to CUDA 12.4.1 in MSVC CI (details / githubweb)
  177. Workaround MSVC compiler issues in Views (details / githubweb)
  178. Avoid lambda in sort_by_key_via_sort (details / githubweb)
  179. Fix using CUDAToolkit for CMake 3.28.4 and higher (details / githubweb)
  180. Move check CMake 3.20.1 with nvhpc (details / githubweb)
  181. Fix closing brackets (details / githubweb)
  182. KOKKOS_CUDA_ERROR->DEFAULT_MSG (details / githubweb)
  183. Workaround icpc "missing return statement at end of non-void function" (details / githubweb)
  184. Enable MDSPAN support by default (#7069) (details / githubweb)
  185. Add SpaceAwareAccessor (details / githubweb)
  186. Add basic tests for SpaceAwareAccessor (details / githubweb)
  187. Use SpaceAwareAccessor in View mdspan-interop (details / githubweb)
  188. Add specialization of SpaceAwareAcc for AnonymousSpace (details / githubweb)
  189. Address review comments (details / githubweb)
  190. Fix incompatible dynamic_extent definition in Kokkos (details / githubweb)
  191. SpaceAwareAccessor: fix issues (no-unique-address, is_empty) (details / githubweb)
  192. Fix using shared libraries and -fvisibility=hidden (#7065) (details / githubweb)
  193. OpenACC: Skip exec_space_thread_safety_range_scan (#7022) (details / githubweb)
  194. Structured binding support for Kokkos::complex (#7040) (details / githubweb)
  195. Fix thread-safety for the Serial backend (details / githubweb)
  196. Update from GCC 13 to 14 and use C++ 26 in Jenkins nightly (details / githubweb)
  197. Move `layout_iterate_type_selector` into Impl namespace (#7076) (details / githubweb)
  198. Use correct includes for spaceawareaccessor (details / githubweb)
  199. Improve performance for deleting an instance. (details / githubweb)
  200. fix refcount exception safety (#6289) (details / githubweb)
  201. Don't delete special member functions explicitly (details / githubweb)
  202. Complex needs a tuple include (details / githubweb)
  203. Update internal desul file copies to 60c1115 (details / githubweb)
  204. Update desul version in github workflow (details / githubweb)
  205. Resolve various bogus icpc -Werror (#7079) (details / githubweb)
  206. Remove support for NVHPC as CUDA device compiler (details / githubweb)
  207. remove cmake options to change default spaces (details / githubweb)
  208. Temporarily disable failing parts in the TestMDSpan.hpp for the OpenACC (details / githubweb)
  209. Add missing space (details / githubweb)
  210. Fix more icpc issues (details / githubweb)
  211. Bump Jimver/cuda-toolkit from 0.2.15 to 0.2.16 (details / githubweb)
  212. Copy print_configuration setting in combination of kokkos settings. (details / githubweb)
  213. Couple more icpc -Werror fixes (details / githubweb)
  214. Cleanup KokkosP hooks in `Profiling::` (#7096) (details / githubweb)
  215. Remove FIXME_NVHPC 23.7 guards (details / githubweb)
  216. Fix SpaceAwareAccessor based on usage experiment in View (#7088) (details / githubweb)
  217. Added width 8 abi for avx2 (details / githubweb)
  218. Added for AVX512 (details / githubweb)
  219. Added for width 4 for NEON (details / githubweb)
  220. clang-formatted (details / githubweb)
  221. Disabling simd unit tests from building for Windows+CUDA build (details / githubweb)
  222. Workaround for the compilation failure for rocm 5.6-6.0 (details / githubweb)
  223. Added a comment about is_type structs (details / githubweb)
  224. clang formating (details / githubweb)
  225. Restrict jenkins CI not to run on hopper for nvcc < 11.8 (details / githubweb)
  226. cmake: fix relative to find kokkos_compiler_launcher (details / githubweb)
  227. Clean up KOKKOS_LIB_INCLUDE_DIRECTORIES, append include directories to associated targets in Trilinos builds (#7103) (details / githubweb)
  228. Workaround to ice with icpc when using -no-ip (#7106) (details / githubweb)
  229. move view allocation related functionality to a new header (#7110) (details / githubweb)
  230. SYCL CI: Manually build oneDPL (details / githubweb)
  231. Add missing concepts include in test (details / githubweb)
  232. Github CI: Test with C++17, C++20, and C++23 (#7082) (details / githubweb)
  233. Update base image for ROCm 5.6 (details / githubweb)
  234. Update mdspan to 98a12b01b51b2 (details / githubweb)
  235. Add AtomicAccessorRelaxed (#7089) (details / githubweb)
  236. Fix div by zero in math ops testing (details / githubweb)
  237. Fix and test with -fsanitize=undefined in GitHub CI (#7104) (details / githubweb)
  238. Bump actions/upload-artifact from 4.3.3 to 4.3.4 (details / githubweb)
  239. Update HIP nightly build base image Ubuntu 20.04 -> 22.04 (details / githubweb)
  240. Deprecate `RawMemoryAllocationFailure::FailureMode::MaximumCudaUVMAllocationsExceeded` (#7120) (details / githubweb)
  241. Refactor: Move logic of `create_mirror*` to `Impl::create_mirror*` (#7061) (details / githubweb)
  242. Define minimum compiler versions for C++20 support (details / githubweb)
  243. Drop (unused) cuda uvm allocations counter (details / githubweb)
  244. Do not bother catching memory allocation failure and rethrow (details / githubweb)
  245. Fix using and, or, xor in desul with MSVC (#7124) (details / githubweb)
  246. Do not swallow errors when deallocating memory with CUDA (details / githubweb)
  247. Let the throwing push finalize hook calls terminate test actually run (details / githubweb)
  248. Drop (unused) KOKKOS_ADD_ADVANCED_TEST TriBITS function (details / githubweb)
  249. Simplify the logic when finalizing and calling the registered functions (details / githubweb)
  250. [ci skip] Fix ROCm version to 6.1.2 in nightly CI (details / githubweb)
  251. Disable the PushFinalizeHookTerminate test on Windows (details / githubweb)
  252. OpenMP: Ensure kernels submitted by multiple threads to the same instance don't run concurrently (#6151) (details / githubweb)
  253. SYCL: Add missing include for std::stringstream (details / githubweb)
  254. SYCL: Add support for Graphs (#6912) (details / githubweb)
  255. [ci skip] rename jenkins build (details / githubweb)
  256. NVHPC: Disable failing NaN tests (details / githubweb)
  257. no_device_stack is unknown (details / githubweb)
  258. Check for LIBCXX 10 or later for C++20 and later (#7123) (details / githubweb)
  259. Get rid of `RawMemoryAllocationFailure::AllocationMechanism` and derived backend-specific exceptions (#7139) (details / githubweb)
  260. Add `likely` and `unlikely` attribute from C++20 to ref counting in views (#6730) (details / githubweb)
  261. Throw bad alloc if omp_target_alloc() returns nullptr (details / githubweb)
  262. Enable test_view_allocation_error with OpenMPTarget (details / githubweb)
  263. Fix gcc-14 C++26 nightly jenkins build (#7137) (details / githubweb)
  264. Fix Trilinos nightly failure due to `create_mirror*` refactor (#7126) (details / githubweb)
  265. fixup! Throw bad alloc if omp_target_alloc() returns nullptr (details / githubweb)
  266. Add CMake options to control compilation flags for AMD GPUs (#7127) (details / githubweb)
  267. Fix using View without corresponding mdspan-type (#7140) (details / githubweb)
Changes
  1. Reenable TestHIP_Memory_Requirements (details / githubweb)
  2. SYCL: Cleanup device selection (details / githubweb)
  3. Redeine deep_copy for UnorderMap (details / githubweb)
  4. Remove unused typedef (details / githubweb)
  5. Adding SECURITY.md file (details / githubweb)
  6. Only call deep_copy_view() from deep_copy(), add deprecation warning (details / githubweb)
  7. Cleanup KOKKOS_CONFIGURE_CORE (details / githubweb)
  8. [4.2.01]: changelog update  (#6656) (details / githubweb)
  9. Clean up shift_{right, left}_team_impl (#6821) (details / githubweb)
  10. Avoid detecting unwanted fences in the parallel_scan_no_fence test (details / githubweb)
  11. Add test for current CTAD support with RangePolicy (#6803) (details / githubweb)
  12. multi-GPU support: Add test for all policies (#6782) (details / githubweb)
  13. Implement sort_by_key (#6801) (details / githubweb)
  14. Fix scorecard workflow (#6831) (details / githubweb)
  15. Remove redundant RangePolicy constructor (#6841) (details / githubweb)
  16. Bump Google Benchmark version v1.{6.2 -> 7.1} in CMake FetchContent (details / githubweb)
  17. Remove variadic range policy constructor (#6845) (details / githubweb)
  18. Avoid unused variable warning in TestRangePolicyCTAD.cpp (details / githubweb)
  19. bytes_and_flops with CMake (details / githubweb)
  20. bytes_and_flops fix a small bug in command line argument (details / githubweb)
  21. add gather benchmark to CMake (details / githubweb)
  22. add atomic benchmark to CMake (details / githubweb)
  23. add policy_performance benchmark to CMake (details / githubweb)
  24. fix some warnings in policy_performance benchmark (details / githubweb)
  25. policy_benchmark: fix indentation (details / githubweb)
  26. Benchmarks: disable 2 benchmarks for OpenMPTarget (details / githubweb)
  27. bytes_and_flops: fix a counter name (details / githubweb)
  28. Fix C-style cast (details / githubweb)
  29. Fix amdclang++ compilation (#6857) (details / githubweb)
  30. Fix fallback implementation for sort_by_key (#6856) (details / githubweb)
  31. Update the unsafe implicit conversion error message in MDRangePolicy (#6855) (details / githubweb)
  32. Split some classes from Kokkos_ViewMapping (#6859) (details / githubweb)
  33. update mdspan tpl (details / githubweb)
  34. CTAD (deduction guides) for RangePolicy (#6850) (details / githubweb)
  35. CTAD (deduction guides) for MDRangePolicy (#5516) (details / githubweb)
  36. Add support for rocThrust in sort when using HIP (#6793) (details / githubweb)
  37. Add HIP specialization for sort-by-key (details / githubweb)
  38. Fix unused variable warning in TestMDRangePolicyCTAD.cpp (details / githubweb)
  39. Improve message on view out of bounds access and always abort (#6861) (details / githubweb)
  40. Fix linking with rothrust in downstream applications (details / githubweb)
  41. SYCL: Make sure to call find_dependency for oneDPL if necessary  (#6870) (details / githubweb)
  42. Don't use Fedora development version in GitHub CI (details / githubweb)
  43. kokkos_tpls.cmake: update default option to enable rocthrust (details / githubweb)
  44. Scratch space fix for MultiGPU (#6866) (details / githubweb)
  45. [ci skip] Bump version number to 4.3.99 (details / githubweb)
  46. Fix deep copy when filling Rank-7 views (details / githubweb)
  47. TestViewCopy_c.hpp: add new unit test for deep copy (ViewFill) (details / githubweb)
  48. TestViewCopy_c.hpp: better handling for OpenMPTarget (details / githubweb)
  49. Use builtin for atomic_fetch in the HIP backend (details / githubweb)
  50. Fix Makefile.kokkos for Threads (details / githubweb)
  51. Fix a bug in Makefile when using AMD GPU architectures (#6892) (details / githubweb)
  52. Cuda: Fix configuring with CMake 3.28.4 (#6898) (details / githubweb)
  53. Update Intel GPU architectures in Makefile (#6895) (details / githubweb)
  54. Drop Experimental::LayoutTiled class template (details / githubweb)
  55. Get rid of now unnecessary use of is_layouttiled trait (details / githubweb)
  56. Deprecate is_layouttiled trait (details / githubweb)
  57. Move `Kokkos::Array` tests to a more suitable place (#6905) (details / githubweb)
  58. Do not return a copy of the input functor for Kokkos::Experimental::for_each (details / githubweb)
  59. Drop specialization of ViewMapping for Kokkos::Array (details / githubweb)
  60. Accommodate users that depend on a code that define silly macros (#6909) (details / githubweb)
  61. SYCL: Fix multi-GPU support and add test (#6887) (details / githubweb)
  62. SYCL: Unroll shuffle loops for top-level parallel_reduce and parallel_scan (#6750) (details / githubweb)
  63. Preparing readme for develop as the default branch (#6796) (details / githubweb)
  64. CHANGELOG.md: 4.3.00 update (details / githubweb)
  65. Refactor: Uniformize `create_mirror*` parameter name for views (#6917) (details / githubweb)
  66. SYCL: Improve team_reduce implementation (#6562) (details / githubweb)
  67. Use recommended/max team size functions in Cuda ParallelFor and Reduce constructors (#6891) (details / githubweb)
  68. SYCL: Fix range in subgroup scan for workgroup_scan (details / githubweb)
  69. Add ROCm 6.0 in the nightly CI (details / githubweb)
  70. MI300 support unified memory support (#6877) (details / githubweb)
  71. cuda: reduction with `RangePolicy`: fix grid dimensions to work for large values and avoid overflow (#6578) (details / githubweb)
  72. Fix nightly CI (details / githubweb)
  73. Deprecate trailing Proxy template argument in Kokkos::Array (details / githubweb)
  74. CUDA: Update nvcc_wrapper (details / githubweb)
  75. bring back --fmad option to nvcc_wrapper (#6931) (details / githubweb)
  76. Remove unnecessary header include (details / githubweb)
  77. Adding converting constructor in Kokkos::RandomAccessIterator (#6929) (details / githubweb)
  78. Prefer standard C++ feature testing to guard the C++20 requires expression (details / githubweb)
  79. Add `kokkos_swap(Array<T, N>)` sepcialization (details / githubweb)
  80. Deprecate specialization of Kokkos::pair for a single element (details / githubweb)
  81. Fix deprecated warning from `Kokkos::Array` specialization (#6945) (details / githubweb)
  82. [ci skip] Enable deprecated code and deprecated warnings in nightly CI (details / githubweb)
  83. Serial: Use the provided execution space instance in TeamPolicy (details / githubweb)
  84. Fixed the link for P6601 (Threads backend change) (details / githubweb)
  85. Add test (details / githubweb)
  86. Use Kokkos::AUTO for OpenMPTarget (details / githubweb)
  87. kokkos_swap(Array) member friend should not be templated on some other type U (details / githubweb)
  88. Fix noexcept specification for kokkos_swap on zero-sized arrays (details / githubweb)
  89. Replace deprecated sycl::device_ptr/sycl::host_ptr (details / githubweb)
  90. Introduce alias based on feature macro (details / githubweb)
  91. Kokkos::Impl::SYCLTypes:: -> Kokkos::Impl::sycl_ (details / githubweb)
  92. Fix wrong macro guards for deprecated Kokkos::pair<T1,void> specialization (details / githubweb)
  93. Fix support for Kokkos::Array of const-qualified element type (details / githubweb)
  94. Try to fix the CUDA 11.0 build (details / githubweb)
  95. Revert "Try to fix the CUDA 11.0 build" (details / githubweb)
  96. Alternate definition of Impl::is_nothrow_swappable_v for NVCC version less than 11.4 (details / githubweb)
  97. Avoid introducing a 2nd definition of the Impl::swappable trait (details / githubweb)
  98. Do not require OpenMP support for languages other than CXX (details / githubweb)
  99. Update version (details / githubweb)
  100. Add Linux Foundation notice and fix C++ standard (details / githubweb)
  101. SYCL: Avoid deprecated floating-point number abs overloads (#6959) (details / githubweb)
  102. OpenMPTarget: Update loop order in MDRange (#6925) (details / githubweb)
  103. Threads: Don't silently allow m_instance to be a nullptr (#6969) (details / githubweb)
  104. Fix enabling OpenMP with HIP and "compile as CMake language" (details / githubweb)
  105. Fix OpenMP+CUDA when `Kokkos_ENABLE_COMPILE_AS_CMAKE_LANGUAGE` is `ON` (details / githubweb)
  106. Fix fedora CI builds with flang-new (details / githubweb)
  107. Fix Copyright file (details / githubweb)
  108. Also use is_nothrow_swappable workaround for Intel Classic Compilers (#6983) (details / githubweb)
  109. Add thread-safety tests (#6938) (details / githubweb)
  110. Fix deprecation warnings with GCC for pair<T1,void> comparison operators (details / githubweb)
  111. Fix TPL_LIBRARY_SUFFIXES for 32-bit build (details / githubweb)
  112. Avoid duplicated definition of KOKKOS_IMPL_32BIT (details / githubweb)
  113. Disable failing parallel_scan_with_reducers test (details / githubweb)
  114. Fix a bug when using realloc on views of non-default constructible element types (#6993) (details / githubweb)
  115. SYCL: Prepare Parallel* for Graphs (#6988) (details / githubweb)
  116. SYCL: Fix deprecation in custom parallel_for RangePolicy implementation (details / githubweb)
  117. [ci skip] update changelog for 4.3.1 (#6995) (details / githubweb)
  118. OpenMPTarget: Use mutex lock for parallel scan. (details / githubweb)
  119. Remove cuSPARSE TPL (details / githubweb)
  120. SYCL: Print submission command queue property (#7004) (details / githubweb)
  121. Test DualView resize/realloc for types without default constructor (details / githubweb)
  122. Suppress deprecated warnings via pragma push/pop in the tests (#6999) (details / githubweb)
  123. This PR changes the default execution behavior of the parallel_for(team-policy) constructs in the OpenACC backend. (details / githubweb)
  124. Use parallel_for instead of parallel_reduce for check (details / githubweb)
  125. Fix unique_any_senders nvcc template deduction (details / githubweb)
  126. Properly guard deprecated `Kokkos_Vector.hpp` header self contained test (#7016) (details / githubweb)
  127. mdspan converting constructors (#6830) (details / githubweb)
  128. SYCL: Don't use shuffles for top-level reductions (#7009) (details / githubweb)
  129. Add a CITATION.cff file (#7008) (details / githubweb)
  130. Always run Graph tests (#7011) (details / githubweb)
  131. Fix undefined behavior in is_zero_byte (#7014) (details / githubweb)
  132. Disable OpenMPTarget Kokkos::Graph test (does not compile) (details / githubweb)
  133. Fixup for disabling deprecation warnings with NVC++ (details / githubweb)
  134. Remove unused CudaInternal::cuda_{malloc,free}_async_wrapper (details / githubweb)
  135. Improve `Impl::is_zero_byte()` (#7017) (details / githubweb)
  136. OpenMPTarget: Fix compiling Graph tests (#7020) (details / githubweb)
  137. SYCL: Skip launch_six Graph test (details / githubweb)
  138. HIP: Use builtin atomic for compare_exchange (#7000) (details / githubweb)
  139. Fix Kokkos README's FENL link (details / githubweb)
  140. Copied the deduction guides and test cases over from branch (details / githubweb)
  141. Refactor: Replace SFINAE by `if constexpr` for `create_mirror*` functions (#6955) (details / githubweb)
  142. Workaround icpc warnings (details / githubweb)
  143. Adding `Kokkos::to_array` (#6375) (details / githubweb)
  144. Fix adjacent difference (#6922) (details / githubweb)
  145. TestArray: add intel guard to to_array implicit conversion test (details / githubweb)
  146. Avoid using "#if not defined" (details / githubweb)
  147. (Rebase) Partial fix to compile time issues w/nvcc + Kokkos_ENABLE_DEBUG_BOUNDS_CHECK (#7013) (details / githubweb)
  148. Add dependabot for GitHub Actions (details / githubweb)
  149. Added static_asserts for out of range tuple_element and get (details / githubweb)
  150. Add tests to `ScopeGuard` (#7028) (details / githubweb)
  151. Improve Kokkos Graphs (#7039) (details / githubweb)
  152. Introduce `KOKKOS_DEDUCTION_GUIDE` macro to allow user-defined deduction guide in device code for clang compiler (#6954) (details / githubweb)
  153. Add nightly CI on Frontier (#7048) (details / githubweb)
  154. Update scorecard GitHub workflow (details / githubweb)
  155. Fix unused symbols in CTAD tests (details / githubweb)
  156. Cuda: Fix nvcc warnings (details / githubweb)
  157. Fix quotation marks in CXX flags (details / githubweb)
  158. Fix kokkos_swap (details / githubweb)
  159. Fix array size (details / githubweb)
  160. Fix gtest (details / githubweb)
  161. Fix .jenkins whitespce (details / githubweb)
  162. Only use -Werror all-warnings with explicit nvcc_wrapper (details / githubweb)
  163. Fix dangling reference (details / githubweb)
  164. Suppress 'long double' is treated as 'double' in device code (details / githubweb)
  165. Use -Xcudafe --diag_suppress=20208 for 11.6 build; nothing else seems to help (details / githubweb)
  166. Try moving pragma suppress to tests (details / githubweb)
  167. Use -Xcudafe --diag_suppress=20208 in Makefile build (details / githubweb)
  168. Move dependabot to .github (details / githubweb)
  169. Bump Jimver/cuda-toolkit from 0.2.14 to 0.2.15 (details / githubweb)
  170. Bump DoozyX/clang-format-lint-action from 0.16.2 to 0.17 (details / githubweb)
  171. Bump version in the readme (details / githubweb)
  172. Restrict to_array subtest to NVCC >= 11.4.0 (details / githubweb)
  173. [ci skip] Use ROCM 6.1 in the nightly CI and disable one test (details / githubweb)
  174. Switch to using functors in sort_by_key_via_sort (#7059) (details / githubweb)
  175. Fix mdspan test (details / githubweb)
  176. Update to CUDA 12.4.1 in MSVC CI (details / githubweb)
  177. Workaround MSVC compiler issues in Views (details / githubweb)
  178. Avoid lambda in sort_by_key_via_sort (details / githubweb)
  179. Fix using CUDAToolkit for CMake 3.28.4 and higher (details / githubweb)
  180. Move check CMake 3.20.1 with nvhpc (details / githubweb)
  181. Fix closing brackets (details / githubweb)
  182. KOKKOS_CUDA_ERROR->DEFAULT_MSG (details / githubweb)
  183. Workaround icpc "missing return statement at end of non-void function" (details / githubweb)
  184. Enable MDSPAN support by default (#7069) (details / githubweb)
  185. Add SpaceAwareAccessor (details / githubweb)
  186. Add basic tests for SpaceAwareAccessor (details / githubweb)
  187. Use SpaceAwareAccessor in View mdspan-interop (details / githubweb)
  188. Add specialization of SpaceAwareAcc for AnonymousSpace (details / githubweb)
  189. Address review comments (details / githubweb)
  190. Fix incompatible dynamic_extent definition in Kokkos (details / githubweb)
  191. SpaceAwareAccessor: fix issues (no-unique-address, is_empty) (details / githubweb)
  192. Fix using shared libraries and -fvisibility=hidden (#7065) (details / githubweb)
  193. OpenACC: Skip exec_space_thread_safety_range_scan (#7022) (details / githubweb)
  194. Structured binding support for Kokkos::complex (#7040) (details / githubweb)
  195. Fix thread-safety for the Serial backend (details / githubweb)
  196. Update from GCC 13 to 14 and use C++ 26 in Jenkins nightly (details / githubweb)
  197. Move `layout_iterate_type_selector` into Impl namespace (#7076) (details / githubweb)
  198. Use correct includes for spaceawareaccessor (details / githubweb)
  199. Improve performance for deleting an instance. (details / githubweb)
  200. fix refcount exception safety (#6289) (details / githubweb)
  201. Don't delete special member functions explicitly (details / githubweb)
  202. Complex needs a tuple include (details / githubweb)
  203. Update internal desul file copies to 60c1115 (details / githubweb)
  204. Update desul version in github workflow (details / githubweb)
  205. Resolve various bogus icpc -Werror (#7079) (details / githubweb)
  206. Remove support for NVHPC as CUDA device compiler (details / githubweb)
  207. remove cmake options to change default spaces (details / githubweb)
  208. Temporarily disable failing parts in the TestMDSpan.hpp for the OpenACC (details / githubweb)
  209. Add missing space (details / githubweb)
  210. Fix more icpc issues (details / githubweb)
  211. Bump Jimver/cuda-toolkit from 0.2.15 to 0.2.16 (details / githubweb)
  212. Copy print_configuration setting in combination of kokkos settings. (details / githubweb)
  213. Couple more icpc -Werror fixes (details / githubweb)
  214. Cleanup KokkosP hooks in `Profiling::` (#7096) (details / githubweb)
  215. Remove FIXME_NVHPC 23.7 guards (details / githubweb)
  216. Fix SpaceAwareAccessor based on usage experiment in View (#7088) (details / githubweb)
  217. Added width 8 abi for avx2 (details / githubweb)
  218. Added for AVX512 (details / githubweb)
  219. Added for width 4 for NEON (details / githubweb)
  220. clang-formatted (details / githubweb)
  221. Disabling simd unit tests from building for Windows+CUDA build (details / githubweb)
  222. Workaround for the compilation failure for rocm 5.6-6.0 (details / githubweb)
  223. Added a comment about is_type structs (details / githubweb)
  224. clang formating (details / githubweb)
  225. Restrict jenkins CI not to run on hopper for nvcc < 11.8 (details / githubweb)
  226. cmake: fix relative to find kokkos_compiler_launcher (details / githubweb)
  227. Clean up KOKKOS_LIB_INCLUDE_DIRECTORIES, append include directories to associated targets in Trilinos builds (#7103) (details / githubweb)
  228. Workaround to ice with icpc when using -no-ip (#7106) (details / githubweb)
  229. move view allocation related functionality to a new header (#7110) (details / githubweb)
  230. SYCL CI: Manually build oneDPL (details / githubweb)
  231. Add missing concepts include in test (details / githubweb)
  232. Github CI: Test with C++17, C++20, and C++23 (#7082) (details / githubweb)
  233. Update base image for ROCm 5.6 (details / githubweb)
  234. Update mdspan to 98a12b01b51b2 (details / githubweb)
  235. Add AtomicAccessorRelaxed (#7089) (details / githubweb)
  236. Fix div by zero in math ops testing (details / githubweb)
  237. Fix and test with -fsanitize=undefined in GitHub CI (#7104) (details / githubweb)
  238. Bump actions/upload-artifact from 4.3.3 to 4.3.4 (details / githubweb)
  239. Update HIP nightly build base image Ubuntu 20.04 -> 22.04 (details / githubweb)
  240. Deprecate `RawMemoryAllocationFailure::FailureMode::MaximumCudaUVMAllocationsExceeded` (#7120) (details / githubweb)
  241. Refactor: Move logic of `create_mirror*` to `Impl::create_mirror*` (#7061) (details / githubweb)
  242. Define minimum compiler versions for C++20 support (details / githubweb)
  243. Drop (unused) cuda uvm allocations counter (details / githubweb)
  244. Do not bother catching memory allocation failure and rethrow (details / githubweb)
  245. Fix using and, or, xor in desul with MSVC (#7124) (details / githubweb)
  246. Do not swallow errors when deallocating memory with CUDA (details / githubweb)
  247. Let the throwing push finalize hook calls terminate test actually run (details / githubweb)
  248. Drop (unused) KOKKOS_ADD_ADVANCED_TEST TriBITS function (details / githubweb)
  249. Simplify the logic when finalizing and calling the registered functions (details / githubweb)
  250. [ci skip] Fix ROCm version to 6.1.2 in nightly CI (details / githubweb)
  251. Disable the PushFinalizeHookTerminate test on Windows (details / githubweb)
  252. OpenMP: Ensure kernels submitted by multiple threads to the same instance don't run concurrently (#6151) (details / githubweb)
  253. SYCL: Add missing include for std::stringstream (details / githubweb)
  254. SYCL: Add support for Graphs (#6912) (details / githubweb)
  255. [ci skip] rename jenkins build (details / githubweb)
  256. NVHPC: Disable failing NaN tests (details / githubweb)
  257. no_device_stack is unknown (details / githubweb)
  258. Check for LIBCXX 10 or later for C++20 and later (#7123) (details / githubweb)
  259. Get rid of `RawMemoryAllocationFailure::AllocationMechanism` and derived backend-specific exceptions (#7139) (details / githubweb)
  260. Add `likely` and `unlikely` attribute from C++20 to ref counting in views (#6730) (details / githubweb)
  261. Throw bad alloc if omp_target_alloc() returns nullptr (details / githubweb)
  262. Enable test_view_allocation_error with OpenMPTarget (details / githubweb)
  263. Fix gcc-14 C++26 nightly jenkins build (#7137) (details / githubweb)
  264. Fix Trilinos nightly failure due to `create_mirror*` refactor (#7126) (details / githubweb)
  265. fixup! Throw bad alloc if omp_target_alloc() returns nullptr (details / githubweb)
  266. Add CMake options to control compilation flags for AMD GPUs (#7127) (details / githubweb)
  267. Fix using View without corresponding mdspan-type (#7140) (details / githubweb)
Changes
  1. Reenable TestHIP_Memory_Requirements (details / githubweb)
  2. SYCL: Cleanup device selection (details / githubweb)
  3. Redeine deep_copy for UnorderMap (details / githubweb)
  4. Remove unused typedef (details / githubweb)
  5. Adding SECURITY.md file (details / githubweb)
  6. Only call deep_copy_view() from deep_copy(), add deprecation warning (details / githubweb)
  7. Cleanup KOKKOS_CONFIGURE_CORE (details / githubweb)
  8. [4.2.01]: changelog update  (#6656) (details / githubweb)
  9. Clean up shift_{right, left}_team_impl (#6821) (details / githubweb)
  10. Avoid detecting unwanted fences in the parallel_scan_no_fence test (details / githubweb)
  11. Add test for current CTAD support with RangePolicy (#6803) (details / githubweb)
  12. multi-GPU support: Add test for all policies (#6782) (details / githubweb)
  13. Implement sort_by_key (#6801) (details / githubweb)
  14. Fix scorecard workflow (#6831) (details / githubweb)
  15. Remove redundant RangePolicy constructor (#6841) (details / githubweb)
  16. Bump Google Benchmark version v1.{6.2 -> 7.1} in CMake FetchContent (details / githubweb)
  17. Remove variadic range policy constructor (#6845) (details / githubweb)
  18. Avoid unused variable warning in TestRangePolicyCTAD.cpp (details / githubweb)
  19. bytes_and_flops with CMake (details / githubweb)
  20. bytes_and_flops fix a small bug in command line argument (details / githubweb)
  21. add gather benchmark to CMake (details / githubweb)
  22. add atomic benchmark to CMake (details / githubweb)
  23. add policy_performance benchmark to CMake (details / githubweb)
  24. fix some warnings in policy_performance benchmark (details / githubweb)
  25. policy_benchmark: fix indentation (details / githubweb)
  26. Benchmarks: disable 2 benchmarks for OpenMPTarget (details / githubweb)
  27. bytes_and_flops: fix a counter name (details / githubweb)
  28. Fix C-style cast (details / githubweb)
  29. Fix amdclang++ compilation (#6857) (details / githubweb)
  30. Fix fallback implementation for sort_by_key (#6856) (details / githubweb)
  31. Update the unsafe implicit conversion error message in MDRangePolicy (#6855) (details / githubweb)
  32. Split some classes from Kokkos_ViewMapping (#6859) (details / githubweb)
  33. update mdspan tpl (details / githubweb)
  34. CTAD (deduction guides) for RangePolicy (#6850) (details / githubweb)
  35. CTAD (deduction guides) for MDRangePolicy (#5516) (details / githubweb)
  36. Add support for rocThrust in sort when using HIP (#6793) (details / githubweb)
  37. Add HIP specialization for sort-by-key (details / githubweb)
  38. Fix unused variable warning in TestMDRangePolicyCTAD.cpp (details / githubweb)
  39. Improve message on view out of bounds access and always abort (#6861) (details / githubweb)
  40. Fix linking with rothrust in downstream applications (details / githubweb)
  41. SYCL: Make sure to call find_dependency for oneDPL if necessary  (#6870) (details / githubweb)
  42. Don't use Fedora development version in GitHub CI (details / githubweb)
  43. kokkos_tpls.cmake: update default option to enable rocthrust (details / githubweb)
  44. Scratch space fix for MultiGPU (#6866) (details / githubweb)
  45. [ci skip] Bump version number to 4.3.99 (details / githubweb)
  46. Fix deep copy when filling Rank-7 views (details / githubweb)
  47. TestViewCopy_c.hpp: add new unit test for deep copy (ViewFill) (details / githubweb)
  48. TestViewCopy_c.hpp: better handling for OpenMPTarget (details / githubweb)
  49. Use builtin for atomic_fetch in the HIP backend (details / githubweb)
  50. Fix Makefile.kokkos for Threads (details / githubweb)
  51. Fix a bug in Makefile when using AMD GPU architectures (#6892) (details / githubweb)
  52. Cuda: Fix configuring with CMake 3.28.4 (#6898) (details / githubweb)
  53. Update Intel GPU architectures in Makefile (#6895) (details / githubweb)
  54. Drop Experimental::LayoutTiled class template (details / githubweb)
  55. Get rid of now unnecessary use of is_layouttiled trait (details / githubweb)
  56. Deprecate is_layouttiled trait (details / githubweb)
  57. Move `Kokkos::Array` tests to a more suitable place (#6905) (details / githubweb)
  58. Do not return a copy of the input functor for Kokkos::Experimental::for_each (details / githubweb)
  59. Drop specialization of ViewMapping for Kokkos::Array (details / githubweb)
  60. Accommodate users that depend on a code that define silly macros (#6909) (details / githubweb)
  61. SYCL: Fix multi-GPU support and add test (#6887) (details / githubweb)
  62. SYCL: Unroll shuffle loops for top-level parallel_reduce and parallel_scan (#6750) (details / githubweb)
  63. Preparing readme for develop as the default branch (#6796) (details / githubweb)
  64. CHANGELOG.md: 4.3.00 update (details / githubweb)
  65. Refactor: Uniformize `create_mirror*` parameter name for views (#6917) (details / githubweb)
  66. SYCL: Improve team_reduce implementation (#6562) (details / githubweb)
  67. Use recommended/max team size functions in Cuda ParallelFor and Reduce constructors (#6891) (details / githubweb)
  68. SYCL: Fix range in subgroup scan for workgroup_scan (details / githubweb)
  69. Add ROCm 6.0 in the nightly CI (details / githubweb)
  70. MI300 support unified memory support (#6877) (details / githubweb)
  71. cuda: reduction with `RangePolicy`: fix grid dimensions to work for large values and avoid overflow (#6578) (details / githubweb)
  72. Fix nightly CI (details / githubweb)
  73. Deprecate trailing Proxy template argument in Kokkos::Array (details / githubweb)
  74. CUDA: Update nvcc_wrapper (details / githubweb)
  75. bring back --fmad option to nvcc_wrapper (#6931) (details / githubweb)
  76. Remove unnecessary header include (details / githubweb)
  77. Adding converting constructor in Kokkos::RandomAccessIterator (#6929) (details / githubweb)
  78. Prefer standard C++ feature testing to guard the C++20 requires expression (details / githubweb)
  79. Add `kokkos_swap(Array<T, N>)` sepcialization (details / githubweb)
  80. Deprecate specialization of Kokkos::pair for a single element (details / githubweb)
  81. Fix deprecated warning from `Kokkos::Array` specialization (#6945) (details / githubweb)
  82. [ci skip] Enable deprecated code and deprecated warnings in nightly CI (details / githubweb)
  83. Serial: Use the provided execution space instance in TeamPolicy (details / githubweb)
  84. Fixed the link for P6601 (Threads backend change) (details / githubweb)
  85. Add test (details / githubweb)
  86. Use Kokkos::AUTO for OpenMPTarget (details / githubweb)
  87. kokkos_swap(Array) member friend should not be templated on some other type U (details / githubweb)
  88. Fix noexcept specification for kokkos_swap on zero-sized arrays (details / githubweb)
  89. Replace deprecated sycl::device_ptr/sycl::host_ptr (details / githubweb)
  90. Introduce alias based on feature macro (details / githubweb)
  91. Kokkos::Impl::SYCLTypes:: -> Kokkos::Impl::sycl_ (details / githubweb)
  92. Fix wrong macro guards for deprecated Kokkos::pair<T1,void> specialization (details / githubweb)
  93. Fix support for Kokkos::Array of const-qualified element type (details / githubweb)
  94. Try to fix the CUDA 11.0 build (details / githubweb)
  95. Revert "Try to fix the CUDA 11.0 build" (details / githubweb)
  96. Alternate definition of Impl::is_nothrow_swappable_v for NVCC version less than 11.4 (details / githubweb)
  97. Avoid introducing a 2nd definition of the Impl::swappable trait (details / githubweb)
  98. Do not require OpenMP support for languages other than CXX (details / githubweb)
  99. Update version (details / githubweb)
  100. Add Linux Foundation notice and fix C++ standard (details / githubweb)
  101. SYCL: Avoid deprecated floating-point number abs overloads (#6959) (details / githubweb)
  102. OpenMPTarget: Update loop order in MDRange (#6925) (details / githubweb)
  103. Threads: Don't silently allow m_instance to be a nullptr (#6969) (details / githubweb)
  104. Fix enabling OpenMP with HIP and "compile as CMake language" (details / githubweb)
  105. Fix OpenMP+CUDA when `Kokkos_ENABLE_COMPILE_AS_CMAKE_LANGUAGE` is `ON` (details / githubweb)
  106. Fix fedora CI builds with flang-new (details / githubweb)
  107. Fix Copyright file (details / githubweb)
  108. Also use is_nothrow_swappable workaround for Intel Classic Compilers (#6983) (details / githubweb)
  109. Add thread-safety tests (#6938) (details / githubweb)
  110. Fix deprecation warnings with GCC for pair<T1,void> comparison operators (details / githubweb)
  111. Fix TPL_LIBRARY_SUFFIXES for 32-bit build (details / githubweb)
  112. Avoid duplicated definition of KOKKOS_IMPL_32BIT (details / githubweb)
  113. Disable failing parallel_scan_with_reducers test (details / githubweb)
  114. Fix a bug when using realloc on views of non-default constructible element types (#6993) (details / githubweb)
  115. SYCL: Prepare Parallel* for Graphs (#6988) (details / githubweb)
  116. SYCL: Fix deprecation in custom parallel_for RangePolicy implementation (details / githubweb)
  117. [ci skip] update changelog for 4.3.1 (#6995) (details / githubweb)
  118. OpenMPTarget: Use mutex lock for parallel scan. (details / githubweb)
  119. Remove cuSPARSE TPL (details / githubweb)
  120. SYCL: Print submission command queue property (#7004) (details / githubweb)
  121. Test DualView resize/realloc for types without default constructor (details / githubweb)
  122. Suppress deprecated warnings via pragma push/pop in the tests (#6999) (details / githubweb)
  123. This PR changes the default execution behavior of the parallel_for(team-policy) constructs in the OpenACC backend. (details / githubweb)
  124. Use parallel_for instead of parallel_reduce for check (details / githubweb)
  125. Fix unique_any_senders nvcc template deduction (details / githubweb)
  126. Properly guard deprecated `Kokkos_Vector.hpp` header self contained test (#7016) (details / githubweb)
  127. mdspan converting constructors (#6830) (details / githubweb)
  128. SYCL: Don't use shuffles for top-level reductions (#7009) (details / githubweb)
  129. Add a CITATION.cff file (#7008) (details / githubweb)
  130. Always run Graph tests (#7011) (details / githubweb)
  131. Fix undefined behavior in is_zero_byte (#7014) (details / githubweb)
  132. Disable OpenMPTarget Kokkos::Graph test (does not compile) (details / githubweb)
  133. Fixup for disabling deprecation warnings with NVC++ (details / githubweb)
  134. Remove unused CudaInternal::cuda_{malloc,free}_async_wrapper (details / githubweb)
  135. Improve `Impl::is_zero_byte()` (#7017) (details / githubweb)
  136. OpenMPTarget: Fix compiling Graph tests (#7020) (details / githubweb)
  137. SYCL: Skip launch_six Graph test (details / githubweb)
  138. HIP: Use builtin atomic for compare_exchange (#7000) (details / githubweb)
  139. Fix Kokkos README's FENL link (details / githubweb)
  140. Copied the deduction guides and test cases over from branch (details / githubweb)
  141. Refactor: Replace SFINAE by `if constexpr` for `create_mirror*` functions (#6955) (details / githubweb)
  142. Workaround icpc warnings (details / githubweb)
  143. Adding `Kokkos::to_array` (#6375) (details / githubweb)
  144. Fix adjacent difference (#6922) (details / githubweb)
  145. TestArray: add intel guard to to_array implicit conversion test (details / githubweb)
  146. Avoid using "#if not defined" (details / githubweb)
  147. (Rebase) Partial fix to compile time issues w/nvcc + Kokkos_ENABLE_DEBUG_BOUNDS_CHECK (#7013) (details / githubweb)
  148. Add dependabot for GitHub Actions (details / githubweb)
  149. Added static_asserts for out of range tuple_element and get (details / githubweb)
  150. Add tests to `ScopeGuard` (#7028) (details / githubweb)
  151. Improve Kokkos Graphs (#7039) (details / githubweb)
  152. Introduce `KOKKOS_DEDUCTION_GUIDE` macro to allow user-defined deduction guide in device code for clang compiler (#6954) (details / githubweb)
  153. Add nightly CI on Frontier (#7048) (details / githubweb)
  154. Update scorecard GitHub workflow (details / githubweb)
  155. Fix unused symbols in CTAD tests (details / githubweb)
  156. Cuda: Fix nvcc warnings (details / githubweb)
  157. Fix quotation marks in CXX flags (details / githubweb)
  158. Fix kokkos_swap (details / githubweb)
  159. Fix array size (details / githubweb)
  160. Fix gtest (details / githubweb)
  161. Fix .jenkins whitespce (details / githubweb)
  162. Only use -Werror all-warnings with explicit nvcc_wrapper (details / githubweb)
  163. Fix dangling reference (details / githubweb)
  164. Suppress 'long double' is treated as 'double' in device code (details / githubweb)
  165. Use -Xcudafe --diag_suppress=20208 for 11.6 build; nothing else seems to help (details / githubweb)
  166. Try moving pragma suppress to tests (details / githubweb)
  167. Use -Xcudafe --diag_suppress=20208 in Makefile build (details / githubweb)
  168. Move dependabot to .github (details / githubweb)
  169. Bump Jimver/cuda-toolkit from 0.2.14 to 0.2.15 (details / githubweb)
  170. Bump DoozyX/clang-format-lint-action from 0.16.2 to 0.17 (details / githubweb)
  171. Bump version in the readme (details / githubweb)
  172. Restrict to_array subtest to NVCC >= 11.4.0 (details / githubweb)
  173. [ci skip] Use ROCM 6.1 in the nightly CI and disable one test (details / githubweb)
  174. Switch to using functors in sort_by_key_via_sort (#7059) (details / githubweb)
  175. Fix mdspan test (details / githubweb)
  176. Update to CUDA 12.4.1 in MSVC CI (details / githubweb)
  177. Workaround MSVC compiler issues in Views (details / githubweb)
  178. Avoid lambda in sort_by_key_via_sort (details / githubweb)
  179. Fix using CUDAToolkit for CMake 3.28.4 and higher (details / githubweb)
  180. Move check CMake 3.20.1 with nvhpc (details / githubweb)
  181. Fix closing brackets (details / githubweb)
  182. KOKKOS_CUDA_ERROR->DEFAULT_MSG (details / githubweb)
  183. Workaround icpc "missing return statement at end of non-void function" (details / githubweb)
  184. Enable MDSPAN support by default (#7069) (details / githubweb)
  185. Add SpaceAwareAccessor (details / githubweb)
  186. Add basic tests for SpaceAwareAccessor (details / githubweb)
  187. Use SpaceAwareAccessor in View mdspan-interop (details / githubweb)
  188. Add specialization of SpaceAwareAcc for AnonymousSpace (details / githubweb)
  189. Address review comments (details / githubweb)
  190. Fix incompatible dynamic_extent definition in Kokkos (details / githubweb)
  191. SpaceAwareAccessor: fix issues (no-unique-address, is_empty) (details / githubweb)
  192. Fix using shared libraries and -fvisibility=hidden (#7065) (details / githubweb)
  193. OpenACC: Skip exec_space_thread_safety_range_scan (#7022) (details / githubweb)
  194. Structured binding support for Kokkos::complex (#7040) (details / githubweb)
  195. Fix thread-safety for the Serial backend (details / githubweb)
  196. Update from GCC 13 to 14 and use C++ 26 in Jenkins nightly (details / githubweb)
  197. Move `layout_iterate_type_selector` into Impl namespace (#7076) (details / githubweb)
  198. Use correct includes for spaceawareaccessor (details / githubweb)
  199. Improve performance for deleting an instance. (details / githubweb)
  200. fix refcount exception safety (#6289) (details / githubweb)
  201. Don't delete special member functions explicitly (details / githubweb)
  202. Complex needs a tuple include (details / githubweb)
  203. Update internal desul file copies to 60c1115 (details / githubweb)
  204. Update desul version in github workflow (details / githubweb)
  205. Resolve various bogus icpc -Werror (#7079) (details / githubweb)
  206. Remove support for NVHPC as CUDA device compiler (details / githubweb)
  207. remove cmake options to change default spaces (details / githubweb)
  208. Temporarily disable failing parts in the TestMDSpan.hpp for the OpenACC (details / githubweb)
  209. Add missing space (details / githubweb)
  210. Fix more icpc issues (details / githubweb)
  211. Bump Jimver/cuda-toolkit from 0.2.15 to 0.2.16 (details / githubweb)
  212. Copy print_configuration setting in combination of kokkos settings. (details / githubweb)
  213. Couple more icpc -Werror fixes (details / githubweb)
  214. Cleanup KokkosP hooks in `Profiling::` (#7096) (details / githubweb)
  215. Remove FIXME_NVHPC 23.7 guards (details / githubweb)
  216. Fix SpaceAwareAccessor based on usage experiment in View (#7088) (details / githubweb)
  217. Added width 8 abi for avx2 (details / githubweb)
  218. Added for AVX512 (details / githubweb)
  219. Added for width 4 for NEON (details / githubweb)
  220. clang-formatted (details / githubweb)
  221. Disabling simd unit tests from building for Windows+CUDA build (details / githubweb)
  222. Workaround for the compilation failure for rocm 5.6-6.0 (details / githubweb)
  223. Added a comment about is_type structs (details / githubweb)
  224. clang formating (details / githubweb)
  225. Restrict jenkins CI not to run on hopper for nvcc < 11.8 (details / githubweb)
  226. cmake: fix relative to find kokkos_compiler_launcher (details / githubweb)
  227. Clean up KOKKOS_LIB_INCLUDE_DIRECTORIES, append include directories to associated targets in Trilinos builds (#7103) (details / githubweb)
  228. Workaround to ice with icpc when using -no-ip (#7106) (details / githubweb)
  229. move view allocation related functionality to a new header (#7110) (details / githubweb)
  230. SYCL CI: Manually build oneDPL (details / githubweb)
  231. Add missing concepts include in test (details / githubweb)
  232. Github CI: Test with C++17, C++20, and C++23 (#7082) (details / githubweb)
  233. Update base image for ROCm 5.6 (details / githubweb)
  234. Update mdspan to 98a12b01b51b2 (details / githubweb)
  235. Add AtomicAccessorRelaxed (#7089) (details / githubweb)
  236. Fix div by zero in math ops testing (details / githubweb)
  237. Fix and test with -fsanitize=undefined in GitHub CI (#7104) (details / githubweb)
  238. Bump actions/upload-artifact from 4.3.3 to 4.3.4 (details / githubweb)
  239. Update HIP nightly build base image Ubuntu 20.04 -> 22.04 (details / githubweb)
  240. Deprecate `RawMemoryAllocationFailure::FailureMode::MaximumCudaUVMAllocationsExceeded` (#7120) (details / githubweb)
  241. Refactor: Move logic of `create_mirror*` to `Impl::create_mirror*` (#7061) (details / githubweb)
  242. Define minimum compiler versions for C++20 support (details / githubweb)
  243. Drop (unused) cuda uvm allocations counter (details / githubweb)
  244. Do not bother catching memory allocation failure and rethrow (details / githubweb)
  245. Fix using and, or, xor in desul with MSVC (#7124) (details / githubweb)
  246. Do not swallow errors when deallocating memory with CUDA (details / githubweb)
  247. Let the throwing push finalize hook calls terminate test actually run (details / githubweb)
  248. Drop (unused) KOKKOS_ADD_ADVANCED_TEST TriBITS function (details / githubweb)
  249. Simplify the logic when finalizing and calling the registered functions (details / githubweb)
  250. [ci skip] Fix ROCm version to 6.1.2 in nightly CI (details / githubweb)
  251. Disable the PushFinalizeHookTerminate test on Windows (details / githubweb)
  252. OpenMP: Ensure kernels submitted by multiple threads to the same instance don't run concurrently (#6151) (details / githubweb)
  253. SYCL: Add missing include for std::stringstream (details / githubweb)
  254. SYCL: Add support for Graphs (#6912) (details / githubweb)
  255. [ci skip] rename jenkins build (details / githubweb)
  256. NVHPC: Disable failing NaN tests (details / githubweb)
  257. no_device_stack is unknown (details / githubweb)
  258. Check for LIBCXX 10 or later for C++20 and later (#7123) (details / githubweb)
  259. Get rid of `RawMemoryAllocationFailure::AllocationMechanism` and derived backend-specific exceptions (#7139) (details / githubweb)
  260. Add `likely` and `unlikely` attribute from C++20 to ref counting in views (#6730) (details / githubweb)
  261. Throw bad alloc if omp_target_alloc() returns nullptr (details / githubweb)
  262. Enable test_view_allocation_error with OpenMPTarget (details / githubweb)
  263. Fix gcc-14 C++26 nightly jenkins build (#7137) (details / githubweb)
  264. Fix Trilinos nightly failure due to `create_mirror*` refactor (#7126) (details / githubweb)
  265. fixup! Throw bad alloc if omp_target_alloc() returns nullptr (details / githubweb)
  266. Add CMake options to control compilation flags for AMD GPUs (#7127) (details / githubweb)
  267. Fix using View without corresponding mdspan-type (#7140) (details / githubweb)
Changes
  1. Reenable TestHIP_Memory_Requirements (details / githubweb)
  2. SYCL: Cleanup device selection (details / githubweb)
  3. Redeine deep_copy for UnorderMap (details / githubweb)
  4. Remove unused typedef (details / githubweb)
  5. Adding SECURITY.md file (details / githubweb)
  6. Only call deep_copy_view() from deep_copy(), add deprecation warning (details / githubweb)
  7. Cleanup KOKKOS_CONFIGURE_CORE (details / githubweb)
  8. [4.2.01]: changelog update  (#6656) (details / githubweb)
  9. Clean up shift_{right, left}_team_impl (#6821) (details / githubweb)
  10. Avoid detecting unwanted fences in the parallel_scan_no_fence test (details / githubweb)
  11. Add test for current CTAD support with RangePolicy (#6803) (details / githubweb)
  12. multi-GPU support: Add test for all policies (#6782) (details / githubweb)
  13. Implement sort_by_key (#6801) (details / githubweb)
  14. Fix scorecard workflow (#6831) (details / githubweb)
  15. Remove redundant RangePolicy constructor (#6841) (details / githubweb)
  16. Bump Google Benchmark version v1.{6.2 -> 7.1} in CMake FetchContent (details / githubweb)
  17. Remove variadic range policy constructor (#6845) (details / githubweb)
  18. Avoid unused variable warning in TestRangePolicyCTAD.cpp (details / githubweb)
  19. bytes_and_flops with CMake (details / githubweb)
  20. bytes_and_flops fix a small bug in command line argument (details / githubweb)
  21. add gather benchmark to CMake (details / githubweb)
  22. add atomic benchmark to CMake (details / githubweb)
  23. add policy_performance benchmark to CMake (details / githubweb)
  24. fix some warnings in policy_performance benchmark (details / githubweb)
  25. policy_benchmark: fix indentation (details / githubweb)
  26. Benchmarks: disable 2 benchmarks for OpenMPTarget (details / githubweb)
  27. bytes_and_flops: fix a counter name (details / githubweb)
  28. Fix C-style cast (details / githubweb)
  29. Fix amdclang++ compilation (#6857) (details / githubweb)
  30. Fix fallback implementation for sort_by_key (#6856) (details / githubweb)
  31. Update the unsafe implicit conversion error message in MDRangePolicy (#6855) (details / githubweb)
  32. Split some classes from Kokkos_ViewMapping (#6859) (details / githubweb)
  33. update mdspan tpl (details / githubweb)
  34. CTAD (deduction guides) for RangePolicy (#6850) (details / githubweb)
  35. CTAD (deduction guides) for MDRangePolicy (#5516) (details / githubweb)
  36. Add support for rocThrust in sort when using HIP (#6793) (details / githubweb)
  37. Add HIP specialization for sort-by-key (details / githubweb)
  38. Fix unused variable warning in TestMDRangePolicyCTAD.cpp (details / githubweb)
  39. Improve message on view out of bounds access and always abort (#6861) (details / githubweb)
  40. Fix linking with rothrust in downstream applications (details / githubweb)
  41. SYCL: Make sure to call find_dependency for oneDPL if necessary  (#6870) (details / githubweb)
  42. Don't use Fedora development version in GitHub CI (details / githubweb)
  43. kokkos_tpls.cmake: update default option to enable rocthrust (details / githubweb)
  44. Scratch space fix for MultiGPU (#6866) (details / githubweb)
  45. [ci skip] Bump version number to 4.3.99 (details / githubweb)
  46. Fix deep copy when filling Rank-7 views (details / githubweb)
  47. TestViewCopy_c.hpp: add new unit test for deep copy (ViewFill) (details / githubweb)
  48. TestViewCopy_c.hpp: better handling for OpenMPTarget (details / githubweb)
  49. Use builtin for atomic_fetch in the HIP backend (details / githubweb)
  50. Fix Makefile.kokkos for Threads (details / githubweb)
  51. Fix a bug in Makefile when using AMD GPU architectures (#6892) (details / githubweb)
  52. Cuda: Fix configuring with CMake 3.28.4 (#6898) (details / githubweb)
  53. Update Intel GPU architectures in Makefile (#6895) (details / githubweb)
  54. Drop Experimental::LayoutTiled class template (details / githubweb)
  55. Get rid of now unnecessary use of is_layouttiled trait (details / githubweb)
  56. Deprecate is_layouttiled trait (details / githubweb)
  57. Move `Kokkos::Array` tests to a more suitable place (#6905) (details / githubweb)
  58. Do not return a copy of the input functor for Kokkos::Experimental::for_each (details / githubweb)
  59. Drop specialization of ViewMapping for Kokkos::Array (details / githubweb)
  60. Accommodate users that depend on a code that define silly macros (#6909) (details / githubweb)
  61. SYCL: Fix multi-GPU support and add test (#6887) (details / githubweb)
  62. SYCL: Unroll shuffle loops for top-level parallel_reduce and parallel_scan (#6750) (details / githubweb)
  63. Preparing readme for develop as the default branch (#6796) (details / githubweb)
  64. CHANGELOG.md: 4.3.00 update (details / githubweb)
  65. Refactor: Uniformize `create_mirror*` parameter name for views (#6917) (details / githubweb)
  66. SYCL: Improve team_reduce implementation (#6562) (details / githubweb)
  67. Use recommended/max team size functions in Cuda ParallelFor and Reduce constructors (#6891) (details / githubweb)
  68. SYCL: Fix range in subgroup scan for workgroup_scan (details / githubweb)
  69. Add ROCm 6.0 in the nightly CI (details / githubweb)
  70. MI300 support unified memory support (#6877) (details / githubweb)
  71. cuda: reduction with `RangePolicy`: fix grid dimensions to work for large values and avoid overflow (#6578) (details / githubweb)
  72. Fix nightly CI (details / githubweb)
  73. Deprecate trailing Proxy template argument in Kokkos::Array (details / githubweb)
  74. CUDA: Update nvcc_wrapper (details / githubweb)
  75. bring back --fmad option to nvcc_wrapper (#6931) (details / githubweb)
  76. Remove unnecessary header include (details / githubweb)
  77. Adding converting constructor in Kokkos::RandomAccessIterator (#6929) (details / githubweb)
  78. Prefer standard C++ feature testing to guard the C++20 requires expression (details / githubweb)
  79. Add `kokkos_swap(Array<T, N>)` sepcialization (details / githubweb)
  80. Deprecate specialization of Kokkos::pair for a single element (details / githubweb)
  81. Fix deprecated warning from `Kokkos::Array` specialization (#6945) (details / githubweb)
  82. [ci skip] Enable deprecated code and deprecated warnings in nightly CI (details / githubweb)
  83. Serial: Use the provided execution space instance in TeamPolicy (details / githubweb)
  84. Fixed the link for P6601 (Threads backend change) (details / githubweb)
  85. Add test (details / githubweb)
  86. Use Kokkos::AUTO for OpenMPTarget (details / githubweb)
  87. kokkos_swap(Array) member friend should not be templated on some other type U (details / githubweb)
  88. Fix noexcept specification for kokkos_swap on zero-sized arrays (details / githubweb)
  89. Replace deprecated sycl::device_ptr/sycl::host_ptr (details / githubweb)
  90. Introduce alias based on feature macro (details / githubweb)
  91. Kokkos::Impl::SYCLTypes:: -> Kokkos::Impl::sycl_ (details / githubweb)
  92. Fix wrong macro guards for deprecated Kokkos::pair<T1,void> specialization (details / githubweb)
  93. Fix support for Kokkos::Array of const-qualified element type (details / githubweb)
  94. Try to fix the CUDA 11.0 build (details / githubweb)
  95. Revert "Try to fix the CUDA 11.0 build" (details / githubweb)
  96. Alternate definition of Impl::is_nothrow_swappable_v for NVCC version less than 11.4 (details / githubweb)
  97. Avoid introducing a 2nd definition of the Impl::swappable trait (details / githubweb)
  98. Do not require OpenMP support for languages other than CXX (details / githubweb)
  99. Update version (details / githubweb)
  100. Add Linux Foundation notice and fix C++ standard (details / githubweb)
  101. SYCL: Avoid deprecated floating-point number abs overloads (#6959) (details / githubweb)
  102. OpenMPTarget: Update loop order in MDRange (#6925) (details / githubweb)
  103. Threads: Don't silently allow m_instance to be a nullptr (#6969) (details / githubweb)
  104. Fix enabling OpenMP with HIP and "compile as CMake language" (details / githubweb)
  105. Fix OpenMP+CUDA when `Kokkos_ENABLE_COMPILE_AS_CMAKE_LANGUAGE` is `ON` (details / githubweb)
  106. Fix fedora CI builds with flang-new (details / githubweb)
  107. Fix Copyright file (details / githubweb)
  108. Also use is_nothrow_swappable workaround for Intel Classic Compilers (#6983) (details / githubweb)
  109. Add thread-safety tests (#6938) (details / githubweb)
  110. Fix deprecation warnings with GCC for pair<T1,void> comparison operators (details / githubweb)
  111. Fix TPL_LIBRARY_SUFFIXES for 32-bit build (details / githubweb)
  112. Avoid duplicated definition of KOKKOS_IMPL_32BIT (details / githubweb)
  113. Disable failing parallel_scan_with_reducers test (details / githubweb)
  114. Fix a bug when using realloc on views of non-default constructible element types (#6993) (details / githubweb)
  115. SYCL: Prepare Parallel* for Graphs (#6988) (details / githubweb)
  116. SYCL: Fix deprecation in custom parallel_for RangePolicy implementation (details / githubweb)
  117. [ci skip] update changelog for 4.3.1 (#6995) (details / githubweb)
  118. OpenMPTarget: Use mutex lock for parallel scan. (details / githubweb)
  119. Remove cuSPARSE TPL (details / githubweb)
  120. SYCL: Print submission command queue property (#7004) (details / githubweb)
  121. Test DualView resize/realloc for types without default constructor (details / githubweb)
  122. Suppress deprecated warnings via pragma push/pop in the tests (#6999) (details / githubweb)
  123. This PR changes the default execution behavior of the parallel_for(team-policy) constructs in the OpenACC backend. (details / githubweb)
  124. Use parallel_for instead of parallel_reduce for check (details / githubweb)
  125. Fix unique_any_senders nvcc template deduction (details / githubweb)
  126. Properly guard deprecated `Kokkos_Vector.hpp` header self contained test (#7016) (details / githubweb)
  127. mdspan converting constructors (#6830) (details / githubweb)
  128. SYCL: Don't use shuffles for top-level reductions (#7009) (details / githubweb)
  129. Add a CITATION.cff file (#7008) (details / githubweb)
  130. Always run Graph tests (#7011) (details / githubweb)
  131. Fix undefined behavior in is_zero_byte (#7014) (details / githubweb)
  132. Disable OpenMPTarget Kokkos::Graph test (does not compile) (details / githubweb)
  133. Fixup for disabling deprecation warnings with NVC++ (details / githubweb)
  134. Remove unused CudaInternal::cuda_{malloc,free}_async_wrapper (details / githubweb)
  135. Improve `Impl::is_zero_byte()` (#7017) (details / githubweb)
  136. OpenMPTarget: Fix compiling Graph tests (#7020) (details / githubweb)
  137. SYCL: Skip launch_six Graph test (details / githubweb)
  138. HIP: Use builtin atomic for compare_exchange (#7000) (details / githubweb)
  139. Fix Kokkos README's FENL link (details / githubweb)
  140. Copied the deduction guides and test cases over from branch (details / githubweb)
  141. Refactor: Replace SFINAE by `if constexpr` for `create_mirror*` functions (#6955) (details / githubweb)
  142. Workaround icpc warnings (details / githubweb)
  143. Adding `Kokkos::to_array` (#6375) (details / githubweb)
  144. Fix adjacent difference (#6922) (details / githubweb)
  145. TestArray: add intel guard to to_array implicit conversion test (details / githubweb)
  146. Avoid using "#if not defined" (details / githubweb)
  147. (Rebase) Partial fix to compile time issues w/nvcc + Kokkos_ENABLE_DEBUG_BOUNDS_CHECK (#7013) (details / githubweb)
  148. Add dependabot for GitHub Actions (details / githubweb)
  149. Added static_asserts for out of range tuple_element and get (details / githubweb)
  150. Add tests to `ScopeGuard` (#7028) (details / githubweb)
  151. Improve Kokkos Graphs (#7039) (details / githubweb)
  152. Introduce `KOKKOS_DEDUCTION_GUIDE` macro to allow user-defined deduction guide in device code for clang compiler (#6954) (details / githubweb)
  153. Add nightly CI on Frontier (#7048) (details / githubweb)
  154. Update scorecard GitHub workflow (details / githubweb)
  155. Fix unused symbols in CTAD tests (details / githubweb)
  156. Cuda: Fix nvcc warnings (details / githubweb)
  157. Fix quotation marks in CXX flags (details / githubweb)
  158. Fix kokkos_swap (details / githubweb)
  159. Fix array size (details / githubweb)
  160. Fix gtest (details / githubweb)
  161. Fix .jenkins whitespce (details / githubweb)
  162. Only use -Werror all-warnings with explicit nvcc_wrapper (details / githubweb)
  163. Fix dangling reference (details / githubweb)
  164. Suppress 'long double' is treated as 'double' in device code (details / githubweb)
  165. Use -Xcudafe --diag_suppress=20208 for 11.6 build; nothing else seems to help (details / githubweb)
  166. Try moving pragma suppress to tests (details / githubweb)
  167. Use -Xcudafe --diag_suppress=20208 in Makefile build (details / githubweb)
  168. Move dependabot to .github (details / githubweb)
  169. Bump Jimver/cuda-toolkit from 0.2.14 to 0.2.15 (details / githubweb)
  170. Bump DoozyX/clang-format-lint-action from 0.16.2 to 0.17 (details / githubweb)
  171. Bump version in the readme (details / githubweb)
  172. Restrict to_array subtest to NVCC >= 11.4.0 (details / githubweb)
  173. [ci skip] Use ROCM 6.1 in the nightly CI and disable one test (details / githubweb)
  174. Switch to using functors in sort_by_key_via_sort (#7059) (details / githubweb)
  175. Fix mdspan test (details / githubweb)
  176. Update to CUDA 12.4.1 in MSVC CI (details / githubweb)
  177. Workaround MSVC compiler issues in Views (details / githubweb)
  178. Avoid lambda in sort_by_key_via_sort (details / githubweb)
  179. Fix using CUDAToolkit for CMake 3.28.4 and higher (details / githubweb)
  180. Move check CMake 3.20.1 with nvhpc (details / githubweb)
  181. Fix closing brackets (details / githubweb)
  182. KOKKOS_CUDA_ERROR->DEFAULT_MSG (details / githubweb)
  183. Workaround icpc "missing return statement at end of non-void function" (details / githubweb)
  184. Enable MDSPAN support by default (#7069) (details / githubweb)
  185. Add SpaceAwareAccessor (details / githubweb)
  186. Add basic tests for SpaceAwareAccessor (details / githubweb)
  187. Use SpaceAwareAccessor in View mdspan-interop (details / githubweb)
  188. Add specialization of SpaceAwareAcc for AnonymousSpace (details / githubweb)
  189. Address review comments (details / githubweb)
  190. Fix incompatible dynamic_extent definition in Kokkos (details / githubweb)
  191. SpaceAwareAccessor: fix issues (no-unique-address, is_empty) (details / githubweb)
  192. Fix using shared libraries and -fvisibility=hidden (#7065) (details / githubweb)
  193. OpenACC: Skip exec_space_thread_safety_range_scan (#7022) (details / githubweb)
  194. Structured binding support for Kokkos::complex (#7040) (details / githubweb)
  195. Fix thread-safety for the Serial backend (details / githubweb)
  196. Update from GCC 13 to 14 and use C++ 26 in Jenkins nightly (details / githubweb)
  197. Move `layout_iterate_type_selector` into Impl namespace (#7076) (details / githubweb)
  198. Use correct includes for spaceawareaccessor (details / githubweb)
  199. Improve performance for deleting an instance. (details / githubweb)
  200. fix refcount exception safety (#6289) (details / githubweb)
  201. Don't delete special member functions explicitly (details / githubweb)
  202. Complex needs a tuple include (details / githubweb)
  203. Update internal desul file copies to 60c1115 (details / githubweb)
  204. Update desul version in github workflow (details / githubweb)
  205. Resolve various bogus icpc -Werror (#7079) (details / githubweb)
  206. Remove support for NVHPC as CUDA device compiler (details / githubweb)
  207. remove cmake options to change default spaces (details / githubweb)
  208. Temporarily disable failing parts in the TestMDSpan.hpp for the OpenACC (details / githubweb)
  209. Add missing space (details / githubweb)
  210. Fix more icpc issues (details / githubweb)
  211. Bump Jimver/cuda-toolkit from 0.2.15 to 0.2.16 (details / githubweb)
  212. Copy print_configuration setting in combination of kokkos settings. (details / githubweb)
  213. Couple more icpc -Werror fixes (details / githubweb)
  214. Cleanup KokkosP hooks in `Profiling::` (#7096) (details / githubweb)
  215. Remove FIXME_NVHPC 23.7 guards (details / githubweb)
  216. Fix SpaceAwareAccessor based on usage experiment in View (#7088) (details / githubweb)
  217. Added width 8 abi for avx2 (details / githubweb)
  218. Added for AVX512 (details / githubweb)
  219. Added for width 4 for NEON (details / githubweb)
  220. clang-formatted (details / githubweb)
  221. Disabling simd unit tests from building for Windows+CUDA build (details / githubweb)
  222. Workaround for the compilation failure for rocm 5.6-6.0 (details / githubweb)
  223. Added a comment about is_type structs (details / githubweb)
  224. clang formating (details / githubweb)
  225. Restrict jenkins CI not to run on hopper for nvcc < 11.8 (details / githubweb)
  226. cmake: fix relative to find kokkos_compiler_launcher (details / githubweb)
  227. Clean up KOKKOS_LIB_INCLUDE_DIRECTORIES, append include directories to associated targets in Trilinos builds (#7103) (details / githubweb)
  228. Workaround to ice with icpc when using -no-ip (#7106) (details / githubweb)
  229. move view allocation related functionality to a new header (#7110) (details / githubweb)
  230. SYCL CI: Manually build oneDPL (details / githubweb)
  231. Add missing concepts include in test (details / githubweb)
  232. Github CI: Test with C++17, C++20, and C++23 (#7082) (details / githubweb)
  233. Update base image for ROCm 5.6 (details / githubweb)
  234. Update mdspan to 98a12b01b51b2 (details / githubweb)
  235. Add AtomicAccessorRelaxed (#7089) (details / githubweb)
  236. Fix div by zero in math ops testing (details / githubweb)
  237. Fix and test with -fsanitize=undefined in GitHub CI (#7104) (details / githubweb)
  238. Bump actions/upload-artifact from 4.3.3 to 4.3.4 (details / githubweb)
  239. Update HIP nightly build base image Ubuntu 20.04 -> 22.04 (details / githubweb)
  240. Deprecate `RawMemoryAllocationFailure::FailureMode::MaximumCudaUVMAllocationsExceeded` (#7120) (details / githubweb)
  241. Refactor: Move logic of `create_mirror*` to `Impl::create_mirror*` (#7061) (details / githubweb)
  242. Define minimum compiler versions for C++20 support (details / githubweb)
  243. Drop (unused) cuda uvm allocations counter (details / githubweb)
  244. Do not bother catching memory allocation failure and rethrow (details / githubweb)
  245. Fix using and, or, xor in desul with MSVC (#7124) (details / githubweb)
  246. Do not swallow errors when deallocating memory with CUDA (details / githubweb)
  247. Let the throwing push finalize hook calls terminate test actually run (details / githubweb)
  248. Drop (unused) KOKKOS_ADD_ADVANCED_TEST TriBITS function (details / githubweb)
  249. Simplify the logic when finalizing and calling the registered functions (details / githubweb)
  250. [ci skip] Fix ROCm version to 6.1.2 in nightly CI (details / githubweb)
  251. Disable the PushFinalizeHookTerminate test on Windows (details / githubweb)
  252. OpenMP: Ensure kernels submitted by multiple threads to the same instance don't run concurrently (#6151) (details / githubweb)
  253. SYCL: Add missing include for std::stringstream (details / githubweb)
  254. SYCL: Add support for Graphs (#6912) (details / githubweb)
  255. [ci skip] rename jenkins build (details / githubweb)
  256. NVHPC: Disable failing NaN tests (details / githubweb)
  257. no_device_stack is unknown (details / githubweb)
  258. Check for LIBCXX 10 or later for C++20 and later (#7123) (details / githubweb)
  259. Get rid of `RawMemoryAllocationFailure::AllocationMechanism` and derived backend-specific exceptions (#7139) (details / githubweb)
  260. Add `likely` and `unlikely` attribute from C++20 to ref counting in views (#6730) (details / githubweb)
  261. Throw bad alloc if omp_target_alloc() returns nullptr (details / githubweb)
  262. Enable test_view_allocation_error with OpenMPTarget (details / githubweb)
  263. Fix gcc-14 C++26 nightly jenkins build (#7137) (details / githubweb)
  264. Fix Trilinos nightly failure due to `create_mirror*` refactor (#7126) (details / githubweb)
  265. fixup! Throw bad alloc if omp_target_alloc() returns nullptr (details / githubweb)
  266. Add CMake options to control compilation flags for AMD GPUs (#7127) (details / githubweb)
  267. Fix using View without corresponding mdspan-type (#7140) (details / githubweb)
Changes
  1. Reenable TestHIP_Memory_Requirements (details / githubweb)
  2. SYCL: Cleanup device selection (details / githubweb)
  3. Redeine deep_copy for UnorderMap (details / githubweb)
  4. Remove unused typedef (details / githubweb)
  5. Adding SECURITY.md file (details / githubweb)
  6. Only call deep_copy_view() from deep_copy(), add deprecation warning (details / githubweb)
  7. Cleanup KOKKOS_CONFIGURE_CORE (details / githubweb)
  8. [4.2.01]: changelog update  (#6656) (details / githubweb)
  9. Clean up shift_{right, left}_team_impl (#6821) (details / githubweb)
  10. Avoid detecting unwanted fences in the parallel_scan_no_fence test (details / githubweb)
  11. Add test for current CTAD support with RangePolicy (#6803) (details / githubweb)
  12. multi-GPU support: Add test for all policies (#6782) (details / githubweb)
  13. Implement sort_by_key (#6801) (details / githubweb)
  14. Fix scorecard workflow (#6831) (details / githubweb)
  15. Remove redundant RangePolicy constructor (#6841) (details / githubweb)
  16. Bump Google Benchmark version v1.{6.2 -> 7.1} in CMake FetchContent (details / githubweb)
  17. Remove variadic range policy constructor (#6845) (details / githubweb)
  18. Avoid unused variable warning in TestRangePolicyCTAD.cpp (details / githubweb)
  19. bytes_and_flops with CMake (details / githubweb)
  20. bytes_and_flops fix a small bug in command line argument (details / githubweb)
  21. add gather benchmark to CMake (details / githubweb)
  22. add atomic benchmark to CMake (details / githubweb)
  23. add policy_performance benchmark to CMake (details / githubweb)
  24. fix some warnings in policy_performance benchmark (details / githubweb)
  25. policy_benchmark: fix indentation (details / githubweb)
  26. Benchmarks: disable 2 benchmarks for OpenMPTarget (details / githubweb)
  27. bytes_and_flops: fix a counter name (details / githubweb)
  28. Fix C-style cast (details / githubweb)
  29. Fix amdclang++ compilation (#6857) (details / githubweb)
  30. Fix fallback implementation for sort_by_key (#6856) (details / githubweb)
  31. Update the unsafe implicit conversion error message in MDRangePolicy (#6855) (details / githubweb)
  32. Split some classes from Kokkos_ViewMapping (#6859) (details / githubweb)
  33. update mdspan tpl (details / githubweb)
  34. CTAD (deduction guides) for RangePolicy (#6850) (details / githubweb)
  35. CTAD (deduction guides) for MDRangePolicy (#5516) (details / githubweb)
  36. Add support for rocThrust in sort when using HIP (#6793) (details / githubweb)
  37. Add HIP specialization for sort-by-key (details / githubweb)
  38. Fix unused variable warning in TestMDRangePolicyCTAD.cpp (details / githubweb)
  39. Improve message on view out of bounds access and always abort (#6861) (details / githubweb)
  40. Fix linking with rothrust in downstream applications (details / githubweb)
  41. SYCL: Make sure to call find_dependency for oneDPL if necessary  (#6870) (details / githubweb)
  42. Don't use Fedora development version in GitHub CI (details / githubweb)
  43. kokkos_tpls.cmake: update default option to enable rocthrust (details / githubweb)
  44. Scratch space fix for MultiGPU (#6866) (details / githubweb)
  45. [ci skip] Bump version number to 4.3.99 (details / githubweb)
  46. Fix deep copy when filling Rank-7 views (details / githubweb)
  47. TestViewCopy_c.hpp: add new unit test for deep copy (ViewFill) (details / githubweb)
  48. TestViewCopy_c.hpp: better handling for OpenMPTarget (details / githubweb)
  49. Use builtin for atomic_fetch in the HIP backend (details / githubweb)
  50. Fix Makefile.kokkos for Threads (details / githubweb)
  51. Fix a bug in Makefile when using AMD GPU architectures (#6892) (details / githubweb)
  52. Cuda: Fix configuring with CMake 3.28.4 (#6898) (details / githubweb)
  53. Update Intel GPU architectures in Makefile (#6895) (details / githubweb)
  54. Drop Experimental::LayoutTiled class template (details / githubweb)
  55. Get rid of now unnecessary use of is_layouttiled trait (details / githubweb)
  56. Deprecate is_layouttiled trait (details / githubweb)
  57. Move `Kokkos::Array` tests to a more suitable place (#6905) (details / githubweb)
  58. Do not return a copy of the input functor for Kokkos::Experimental::for_each (details / githubweb)
  59. Drop specialization of ViewMapping for Kokkos::Array (details / githubweb)
  60. Accommodate users that depend on a code that define silly macros (#6909) (details / githubweb)
  61. SYCL: Fix multi-GPU support and add test (#6887) (details / githubweb)
  62. SYCL: Unroll shuffle loops for top-level parallel_reduce and parallel_scan (#6750) (details / githubweb)
  63. Preparing readme for develop as the default branch (#6796) (details / githubweb)
  64. CHANGELOG.md: 4.3.00 update (details / githubweb)
  65. Refactor: Uniformize `create_mirror*` parameter name for views (#6917) (details / githubweb)
  66. SYCL: Improve team_reduce implementation (#6562) (details / githubweb)
  67. Use recommended/max team size functions in Cuda ParallelFor and Reduce constructors (#6891) (details / githubweb)
  68. SYCL: Fix range in subgroup scan for workgroup_scan (details / githubweb)
  69. Add ROCm 6.0 in the nightly CI (details / githubweb)
  70. MI300 support unified memory support (#6877) (details / githubweb)
  71. cuda: reduction with `RangePolicy`: fix grid dimensions to work for large values and avoid overflow (#6578) (details / githubweb)
  72. Fix nightly CI (details / githubweb)
  73. Deprecate trailing Proxy template argument in Kokkos::Array (details / githubweb)
  74. CUDA: Update nvcc_wrapper (details / githubweb)
  75. bring back --fmad option to nvcc_wrapper (#6931) (details / githubweb)
  76. Remove unnecessary header include (details / githubweb)
  77. Adding converting constructor in Kokkos::RandomAccessIterator (#6929) (details / githubweb)
  78. Prefer standard C++ feature testing to guard the C++20 requires expression (details / githubweb)
  79. Add `kokkos_swap(Array<T, N>)` sepcialization (details / githubweb)
  80. Deprecate specialization of Kokkos::pair for a single element (details / githubweb)
  81. Fix deprecated warning from `Kokkos::Array` specialization (#6945) (details / githubweb)
  82. [ci skip] Enable deprecated code and deprecated warnings in nightly CI (details / githubweb)
  83. Serial: Use the provided execution space instance in TeamPolicy (details / githubweb)
  84. Fixed the link for P6601 (Threads backend change) (details / githubweb)
  85. Add test (details / githubweb)
  86. Use Kokkos::AUTO for OpenMPTarget (details / githubweb)
  87. kokkos_swap(Array) member friend should not be templated on some other type U (details / githubweb)
  88. Fix noexcept specification for kokkos_swap on zero-sized arrays (details / githubweb)
  89. Replace deprecated sycl::device_ptr/sycl::host_ptr (details / githubweb)
  90. Introduce alias based on feature macro (details / githubweb)
  91. Kokkos::Impl::SYCLTypes:: -> Kokkos::Impl::sycl_ (details / githubweb)
  92. Fix wrong macro guards for deprecated Kokkos::pair<T1,void> specialization (details / githubweb)
  93. Fix support for Kokkos::Array of const-qualified element type (details / githubweb)
  94. Try to fix the CUDA 11.0 build (details / githubweb)
  95. Revert "Try to fix the CUDA 11.0 build" (details / githubweb)
  96. Alternate definition of Impl::is_nothrow_swappable_v for NVCC version less than 11.4 (details / githubweb)
  97. Avoid introducing a 2nd definition of the Impl::swappable trait (details / githubweb)
  98. Do not require OpenMP support for languages other than CXX (details / githubweb)
  99. Update version (details / githubweb)
  100. Add Linux Foundation notice and fix C++ standard (details / githubweb)
  101. SYCL: Avoid deprecated floating-point number abs overloads (#6959) (details / githubweb)
  102. OpenMPTarget: Update loop order in MDRange (#6925) (details / githubweb)
  103. Threads: Don't silently allow m_instance to be a nullptr (#6969) (details / githubweb)
  104. Fix enabling OpenMP with HIP and "compile as CMake language" (details / githubweb)
  105. Fix OpenMP+CUDA when `Kokkos_ENABLE_COMPILE_AS_CMAKE_LANGUAGE` is `ON` (details / githubweb)
  106. Fix fedora CI builds with flang-new (details / githubweb)
  107. Fix Copyright file (details / githubweb)
  108. Also use is_nothrow_swappable workaround for Intel Classic Compilers (#6983) (details / githubweb)
  109. Add thread-safety tests (#6938) (details / githubweb)
  110. Fix deprecation warnings with GCC for pair<T1,void> comparison operators (details / githubweb)
  111. Fix TPL_LIBRARY_SUFFIXES for 32-bit build (details / githubweb)
  112. Avoid duplicated definition of KOKKOS_IMPL_32BIT (details / githubweb)
  113. Disable failing parallel_scan_with_reducers test (details / githubweb)
  114. Fix a bug when using realloc on views of non-default constructible element types (#6993) (details / githubweb)
  115. SYCL: Prepare Parallel* for Graphs (#6988) (details / githubweb)
  116. SYCL: Fix deprecation in custom parallel_for RangePolicy implementation (details / githubweb)
  117. [ci skip] update changelog for 4.3.1 (#6995) (details / githubweb)
  118. OpenMPTarget: Use mutex lock for parallel scan. (details / githubweb)
  119. Remove cuSPARSE TPL (details / githubweb)
  120. SYCL: Print submission command queue property (#7004) (details / githubweb)
  121. Test DualView resize/realloc for types without default constructor (details / githubweb)
  122. Suppress deprecated warnings via pragma push/pop in the tests (#6999) (details / githubweb)
  123. This PR changes the default execution behavior of the parallel_for(team-policy) constructs in the OpenACC backend. (details / githubweb)
  124. Use parallel_for instead of parallel_reduce for check (details / githubweb)
  125. Fix unique_any_senders nvcc template deduction (details / githubweb)
  126. Properly guard deprecated `Kokkos_Vector.hpp` header self contained test (#7016) (details / githubweb)
  127. mdspan converting constructors (#6830) (details / githubweb)
  128. SYCL: Don't use shuffles for top-level reductions (#7009) (details / githubweb)
  129. Add a CITATION.cff file (#7008) (details / githubweb)
  130. Always run Graph tests (#7011) (details / githubweb)
  131. Fix undefined behavior in is_zero_byte (#7014) (details / githubweb)
  132. Disable OpenMPTarget Kokkos::Graph test (does not compile) (details / githubweb)
  133. Fixup for disabling deprecation warnings with NVC++ (details / githubweb)
  134. Remove unused CudaInternal::cuda_{malloc,free}_async_wrapper (details / githubweb)
  135. Improve `Impl::is_zero_byte()` (#7017) (details / githubweb)
  136. OpenMPTarget: Fix compiling Graph tests (#7020) (details / githubweb)
  137. SYCL: Skip launch_six Graph test (details / githubweb)
  138. HIP: Use builtin atomic for compare_exchange (#7000) (details / githubweb)
  139. Fix Kokkos README's FENL link (details / githubweb)
  140. Copied the deduction guides and test cases over from branch (details / githubweb)
  141. Refactor: Replace SFINAE by `if constexpr` for `create_mirror*` functions (#6955) (details / githubweb)
  142. Workaround icpc warnings (details / githubweb)
  143. Adding `Kokkos::to_array` (#6375) (details / githubweb)
  144. Fix adjacent difference (#6922) (details / githubweb)
  145. TestArray: add intel guard to to_array implicit conversion test (details / githubweb)
  146. Avoid using "#if not defined" (details / githubweb)
  147. (Rebase) Partial fix to compile time issues w/nvcc + Kokkos_ENABLE_DEBUG_BOUNDS_CHECK (#7013) (details / githubweb)
  148. Add dependabot for GitHub Actions (details / githubweb)
  149. Added static_asserts for out of range tuple_element and get (details / githubweb)
  150. Add tests to `ScopeGuard` (#7028) (details / githubweb)
  151. Improve Kokkos Graphs (#7039) (details / githubweb)
  152. Introduce `KOKKOS_DEDUCTION_GUIDE` macro to allow user-defined deduction guide in device code for clang compiler (#6954) (details / githubweb)
  153. Add nightly CI on Frontier (#7048) (details / githubweb)
  154. Update scorecard GitHub workflow (details / githubweb)
  155. Fix unused symbols in CTAD tests (details / githubweb)
  156. Cuda: Fix nvcc warnings (details / githubweb)
  157. Fix quotation marks in CXX flags (details / githubweb)
  158. Fix kokkos_swap (details / githubweb)
  159. Fix array size (details / githubweb)
  160. Fix gtest (details / githubweb)
  161. Fix .jenkins whitespce (details / githubweb)
  162. Only use -Werror all-warnings with explicit nvcc_wrapper (details / githubweb)
  163. Fix dangling reference (details / githubweb)
  164. Suppress 'long double' is treated as 'double' in device code (details / githubweb)
  165. Use -Xcudafe --diag_suppress=20208 for 11.6 build; nothing else seems to help (details / githubweb)
  166. Try moving pragma suppress to tests (details / githubweb)
  167. Use -Xcudafe --diag_suppress=20208 in Makefile build (details / githubweb)
  168. Move dependabot to .github (details / githubweb)
  169. Bump Jimver/cuda-toolkit from 0.2.14 to 0.2.15 (details / githubweb)
  170. Bump DoozyX/clang-format-lint-action from 0.16.2 to 0.17 (details / githubweb)
  171. Bump version in the readme (details / githubweb)
  172. Restrict to_array subtest to NVCC >= 11.4.0 (details / githubweb)
  173. [ci skip] Use ROCM 6.1 in the nightly CI and disable one test (details / githubweb)
  174. Switch to using functors in sort_by_key_via_sort (#7059) (details / githubweb)
  175. Fix mdspan test (details / githubweb)
  176. Update to CUDA 12.4.1 in MSVC CI (details / githubweb)
  177. Workaround MSVC compiler issues in Views (details / githubweb)
  178. Avoid lambda in sort_by_key_via_sort (details / githubweb)
  179. Fix using CUDAToolkit for CMake 3.28.4 and higher (details / githubweb)
  180. Move check CMake 3.20.1 with nvhpc (details / githubweb)
  181. Fix closing brackets (details / githubweb)
  182. KOKKOS_CUDA_ERROR->DEFAULT_MSG (details / githubweb)
  183. Workaround icpc "missing return statement at end of non-void function" (details / githubweb)
  184. Enable MDSPAN support by default (#7069) (details / githubweb)
  185. Add SpaceAwareAccessor (details / githubweb)
  186. Add basic tests for SpaceAwareAccessor (details / githubweb)
  187. Use SpaceAwareAccessor in View mdspan-interop (details / githubweb)
  188. Add specialization of SpaceAwareAcc for AnonymousSpace (details / githubweb)
  189. Address review comments (details / githubweb)
  190. Fix incompatible dynamic_extent definition in Kokkos (details / githubweb)
  191. SpaceAwareAccessor: fix issues (no-unique-address, is_empty) (details / githubweb)
  192. Fix using shared libraries and -fvisibility=hidden (#7065) (details / githubweb)
  193. OpenACC: Skip exec_space_thread_safety_range_scan (#7022) (details / githubweb)
  194. Structured binding support for Kokkos::complex (#7040) (details / githubweb)
  195. Fix thread-safety for the Serial backend (details / githubweb)
  196. Update from GCC 13 to 14 and use C++ 26 in Jenkins nightly (details / githubweb)
  197. Move `layout_iterate_type_selector` into Impl namespace (#7076) (details / githubweb)
  198. Use correct includes for spaceawareaccessor (details / githubweb)
  199. Improve performance for deleting an instance. (details / githubweb)
  200. fix refcount exception safety (#6289) (details / githubweb)
  201. Don't delete special member functions explicitly (details / githubweb)
  202. Complex needs a tuple include (details / githubweb)
  203. Update internal desul file copies to 60c1115 (details / githubweb)
  204. Update desul version in github workflow (details / githubweb)
  205. Resolve various bogus icpc -Werror (#7079) (details / githubweb)
  206. Remove support for NVHPC as CUDA device compiler (details / githubweb)
  207. remove cmake options to change default spaces (details / githubweb)
  208. Temporarily disable failing parts in the TestMDSpan.hpp for the OpenACC (details / githubweb)
  209. Add missing space (details / githubweb)
  210. Fix more icpc issues (details / githubweb)
  211. Bump Jimver/cuda-toolkit from 0.2.15 to 0.2.16 (details / githubweb)
  212. Copy print_configuration setting in combination of kokkos settings. (details / githubweb)
  213. Couple more icpc -Werror fixes (details / githubweb)
  214. Cleanup KokkosP hooks in `Profiling::` (#7096) (details / githubweb)
  215. Remove FIXME_NVHPC 23.7 guards (details / githubweb)
  216. Fix SpaceAwareAccessor based on usage experiment in View (#7088) (details / githubweb)
  217. Added width 8 abi for avx2 (details / githubweb)
  218. Added for AVX512 (details / githubweb)
  219. Added for width 4 for NEON (details / githubweb)
  220. clang-formatted (details / githubweb)
  221. Disabling simd unit tests from building for Windows+CUDA build (details / githubweb)
  222. Workaround for the compilation failure for rocm 5.6-6.0 (details / githubweb)
  223. Added a comment about is_type structs (details / githubweb)
  224. clang formating (details / githubweb)
  225. Restrict jenkins CI not to run on hopper for nvcc < 11.8 (details / githubweb)
  226. cmake: fix relative to find kokkos_compiler_launcher (details / githubweb)
  227. Clean up KOKKOS_LIB_INCLUDE_DIRECTORIES, append include directories to associated targets in Trilinos builds (#7103) (details / githubweb)
  228. Workaround to ice with icpc when using -no-ip (#7106) (details / githubweb)
  229. move view allocation related functionality to a new header (#7110) (details / githubweb)
  230. SYCL CI: Manually build oneDPL (details / githubweb)
  231. Add missing concepts include in test (details / githubweb)
  232. Github CI: Test with C++17, C++20, and C++23 (#7082) (details / githubweb)
  233. Update base image for ROCm 5.6 (details / githubweb)
  234. Update mdspan to 98a12b01b51b2 (details / githubweb)
  235. Add AtomicAccessorRelaxed (#7089) (details / githubweb)
  236. Fix div by zero in math ops testing (details / githubweb)
  237. Fix and test with -fsanitize=undefined in GitHub CI (#7104) (details / githubweb)
  238. Bump actions/upload-artifact from 4.3.3 to 4.3.4 (details / githubweb)
  239. Update HIP nightly build base image Ubuntu 20.04 -> 22.04 (details / githubweb)
  240. Deprecate `RawMemoryAllocationFailure::FailureMode::MaximumCudaUVMAllocationsExceeded` (#7120) (details / githubweb)
  241. Refactor: Move logic of `create_mirror*` to `Impl::create_mirror*` (#7061) (details / githubweb)
  242. Define minimum compiler versions for C++20 support (details / githubweb)
  243. Drop (unused) cuda uvm allocations counter (details / githubweb)
  244. Do not bother catching memory allocation failure and rethrow (details / githubweb)
  245. Fix using and, or, xor in desul with MSVC (#7124) (details / githubweb)
  246. Do not swallow errors when deallocating memory with CUDA (details / githubweb)
  247. Let the throwing push finalize hook calls terminate test actually run (details / githubweb)
  248. Drop (unused) KOKKOS_ADD_ADVANCED_TEST TriBITS function (details / githubweb)
  249. Simplify the logic when finalizing and calling the registered functions (details / githubweb)
  250. [ci skip] Fix ROCm version to 6.1.2 in nightly CI (details / githubweb)
  251. Disable the PushFinalizeHookTerminate test on Windows (details / githubweb)
  252. OpenMP: Ensure kernels submitted by multiple threads to the same instance don't run concurrently (#6151) (details / githubweb)
  253. SYCL: Add missing include for std::stringstream (details / githubweb)
  254. SYCL: Add support for Graphs (#6912) (details / githubweb)
  255. [ci skip] rename jenkins build (details / githubweb)
  256. NVHPC: Disable failing NaN tests (details / githubweb)
  257. no_device_stack is unknown (details / githubweb)
  258. Check for LIBCXX 10 or later for C++20 and later (#7123) (details / githubweb)
  259. Get rid of `RawMemoryAllocationFailure::AllocationMechanism` and derived backend-specific exceptions (#7139) (details / githubweb)
  260. Add `likely` and `unlikely` attribute from C++20 to ref counting in views (#6730) (details / githubweb)
  261. Throw bad alloc if omp_target_alloc() returns nullptr (details / githubweb)
  262. Enable test_view_allocation_error with OpenMPTarget (details / githubweb)
  263. Fix gcc-14 C++26 nightly jenkins build (#7137) (details / githubweb)
  264. Fix Trilinos nightly failure due to `create_mirror*` refactor (#7126) (details / githubweb)
  265. fixup! Throw bad alloc if omp_target_alloc() returns nullptr (details / githubweb)
  266. Add CMake options to control compilation flags for AMD GPUs (#7127) (details / githubweb)
  267. Fix using View without corresponding mdspan-type (#7140) (details / githubweb)
Changes
  1. Reenable TestHIP_Memory_Requirements (details / githubweb)
  2. SYCL: Cleanup device selection (details / githubweb)
  3. Redeine deep_copy for UnorderMap (details / githubweb)
  4. Remove unused typedef (details / githubweb)
  5. Adding SECURITY.md file (details / githubweb)
  6. Only call deep_copy_view() from deep_copy(), add deprecation warning (details / githubweb)
  7. Cleanup KOKKOS_CONFIGURE_CORE (details / githubweb)
  8. [4.2.01]: changelog update  (#6656) (details / githubweb)
  9. Clean up shift_{right, left}_team_impl (#6821) (details / githubweb)
  10. Avoid detecting unwanted fences in the parallel_scan_no_fence test (details / githubweb)
  11. Add test for current CTAD support with RangePolicy (#6803) (details / githubweb)
  12. multi-GPU support: Add test for all policies (#6782) (details / githubweb)
  13. Implement sort_by_key (#6801) (details / githubweb)
  14. Fix scorecard workflow (#6831) (details / githubweb)
  15. Remove redundant RangePolicy constructor (#6841) (details / githubweb)
  16. Bump Google Benchmark version v1.{6.2 -> 7.1} in CMake FetchContent (details / githubweb)
  17. Remove variadic range policy constructor (#6845) (details / githubweb)
  18. Avoid unused variable warning in TestRangePolicyCTAD.cpp (details / githubweb)
  19. bytes_and_flops with CMake (details / githubweb)
  20. bytes_and_flops fix a small bug in command line argument (details / githubweb)
  21. add gather benchmark to CMake (details / githubweb)
  22. add atomic benchmark to CMake (details / githubweb)
  23. add policy_performance benchmark to CMake (details / githubweb)
  24. fix some warnings in policy_performance benchmark (details / githubweb)
  25. policy_benchmark: fix indentation (details / githubweb)
  26. Benchmarks: disable 2 benchmarks for OpenMPTarget (details / githubweb)
  27. bytes_and_flops: fix a counter name (details / githubweb)
  28. Fix C-style cast (details / githubweb)
  29. Fix amdclang++ compilation (#6857) (details / githubweb)
  30. Fix fallback implementation for sort_by_key (#6856) (details / githubweb)
  31. Update the unsafe implicit conversion error message in MDRangePolicy (#6855) (details / githubweb)
  32. Split some classes from Kokkos_ViewMapping (#6859) (details / githubweb)
  33. update mdspan tpl (details / githubweb)
  34. CTAD (deduction guides) for RangePolicy (#6850) (details / githubweb)
  35. CTAD (deduction guides) for MDRangePolicy (#5516) (details / githubweb)
  36. Add support for rocThrust in sort when using HIP (#6793) (details / githubweb)
  37. Add HIP specialization for sort-by-key (details / githubweb)
  38. Fix unused variable warning in TestMDRangePolicyCTAD.cpp (details / githubweb)
  39. Improve message on view out of bounds access and always abort (#6861) (details / githubweb)
  40. Fix linking with rothrust in downstream applications (details / githubweb)
  41. SYCL: Make sure to call find_dependency for oneDPL if necessary  (#6870) (details / githubweb)
  42. Don't use Fedora development version in GitHub CI (details / githubweb)
  43. kokkos_tpls.cmake: update default option to enable rocthrust (details / githubweb)
  44. Scratch space fix for MultiGPU (#6866) (details / githubweb)
  45. [ci skip] Bump version number to 4.3.99 (details / githubweb)
  46. Fix deep copy when filling Rank-7 views (details / githubweb)
  47. TestViewCopy_c.hpp: add new unit test for deep copy (ViewFill) (details / githubweb)
  48. TestViewCopy_c.hpp: better handling for OpenMPTarget (details / githubweb)
  49. Use builtin for atomic_fetch in the HIP backend (details / githubweb)
  50. Fix Makefile.kokkos for Threads (details / githubweb)
  51. Fix a bug in Makefile when using AMD GPU architectures (#6892) (details / githubweb)
  52. Cuda: Fix configuring with CMake 3.28.4 (#6898) (details / githubweb)
  53. Update Intel GPU architectures in Makefile (#6895) (details / githubweb)
  54. Drop Experimental::LayoutTiled class template (details / githubweb)
  55. Get rid of now unnecessary use of is_layouttiled trait (details / githubweb)
  56. Deprecate is_layouttiled trait (details / githubweb)
  57. Move `Kokkos::Array` tests to a more suitable place (#6905) (details / githubweb)
  58. Do not return a copy of the input functor for Kokkos::Experimental::for_each (details / githubweb)
  59. Drop specialization of ViewMapping for Kokkos::Array (details / githubweb)
  60. Accommodate users that depend on a code that define silly macros (#6909) (details / githubweb)
  61. SYCL: Fix multi-GPU support and add test (#6887) (details / githubweb)
  62. SYCL: Unroll shuffle loops for top-level parallel_reduce and parallel_scan (#6750) (details / githubweb)
  63. Preparing readme for develop as the default branch (#6796) (details / githubweb)
  64. CHANGELOG.md: 4.3.00 update (details / githubweb)
  65. Refactor: Uniformize `create_mirror*` parameter name for views (#6917) (details / githubweb)
  66. SYCL: Improve team_reduce implementation (#6562) (details / githubweb)
  67. Use recommended/max team size functions in Cuda ParallelFor and Reduce constructors (#6891) (details / githubweb)
  68. SYCL: Fix range in subgroup scan for workgroup_scan (details / githubweb)
  69. Add ROCm 6.0 in the nightly CI (details / githubweb)
  70. MI300 support unified memory support (#6877) (details / githubweb)
  71. cuda: reduction with `RangePolicy`: fix grid dimensions to work for large values and avoid overflow (#6578) (details / githubweb)
  72. Fix nightly CI (details / githubweb)
  73. Deprecate trailing Proxy template argument in Kokkos::Array (details / githubweb)
  74. CUDA: Update nvcc_wrapper (details / githubweb)
  75. bring back --fmad option to nvcc_wrapper (#6931) (details / githubweb)
  76. Remove unnecessary header include (details / githubweb)
  77. Adding converting constructor in Kokkos::RandomAccessIterator (#6929) (details / githubweb)
  78. Prefer standard C++ feature testing to guard the C++20 requires expression (details / githubweb)
  79. Add `kokkos_swap(Array<T, N>)` sepcialization (details / githubweb)
  80. Deprecate specialization of Kokkos::pair for a single element (details / githubweb)
  81. Fix deprecated warning from `Kokkos::Array` specialization (#6945) (details / githubweb)
  82. [ci skip] Enable deprecated code and deprecated warnings in nightly CI (details / githubweb)
  83. Serial: Use the provided execution space instance in TeamPolicy (details / githubweb)
  84. Fixed the link for P6601 (Threads backend change) (details / githubweb)
  85. Add test (details / githubweb)
  86. Use Kokkos::AUTO for OpenMPTarget (details / githubweb)
  87. kokkos_swap(Array) member friend should not be templated on some other type U (details / githubweb)
  88. Fix noexcept specification for kokkos_swap on zero-sized arrays (details / githubweb)
  89. Replace deprecated sycl::device_ptr/sycl::host_ptr (details / githubweb)
  90. Introduce alias based on feature macro (details / githubweb)
  91. Kokkos::Impl::SYCLTypes:: -> Kokkos::Impl::sycl_ (details / githubweb)
  92. Fix wrong macro guards for deprecated Kokkos::pair<T1,void> specialization (details / githubweb)
  93. Fix support for Kokkos::Array of const-qualified element type (details / githubweb)
  94. Try to fix the CUDA 11.0 build (details / githubweb)
  95. Revert "Try to fix the CUDA 11.0 build" (details / githubweb)
  96. Alternate definition of Impl::is_nothrow_swappable_v for NVCC version less than 11.4 (details / githubweb)
  97. Avoid introducing a 2nd definition of the Impl::swappable trait (details / githubweb)
  98. Do not require OpenMP support for languages other than CXX (details / githubweb)
  99. Update version (details / githubweb)
  100. Add Linux Foundation notice and fix C++ standard (details / githubweb)
  101. SYCL: Avoid deprecated floating-point number abs overloads (#6959) (details / githubweb)
  102. OpenMPTarget: Update loop order in MDRange (#6925) (details / githubweb)
  103. Threads: Don't silently allow m_instance to be a nullptr (#6969) (details / githubweb)
  104. Fix enabling OpenMP with HIP and "compile as CMake language" (details / githubweb)
  105. Fix OpenMP+CUDA when `Kokkos_ENABLE_COMPILE_AS_CMAKE_LANGUAGE` is `ON` (details / githubweb)
  106. Fix fedora CI builds with flang-new (details / githubweb)
  107. Fix Copyright file (details / githubweb)
  108. Also use is_nothrow_swappable workaround for Intel Classic Compilers (#6983) (details / githubweb)
  109. Add thread-safety tests (#6938) (details / githubweb)
  110. Fix deprecation warnings with GCC for pair<T1,void> comparison operators (details / githubweb)
  111. Fix TPL_LIBRARY_SUFFIXES for 32-bit build (details / githubweb)
  112. Avoid duplicated definition of KOKKOS_IMPL_32BIT (details / githubweb)
  113. Disable failing parallel_scan_with_reducers test (details / githubweb)
  114. Fix a bug when using realloc on views of non-default constructible element types (#6993) (details / githubweb)
  115. SYCL: Prepare Parallel* for Graphs (#6988) (details / githubweb)
  116. SYCL: Fix deprecation in custom parallel_for RangePolicy implementation (details / githubweb)
  117. [ci skip] update changelog for 4.3.1 (#6995) (details / githubweb)
  118. OpenMPTarget: Use mutex lock for parallel scan. (details / githubweb)
  119. Remove cuSPARSE TPL (details / githubweb)
  120. SYCL: Print submission command queue property (#7004) (details / githubweb)
  121. Test DualView resize/realloc for types without default constructor (details / githubweb)
  122. Suppress deprecated warnings via pragma push/pop in the tests (#6999) (details / githubweb)
  123. This PR changes the default execution behavior of the parallel_for(team-policy) constructs in the OpenACC backend. (details / githubweb)
  124. Use parallel_for instead of parallel_reduce for check (details / githubweb)
  125. Fix unique_any_senders nvcc template deduction (details / githubweb)
  126. Properly guard deprecated `Kokkos_Vector.hpp` header self contained test (#7016) (details / githubweb)
  127. mdspan converting constructors (#6830) (details / githubweb)
  128. SYCL: Don't use shuffles for top-level reductions (#7009) (details / githubweb)
  129. Add a CITATION.cff file (#7008) (details / githubweb)
  130. Always run Graph tests (#7011) (details / githubweb)
  131. Fix undefined behavior in is_zero_byte (#7014) (details / githubweb)
  132. Disable OpenMPTarget Kokkos::Graph test (does not compile) (details / githubweb)
  133. Fixup for disabling deprecation warnings with NVC++ (details / githubweb)
  134. Remove unused CudaInternal::cuda_{malloc,free}_async_wrapper (details / githubweb)
  135. Improve `Impl::is_zero_byte()` (#7017) (details / githubweb)
  136. OpenMPTarget: Fix compiling Graph tests (#7020) (details / githubweb)
  137. SYCL: Skip launch_six Graph test (details / githubweb)
  138. HIP: Use builtin atomic for compare_exchange (#7000) (details / githubweb)
  139. Fix Kokkos README's FENL link (details / githubweb)
  140. Copied the deduction guides and test cases over from branch (details / githubweb)
  141. Refactor: Replace SFINAE by `if constexpr` for `create_mirror*` functions (#6955) (details / githubweb)
  142. Workaround icpc warnings (details / githubweb)
  143. Adding `Kokkos::to_array` (#6375) (details / githubweb)
  144. Fix adjacent difference (#6922) (details / githubweb)
  145. TestArray: add intel guard to to_array implicit conversion test (details / githubweb)
  146. Avoid using "#if not defined" (details / githubweb)
  147. (Rebase) Partial fix to compile time issues w/nvcc + Kokkos_ENABLE_DEBUG_BOUNDS_CHECK (#7013) (details / githubweb)
  148. Add dependabot for GitHub Actions (details / githubweb)
  149. Added static_asserts for out of range tuple_element and get (details / githubweb)
  150. Add tests to `ScopeGuard` (#7028) (details / githubweb)
  151. Improve Kokkos Graphs (#7039) (details / githubweb)
  152. Introduce `KOKKOS_DEDUCTION_GUIDE` macro to allow user-defined deduction guide in device code for clang compiler (#6954) (details / githubweb)
  153. Add nightly CI on Frontier (#7048) (details / githubweb)
  154. Update scorecard GitHub workflow (details / githubweb)
  155. Fix unused symbols in CTAD tests (details / githubweb)
  156. Cuda: Fix nvcc warnings (details / githubweb)
  157. Fix quotation marks in CXX flags (details / githubweb)
  158. Fix kokkos_swap (details / githubweb)
  159. Fix array size (details / githubweb)
  160. Fix gtest (details / githubweb)
  161. Fix .jenkins whitespce (details / githubweb)
  162. Only use -Werror all-warnings with explicit nvcc_wrapper (details / githubweb)
  163. Fix dangling reference (details / githubweb)
  164. Suppress 'long double' is treated as 'double' in device code (details / githubweb)
  165. Use -Xcudafe --diag_suppress=20208 for 11.6 build; nothing else seems to help (details / githubweb)
  166. Try moving pragma suppress to tests (details / githubweb)
  167. Use -Xcudafe --diag_suppress=20208 in Makefile build (details / githubweb)
  168. Move dependabot to .github (details / githubweb)
  169. Bump Jimver/cuda-toolkit from 0.2.14 to 0.2.15 (details / githubweb)
  170. Bump DoozyX/clang-format-lint-action from 0.16.2 to 0.17 (details / githubweb)
  171. Bump version in the readme (details / githubweb)
  172. Restrict to_array subtest to NVCC >= 11.4.0 (details / githubweb)
  173. [ci skip] Use ROCM 6.1 in the nightly CI and disable one test (details / githubweb)
  174. Switch to using functors in sort_by_key_via_sort (#7059) (details / githubweb)
  175. Fix mdspan test (details / githubweb)
  176. Update to CUDA 12.4.1 in MSVC CI (details / githubweb)
  177. Workaround MSVC compiler issues in Views (details / githubweb)
  178. Avoid lambda in sort_by_key_via_sort (details / githubweb)
  179. Fix using CUDAToolkit for CMake 3.28.4 and higher (details / githubweb)
  180. Move check CMake 3.20.1 with nvhpc (details / githubweb)
  181. Fix closing brackets (details / githubweb)
  182. KOKKOS_CUDA_ERROR->DEFAULT_MSG (details / githubweb)
  183. Workaround icpc "missing return statement at end of non-void function" (details / githubweb)
  184. Enable MDSPAN support by default (#7069) (details / githubweb)
  185. Add SpaceAwareAccessor (details / githubweb)
  186. Add basic tests for SpaceAwareAccessor (details / githubweb)
  187. Use SpaceAwareAccessor in View mdspan-interop (details / githubweb)
  188. Add specialization of SpaceAwareAcc for AnonymousSpace (details / githubweb)
  189. Address review comments (details / githubweb)
  190. Fix incompatible dynamic_extent definition in Kokkos (details / githubweb)
  191. SpaceAwareAccessor: fix issues (no-unique-address, is_empty) (details / githubweb)
  192. Fix using shared libraries and -fvisibility=hidden (#7065) (details / githubweb)
  193. OpenACC: Skip exec_space_thread_safety_range_scan (#7022) (details / githubweb)
  194. Structured binding support for Kokkos::complex (#7040) (details / githubweb)
  195. Fix thread-safety for the Serial backend (details / githubweb)
  196. Update from GCC 13 to 14 and use C++ 26 in Jenkins nightly (details / githubweb)
  197. Move `layout_iterate_type_selector` into Impl namespace (#7076) (details / githubweb)
  198. Use correct includes for spaceawareaccessor (details / githubweb)
  199. Improve performance for deleting an instance. (details / githubweb)
  200. fix refcount exception safety (#6289) (details / githubweb)
  201. Don't delete special member functions explicitly (details / githubweb)
  202. Complex needs a tuple include (details / githubweb)
  203. Update internal desul file copies to 60c1115 (details / githubweb)
  204. Update desul version in github workflow (details / githubweb)
  205. Resolve various bogus icpc -Werror (#7079) (details / githubweb)
  206. Remove support for NVHPC as CUDA device compiler (details / githubweb)
  207. remove cmake options to change default spaces (details / githubweb)
  208. Temporarily disable failing parts in the TestMDSpan.hpp for the OpenACC (details / githubweb)
  209. Add missing space (details / githubweb)
  210. Fix more icpc issues (details / githubweb)
  211. Bump Jimver/cuda-toolkit from 0.2.15 to 0.2.16 (details / githubweb)
  212. Copy print_configuration setting in combination of kokkos settings. (details / githubweb)
  213. Couple more icpc -Werror fixes (details / githubweb)
  214. Cleanup KokkosP hooks in `Profiling::` (#7096) (details / githubweb)
  215. Remove FIXME_NVHPC 23.7 guards (details / githubweb)
  216. Fix SpaceAwareAccessor based on usage experiment in View (#7088) (details / githubweb)
  217. Added width 8 abi for avx2 (details / githubweb)
  218. Added for AVX512 (details / githubweb)
  219. Added for width 4 for NEON (details / githubweb)
  220. clang-formatted (details / githubweb)
  221. Disabling simd unit tests from building for Windows+CUDA build (details / githubweb)
  222. Workaround for the compilation failure for rocm 5.6-6.0 (details / githubweb)
  223. Added a comment about is_type structs (details / githubweb)
  224. clang formating (details / githubweb)
  225. Restrict jenkins CI not to run on hopper for nvcc < 11.8 (details / githubweb)
  226. cmake: fix relative to find kokkos_compiler_launcher (details / githubweb)
  227. Clean up KOKKOS_LIB_INCLUDE_DIRECTORIES, append include directories to associated targets in Trilinos builds (#7103) (details / githubweb)
  228. Workaround to ice with icpc when using -no-ip (#7106) (details / githubweb)
  229. move view allocation related functionality to a new header (#7110) (details / githubweb)
  230. SYCL CI: Manually build oneDPL (details / githubweb)
  231. Add missing concepts include in test (details / githubweb)
  232. Github CI: Test with C++17, C++20, and C++23 (#7082) (details / githubweb)
  233. Update base image for ROCm 5.6 (details / githubweb)
  234. Update mdspan to 98a12b01b51b2 (details / githubweb)
  235. Add AtomicAccessorRelaxed (#7089) (details / githubweb)
  236. Fix div by zero in math ops testing (details / githubweb)
  237. Fix and test with -fsanitize=undefined in GitHub CI (#7104) (details / githubweb)
  238. Bump actions/upload-artifact from 4.3.3 to 4.3.4 (details / githubweb)
  239. Update HIP nightly build base image Ubuntu 20.04 -> 22.04 (details / githubweb)
  240. Deprecate `RawMemoryAllocationFailure::FailureMode::MaximumCudaUVMAllocationsExceeded` (#7120) (details / githubweb)
  241. Refactor: Move logic of `create_mirror*` to `Impl::create_mirror*` (#7061) (details / githubweb)
  242. Define minimum compiler versions for C++20 support (details / githubweb)
  243. Drop (unused) cuda uvm allocations counter (details / githubweb)
  244. Do not bother catching memory allocation failure and rethrow (details / githubweb)
  245. Fix using and, or, xor in desul with MSVC (#7124) (details / githubweb)
  246. Do not swallow errors when deallocating memory with CUDA (details / githubweb)
  247. Let the throwing push finalize hook calls terminate test actually run (details / githubweb)
  248. Drop (unused) KOKKOS_ADD_ADVANCED_TEST TriBITS function (details / githubweb)
  249. Simplify the logic when finalizing and calling the registered functions (details / githubweb)
  250. [ci skip] Fix ROCm version to 6.1.2 in nightly CI (details / githubweb)
  251. Disable the PushFinalizeHookTerminate test on Windows (details / githubweb)
  252. OpenMP: Ensure kernels submitted by multiple threads to the same instance don't run concurrently (#6151) (details / githubweb)
  253. SYCL: Add missing include for std::stringstream (details / githubweb)
  254. SYCL: Add support for Graphs (#6912) (details / githubweb)
  255. [ci skip] rename jenkins build (details / githubweb)
  256. NVHPC: Disable failing NaN tests (details / githubweb)
  257. no_device_stack is unknown (details / githubweb)
  258. Check for LIBCXX 10 or later for C++20 and later (#7123) (details / githubweb)
  259. Get rid of `RawMemoryAllocationFailure::AllocationMechanism` and derived backend-specific exceptions (#7139) (details / githubweb)
  260. Add `likely` and `unlikely` attribute from C++20 to ref counting in views (#6730) (details / githubweb)
  261. Throw bad alloc if omp_target_alloc() returns nullptr (details / githubweb)
  262. Enable test_view_allocation_error with OpenMPTarget (details / githubweb)
  263. Fix gcc-14 C++26 nightly jenkins build (#7137) (details / githubweb)
  264. Fix Trilinos nightly failure due to `create_mirror*` refactor (#7126) (details / githubweb)
  265. fixup! Throw bad alloc if omp_target_alloc() returns nullptr (details / githubweb)
  266. Add CMake options to control compilation flags for AMD GPUs (#7127) (details / githubweb)
  267. Fix using View without corresponding mdspan-type (#7140) (details / githubweb)
Changes
  1. Reenable TestHIP_Memory_Requirements (details / githubweb)
  2. SYCL: Cleanup device selection (details / githubweb)
  3. Redeine deep_copy for UnorderMap (details / githubweb)
  4. Remove unused typedef (details / githubweb)
  5. Adding SECURITY.md file (details / githubweb)
  6. Only call deep_copy_view() from deep_copy(), add deprecation warning (details / githubweb)
  7. Cleanup KOKKOS_CONFIGURE_CORE (details / githubweb)
  8. [4.2.01]: changelog update  (#6656) (details / githubweb)
  9. Clean up shift_{right, left}_team_impl (#6821) (details / githubweb)
  10. Avoid detecting unwanted fences in the parallel_scan_no_fence test (details / githubweb)
  11. Add test for current CTAD support with RangePolicy (#6803) (details / githubweb)
  12. multi-GPU support: Add test for all policies (#6782) (details / githubweb)
  13. Implement sort_by_key (#6801) (details / githubweb)
  14. Fix scorecard workflow (#6831) (details / githubweb)
  15. Remove redundant RangePolicy constructor (#6841) (details / githubweb)
  16. Bump Google Benchmark version v1.{6.2 -> 7.1} in CMake FetchContent (details / githubweb)
  17. Remove variadic range policy constructor (#6845) (details / githubweb)
  18. Avoid unused variable warning in TestRangePolicyCTAD.cpp (details / githubweb)
  19. bytes_and_flops with CMake (details / githubweb)
  20. bytes_and_flops fix a small bug in command line argument (details / githubweb)
  21. add gather benchmark to CMake (details / githubweb)
  22. add atomic benchmark to CMake (details / githubweb)
  23. add policy_performance benchmark to CMake (details / githubweb)
  24. fix some warnings in policy_performance benchmark (details / githubweb)
  25. policy_benchmark: fix indentation (details / githubweb)
  26. Benchmarks: disable 2 benchmarks for OpenMPTarget (details / githubweb)
  27. bytes_and_flops: fix a counter name (details / githubweb)
  28. Fix C-style cast (details / githubweb)
  29. Fix amdclang++ compilation (#6857) (details / githubweb)
  30. Fix fallback implementation for sort_by_key (#6856) (details / githubweb)
  31. Update the unsafe implicit conversion error message in MDRangePolicy (#6855) (details / githubweb)
  32. Split some classes from Kokkos_ViewMapping (#6859) (details / githubweb)
  33. update mdspan tpl (details / githubweb)
  34. CTAD (deduction guides) for RangePolicy (#6850) (details / githubweb)
  35. CTAD (deduction guides) for MDRangePolicy (#5516) (details / githubweb)
  36. Add support for rocThrust in sort when using HIP (#6793) (details / githubweb)
  37. Add HIP specialization for sort-by-key (details / githubweb)
  38. Fix unused variable warning in TestMDRangePolicyCTAD.cpp (details / githubweb)
  39. Improve message on view out of bounds access and always abort (#6861) (details / githubweb)
  40. Fix linking with rothrust in downstream applications (details / githubweb)
  41. SYCL: Make sure to call find_dependency for oneDPL if necessary  (#6870) (details / githubweb)
  42. Don't use Fedora development version in GitHub CI (details / githubweb)
  43. kokkos_tpls.cmake: update default option to enable rocthrust (details / githubweb)
  44. Scratch space fix for MultiGPU (#6866) (details / githubweb)
  45. [ci skip] Bump version number to 4.3.99 (details / githubweb)
  46. Fix deep copy when filling Rank-7 views (details / githubweb)
  47. TestViewCopy_c.hpp: add new unit test for deep copy (ViewFill) (details / githubweb)
  48. TestViewCopy_c.hpp: better handling for OpenMPTarget (details / githubweb)
  49. Use builtin for atomic_fetch in the HIP backend (details / githubweb)
  50. Fix Makefile.kokkos for Threads (details / githubweb)
  51. Fix a bug in Makefile when using AMD GPU architectures (#6892) (details / githubweb)
  52. Cuda: Fix configuring with CMake 3.28.4 (#6898) (details / githubweb)
  53. Update Intel GPU architectures in Makefile (#6895) (details / githubweb)
  54. Drop Experimental::LayoutTiled class template (details / githubweb)
  55. Get rid of now unnecessary use of is_layouttiled trait (details / githubweb)
  56. Deprecate is_layouttiled trait (details / githubweb)
  57. Move `Kokkos::Array` tests to a more suitable place (#6905) (details / githubweb)
  58. Do not return a copy of the input functor for Kokkos::Experimental::for_each (details / githubweb)
  59. Drop specialization of ViewMapping for Kokkos::Array (details / githubweb)
  60. Accommodate users that depend on a code that define silly macros (#6909) (details / githubweb)
  61. SYCL: Fix multi-GPU support and add test (#6887) (details / githubweb)
  62. SYCL: Unroll shuffle loops for top-level parallel_reduce and parallel_scan (#6750) (details / githubweb)
  63. Preparing readme for develop as the default branch (#6796) (details / githubweb)
  64. CHANGELOG.md: 4.3.00 update (details / githubweb)
  65. Refactor: Uniformize `create_mirror*` parameter name for views (#6917) (details / githubweb)
  66. SYCL: Improve team_reduce implementation (#6562) (details / githubweb)
  67. Use recommended/max team size functions in Cuda ParallelFor and Reduce constructors (#6891) (details / githubweb)
  68. SYCL: Fix range in subgroup scan for workgroup_scan (details / githubweb)
  69. Add ROCm 6.0 in the nightly CI (details / githubweb)
  70. MI300 support unified memory support (#6877) (details / githubweb)
  71. cuda: reduction with `RangePolicy`: fix grid dimensions to work for large values and avoid overflow (#6578) (details / githubweb)
  72. Fix nightly CI (details / githubweb)
  73. Deprecate trailing Proxy template argument in Kokkos::Array (details / githubweb)
  74. CUDA: Update nvcc_wrapper (details / githubweb)
  75. bring back --fmad option to nvcc_wrapper (#6931) (details / githubweb)
  76. Remove unnecessary header include (details / githubweb)
  77. Adding converting constructor in Kokkos::RandomAccessIterator (#6929) (details / githubweb)
  78. Prefer standard C++ feature testing to guard the C++20 requires expression (details / githubweb)
  79. Add `kokkos_swap(Array<T, N>)` sepcialization (details / githubweb)
  80. Deprecate specialization of Kokkos::pair for a single element (details / githubweb)
  81. Fix deprecated warning from `Kokkos::Array` specialization (#6945) (details / githubweb)
  82. [ci skip] Enable deprecated code and deprecated warnings in nightly CI (details / githubweb)
  83. Serial: Use the provided execution space instance in TeamPolicy (details / githubweb)
  84. Fixed the link for P6601 (Threads backend change) (details / githubweb)
  85. Add test (details / githubweb)
  86. Use Kokkos::AUTO for OpenMPTarget (details / githubweb)
  87. kokkos_swap(Array) member friend should not be templated on some other type U (details / githubweb)
  88. Fix noexcept specification for kokkos_swap on zero-sized arrays (details / githubweb)
  89. Replace deprecated sycl::device_ptr/sycl::host_ptr (details / githubweb)
  90. Introduce alias based on feature macro (details / githubweb)
  91. Kokkos::Impl::SYCLTypes:: -> Kokkos::Impl::sycl_ (details / githubweb)
  92. Fix wrong macro guards for deprecated Kokkos::pair<T1,void> specialization (details / githubweb)
  93. Fix support for Kokkos::Array of const-qualified element type (details / githubweb)
  94. Try to fix the CUDA 11.0 build (details / githubweb)
  95. Revert "Try to fix the CUDA 11.0 build" (details / githubweb)
  96. Alternate definition of Impl::is_nothrow_swappable_v for NVCC version less than 11.4 (details / githubweb)
  97. Avoid introducing a 2nd definition of the Impl::swappable trait (details / githubweb)
  98. Do not require OpenMP support for languages other than CXX (details / githubweb)
  99. Update version (details / githubweb)
  100. Add Linux Foundation notice and fix C++ standard (details / githubweb)
  101. SYCL: Avoid deprecated floating-point number abs overloads (#6959) (details / githubweb)
  102. OpenMPTarget: Update loop order in MDRange (#6925) (details / githubweb)
  103. Threads: Don't silently allow m_instance to be a nullptr (#6969) (details / githubweb)
  104. Fix enabling OpenMP with HIP and "compile as CMake language" (details / githubweb)
  105. Fix OpenMP+CUDA when `Kokkos_ENABLE_COMPILE_AS_CMAKE_LANGUAGE` is `ON` (details / githubweb)
  106. Fix fedora CI builds with flang-new (details / githubweb)
  107. Fix Copyright file (details / githubweb)
  108. Also use is_nothrow_swappable workaround for Intel Classic Compilers (#6983) (details / githubweb)
  109. Add thread-safety tests (#6938) (details / githubweb)
  110. Fix deprecation warnings with GCC for pair<T1,void> comparison operators (details / githubweb)
  111. Fix TPL_LIBRARY_SUFFIXES for 32-bit build (details / githubweb)
  112. Avoid duplicated definition of KOKKOS_IMPL_32BIT (details / githubweb)
  113. Disable failing parallel_scan_with_reducers test (details / githubweb)
  114. Fix a bug when using realloc on views of non-default constructible element types (#6993) (details / githubweb)
  115. SYCL: Prepare Parallel* for Graphs (#6988) (details / githubweb)
  116. SYCL: Fix deprecation in custom parallel_for RangePolicy implementation (details / githubweb)
  117. [ci skip] update changelog for 4.3.1 (#6995) (details / githubweb)
  118. OpenMPTarget: Use mutex lock for parallel scan. (details / githubweb)
  119. Remove cuSPARSE TPL (details / githubweb)
  120. SYCL: Print submission command queue property (#7004) (details / githubweb)
  121. Test DualView resize/realloc for types without default constructor (details / githubweb)
  122. Suppress deprecated warnings via pragma push/pop in the tests (#6999) (details / githubweb)
  123. This PR changes the default execution behavior of the parallel_for(team-policy) constructs in the OpenACC backend. (details / githubweb)
  124. Use parallel_for instead of parallel_reduce for check (details / githubweb)
  125. Fix unique_any_senders nvcc template deduction (details / githubweb)
  126. Properly guard deprecated `Kokkos_Vector.hpp` header self contained test (#7016) (details / githubweb)
  127. mdspan converting constructors (#6830) (details / githubweb)
  128. SYCL: Don't use shuffles for top-level reductions (#7009) (details / githubweb)
  129. Add a CITATION.cff file (#7008) (details / githubweb)
  130. Always run Graph tests (#7011) (details / githubweb)
  131. Fix undefined behavior in is_zero_byte (#7014) (details / githubweb)
  132. Disable OpenMPTarget Kokkos::Graph test (does not compile) (details / githubweb)
  133. Fixup for disabling deprecation warnings with NVC++ (details / githubweb)
  134. Remove unused CudaInternal::cuda_{malloc,free}_async_wrapper (details / githubweb)
  135. Improve `Impl::is_zero_byte()` (#7017) (details / githubweb)
  136. OpenMPTarget: Fix compiling Graph tests (#7020) (details / githubweb)
  137. SYCL: Skip launch_six Graph test (details / githubweb)
  138. HIP: Use builtin atomic for compare_exchange (#7000) (details / githubweb)
  139. Fix Kokkos README's FENL link (details / githubweb)
  140. Copied the deduction guides and test cases over from branch (details / githubweb)
  141. Refactor: Replace SFINAE by `if constexpr` for `create_mirror*` functions (#6955) (details / githubweb)
  142. Workaround icpc warnings (details / githubweb)
  143. Adding `Kokkos::to_array` (#6375) (details / githubweb)
  144. Fix adjacent difference (#6922) (details / githubweb)
  145. TestArray: add intel guard to to_array implicit conversion test (details / githubweb)
  146. Avoid using "#if not defined" (details / githubweb)
  147. (Rebase) Partial fix to compile time issues w/nvcc + Kokkos_ENABLE_DEBUG_BOUNDS_CHECK (#7013) (details / githubweb)
  148. Add dependabot for GitHub Actions (details / githubweb)
  149. Added static_asserts for out of range tuple_element and get (details / githubweb)
  150. Add tests to `ScopeGuard` (#7028) (details / githubweb)
  151. Improve Kokkos Graphs (#7039) (details / githubweb)
  152. Introduce `KOKKOS_DEDUCTION_GUIDE` macro to allow user-defined deduction guide in device code for clang compiler (#6954) (details / githubweb)
  153. Add nightly CI on Frontier (#7048) (details / githubweb)
  154. Update scorecard GitHub workflow (details / githubweb)
  155. Fix unused symbols in CTAD tests (details / githubweb)
  156. Cuda: Fix nvcc warnings (details / githubweb)
  157. Fix quotation marks in CXX flags (details / githubweb)
  158. Fix kokkos_swap (details / githubweb)
  159. Fix array size (details / githubweb)
  160. Fix gtest (details / githubweb)
  161. Fix .jenkins whitespce (details / githubweb)
  162. Only use -Werror all-warnings with explicit nvcc_wrapper (details / githubweb)
  163. Fix dangling reference (details / githubweb)
  164. Suppress 'long double' is treated as 'double' in device code (details / githubweb)
  165. Use -Xcudafe --diag_suppress=20208 for 11.6 build; nothing else seems to help (details / githubweb)
  166. Try moving pragma suppress to tests (details / githubweb)
  167. Use -Xcudafe --diag_suppress=20208 in Makefile build (details / githubweb)
  168. Move dependabot to .github (details / githubweb)
  169. Bump Jimver/cuda-toolkit from 0.2.14 to 0.2.15 (details / githubweb)
  170. Bump DoozyX/clang-format-lint-action from 0.16.2 to 0.17 (details / githubweb)
  171. Bump version in the readme (details / githubweb)
  172. Restrict to_array subtest to NVCC >= 11.4.0 (details / githubweb)
  173. [ci skip] Use ROCM 6.1 in the nightly CI and disable one test (details / githubweb)
  174. Switch to using functors in sort_by_key_via_sort (#7059) (details / githubweb)
  175. Fix mdspan test (details / githubweb)
  176. Update to CUDA 12.4.1 in MSVC CI (details / githubweb)
  177. Workaround MSVC compiler issues in Views (details / githubweb)
  178. Avoid lambda in sort_by_key_via_sort (details / githubweb)
  179. Fix using CUDAToolkit for CMake 3.28.4 and higher (details / githubweb)
  180. Move check CMake 3.20.1 with nvhpc (details / githubweb)
  181. Fix closing brackets (details / githubweb)
  182. KOKKOS_CUDA_ERROR->DEFAULT_MSG (details / githubweb)
  183. Workaround icpc "missing return statement at end of non-void function" (details / githubweb)
  184. Enable MDSPAN support by default (#7069) (details / githubweb)
  185. Add SpaceAwareAccessor (details / githubweb)
  186. Add basic tests for SpaceAwareAccessor (details / githubweb)
  187. Use SpaceAwareAccessor in View mdspan-interop (details / githubweb)
  188. Add specialization of SpaceAwareAcc for AnonymousSpace (details / githubweb)
  189. Address review comments (details / githubweb)
  190. Fix incompatible dynamic_extent definition in Kokkos (details / githubweb)
  191. SpaceAwareAccessor: fix issues (no-unique-address, is_empty) (details / githubweb)
  192. Fix using shared libraries and -fvisibility=hidden (#7065) (details / githubweb)
  193. OpenACC: Skip exec_space_thread_safety_range_scan (#7022) (details / githubweb)
  194. Structured binding support for Kokkos::complex (#7040) (details / githubweb)
  195. Fix thread-safety for the Serial backend (details / githubweb)
  196. Update from GCC 13 to 14 and use C++ 26 in Jenkins nightly (details / githubweb)
  197. Move `layout_iterate_type_selector` into Impl namespace (#7076) (details / githubweb)
  198. Use correct includes for spaceawareaccessor (details / githubweb)
  199. Improve performance for deleting an instance. (details / githubweb)
  200. fix refcount exception safety (#6289) (details / githubweb)
  201. Don't delete special member functions explicitly (details / githubweb)
  202. Complex needs a tuple include (details / githubweb)
  203. Update internal desul file copies to 60c1115 (details / githubweb)
  204. Update desul version in github workflow (details / githubweb)
  205. Resolve various bogus icpc -Werror (#7079) (details / githubweb)
  206. Remove support for NVHPC as CUDA device compiler (details / githubweb)
  207. remove cmake options to change default spaces (details / githubweb)
  208. Temporarily disable failing parts in the TestMDSpan.hpp for the OpenACC (details / githubweb)
  209. Add missing space (details / githubweb)
  210. Fix more icpc issues (details / githubweb)
  211. Bump Jimver/cuda-toolkit from 0.2.15 to 0.2.16 (details / githubweb)
  212. Copy print_configuration setting in combination of kokkos settings. (details / githubweb)
  213. Couple more icpc -Werror fixes (details / githubweb)
  214. Cleanup KokkosP hooks in `Profiling::` (#7096) (details / githubweb)
  215. Remove FIXME_NVHPC 23.7 guards (details / githubweb)
  216. Fix SpaceAwareAccessor based on usage experiment in View (#7088) (details / githubweb)
  217. Added width 8 abi for avx2 (details / githubweb)
  218. Added for AVX512 (details / githubweb)
  219. Added for width 4 for NEON (details / githubweb)
  220. clang-formatted (details / githubweb)
  221. Disabling simd unit tests from building for Windows+CUDA build (details / githubweb)
  222. Workaround for the compilation failure for rocm 5.6-6.0 (details / githubweb)
  223. Added a comment about is_type structs (details / githubweb)
  224. clang formating (details / githubweb)
  225. Restrict jenkins CI not to run on hopper for nvcc < 11.8 (details / githubweb)
  226. cmake: fix relative to find kokkos_compiler_launcher (details / githubweb)
  227. Clean up KOKKOS_LIB_INCLUDE_DIRECTORIES, append include directories to associated targets in Trilinos builds (#7103) (details / githubweb)
  228. Workaround to ice with icpc when using -no-ip (#7106) (details / githubweb)
  229. move view allocation related functionality to a new header (#7110) (details / githubweb)
  230. SYCL CI: Manually build oneDPL (details / githubweb)
  231. Add missing concepts include in test (details / githubweb)
  232. Github CI: Test with C++17, C++20, and C++23 (#7082) (details / githubweb)
  233. Update base image for ROCm 5.6 (details / githubweb)
  234. Update mdspan to 98a12b01b51b2 (details / githubweb)
  235. Add AtomicAccessorRelaxed (#7089) (details / githubweb)
  236. Fix div by zero in math ops testing (details / githubweb)
  237. Fix and test with -fsanitize=undefined in GitHub CI (#7104) (details / githubweb)
  238. Bump actions/upload-artifact from 4.3.3 to 4.3.4 (details / githubweb)
  239. Update HIP nightly build base image Ubuntu 20.04 -> 22.04 (details / githubweb)
  240. Deprecate `RawMemoryAllocationFailure::FailureMode::MaximumCudaUVMAllocationsExceeded` (#7120) (details / githubweb)
  241. Refactor: Move logic of `create_mirror*` to `Impl::create_mirror*` (#7061) (details / githubweb)
  242. Define minimum compiler versions for C++20 support (details / githubweb)
  243. Drop (unused) cuda uvm allocations counter (details / githubweb)
  244. Do not bother catching memory allocation failure and rethrow (details / githubweb)
  245. Fix using and, or, xor in desul with MSVC (#7124) (details / githubweb)
  246. Do not swallow errors when deallocating memory with CUDA (details / githubweb)
  247. Let the throwing push finalize hook calls terminate test actually run (details / githubweb)
  248. Drop (unused) KOKKOS_ADD_ADVANCED_TEST TriBITS function (details / githubweb)
  249. Simplify the logic when finalizing and calling the registered functions (details / githubweb)
  250. [ci skip] Fix ROCm version to 6.1.2 in nightly CI (details / githubweb)
  251. Disable the PushFinalizeHookTerminate test on Windows (details / githubweb)
  252. OpenMP: Ensure kernels submitted by multiple threads to the same instance don't run concurrently (#6151) (details / githubweb)
  253. SYCL: Add missing include for std::stringstream (details / githubweb)
  254. SYCL: Add support for Graphs (#6912) (details / githubweb)
  255. [ci skip] rename jenkins build (details / githubweb)
  256. NVHPC: Disable failing NaN tests (details / githubweb)
  257. no_device_stack is unknown (details / githubweb)
  258. Check for LIBCXX 10 or later for C++20 and later (#7123) (details / githubweb)
  259. Get rid of `RawMemoryAllocationFailure::AllocationMechanism` and derived backend-specific exceptions (#7139) (details / githubweb)
  260. Add `likely` and `unlikely` attribute from C++20 to ref counting in views (#6730) (details / githubweb)
  261. Throw bad alloc if omp_target_alloc() returns nullptr (details / githubweb)
  262. Enable test_view_allocation_error with OpenMPTarget (details / githubweb)
  263. Fix gcc-14 C++26 nightly jenkins build (#7137) (details / githubweb)
  264. Fix Trilinos nightly failure due to `create_mirror*` refactor (#7126) (details / githubweb)
  265. fixup! Throw bad alloc if omp_target_alloc() returns nullptr (details / githubweb)
  266. Add CMake options to control compilation flags for AMD GPUs (#7127) (details / githubweb)
  267. Fix using View without corresponding mdspan-type (#7140) (details / githubweb)
Changes
  1. Reenable TestHIP_Memory_Requirements (details / githubweb)
  2. SYCL: Cleanup device selection (details / githubweb)
  3. Redeine deep_copy for UnorderMap (details / githubweb)
  4. Remove unused typedef (details / githubweb)
  5. Adding SECURITY.md file (details / githubweb)
  6. Only call deep_copy_view() from deep_copy(), add deprecation warning (details / githubweb)
  7. Cleanup KOKKOS_CONFIGURE_CORE (details / githubweb)
  8. [4.2.01]: changelog update  (#6656) (details / githubweb)
  9. Clean up shift_{right, left}_team_impl (#6821) (details / githubweb)
  10. Avoid detecting unwanted fences in the parallel_scan_no_fence test (details / githubweb)
  11. Add test for current CTAD support with RangePolicy (#6803) (details / githubweb)
  12. multi-GPU support: Add test for all policies (#6782) (details / githubweb)
  13. Implement sort_by_key (#6801) (details / githubweb)
  14. Fix scorecard workflow (#6831) (details / githubweb)
  15. Remove redundant RangePolicy constructor (#6841) (details / githubweb)
  16. Bump Google Benchmark version v1.{6.2 -> 7.1} in CMake FetchContent (details / githubweb)
  17. Remove variadic range policy constructor (#6845) (details / githubweb)
  18. Avoid unused variable warning in TestRangePolicyCTAD.cpp (details / githubweb)
  19. bytes_and_flops with CMake (details / githubweb)
  20. bytes_and_flops fix a small bug in command line argument (details / githubweb)
  21. add gather benchmark to CMake (details / githubweb)
  22. add atomic benchmark to CMake (details / githubweb)
  23. add policy_performance benchmark to CMake (details / githubweb)
  24. fix some warnings in policy_performance benchmark (details / githubweb)
  25. policy_benchmark: fix indentation (details / githubweb)
  26. Benchmarks: disable 2 benchmarks for OpenMPTarget (details / githubweb)
  27. bytes_and_flops: fix a counter name (details / githubweb)
  28. Fix C-style cast (details / githubweb)
  29. Fix amdclang++ compilation (#6857) (details / githubweb)
  30. Fix fallback implementation for sort_by_key (#6856) (details / githubweb)
  31. Update the unsafe implicit conversion error message in MDRangePolicy (#6855) (details / githubweb)
  32. Split some classes from Kokkos_ViewMapping (#6859) (details / githubweb)
  33. update mdspan tpl (details / githubweb)
  34. CTAD (deduction guides) for RangePolicy (#6850) (details / githubweb)
  35. CTAD (deduction guides) for MDRangePolicy (#5516) (details / githubweb)
  36. Add support for rocThrust in sort when using HIP (#6793) (details / githubweb)
  37. Add HIP specialization for sort-by-key (details / githubweb)
  38. Fix unused variable warning in TestMDRangePolicyCTAD.cpp (details / githubweb)
  39. Improve message on view out of bounds access and always abort (#6861) (details / githubweb)
  40. Fix linking with rothrust in downstream applications (details / githubweb)
  41. SYCL: Make sure to call find_dependency for oneDPL if necessary  (#6870) (details / githubweb)
  42. Don't use Fedora development version in GitHub CI (details / githubweb)
  43. kokkos_tpls.cmake: update default option to enable rocthrust (details / githubweb)
  44. Scratch space fix for MultiGPU (#6866) (details / githubweb)
  45. [ci skip] Bump version number to 4.3.99 (details / githubweb)
  46. Fix deep copy when filling Rank-7 views (details / githubweb)
  47. TestViewCopy_c.hpp: add new unit test for deep copy (ViewFill) (details / githubweb)
  48. TestViewCopy_c.hpp: better handling for OpenMPTarget (details / githubweb)
  49. Use builtin for atomic_fetch in the HIP backend (details / githubweb)
  50. Fix Makefile.kokkos for Threads (details / githubweb)
  51. Fix a bug in Makefile when using AMD GPU architectures (#6892) (details / githubweb)
  52. Cuda: Fix configuring with CMake 3.28.4 (#6898) (details / githubweb)
  53. Update Intel GPU architectures in Makefile (#6895) (details / githubweb)
  54. Drop Experimental::LayoutTiled class template (details / githubweb)
  55. Get rid of now unnecessary use of is_layouttiled trait (details / githubweb)
  56. Deprecate is_layouttiled trait (details / githubweb)
  57. Move `Kokkos::Array` tests to a more suitable place (#6905) (details / githubweb)
  58. Do not return a copy of the input functor for Kokkos::Experimental::for_each (details / githubweb)
  59. Drop specialization of ViewMapping for Kokkos::Array (details / githubweb)
  60. Accommodate users that depend on a code that define silly macros (#6909) (details / githubweb)
  61. SYCL: Fix multi-GPU support and add test (#6887) (details / githubweb)
  62. SYCL: Unroll shuffle loops for top-level parallel_reduce and parallel_scan (#6750) (details / githubweb)
  63. Preparing readme for develop as the default branch (#6796) (details / githubweb)
  64. CHANGELOG.md: 4.3.00 update (details / githubweb)
  65. Refactor: Uniformize `create_mirror*` parameter name for views (#6917) (details / githubweb)
  66. SYCL: Improve team_reduce implementation (#6562) (details / githubweb)
  67. Use recommended/max team size functions in Cuda ParallelFor and Reduce constructors (#6891) (details / githubweb)
  68. SYCL: Fix range in subgroup scan for workgroup_scan (details / githubweb)
  69. Add ROCm 6.0 in the nightly CI (details / githubweb)
  70. MI300 support unified memory support (#6877) (details / githubweb)
  71. cuda: reduction with `RangePolicy`: fix grid dimensions to work for large values and avoid overflow (#6578) (details / githubweb)
  72. Fix nightly CI (details / githubweb)
  73. Deprecate trailing Proxy template argument in Kokkos::Array (details / githubweb)
  74. CUDA: Update nvcc_wrapper (details / githubweb)
  75. bring back --fmad option to nvcc_wrapper (#6931) (details / githubweb)
  76. Remove unnecessary header include (details / githubweb)
  77. Adding converting constructor in Kokkos::RandomAccessIterator (#6929) (details / githubweb)
  78. Prefer standard C++ feature testing to guard the C++20 requires expression (details / githubweb)
  79. Add `kokkos_swap(Array<T, N>)` sepcialization (details / githubweb)
  80. Deprecate specialization of Kokkos::pair for a single element (details / githubweb)
  81. Fix deprecated warning from `Kokkos::Array` specialization (#6945) (details / githubweb)
  82. [ci skip] Enable deprecated code and deprecated warnings in nightly CI (details / githubweb)
  83. Serial: Use the provided execution space instance in TeamPolicy (details / githubweb)
  84. Fixed the link for P6601 (Threads backend change) (details / githubweb)
  85. Add test (details / githubweb)
  86. Use Kokkos::AUTO for OpenMPTarget (details / githubweb)
  87. kokkos_swap(Array) member friend should not be templated on some other type U (details / githubweb)
  88. Fix noexcept specification for kokkos_swap on zero-sized arrays (details / githubweb)
  89. Replace deprecated sycl::device_ptr/sycl::host_ptr (details / githubweb)
  90. Introduce alias based on feature macro (details / githubweb)
  91. Kokkos::Impl::SYCLTypes:: -> Kokkos::Impl::sycl_ (details / githubweb)
  92. Fix wrong macro guards for deprecated Kokkos::pair<T1,void> specialization (details / githubweb)
  93. Fix support for Kokkos::Array of const-qualified element type (details / githubweb)
  94. Try to fix the CUDA 11.0 build (details / githubweb)
  95. Revert "Try to fix the CUDA 11.0 build" (details / githubweb)
  96. Alternate definition of Impl::is_nothrow_swappable_v for NVCC version less than 11.4 (details / githubweb)
  97. Avoid introducing a 2nd definition of the Impl::swappable trait (details / githubweb)
  98. Do not require OpenMP support for languages other than CXX (details / githubweb)
  99. Update version (details / githubweb)
  100. Add Linux Foundation notice and fix C++ standard (details / githubweb)
  101. SYCL: Avoid deprecated floating-point number abs overloads (#6959) (details / githubweb)
  102. OpenMPTarget: Update loop order in MDRange (#6925) (details / githubweb)
  103. Threads: Don't silently allow m_instance to be a nullptr (#6969) (details / githubweb)
  104. Fix enabling OpenMP with HIP and "compile as CMake language" (details / githubweb)
  105. Fix OpenMP+CUDA when `Kokkos_ENABLE_COMPILE_AS_CMAKE_LANGUAGE` is `ON` (details / githubweb)
  106. Fix fedora CI builds with flang-new (details / githubweb)
  107. Fix Copyright file (details / githubweb)
  108. Also use is_nothrow_swappable workaround for Intel Classic Compilers (#6983) (details / githubweb)
  109. Add thread-safety tests (#6938) (details / githubweb)
  110. Fix deprecation warnings with GCC for pair<T1,void> comparison operators (details / githubweb)
  111. Fix TPL_LIBRARY_SUFFIXES for 32-bit build (details / githubweb)
  112. Avoid duplicated definition of KOKKOS_IMPL_32BIT (details / githubweb)
  113. Disable failing parallel_scan_with_reducers test (details / githubweb)
  114. Fix a bug when using realloc on views of non-default constructible element types (#6993) (details / githubweb)
  115. SYCL: Prepare Parallel* for Graphs (#6988) (details / githubweb)
  116. SYCL: Fix deprecation in custom parallel_for RangePolicy implementation (details / githubweb)
  117. [ci skip] update changelog for 4.3.1 (#6995) (details / githubweb)
  118. OpenMPTarget: Use mutex lock for parallel scan. (details / githubweb)
  119. Remove cuSPARSE TPL (details / githubweb)
  120. SYCL: Print submission command queue property (#7004) (details / githubweb)
  121. Test DualView resize/realloc for types without default constructor (details / githubweb)
  122. Suppress deprecated warnings via pragma push/pop in the tests (#6999) (details / githubweb)
  123. This PR changes the default execution behavior of the parallel_for(team-policy) constructs in the OpenACC backend. (details / githubweb)
  124. Use parallel_for instead of parallel_reduce for check (details / githubweb)
  125. Fix unique_any_senders nvcc template deduction (details / githubweb)
  126. Properly guard deprecated `Kokkos_Vector.hpp` header self contained test (#7016) (details / githubweb)
  127. mdspan converting constructors (#6830) (details / githubweb)
  128. SYCL: Don't use shuffles for top-level reductions (#7009) (details / githubweb)
  129. Add a CITATION.cff file (#7008) (details / githubweb)
  130. Always run Graph tests (#7011) (details / githubweb)
  131. Fix undefined behavior in is_zero_byte (#7014) (details / githubweb)
  132. Disable OpenMPTarget Kokkos::Graph test (does not compile) (details / githubweb)
  133. Fixup for disabling deprecation warnings with NVC++ (details / githubweb)
  134. Remove unused CudaInternal::cuda_{malloc,free}_async_wrapper (details / githubweb)
  135. Improve `Impl::is_zero_byte()` (#7017) (details / githubweb)
  136. OpenMPTarget: Fix compiling Graph tests (#7020) (details / githubweb)
  137. SYCL: Skip launch_six Graph test (details / githubweb)
  138. HIP: Use builtin atomic for compare_exchange (#7000) (details / githubweb)
  139. Fix Kokkos README's FENL link (details / githubweb)
  140. Copied the deduction guides and test cases over from branch (details / githubweb)
  141. Refactor: Replace SFINAE by `if constexpr` for `create_mirror*` functions (#6955) (details / githubweb)
  142. Workaround icpc warnings (details / githubweb)
  143. Adding `Kokkos::to_array` (#6375) (details / githubweb)
  144. Fix adjacent difference (#6922) (details / githubweb)
  145. TestArray: add intel guard to to_array implicit conversion test (details / githubweb)
  146. Avoid using "#if not defined" (details / githubweb)
  147. (Rebase) Partial fix to compile time issues w/nvcc + Kokkos_ENABLE_DEBUG_BOUNDS_CHECK (#7013) (details / githubweb)
  148. Add dependabot for GitHub Actions (details / githubweb)
  149. Added static_asserts for out of range tuple_element and get (details / githubweb)
  150. Add tests to `ScopeGuard` (#7028) (details / githubweb)
  151. Improve Kokkos Graphs (#7039) (details / githubweb)
  152. Introduce `KOKKOS_DEDUCTION_GUIDE` macro to allow user-defined deduction guide in device code for clang compiler (#6954) (details / githubweb)
  153. Add nightly CI on Frontier (#7048) (details / githubweb)
  154. Update scorecard GitHub workflow (details / githubweb)
  155. Fix unused symbols in CTAD tests (details / githubweb)
  156. Cuda: Fix nvcc warnings (details / githubweb)
  157. Fix quotation marks in CXX flags (details / githubweb)
  158. Fix kokkos_swap (details / githubweb)
  159. Fix array size (details / githubweb)
  160. Fix gtest (details / githubweb)
  161. Fix .jenkins whitespce (details / githubweb)
  162. Only use -Werror all-warnings with explicit nvcc_wrapper (details / githubweb)
  163. Fix dangling reference (details / githubweb)
  164. Suppress 'long double' is treated as 'double' in device code (details / githubweb)
  165. Use -Xcudafe --diag_suppress=20208 for 11.6 build; nothing else seems to help (details / githubweb)
  166. Try moving pragma suppress to tests (details / githubweb)
  167. Use -Xcudafe --diag_suppress=20208 in Makefile build (details / githubweb)
  168. Move dependabot to .github (details / githubweb)
  169. Bump Jimver/cuda-toolkit from 0.2.14 to 0.2.15 (details / githubweb)
  170. Bump DoozyX/clang-format-lint-action from 0.16.2 to 0.17 (details / githubweb)
  171. Bump version in the readme (details / githubweb)
  172. Restrict to_array subtest to NVCC >= 11.4.0 (details / githubweb)
  173. [ci skip] Use ROCM 6.1 in the nightly CI and disable one test (details / githubweb)
  174. Switch to using functors in sort_by_key_via_sort (#7059) (details / githubweb)
  175. Fix mdspan test (details / githubweb)
  176. Update to CUDA 12.4.1 in MSVC CI (details / githubweb)
  177. Workaround MSVC compiler issues in Views (details / githubweb)
  178. Avoid lambda in sort_by_key_via_sort (details / githubweb)
  179. Fix using CUDAToolkit for CMake 3.28.4 and higher (details / githubweb)
  180. Move check CMake 3.20.1 with nvhpc (details / githubweb)
  181. Fix closing brackets (details / githubweb)
  182. KOKKOS_CUDA_ERROR->DEFAULT_MSG (details / githubweb)
  183. Workaround icpc "missing return statement at end of non-void function" (details / githubweb)
  184. Enable MDSPAN support by default (#7069) (details / githubweb)
  185. Add SpaceAwareAccessor (details / githubweb)
  186. Add basic tests for SpaceAwareAccessor (details / githubweb)
  187. Use SpaceAwareAccessor in View mdspan-interop (details / githubweb)
  188. Add specialization of SpaceAwareAcc for AnonymousSpace (details / githubweb)
  189. Address review comments (details / githubweb)
  190. Fix incompatible dynamic_extent definition in Kokkos (details / githubweb)
  191. SpaceAwareAccessor: fix issues (no-unique-address, is_empty) (details / githubweb)
  192. Fix using shared libraries and -fvisibility=hidden (#7065) (details / githubweb)
  193. OpenACC: Skip exec_space_thread_safety_range_scan (#7022) (details / githubweb)
  194. Structured binding support for Kokkos::complex (#7040) (details / githubweb)
  195. Fix thread-safety for the Serial backend (details / githubweb)
  196. Update from GCC 13 to 14 and use C++ 26 in Jenkins nightly (details / githubweb)
  197. Move `layout_iterate_type_selector` into Impl namespace (#7076) (details / githubweb)
  198. Use correct includes for spaceawareaccessor (details / githubweb)
  199. Improve performance for deleting an instance. (details / githubweb)
  200. fix refcount exception safety (#6289) (details / githubweb)
  201. Don't delete special member functions explicitly (details / githubweb)
  202. Complex needs a tuple include (details / githubweb)
  203. Update internal desul file copies to 60c1115 (details / githubweb)
  204. Update desul version in github workflow (details / githubweb)
  205. Resolve various bogus icpc -Werror (#7079) (details / githubweb)
  206. Remove support for NVHPC as CUDA device compiler (details / githubweb)
  207. remove cmake options to change default spaces (details / githubweb)
  208. Temporarily disable failing parts in the TestMDSpan.hpp for the OpenACC (details / githubweb)
  209. Add missing space (details / githubweb)
  210. Fix more icpc issues (details / githubweb)
  211. Bump Jimver/cuda-toolkit from 0.2.15 to 0.2.16 (details / githubweb)
  212. Copy print_configuration setting in combination of kokkos settings. (details / githubweb)
  213. Couple more icpc -Werror fixes (details / githubweb)
  214. Cleanup KokkosP hooks in `Profiling::` (#7096) (details / githubweb)
  215. Remove FIXME_NVHPC 23.7 guards (details / githubweb)
  216. Fix SpaceAwareAccessor based on usage experiment in View (#7088) (details / githubweb)
  217. Added width 8 abi for avx2 (details / githubweb)
  218. Added for AVX512 (details / githubweb)
  219. Added for width 4 for NEON (details / githubweb)
  220. clang-formatted (details / githubweb)
  221. Disabling simd unit tests from building for Windows+CUDA build (details / githubweb)
  222. Workaround for the compilation failure for rocm 5.6-6.0 (details / githubweb)
  223. Added a comment about is_type structs (details / githubweb)
  224. clang formating (details / githubweb)
  225. Restrict jenkins CI not to run on hopper for nvcc < 11.8 (details / githubweb)
  226. cmake: fix relative to find kokkos_compiler_launcher (details / githubweb)
  227. Clean up KOKKOS_LIB_INCLUDE_DIRECTORIES, append include directories to associated targets in Trilinos builds (#7103) (details / githubweb)
  228. Workaround to ice with icpc when using -no-ip (#7106) (details / githubweb)
  229. move view allocation related functionality to a new header (#7110) (details / githubweb)
  230. SYCL CI: Manually build oneDPL (details / githubweb)
  231. Add missing concepts include in test (details / githubweb)
  232. Github CI: Test with C++17, C++20, and C++23 (#7082) (details / githubweb)
  233. Update base image for ROCm 5.6 (details / githubweb)
  234. Update mdspan to 98a12b01b51b2 (details / githubweb)
  235. Add AtomicAccessorRelaxed (#7089) (details / githubweb)
  236. Fix div by zero in math ops testing (details / githubweb)
  237. Fix and test with -fsanitize=undefined in GitHub CI (#7104) (details / githubweb)
  238. Bump actions/upload-artifact from 4.3.3 to 4.3.4 (details / githubweb)
  239. Update HIP nightly build base image Ubuntu 20.04 -> 22.04 (details / githubweb)
  240. Deprecate `RawMemoryAllocationFailure::FailureMode::MaximumCudaUVMAllocationsExceeded` (#7120) (details / githubweb)
  241. Refactor: Move logic of `create_mirror*` to `Impl::create_mirror*` (#7061) (details / githubweb)
  242. Define minimum compiler versions for C++20 support (details / githubweb)
  243. Drop (unused) cuda uvm allocations counter (details / githubweb)
  244. Do not bother catching memory allocation failure and rethrow (details / githubweb)
  245. Fix using and, or, xor in desul with MSVC (#7124) (details / githubweb)
  246. Do not swallow errors when deallocating memory with CUDA (details / githubweb)
  247. Let the throwing push finalize hook calls terminate test actually run (details / githubweb)
  248. Drop (unused) KOKKOS_ADD_ADVANCED_TEST TriBITS function (details / githubweb)
  249. Simplify the logic when finalizing and calling the registered functions (details / githubweb)
  250. [ci skip] Fix ROCm version to 6.1.2 in nightly CI (details / githubweb)
  251. Disable the PushFinalizeHookTerminate test on Windows (details / githubweb)
  252. OpenMP: Ensure kernels submitted by multiple threads to the same instance don't run concurrently (#6151) (details / githubweb)
  253. SYCL: Add missing include for std::stringstream (details / githubweb)
  254. SYCL: Add support for Graphs (#6912) (details / githubweb)
  255. [ci skip] rename jenkins build (details / githubweb)
  256. NVHPC: Disable failing NaN tests (details / githubweb)
  257. no_device_stack is unknown (details / githubweb)
  258. Check for LIBCXX 10 or later for C++20 and later (#7123) (details / githubweb)
  259. Get rid of `RawMemoryAllocationFailure::AllocationMechanism` and derived backend-specific exceptions (#7139) (details / githubweb)
  260. Add `likely` and `unlikely` attribute from C++20 to ref counting in views (#6730) (details / githubweb)
  261. Throw bad alloc if omp_target_alloc() returns nullptr (details / githubweb)
  262. Enable test_view_allocation_error with OpenMPTarget (details / githubweb)
  263. Fix gcc-14 C++26 nightly jenkins build (#7137) (details / githubweb)
  264. Fix Trilinos nightly failure due to `create_mirror*` refactor (#7126) (details / githubweb)
  265. fixup! Throw bad alloc if omp_target_alloc() returns nullptr (details / githubweb)
  266. Add CMake options to control compilation flags for AMD GPUs (#7127) (details / githubweb)
  267. Fix using View without corresponding mdspan-type (#7140) (details / githubweb)
Changes
  1. Reenable TestHIP_Memory_Requirements (details / githubweb)
  2. SYCL: Cleanup device selection (details / githubweb)
  3. Redeine deep_copy for UnorderMap (details / githubweb)
  4. Remove unused typedef (details / githubweb)
  5. Adding SECURITY.md file (details / githubweb)
  6. Only call deep_copy_view() from deep_copy(), add deprecation warning (details / githubweb)
  7. Cleanup KOKKOS_CONFIGURE_CORE (details / githubweb)
  8. [4.2.01]: changelog update  (#6656) (details / githubweb)
  9. Clean up shift_{right, left}_team_impl (#6821) (details / githubweb)
  10. Avoid detecting unwanted fences in the parallel_scan_no_fence test (details / githubweb)
  11. Add test for current CTAD support with RangePolicy (#6803) (details / githubweb)
  12. multi-GPU support: Add test for all policies (#6782) (details / githubweb)
  13. Implement sort_by_key (#6801) (details / githubweb)
  14. Fix scorecard workflow (#6831) (details / githubweb)
  15. Remove redundant RangePolicy constructor (#6841) (details / githubweb)
  16. Bump Google Benchmark version v1.{6.2 -> 7.1} in CMake FetchContent (details / githubweb)
  17. Remove variadic range policy constructor (#6845) (details / githubweb)
  18. Avoid unused variable warning in TestRangePolicyCTAD.cpp (details / githubweb)
  19. bytes_and_flops with CMake (details / githubweb)
  20. bytes_and_flops fix a small bug in command line argument (details / githubweb)
  21. add gather benchmark to CMake (details / githubweb)
  22. add atomic benchmark to CMake (details / githubweb)
  23. add policy_performance benchmark to CMake (details / githubweb)
  24. fix some warnings in policy_performance benchmark (details / githubweb)
  25. policy_benchmark: fix indentation (details / githubweb)
  26. Benchmarks: disable 2 benchmarks for OpenMPTarget (details / githubweb)
  27. bytes_and_flops: fix a counter name (details / githubweb)
  28. Fix C-style cast (details / githubweb)
  29. Fix amdclang++ compilation (#6857) (details / githubweb)
  30. Fix fallback implementation for sort_by_key (#6856) (details / githubweb)
  31. Update the unsafe implicit conversion error message in MDRangePolicy (#6855) (details / githubweb)
  32. Split some classes from Kokkos_ViewMapping (#6859) (details / githubweb)
  33. update mdspan tpl (details / githubweb)
  34. CTAD (deduction guides) for RangePolicy (#6850) (details / githubweb)
  35. CTAD (deduction guides) for MDRangePolicy (#5516) (details / githubweb)
  36. Add support for rocThrust in sort when using HIP (#6793) (details / githubweb)
  37. Add HIP specialization for sort-by-key (details / githubweb)
  38. Fix unused variable warning in TestMDRangePolicyCTAD.cpp (details / githubweb)
  39. Improve message on view out of bounds access and always abort (#6861) (details / githubweb)
  40. Fix linking with rothrust in downstream applications (details / githubweb)
  41. SYCL: Make sure to call find_dependency for oneDPL if necessary  (#6870) (details / githubweb)
  42. Don't use Fedora development version in GitHub CI (details / githubweb)
  43. kokkos_tpls.cmake: update default option to enable rocthrust (details / githubweb)
  44. Scratch space fix for MultiGPU (#6866) (details / githubweb)
  45. [ci skip] Bump version number to 4.3.99 (details / githubweb)
  46. Fix deep copy when filling Rank-7 views (details / githubweb)
  47. TestViewCopy_c.hpp: add new unit test for deep copy (ViewFill) (details / githubweb)
  48. TestViewCopy_c.hpp: better handling for OpenMPTarget (details / githubweb)
  49. Use builtin for atomic_fetch in the HIP backend (details / githubweb)
  50. Fix Makefile.kokkos for Threads (details / githubweb)
  51. Fix a bug in Makefile when using AMD GPU architectures (#6892) (details / githubweb)
  52. Cuda: Fix configuring with CMake 3.28.4 (#6898) (details / githubweb)
  53. Update Intel GPU architectures in Makefile (#6895) (details / githubweb)
  54. Drop Experimental::LayoutTiled class template (details / githubweb)
  55. Get rid of now unnecessary use of is_layouttiled trait (details / githubweb)
  56. Deprecate is_layouttiled trait (details / githubweb)
  57. Move `Kokkos::Array` tests to a more suitable place (#6905) (details / githubweb)
  58. Do not return a copy of the input functor for Kokkos::Experimental::for_each (details / githubweb)
  59. Drop specialization of ViewMapping for Kokkos::Array (details / githubweb)
  60. Accommodate users that depend on a code that define silly macros (#6909) (details / githubweb)
  61. SYCL: Fix multi-GPU support and add test (#6887) (details / githubweb)
  62. SYCL: Unroll shuffle loops for top-level parallel_reduce and parallel_scan (#6750) (details / githubweb)
  63. Preparing readme for develop as the default branch (#6796) (details / githubweb)
  64. CHANGELOG.md: 4.3.00 update (details / githubweb)
  65. Refactor: Uniformize `create_mirror*` parameter name for views (#6917) (details / githubweb)
  66. SYCL: Improve team_reduce implementation (#6562) (details / githubweb)
  67. Use recommended/max team size functions in Cuda ParallelFor and Reduce constructors (#6891) (details / githubweb)
  68. SYCL: Fix range in subgroup scan for workgroup_scan (details / githubweb)
  69. Add ROCm 6.0 in the nightly CI (details / githubweb)
  70. MI300 support unified memory support (#6877) (details / githubweb)
  71. cuda: reduction with `RangePolicy`: fix grid dimensions to work for large values and avoid overflow (#6578) (details / githubweb)
  72. Fix nightly CI (details / githubweb)
  73. Deprecate trailing Proxy template argument in Kokkos::Array (details / githubweb)
  74. CUDA: Update nvcc_wrapper (details / githubweb)
  75. bring back --fmad option to nvcc_wrapper (#6931) (details / githubweb)
  76. Remove unnecessary header include (details / githubweb)
  77. Adding converting constructor in Kokkos::RandomAccessIterator (#6929) (details / githubweb)
  78. Prefer standard C++ feature testing to guard the C++20 requires expression (details / githubweb)
  79. Add `kokkos_swap(Array<T, N>)` sepcialization (details / githubweb)
  80. Deprecate specialization of Kokkos::pair for a single element (details / githubweb)
  81. Fix deprecated warning from `Kokkos::Array` specialization (#6945) (details / githubweb)
  82. [ci skip] Enable deprecated code and deprecated warnings in nightly CI (details / githubweb)
  83. Serial: Use the provided execution space instance in TeamPolicy (details / githubweb)
  84. Fixed the link for P6601 (Threads backend change) (details / githubweb)
  85. Add test (details / githubweb)
  86. Use Kokkos::AUTO for OpenMPTarget (details / githubweb)
  87. kokkos_swap(Array) member friend should not be templated on some other type U (details / githubweb)
  88. Fix noexcept specification for kokkos_swap on zero-sized arrays (details / githubweb)
  89. Replace deprecated sycl::device_ptr/sycl::host_ptr (details / githubweb)
  90. Introduce alias based on feature macro (details / githubweb)
  91. Kokkos::Impl::SYCLTypes:: -> Kokkos::Impl::sycl_ (details / githubweb)
  92. Fix wrong macro guards for deprecated Kokkos::pair<T1,void> specialization (details / githubweb)
  93. Fix support for Kokkos::Array of const-qualified element type (details / githubweb)
  94. Try to fix the CUDA 11.0 build (details / githubweb)
  95. Revert "Try to fix the CUDA 11.0 build" (details / githubweb)
  96. Alternate definition of Impl::is_nothrow_swappable_v for NVCC version less than 11.4 (details / githubweb)
  97. Avoid introducing a 2nd definition of the Impl::swappable trait (details / githubweb)
  98. Do not require OpenMP support for languages other than CXX (details / githubweb)
  99. Update version (details / githubweb)
  100. Add Linux Foundation notice and fix C++ standard (details / githubweb)
  101. SYCL: Avoid deprecated floating-point number abs overloads (#6959) (details / githubweb)
  102. OpenMPTarget: Update loop order in MDRange (#6925) (details / githubweb)
  103. Threads: Don't silently allow m_instance to be a nullptr (#6969) (details / githubweb)
  104. Fix enabling OpenMP with HIP and "compile as CMake language" (details / githubweb)
  105. Fix OpenMP+CUDA when `Kokkos_ENABLE_COMPILE_AS_CMAKE_LANGUAGE` is `ON` (details / githubweb)
  106. Fix fedora CI builds with flang-new (details / githubweb)
  107. Fix Copyright file (details / githubweb)
  108. Also use is_nothrow_swappable workaround for Intel Classic Compilers (#6983) (details / githubweb)
  109. Add thread-safety tests (#6938) (details / githubweb)
  110. Fix deprecation warnings with GCC for pair<T1,void> comparison operators (details / githubweb)
  111. Fix TPL_LIBRARY_SUFFIXES for 32-bit build (details / githubweb)
  112. Avoid duplicated definition of KOKKOS_IMPL_32BIT (details / githubweb)
  113. Disable failing parallel_scan_with_reducers test (details / githubweb)
  114. Fix a bug when using realloc on views of non-default constructible element types (#6993) (details / githubweb)
  115. SYCL: Prepare Parallel* for Graphs (#6988) (details / githubweb)
  116. SYCL: Fix deprecation in custom parallel_for RangePolicy implementation (details / githubweb)
  117. [ci skip] update changelog for 4.3.1 (#6995) (details / githubweb)
  118. OpenMPTarget: Use mutex lock for parallel scan. (details / githubweb)
  119. Remove cuSPARSE TPL (details / githubweb)
  120. SYCL: Print submission command queue property (#7004) (details / githubweb)
  121. Test DualView resize/realloc for types without default constructor (details / githubweb)
  122. Suppress deprecated warnings via pragma push/pop in the tests (#6999) (details / githubweb)
  123. This PR changes the default execution behavior of the parallel_for(team-policy) constructs in the OpenACC backend. (details / githubweb)
  124. Use parallel_for instead of parallel_reduce for check (details / githubweb)
  125. Fix unique_any_senders nvcc template deduction (details / githubweb)
  126. Properly guard deprecated `Kokkos_Vector.hpp` header self contained test (#7016) (details / githubweb)
  127. mdspan converting constructors (#6830) (details / githubweb)
  128. SYCL: Don't use shuffles for top-level reductions (#7009) (details / githubweb)
  129. Add a CITATION.cff file (#7008) (details / githubweb)
  130. Always run Graph tests (#7011) (details / githubweb)
  131. Fix undefined behavior in is_zero_byte (#7014) (details / githubweb)
  132. Disable OpenMPTarget Kokkos::Graph test (does not compile) (details / githubweb)
  133. Fixup for disabling deprecation warnings with NVC++ (details / githubweb)
  134. Remove unused CudaInternal::cuda_{malloc,free}_async_wrapper (details / githubweb)
  135. Improve `Impl::is_zero_byte()` (#7017) (details / githubweb)
  136. OpenMPTarget: Fix compiling Graph tests (#7020) (details / githubweb)
  137. SYCL: Skip launch_six Graph test (details / githubweb)
  138. HIP: Use builtin atomic for compare_exchange (#7000) (details / githubweb)
  139. Fix Kokkos README's FENL link (details / githubweb)
  140. Copied the deduction guides and test cases over from branch (details / githubweb)
  141. Refactor: Replace SFINAE by `if constexpr` for `create_mirror*` functions (#6955) (details / githubweb)
  142. Workaround icpc warnings (details / githubweb)
  143. Adding `Kokkos::to_array` (#6375) (details / githubweb)
  144. Fix adjacent difference (#6922) (details / githubweb)
  145. TestArray: add intel guard to to_array implicit conversion test (details / githubweb)
  146. Avoid using "#if not defined" (details / githubweb)
  147. (Rebase) Partial fix to compile time issues w/nvcc + Kokkos_ENABLE_DEBUG_BOUNDS_CHECK (#7013) (details / githubweb)
  148. Add dependabot for GitHub Actions (details / githubweb)
  149. Added static_asserts for out of range tuple_element and get (details / githubweb)
  150. Add tests to `ScopeGuard` (#7028) (details / githubweb)
  151. Improve Kokkos Graphs (#7039) (details / githubweb)
  152. Introduce `KOKKOS_DEDUCTION_GUIDE` macro to allow user-defined deduction guide in device code for clang compiler (#6954) (details / githubweb)
  153. Add nightly CI on Frontier (#7048) (details / githubweb)
  154. Update scorecard GitHub workflow (details / githubweb)
  155. Fix unused symbols in CTAD tests (details / githubweb)
  156. Cuda: Fix nvcc warnings (details / githubweb)
  157. Fix quotation marks in CXX flags (details / githubweb)
  158. Fix kokkos_swap (details / githubweb)
  159. Fix array size (details / githubweb)
  160. Fix gtest (details / githubweb)
  161. Fix .jenkins whitespce (details / githubweb)
  162. Only use -Werror all-warnings with explicit nvcc_wrapper (details / githubweb)
  163. Fix dangling reference (details / githubweb)
  164. Suppress 'long double' is treated as 'double' in device code (details / githubweb)
  165. Use -Xcudafe --diag_suppress=20208 for 11.6 build; nothing else seems to help (details / githubweb)
  166. Try moving pragma suppress to tests (details / githubweb)
  167. Use -Xcudafe --diag_suppress=20208 in Makefile build (details / githubweb)
  168. Move dependabot to .github (details / githubweb)
  169. Bump Jimver/cuda-toolkit from 0.2.14 to 0.2.15 (details / githubweb)
  170. Bump DoozyX/clang-format-lint-action from 0.16.2 to 0.17 (details / githubweb)
  171. Bump version in the readme (details / githubweb)
  172. Restrict to_array subtest to NVCC >= 11.4.0 (details / githubweb)
  173. [ci skip] Use ROCM 6.1 in the nightly CI and disable one test (details / githubweb)
  174. Switch to using functors in sort_by_key_via_sort (#7059) (details / githubweb)
  175. Fix mdspan test (details / githubweb)
  176. Update to CUDA 12.4.1 in MSVC CI (details / githubweb)
  177. Workaround MSVC compiler issues in Views (details / githubweb)
  178. Avoid lambda in sort_by_key_via_sort (details / githubweb)
  179. Fix using CUDAToolkit for CMake 3.28.4 and higher (details / githubweb)
  180. Move check CMake 3.20.1 with nvhpc (details / githubweb)
  181. Fix closing brackets (details / githubweb)
  182. KOKKOS_CUDA_ERROR->DEFAULT_MSG (details / githubweb)
  183. Workaround icpc "missing return statement at end of non-void function" (details / githubweb)
  184. Enable MDSPAN support by default (#7069) (details / githubweb)
  185. Add SpaceAwareAccessor (details / githubweb)
  186. Add basic tests for SpaceAwareAccessor (details / githubweb)
  187. Use SpaceAwareAccessor in View mdspan-interop (details / githubweb)
  188. Add specialization of SpaceAwareAcc for AnonymousSpace (details / githubweb)
  189. Address review comments (details / githubweb)
  190. Fix incompatible dynamic_extent definition in Kokkos (details / githubweb)
  191. SpaceAwareAccessor: fix issues (no-unique-address, is_empty) (details / githubweb)
  192. Fix using shared libraries and -fvisibility=hidden (#7065) (details / githubweb)
  193. OpenACC: Skip exec_space_thread_safety_range_scan (#7022) (details / githubweb)
  194. Structured binding support for Kokkos::complex (#7040) (details / githubweb)
  195. Fix thread-safety for the Serial backend (details / githubweb)
  196. Update from GCC 13 to 14 and use C++ 26 in Jenkins nightly (details / githubweb)
  197. Move `layout_iterate_type_selector` into Impl namespace (#7076) (details / githubweb)
  198. Use correct includes for spaceawareaccessor (details / githubweb)
  199. Improve performance for deleting an instance. (details / githubweb)
  200. fix refcount exception safety (#6289) (details / githubweb)
  201. Don't delete special member functions explicitly (details / githubweb)
  202. Complex needs a tuple include (details / githubweb)
  203. Update internal desul file copies to 60c1115 (details / githubweb)
  204. Update desul version in github workflow (details / githubweb)
  205. Resolve various bogus icpc -Werror (#7079) (details / githubweb)
  206. Remove support for NVHPC as CUDA device compiler (details / githubweb)
  207. remove cmake options to change default spaces (details / githubweb)
  208. Temporarily disable failing parts in the TestMDSpan.hpp for the OpenACC (details / githubweb)
  209. Add missing space (details / githubweb)
  210. Fix more icpc issues (details / githubweb)
  211. Bump Jimver/cuda-toolkit from 0.2.15 to 0.2.16 (details / githubweb)
  212. Copy print_configuration setting in combination of kokkos settings. (details / githubweb)
  213. Couple more icpc -Werror fixes (details / githubweb)
  214. Cleanup KokkosP hooks in `Profiling::` (#7096) (details / githubweb)
  215. Remove FIXME_NVHPC 23.7 guards (details / githubweb)
  216. Fix SpaceAwareAccessor based on usage experiment in View (#7088) (details / githubweb)
  217. Added width 8 abi for avx2 (details / githubweb)
  218. Added for AVX512 (details / githubweb)
  219. Added for width 4 for NEON (details / githubweb)
  220. clang-formatted (details / githubweb)
  221. Disabling simd unit tests from building for Windows+CUDA build (details / githubweb)
  222. Workaround for the compilation failure for rocm 5.6-6.0 (details / githubweb)
  223. Added a comment about is_type structs (details / githubweb)
  224. clang formating (details / githubweb)
  225. Restrict jenkins CI not to run on hopper for nvcc < 11.8 (details / githubweb)
  226. cmake: fix relative to find kokkos_compiler_launcher (details / githubweb)
  227. Clean up KOKKOS_LIB_INCLUDE_DIRECTORIES, append include directories to associated targets in Trilinos builds (#7103) (details / githubweb)
  228. Workaround to ice with icpc when using -no-ip (#7106) (details / githubweb)
  229. move view allocation related functionality to a new header (#7110) (details / githubweb)
  230. SYCL CI: Manually build oneDPL (details / githubweb)
  231. Add missing concepts include in test (details / githubweb)
  232. Github CI: Test with C++17, C++20, and C++23 (#7082) (details / githubweb)
  233. Update base image for ROCm 5.6 (details / githubweb)
  234. Update mdspan to 98a12b01b51b2 (details / githubweb)
  235. Add AtomicAccessorRelaxed (#7089) (details / githubweb)
  236. Fix div by zero in math ops testing (details / githubweb)
  237. Fix and test with -fsanitize=undefined in GitHub CI (#7104) (details / githubweb)
  238. Bump actions/upload-artifact from 4.3.3 to 4.3.4 (details / githubweb)
  239. Update HIP nightly build base image Ubuntu 20.04 -> 22.04 (details / githubweb)
  240. Deprecate `RawMemoryAllocationFailure::FailureMode::MaximumCudaUVMAllocationsExceeded` (#7120) (details / githubweb)
  241. Refactor: Move logic of `create_mirror*` to `Impl::create_mirror*` (#7061) (details / githubweb)
  242. Define minimum compiler versions for C++20 support (details / githubweb)
  243. Drop (unused) cuda uvm allocations counter (details / githubweb)
  244. Do not bother catching memory allocation failure and rethrow (details / githubweb)
  245. Fix using and, or, xor in desul with MSVC (#7124) (details / githubweb)
  246. Do not swallow errors when deallocating memory with CUDA (details / githubweb)
  247. Let the throwing push finalize hook calls terminate test actually run (details / githubweb)
  248. Drop (unused) KOKKOS_ADD_ADVANCED_TEST TriBITS function (details / githubweb)
  249. Simplify the logic when finalizing and calling the registered functions (details / githubweb)
  250. [ci skip] Fix ROCm version to 6.1.2 in nightly CI (details / githubweb)
  251. Disable the PushFinalizeHookTerminate test on Windows (details / githubweb)
  252. OpenMP: Ensure kernels submitted by multiple threads to the same instance don't run concurrently (#6151) (details / githubweb)
  253. SYCL: Add missing include for std::stringstream (details / githubweb)
  254. SYCL: Add support for Graphs (#6912) (details / githubweb)
  255. [ci skip] rename jenkins build (details / githubweb)
  256. NVHPC: Disable failing NaN tests (details / githubweb)
  257. no_device_stack is unknown (details / githubweb)
  258. Check for LIBCXX 10 or later for C++20 and later (#7123) (details / githubweb)
  259. Get rid of `RawMemoryAllocationFailure::AllocationMechanism` and derived backend-specific exceptions (#7139) (details / githubweb)
  260. Add `likely` and `unlikely` attribute from C++20 to ref counting in views (#6730) (details / githubweb)
  261. Throw bad alloc if omp_target_alloc() returns nullptr (details / githubweb)
  262. Enable test_view_allocation_error with OpenMPTarget (details / githubweb)
  263. Fix gcc-14 C++26 nightly jenkins build (#7137) (details / githubweb)
  264. Fix Trilinos nightly failure due to `create_mirror*` refactor (#7126) (details / githubweb)
  265. fixup! Throw bad alloc if omp_target_alloc() returns nullptr (details / githubweb)
  266. Add CMake options to control compilation flags for AMD GPUs (#7127) (details / githubweb)
  267. Fix using View without corresponding mdspan-type (#7140) (details / githubweb)
Changes
  1. Reenable TestHIP_Memory_Requirements (details / githubweb)
  2. SYCL: Cleanup device selection (details / githubweb)
  3. Redeine deep_copy for UnorderMap (details / githubweb)
  4. Remove unused typedef (details / githubweb)
  5. Adding SECURITY.md file (details / githubweb)
  6. Only call deep_copy_view() from deep_copy(), add deprecation warning (details / githubweb)
  7. Cleanup KOKKOS_CONFIGURE_CORE (details / githubweb)
  8. [4.2.01]: changelog update  (#6656) (details / githubweb)
  9. Clean up shift_{right, left}_team_impl (#6821) (details / githubweb)
  10. Avoid detecting unwanted fences in the parallel_scan_no_fence test (details / githubweb)
  11. Add test for current CTAD support with RangePolicy (#6803) (details / githubweb)
  12. multi-GPU support: Add test for all policies (#6782) (details / githubweb)
  13. Implement sort_by_key (#6801) (details / githubweb)
  14. Fix scorecard workflow (#6831) (details / githubweb)
  15. Remove redundant RangePolicy constructor (#6841) (details / githubweb)
  16. Bump Google Benchmark version v1.{6.2 -> 7.1} in CMake FetchContent (details / githubweb)
  17. Remove variadic range policy constructor (#6845) (details / githubweb)
  18. Avoid unused variable warning in TestRangePolicyCTAD.cpp (details / githubweb)
  19. bytes_and_flops with CMake (details / githubweb)
  20. bytes_and_flops fix a small bug in command line argument (details / githubweb)
  21. add gather benchmark to CMake (details / githubweb)
  22. add atomic benchmark to CMake (details / githubweb)
  23. add policy_performance benchmark to CMake (details / githubweb)
  24. fix some warnings in policy_performance benchmark (details / githubweb)
  25. policy_benchmark: fix indentation (details / githubweb)
  26. Benchmarks: disable 2 benchmarks for OpenMPTarget (details / githubweb)
  27. bytes_and_flops: fix a counter name (details / githubweb)
  28. Fix C-style cast (details / githubweb)
  29. Fix amdclang++ compilation (#6857) (details / githubweb)
  30. Fix fallback implementation for sort_by_key (#6856) (details / githubweb)
  31. Update the unsafe implicit conversion error message in MDRangePolicy (#6855) (details / githubweb)
  32. Split some classes from Kokkos_ViewMapping (#6859) (details / githubweb)
  33. update mdspan tpl (details / githubweb)
  34. CTAD (deduction guides) for RangePolicy (#6850) (details / githubweb)
  35. CTAD (deduction guides) for MDRangePolicy (#5516) (details / githubweb)
  36. Add support for rocThrust in sort when using HIP (#6793) (details / githubweb)
  37. Add HIP specialization for sort-by-key (details / githubweb)
  38. Fix unused variable warning in TestMDRangePolicyCTAD.cpp (details / githubweb)
  39. Improve message on view out of bounds access and always abort (#6861) (details / githubweb)
  40. Fix linking with rothrust in downstream applications (details / githubweb)
  41. SYCL: Make sure to call find_dependency for oneDPL if necessary  (#6870) (details / githubweb)
  42. Don't use Fedora development version in GitHub CI (details / githubweb)
  43. kokkos_tpls.cmake: update default option to enable rocthrust (details / githubweb)
  44. Scratch space fix for MultiGPU (#6866) (details / githubweb)
  45. [ci skip] Bump version number to 4.3.99 (details / githubweb)
  46. Fix deep copy when filling Rank-7 views (details / githubweb)
  47. TestViewCopy_c.hpp: add new unit test for deep copy (ViewFill) (details / githubweb)
  48. TestViewCopy_c.hpp: better handling for OpenMPTarget (details / githubweb)
  49. Use builtin for atomic_fetch in the HIP backend (details / githubweb)
  50. Fix Makefile.kokkos for Threads (details / githubweb)
  51. Fix a bug in Makefile when using AMD GPU architectures (#6892) (details / githubweb)
  52. Cuda: Fix configuring with CMake 3.28.4 (#6898) (details / githubweb)
  53. Update Intel GPU architectures in Makefile (#6895) (details / githubweb)
  54. Drop Experimental::LayoutTiled class template (details / githubweb)
  55. Get rid of now unnecessary use of is_layouttiled trait (details / githubweb)
  56. Deprecate is_layouttiled trait (details / githubweb)
  57. Move `Kokkos::Array` tests to a more suitable place (#6905) (details / githubweb)
  58. Do not return a copy of the input functor for Kokkos::Experimental::for_each (details / githubweb)
  59. Drop specialization of ViewMapping for Kokkos::Array (details / githubweb)
  60. Accommodate users that depend on a code that define silly macros (#6909) (details / githubweb)
  61. SYCL: Fix multi-GPU support and add test (#6887) (details / githubweb)
  62. SYCL: Unroll shuffle loops for top-level parallel_reduce and parallel_scan (#6750) (details / githubweb)
  63. Preparing readme for develop as the default branch (#6796) (details / githubweb)
  64. CHANGELOG.md: 4.3.00 update (details / githubweb)
  65. Refactor: Uniformize `create_mirror*` parameter name for views (#6917) (details / githubweb)
  66. SYCL: Improve team_reduce implementation (#6562) (details / githubweb)
  67. Use recommended/max team size functions in Cuda ParallelFor and Reduce constructors (#6891) (details / githubweb)
  68. SYCL: Fix range in subgroup scan for workgroup_scan (details / githubweb)
  69. Add ROCm 6.0 in the nightly CI (details / githubweb)
  70. MI300 support unified memory support (#6877) (details / githubweb)
  71. cuda: reduction with `RangePolicy`: fix grid dimensions to work for large values and avoid overflow (#6578) (details / githubweb)
  72. Fix nightly CI (details / githubweb)
  73. Deprecate trailing Proxy template argument in Kokkos::Array (details / githubweb)
  74. CUDA: Update nvcc_wrapper (details / githubweb)
  75. bring back --fmad option to nvcc_wrapper (#6931) (details / githubweb)
  76. Remove unnecessary header include (details / githubweb)
  77. Adding converting constructor in Kokkos::RandomAccessIterator (#6929) (details / githubweb)
  78. Prefer standard C++ feature testing to guard the C++20 requires expression (details / githubweb)
  79. Add `kokkos_swap(Array<T, N>)` sepcialization (details / githubweb)
  80. Deprecate specialization of Kokkos::pair for a single element (details / githubweb)
  81. Fix deprecated warning from `Kokkos::Array` specialization (#6945) (details / githubweb)
  82. [ci skip] Enable deprecated code and deprecated warnings in nightly CI (details / githubweb)
  83. Serial: Use the provided execution space instance in TeamPolicy (details / githubweb)
  84. Fixed the link for P6601 (Threads backend change) (details / githubweb)
  85. Add test (details / githubweb)
  86. Use Kokkos::AUTO for OpenMPTarget (details / githubweb)
  87. kokkos_swap(Array) member friend should not be templated on some other type U (details / githubweb)
  88. Fix noexcept specification for kokkos_swap on zero-sized arrays (details / githubweb)
  89. Replace deprecated sycl::device_ptr/sycl::host_ptr (details / githubweb)
  90. Introduce alias based on feature macro (details / githubweb)
  91. Kokkos::Impl::SYCLTypes:: -> Kokkos::Impl::sycl_ (details / githubweb)
  92. Fix wrong macro guards for deprecated Kokkos::pair<T1,void> specialization (details / githubweb)
  93. Fix support for Kokkos::Array of const-qualified element type (details / githubweb)
  94. Try to fix the CUDA 11.0 build (details / githubweb)
  95. Revert "Try to fix the CUDA 11.0 build" (details / githubweb)
  96. Alternate definition of Impl::is_nothrow_swappable_v for NVCC version less than 11.4 (details / githubweb)
  97. Avoid introducing a 2nd definition of the Impl::swappable trait (details / githubweb)
  98. Do not require OpenMP support for languages other than CXX (details / githubweb)
  99. Update version (details / githubweb)
  100. Add Linux Foundation notice and fix C++ standard (details / githubweb)
  101. SYCL: Avoid deprecated floating-point number abs overloads (#6959) (details / githubweb)
  102. OpenMPTarget: Update loop order in MDRange (#6925) (details / githubweb)
  103. Threads: Don't silently allow m_instance to be a nullptr (#6969) (details / githubweb)
  104. Fix enabling OpenMP with HIP and "compile as CMake language" (details / githubweb)
  105. Fix OpenMP+CUDA when `Kokkos_ENABLE_COMPILE_AS_CMAKE_LANGUAGE` is `ON` (details / githubweb)
  106. Fix fedora CI builds with flang-new (details / githubweb)
  107. Fix Copyright file (details / githubweb)
  108. Also use is_nothrow_swappable workaround for Intel Classic Compilers (#6983) (details / githubweb)
  109. Add thread-safety tests (#6938) (details / githubweb)
  110. Fix deprecation warnings with GCC for pair<T1,void> comparison operators (details / githubweb)
  111. Fix TPL_LIBRARY_SUFFIXES for 32-bit build (details / githubweb)
  112. Avoid duplicated definition of KOKKOS_IMPL_32BIT (details / githubweb)
  113. Disable failing parallel_scan_with_reducers test (details / githubweb)
  114. Fix a bug when using realloc on views of non-default constructible element types (#6993) (details / githubweb)
  115. SYCL: Prepare Parallel* for Graphs (#6988) (details / githubweb)
  116. SYCL: Fix deprecation in custom parallel_for RangePolicy implementation (details / githubweb)
  117. [ci skip] update changelog for 4.3.1 (#6995) (details / githubweb)
  118. OpenMPTarget: Use mutex lock for parallel scan. (details / githubweb)
  119. Remove cuSPARSE TPL (details / githubweb)
  120. SYCL: Print submission command queue property (#7004) (details / githubweb)
  121. Test DualView resize/realloc for types without default constructor (details / githubweb)
  122. Suppress deprecated warnings via pragma push/pop in the tests (#6999) (details / githubweb)
  123. This PR changes the default execution behavior of the parallel_for(team-policy) constructs in the OpenACC backend. (details / githubweb)
  124. Use parallel_for instead of parallel_reduce for check (details / githubweb)
  125. Fix unique_any_senders nvcc template deduction (details / githubweb)
  126. Properly guard deprecated `Kokkos_Vector.hpp` header self contained test (#7016) (details / githubweb)
  127. mdspan converting constructors (#6830) (details / githubweb)
  128. SYCL: Don't use shuffles for top-level reductions (#7009) (details / githubweb)
  129. Add a CITATION.cff file (#7008) (details / githubweb)
  130. Always run Graph tests (#7011) (details / githubweb)
  131. Fix undefined behavior in is_zero_byte (#7014) (details / githubweb)
  132. Disable OpenMPTarget Kokkos::Graph test (does not compile) (details / githubweb)
  133. Fixup for disabling deprecation warnings with NVC++ (details / githubweb)
  134. Remove unused CudaInternal::cuda_{malloc,free}_async_wrapper (details / githubweb)
  135. Improve `Impl::is_zero_byte()` (#7017) (details / githubweb)
  136. OpenMPTarget: Fix compiling Graph tests (#7020) (details / githubweb)
  137. SYCL: Skip launch_six Graph test (details / githubweb)
  138. HIP: Use builtin atomic for compare_exchange (#7000) (details / githubweb)
  139. Fix Kokkos README's FENL link (details / githubweb)
  140. Copied the deduction guides and test cases over from branch (details / githubweb)
  141. Refactor: Replace SFINAE by `if constexpr` for `create_mirror*` functions (#6955) (details / githubweb)
  142. Workaround icpc warnings (details / githubweb)
  143. Adding `Kokkos::to_array` (#6375) (details / githubweb)
  144. Fix adjacent difference (#6922) (details / githubweb)
  145. TestArray: add intel guard to to_array implicit conversion test (details / githubweb)
  146. Avoid using "#if not defined" (details / githubweb)
  147. (Rebase) Partial fix to compile time issues w/nvcc + Kokkos_ENABLE_DEBUG_BOUNDS_CHECK (#7013) (details / githubweb)
  148. Add dependabot for GitHub Actions (details / githubweb)
  149. Added static_asserts for out of range tuple_element and get (details / githubweb)
  150. Add tests to `ScopeGuard` (#7028) (details / githubweb)
  151. Improve Kokkos Graphs (#7039) (details / githubweb)
  152. Introduce `KOKKOS_DEDUCTION_GUIDE` macro to allow user-defined deduction guide in device code for clang compiler (#6954) (details / githubweb)
  153. Add nightly CI on Frontier (#7048) (details / githubweb)
  154. Update scorecard GitHub workflow (details / githubweb)
  155. Fix unused symbols in CTAD tests (details / githubweb)
  156. Cuda: Fix nvcc warnings (details / githubweb)
  157. Fix quotation marks in CXX flags (details / githubweb)
  158. Fix kokkos_swap (details / githubweb)
  159. Fix array size (details / githubweb)
  160. Fix gtest (details / githubweb)
  161. Fix .jenkins whitespce (details / githubweb)
  162. Only use -Werror all-warnings with explicit nvcc_wrapper (details / githubweb)
  163. Fix dangling reference (details / githubweb)
  164. Suppress 'long double' is treated as 'double' in device code (details / githubweb)
  165. Use -Xcudafe --diag_suppress=20208 for 11.6 build; nothing else seems to help (details / githubweb)
  166. Try moving pragma suppress to tests (details / githubweb)
  167. Use -Xcudafe --diag_suppress=20208 in Makefile build (details / githubweb)
  168. Move dependabot to .github (details / githubweb)
  169. Bump Jimver/cuda-toolkit from 0.2.14 to 0.2.15 (details / githubweb)
  170. Bump DoozyX/clang-format-lint-action from 0.16.2 to 0.17 (details / githubweb)
  171. Bump version in the readme (details / githubweb)
  172. Restrict to_array subtest to NVCC >= 11.4.0 (details / githubweb)
  173. [ci skip] Use ROCM 6.1 in the nightly CI and disable one test (details / githubweb)
  174. Switch to using functors in sort_by_key_via_sort (#7059) (details / githubweb)
  175. Fix mdspan test (details / githubweb)
  176. Update to CUDA 12.4.1 in MSVC CI (details / githubweb)
  177. Workaround MSVC compiler issues in Views (details / githubweb)
  178. Avoid lambda in sort_by_key_via_sort (details / githubweb)
  179. Fix using CUDAToolkit for CMake 3.28.4 and higher (details / githubweb)
  180. Move check CMake 3.20.1 with nvhpc (details / githubweb)
  181. Fix closing brackets (details / githubweb)
  182. KOKKOS_CUDA_ERROR->DEFAULT_MSG (details / githubweb)
  183. Workaround icpc "missing return statement at end of non-void function" (details / githubweb)
  184. Enable MDSPAN support by default (#7069) (details / githubweb)
  185. Add SpaceAwareAccessor (details / githubweb)
  186. Add basic tests for SpaceAwareAccessor (details / githubweb)
  187. Use SpaceAwareAccessor in View mdspan-interop (details / githubweb)
  188. Add specialization of SpaceAwareAcc for AnonymousSpace (details / githubweb)
  189. Address review comments (details / githubweb)
  190. Fix incompatible dynamic_extent definition in Kokkos (details / githubweb)
  191. SpaceAwareAccessor: fix issues (no-unique-address, is_empty) (details / githubweb)
  192. Fix using shared libraries and -fvisibility=hidden (#7065) (details / githubweb)
  193. OpenACC: Skip exec_space_thread_safety_range_scan (#7022) (details / githubweb)
  194. Structured binding support for Kokkos::complex (#7040) (details / githubweb)
  195. Fix thread-safety for the Serial backend (details / githubweb)
  196. Update from GCC 13 to 14 and use C++ 26 in Jenkins nightly (details / githubweb)
  197. Move `layout_iterate_type_selector` into Impl namespace (#7076) (details / githubweb)
  198. Use correct includes for spaceawareaccessor (details / githubweb)
  199. Improve performance for deleting an instance. (details / githubweb)
  200. fix refcount exception safety (#6289) (details / githubweb)
  201. Don't delete special member functions explicitly (details / githubweb)
  202. Complex needs a tuple include (details / githubweb)
  203. Update internal desul file copies to 60c1115 (details / githubweb)
  204. Update desul version in github workflow (details / githubweb)
  205. Resolve various bogus icpc -Werror (#7079) (details / githubweb)
  206. Remove support for NVHPC as CUDA device compiler (details / githubweb)
  207. remove cmake options to change default spaces (details / githubweb)
  208. Temporarily disable failing parts in the TestMDSpan.hpp for the OpenACC (details / githubweb)
  209. Add missing space (details / githubweb)
  210. Fix more icpc issues (details / githubweb)
  211. Bump Jimver/cuda-toolkit from 0.2.15 to 0.2.16 (details / githubweb)
  212. Copy print_configuration setting in combination of kokkos settings. (details / githubweb)
  213. Couple more icpc -Werror fixes (details / githubweb)
  214. Cleanup KokkosP hooks in `Profiling::` (#7096) (details / githubweb)
  215. Remove FIXME_NVHPC 23.7 guards (details / githubweb)
  216. Fix SpaceAwareAccessor based on usage experiment in View (#7088) (details / githubweb)
  217. Added width 8 abi for avx2 (details / githubweb)
  218. Added for AVX512 (details / githubweb)
  219. Added for width 4 for NEON (details / githubweb)
  220. clang-formatted (details / githubweb)
  221. Disabling simd unit tests from building for Windows+CUDA build (details / githubweb)
  222. Workaround for the compilation failure for rocm 5.6-6.0 (details / githubweb)
  223. Added a comment about is_type structs (details / githubweb)
  224. clang formating (details / githubweb)
  225. Restrict jenkins CI not to run on hopper for nvcc < 11.8 (details / githubweb)
  226. cmake: fix relative to find kokkos_compiler_launcher (details / githubweb)
  227. Clean up KOKKOS_LIB_INCLUDE_DIRECTORIES, append include directories to associated targets in Trilinos builds (#7103) (details / githubweb)
  228. Workaround to ice with icpc when using -no-ip (#7106) (details / githubweb)
  229. move view allocation related functionality to a new header (#7110) (details / githubweb)
  230. SYCL CI: Manually build oneDPL (details / githubweb)
  231. Add missing concepts include in test (details / githubweb)
  232. Github CI: Test with C++17, C++20, and C++23 (#7082) (details / githubweb)
  233. Update base image for ROCm 5.6 (details / githubweb)
  234. Update mdspan to 98a12b01b51b2 (details / githubweb)
  235. Add AtomicAccessorRelaxed (#7089) (details / githubweb)
  236. Fix div by zero in math ops testing (details / githubweb)
  237. Fix and test with -fsanitize=undefined in GitHub CI (#7104) (details / githubweb)
  238. Bump actions/upload-artifact from 4.3.3 to 4.3.4 (details / githubweb)
  239. Update HIP nightly build base image Ubuntu 20.04 -> 22.04 (details / githubweb)
  240. Deprecate `RawMemoryAllocationFailure::FailureMode::MaximumCudaUVMAllocationsExceeded` (#7120) (details / githubweb)
  241. Refactor: Move logic of `create_mirror*` to `Impl::create_mirror*` (#7061) (details / githubweb)
  242. Define minimum compiler versions for C++20 support (details / githubweb)
  243. Drop (unused) cuda uvm allocations counter (details / githubweb)
  244. Do not bother catching memory allocation failure and rethrow (details / githubweb)
  245. Fix using and, or, xor in desul with MSVC (#7124) (details / githubweb)
  246. Do not swallow errors when deallocating memory with CUDA (details / githubweb)
  247. Let the throwing push finalize hook calls terminate test actually run (details / githubweb)
  248. Drop (unused) KOKKOS_ADD_ADVANCED_TEST TriBITS function (details / githubweb)
  249. Simplify the logic when finalizing and calling the registered functions (details / githubweb)
  250. [ci skip] Fix ROCm version to 6.1.2 in nightly CI (details / githubweb)
  251. Disable the PushFinalizeHookTerminate test on Windows (details / githubweb)
  252. OpenMP: Ensure kernels submitted by multiple threads to the same instance don't run concurrently (#6151) (details / githubweb)
  253. SYCL: Add missing include for std::stringstream (details / githubweb)
  254. SYCL: Add support for Graphs (#6912) (details / githubweb)
  255. [ci skip] rename jenkins build (details / githubweb)
  256. NVHPC: Disable failing NaN tests (details / githubweb)
  257. no_device_stack is unknown (details / githubweb)
  258. Check for LIBCXX 10 or later for C++20 and later (#7123) (details / githubweb)
  259. Get rid of `RawMemoryAllocationFailure::AllocationMechanism` and derived backend-specific exceptions (#7139) (details / githubweb)
  260. Add `likely` and `unlikely` attribute from C++20 to ref counting in views (#6730) (details / githubweb)
  261. Throw bad alloc if omp_target_alloc() returns nullptr (details / githubweb)
  262. Enable test_view_allocation_error with OpenMPTarget (details / githubweb)
  263. Fix gcc-14 C++26 nightly jenkins build (#7137) (details / githubweb)
  264. Fix Trilinos nightly failure due to `create_mirror*` refactor (#7126) (details / githubweb)
  265. fixup! Throw bad alloc if omp_target_alloc() returns nullptr (details / githubweb)
  266. Add CMake options to control compilation flags for AMD GPUs (#7127) (details / githubweb)
  267. Fix using View without corresponding mdspan-type (#7140) (details / githubweb)
Changes
  1. Reenable TestHIP_Memory_Requirements (details / githubweb)
  2. SYCL: Cleanup device selection (details / githubweb)
  3. Redeine deep_copy for UnorderMap (details / githubweb)
  4. Remove unused typedef (details / githubweb)
  5. Adding SECURITY.md file (details / githubweb)
  6. Only call deep_copy_view() from deep_copy(), add deprecation warning (details / githubweb)
  7. Cleanup KOKKOS_CONFIGURE_CORE (details / githubweb)
  8. [4.2.01]: changelog update  (#6656) (details / githubweb)
  9. Clean up shift_{right, left}_team_impl (#6821) (details / githubweb)
  10. Avoid detecting unwanted fences in the parallel_scan_no_fence test (details / githubweb)
  11. Add test for current CTAD support with RangePolicy (#6803) (details / githubweb)
  12. multi-GPU support: Add test for all policies (#6782) (details / githubweb)
  13. Implement sort_by_key (#6801) (details / githubweb)
  14. Fix scorecard workflow (#6831) (details / githubweb)
  15. Remove redundant RangePolicy constructor (#6841) (details / githubweb)
  16. Bump Google Benchmark version v1.{6.2 -> 7.1} in CMake FetchContent (details / githubweb)
  17. Remove variadic range policy constructor (#6845) (details / githubweb)
  18. Avoid unused variable warning in TestRangePolicyCTAD.cpp (details / githubweb)
  19. bytes_and_flops with CMake (details / githubweb)
  20. bytes_and_flops fix a small bug in command line argument (details / githubweb)
  21. add gather benchmark to CMake (details / githubweb)
  22. add atomic benchmark to CMake (details / githubweb)
  23. add policy_performance benchmark to CMake (details / githubweb)
  24. fix some warnings in policy_performance benchmark (details / githubweb)
  25. policy_benchmark: fix indentation (details / githubweb)
  26. Benchmarks: disable 2 benchmarks for OpenMPTarget (details / githubweb)
  27. bytes_and_flops: fix a counter name (details / githubweb)
  28. Fix C-style cast (details / githubweb)
  29. Fix amdclang++ compilation (#6857) (details / githubweb)
  30. Fix fallback implementation for sort_by_key (#6856) (details / githubweb)
  31. Update the unsafe implicit conversion error message in MDRangePolicy (#6855) (details / githubweb)
  32. Split some classes from Kokkos_ViewMapping (#6859) (details / githubweb)
  33. update mdspan tpl (details / githubweb)
  34. CTAD (deduction guides) for RangePolicy (#6850) (details / githubweb)
  35. CTAD (deduction guides) for MDRangePolicy (#5516) (details / githubweb)
  36. Add support for rocThrust in sort when using HIP (#6793) (details / githubweb)
  37. Add HIP specialization for sort-by-key (details / githubweb)
  38. Fix unused variable warning in TestMDRangePolicyCTAD.cpp (details / githubweb)
  39. Improve message on view out of bounds access and always abort (#6861) (details / githubweb)
  40. Fix linking with rothrust in downstream applications (details / githubweb)
  41. SYCL: Make sure to call find_dependency for oneDPL if necessary  (#6870) (details / githubweb)
  42. Don't use Fedora development version in GitHub CI (details / githubweb)
  43. kokkos_tpls.cmake: update default option to enable rocthrust (details / githubweb)
  44. Scratch space fix for MultiGPU (#6866) (details / githubweb)
  45. [ci skip] Bump version number to 4.3.99 (details / githubweb)
  46. Fix deep copy when filling Rank-7 views (details / githubweb)
  47. TestViewCopy_c.hpp: add new unit test for deep copy (ViewFill) (details / githubweb)
  48. TestViewCopy_c.hpp: better handling for OpenMPTarget (details / githubweb)
  49. Use builtin for atomic_fetch in the HIP backend (details / githubweb)
  50. Fix Makefile.kokkos for Threads (details / githubweb)
  51. Fix a bug in Makefile when using AMD GPU architectures (#6892) (details / githubweb)
  52. Cuda: Fix configuring with CMake 3.28.4 (#6898) (details / githubweb)
  53. Update Intel GPU architectures in Makefile (#6895) (details / githubweb)
  54. Drop Experimental::LayoutTiled class template (details / githubweb)
  55. Get rid of now unnecessary use of is_layouttiled trait (details / githubweb)
  56. Deprecate is_layouttiled trait (details / githubweb)
  57. Move `Kokkos::Array` tests to a more suitable place (#6905) (details / githubweb)
  58. Do not return a copy of the input functor for Kokkos::Experimental::for_each (details / githubweb)
  59. Drop specialization of ViewMapping for Kokkos::Array (details / githubweb)
  60. Accommodate users that depend on a code that define silly macros (#6909) (details / githubweb)
  61. SYCL: Fix multi-GPU support and add test (#6887) (details / githubweb)
  62. SYCL: Unroll shuffle loops for top-level parallel_reduce and parallel_scan (#6750) (details / githubweb)
  63. Preparing readme for develop as the default branch (#6796) (details / githubweb)
  64. CHANGELOG.md: 4.3.00 update (details / githubweb)
  65. Refactor: Uniformize `create_mirror*` parameter name for views (#6917) (details / githubweb)
  66. SYCL: Improve team_reduce implementation (#6562) (details / githubweb)
  67. Use recommended/max team size functions in Cuda ParallelFor and Reduce constructors (#6891) (details / githubweb)
  68. SYCL: Fix range in subgroup scan for workgroup_scan (details / githubweb)
  69. Add ROCm 6.0 in the nightly CI (details / githubweb)
  70. MI300 support unified memory support (#6877) (details / githubweb)
  71. cuda: reduction with `RangePolicy`: fix grid dimensions to work for large values and avoid overflow (#6578) (details / githubweb)
  72. Fix nightly CI (details / githubweb)
  73. Deprecate trailing Proxy template argument in Kokkos::Array (details / githubweb)
  74. CUDA: Update nvcc_wrapper (details / githubweb)
  75. bring back --fmad option to nvcc_wrapper (#6931) (details / githubweb)
  76. Remove unnecessary header include (details / githubweb)
  77. Adding converting constructor in Kokkos::RandomAccessIterator (#6929) (details / githubweb)
  78. Prefer standard C++ feature testing to guard the C++20 requires expression (details / githubweb)
  79. Add `kokkos_swap(Array<T, N>)` sepcialization (details / githubweb)
  80. Deprecate specialization of Kokkos::pair for a single element (details / githubweb)
  81. Fix deprecated warning from `Kokkos::Array` specialization (#6945) (details / githubweb)
  82. [ci skip] Enable deprecated code and deprecated warnings in nightly CI (details / githubweb)
  83. Serial: Use the provided execution space instance in TeamPolicy (details / githubweb)
  84. Fixed the link for P6601 (Threads backend change) (details / githubweb)
  85. Add test (details / githubweb)
  86. Use Kokkos::AUTO for OpenMPTarget (details / githubweb)
  87. kokkos_swap(Array) member friend should not be templated on some other type U (details / githubweb)
  88. Fix noexcept specification for kokkos_swap on zero-sized arrays (details / githubweb)
  89. Replace deprecated sycl::device_ptr/sycl::host_ptr (details / githubweb)
  90. Introduce alias based on feature macro (details / githubweb)
  91. Kokkos::Impl::SYCLTypes:: -> Kokkos::Impl::sycl_ (details / githubweb)
  92. Fix wrong macro guards for deprecated Kokkos::pair<T1,void> specialization (details / githubweb)
  93. Fix support for Kokkos::Array of const-qualified element type (details / githubweb)
  94. Try to fix the CUDA 11.0 build (details / githubweb)
  95. Revert "Try to fix the CUDA 11.0 build" (details / githubweb)
  96. Alternate definition of Impl::is_nothrow_swappable_v for NVCC version less than 11.4 (details / githubweb)
  97. Avoid introducing a 2nd definition of the Impl::swappable trait (details / githubweb)
  98. Do not require OpenMP support for languages other than CXX (details / githubweb)
  99. Update version (details / githubweb)
  100. Add Linux Foundation notice and fix C++ standard (details / githubweb)
  101. SYCL: Avoid deprecated floating-point number abs overloads (#6959) (details / githubweb)
  102. OpenMPTarget: Update loop order in MDRange (#6925) (details / githubweb)
  103. Threads: Don't silently allow m_instance to be a nullptr (#6969) (details / githubweb)
  104. Fix enabling OpenMP with HIP and "compile as CMake language" (details / githubweb)
  105. Fix OpenMP+CUDA when `Kokkos_ENABLE_COMPILE_AS_CMAKE_LANGUAGE` is `ON` (details / githubweb)
  106. Fix fedora CI builds with flang-new (details / githubweb)
  107. Fix Copyright file (details / githubweb)
  108. Also use is_nothrow_swappable workaround for Intel Classic Compilers (#6983) (details / githubweb)
  109. Add thread-safety tests (#6938) (details / githubweb)
  110. Fix deprecation warnings with GCC for pair<T1,void> comparison operators (details / githubweb)
  111. Fix TPL_LIBRARY_SUFFIXES for 32-bit build (details / githubweb)
  112. Avoid duplicated definition of KOKKOS_IMPL_32BIT (details / githubweb)
  113. Disable failing parallel_scan_with_reducers test (details / githubweb)
  114. Fix a bug when using realloc on views of non-default constructible element types (#6993) (details / githubweb)
  115. SYCL: Prepare Parallel* for Graphs (#6988) (details / githubweb)
  116. SYCL: Fix deprecation in custom parallel_for RangePolicy implementation (details / githubweb)
  117. [ci skip] update changelog for 4.3.1 (#6995) (details / githubweb)
  118. OpenMPTarget: Use mutex lock for parallel scan. (details / githubweb)
  119. Remove cuSPARSE TPL (details / githubweb)
  120. SYCL: Print submission command queue property (#7004) (details / githubweb)
  121. Test DualView resize/realloc for types without default constructor (details / githubweb)
  122. Suppress deprecated warnings via pragma push/pop in the tests (#6999) (details / githubweb)
  123. This PR changes the default execution behavior of the parallel_for(team-policy) constructs in the OpenACC backend. (details / githubweb)
  124. Use parallel_for instead of parallel_reduce for check (details / githubweb)
  125. Fix unique_any_senders nvcc template deduction (details / githubweb)
  126. Properly guard deprecated `Kokkos_Vector.hpp` header self contained test (#7016) (details / githubweb)
  127. mdspan converting constructors (#6830) (details / githubweb)
  128. SYCL: Don't use shuffles for top-level reductions (#7009) (details / githubweb)
  129. Add a CITATION.cff file (#7008) (details / githubweb)
  130. Always run Graph tests (#7011) (details / githubweb)
  131. Fix undefined behavior in is_zero_byte (#7014) (details / githubweb)
  132. Disable OpenMPTarget Kokkos::Graph test (does not compile) (details / githubweb)
  133. Fixup for disabling deprecation warnings with NVC++ (details / githubweb)
  134. Remove unused CudaInternal::cuda_{malloc,free}_async_wrapper (details / githubweb)
  135. Improve `Impl::is_zero_byte()` (#7017) (details / githubweb)
  136. OpenMPTarget: Fix compiling Graph tests (#7020) (details / githubweb)
  137. SYCL: Skip launch_six Graph test (details / githubweb)
  138. HIP: Use builtin atomic for compare_exchange (#7000) (details / githubweb)
  139. Fix Kokkos README's FENL link (details / githubweb)
  140. Copied the deduction guides and test cases over from branch (details / githubweb)
  141. Refactor: Replace SFINAE by `if constexpr` for `create_mirror*` functions (#6955) (details / githubweb)
  142. Workaround icpc warnings (details / githubweb)
  143. Adding `Kokkos::to_array` (#6375) (details / githubweb)
  144. Fix adjacent difference (#6922) (details / githubweb)
  145. TestArray: add intel guard to to_array implicit conversion test (details / githubweb)
  146. Avoid using "#if not defined" (details / githubweb)
  147. (Rebase) Partial fix to compile time issues w/nvcc + Kokkos_ENABLE_DEBUG_BOUNDS_CHECK (#7013) (details / githubweb)
  148. Add dependabot for GitHub Actions (details / githubweb)
  149. Added static_asserts for out of range tuple_element and get (details / githubweb)
  150. Add tests to `ScopeGuard` (#7028) (details / githubweb)
  151. Improve Kokkos Graphs (#7039) (details / githubweb)
  152. Introduce `KOKKOS_DEDUCTION_GUIDE` macro to allow user-defined deduction guide in device code for clang compiler (#6954) (details / githubweb)
  153. Add nightly CI on Frontier (#7048) (details / githubweb)
  154. Update scorecard GitHub workflow (details / githubweb)
  155. Fix unused symbols in CTAD tests (details / githubweb)
  156. Cuda: Fix nvcc warnings (details / githubweb)
  157. Fix quotation marks in CXX flags (details / githubweb)
  158. Fix kokkos_swap (details / githubweb)
  159. Fix array size (details / githubweb)
  160. Fix gtest (details / githubweb)
  161. Fix .jenkins whitespce (details / githubweb)
  162. Only use -Werror all-warnings with explicit nvcc_wrapper (details / githubweb)
  163. Fix dangling reference (details / githubweb)
  164. Suppress 'long double' is treated as 'double' in device code (details / githubweb)
  165. Use -Xcudafe --diag_suppress=20208 for 11.6 build; nothing else seems to help (details / githubweb)
  166. Try moving pragma suppress to tests (details / githubweb)
  167. Use -Xcudafe --diag_suppress=20208 in Makefile build (details / githubweb)
  168. Move dependabot to .github (details / githubweb)
  169. Bump Jimver/cuda-toolkit from 0.2.14 to 0.2.15 (details / githubweb)
  170. Bump DoozyX/clang-format-lint-action from 0.16.2 to 0.17 (details / githubweb)
  171. Bump version in the readme (details / githubweb)
  172. Restrict to_array subtest to NVCC >= 11.4.0 (details / githubweb)
  173. [ci skip] Use ROCM 6.1 in the nightly CI and disable one test (details / githubweb)
  174. Switch to using functors in sort_by_key_via_sort (#7059) (details / githubweb)
  175. Fix mdspan test (details / githubweb)
  176. Update to CUDA 12.4.1 in MSVC CI (details / githubweb)
  177. Workaround MSVC compiler issues in Views (details / githubweb)
  178. Avoid lambda in sort_by_key_via_sort (details / githubweb)
  179. Fix using CUDAToolkit for CMake 3.28.4 and higher (details / githubweb)
  180. Move check CMake 3.20.1 with nvhpc (details / githubweb)
  181. Fix closing brackets (details / githubweb)
  182. KOKKOS_CUDA_ERROR->DEFAULT_MSG (details / githubweb)
  183. Workaround icpc "missing return statement at end of non-void function" (details / githubweb)
  184. Enable MDSPAN support by default (#7069) (details / githubweb)
  185. Add SpaceAwareAccessor (details / githubweb)
  186. Add basic tests for SpaceAwareAccessor (details / githubweb)
  187. Use SpaceAwareAccessor in View mdspan-interop (details / githubweb)
  188. Add specialization of SpaceAwareAcc for AnonymousSpace (details / githubweb)
  189. Address review comments (details / githubweb)
  190. Fix incompatible dynamic_extent definition in Kokkos (details / githubweb)
  191. SpaceAwareAccessor: fix issues (no-unique-address, is_empty) (details / githubweb)
  192. Fix using shared libraries and -fvisibility=hidden (#7065) (details / githubweb)
  193. OpenACC: Skip exec_space_thread_safety_range_scan (#7022) (details / githubweb)
  194. Structured binding support for Kokkos::complex (#7040) (details / githubweb)
  195. Fix thread-safety for the Serial backend (details / githubweb)
  196. Update from GCC 13 to 14 and use C++ 26 in Jenkins nightly (details / githubweb)
  197. Move `layout_iterate_type_selector` into Impl namespace (#7076) (details / githubweb)
  198. Use correct includes for spaceawareaccessor (details / githubweb)
  199. Improve performance for deleting an instance. (details / githubweb)
  200. fix refcount exception safety (#6289) (details / githubweb)
  201. Don't delete special member functions explicitly (details / githubweb)
  202. Complex needs a tuple include (details / githubweb)
  203. Update internal desul file copies to 60c1115 (details / githubweb)
  204. Update desul version in github workflow (details / githubweb)
  205. Resolve various bogus icpc -Werror (#7079) (details / githubweb)
  206. Remove support for NVHPC as CUDA device compiler (details / githubweb)
  207. remove cmake options to change default spaces (details / githubweb)
  208. Temporarily disable failing parts in the TestMDSpan.hpp for the OpenACC (details / githubweb)
  209. Add missing space (details / githubweb)
  210. Fix more icpc issues (details / githubweb)
  211. Bump Jimver/cuda-toolkit from 0.2.15 to 0.2.16 (details / githubweb)
  212. Copy print_configuration setting in combination of kokkos settings. (details / githubweb)
  213. Couple more icpc -Werror fixes (details / githubweb)
  214. Cleanup KokkosP hooks in `Profiling::` (#7096) (details / githubweb)
  215. Remove FIXME_NVHPC 23.7 guards (details / githubweb)
  216. Fix SpaceAwareAccessor based on usage experiment in View (#7088) (details / githubweb)
  217. Added width 8 abi for avx2 (details / githubweb)
  218. Added for AVX512 (details / githubweb)
  219. Added for width 4 for NEON (details / githubweb)
  220. clang-formatted (details / githubweb)
  221. Disabling simd unit tests from building for Windows+CUDA build (details / githubweb)
  222. Workaround for the compilation failure for rocm 5.6-6.0 (details / githubweb)
  223. Added a comment about is_type structs (details / githubweb)
  224. clang formating (details / githubweb)
  225. Restrict jenkins CI not to run on hopper for nvcc < 11.8 (details / githubweb)
  226. cmake: fix relative to find kokkos_compiler_launcher (details / githubweb)
  227. Clean up KOKKOS_LIB_INCLUDE_DIRECTORIES, append include directories to associated targets in Trilinos builds (#7103) (details / githubweb)
  228. Workaround to ice with icpc when using -no-ip (#7106) (details / githubweb)
  229. move view allocation related functionality to a new header (#7110) (details / githubweb)
  230. SYCL CI: Manually build oneDPL (details / githubweb)
  231. Add missing concepts include in test (details / githubweb)
  232. Github CI: Test with C++17, C++20, and C++23 (#7082) (details / githubweb)
  233. Update base image for ROCm 5.6 (details / githubweb)
  234. Update mdspan to 98a12b01b51b2 (details / githubweb)
  235. Add AtomicAccessorRelaxed (#7089) (details / githubweb)
  236. Fix div by zero in math ops testing (details / githubweb)
  237. Fix and test with -fsanitize=undefined in GitHub CI (#7104) (details / githubweb)
  238. Bump actions/upload-artifact from 4.3.3 to 4.3.4 (details / githubweb)
  239. Update HIP nightly build base image Ubuntu 20.04 -> 22.04 (details / githubweb)
  240. Deprecate `RawMemoryAllocationFailure::FailureMode::MaximumCudaUVMAllocationsExceeded` (#7120) (details / githubweb)
  241. Refactor: Move logic of `create_mirror*` to `Impl::create_mirror*` (#7061) (details / githubweb)
  242. Define minimum compiler versions for C++20 support (details / githubweb)
  243. Drop (unused) cuda uvm allocations counter (details / githubweb)
  244. Do not bother catching memory allocation failure and rethrow (details / githubweb)
  245. Fix using and, or, xor in desul with MSVC (#7124) (details / githubweb)
  246. Do not swallow errors when deallocating memory with CUDA (details / githubweb)
  247. Let the throwing push finalize hook calls terminate test actually run (details / githubweb)
  248. Drop (unused) KOKKOS_ADD_ADVANCED_TEST TriBITS function (details / githubweb)
  249. Simplify the logic when finalizing and calling the registered functions (details / githubweb)
  250. [ci skip] Fix ROCm version to 6.1.2 in nightly CI (details / githubweb)
  251. Disable the PushFinalizeHookTerminate test on Windows (details / githubweb)
  252. OpenMP: Ensure kernels submitted by multiple threads to the same instance don't run concurrently (#6151) (details / githubweb)
  253. SYCL: Add missing include for std::stringstream (details / githubweb)
  254. SYCL: Add support for Graphs (#6912) (details / githubweb)
  255. [ci skip] rename jenkins build (details / githubweb)
  256. NVHPC: Disable failing NaN tests (details / githubweb)
  257. no_device_stack is unknown (details / githubweb)
  258. Check for LIBCXX 10 or later for C++20 and later (#7123) (details / githubweb)
  259. Get rid of `RawMemoryAllocationFailure::AllocationMechanism` and derived backend-specific exceptions (#7139) (details / githubweb)
  260. Add `likely` and `unlikely` attribute from C++20 to ref counting in views (#6730) (details / githubweb)
  261. Throw bad alloc if omp_target_alloc() returns nullptr (details / githubweb)
  262. Enable test_view_allocation_error with OpenMPTarget (details / githubweb)
  263. Fix gcc-14 C++26 nightly jenkins build (#7137) (details / githubweb)
  264. Fix Trilinos nightly failure due to `create_mirror*` refactor (#7126) (details / githubweb)
  265. fixup! Throw bad alloc if omp_target_alloc() returns nullptr (details / githubweb)
  266. Add CMake options to control compilation flags for AMD GPUs (#7127) (details / githubweb)
  267. Fix using View without corresponding mdspan-type (#7140) (details / githubweb)
Changes
  1. Reenable TestHIP_Memory_Requirements (details / githubweb)
  2. SYCL: Cleanup device selection (details / githubweb)
  3. Redeine deep_copy for UnorderMap (details / githubweb)
  4. Remove unused typedef (details / githubweb)
  5. Adding SECURITY.md file (details / githubweb)
  6. Only call deep_copy_view() from deep_copy(), add deprecation warning (details / githubweb)
  7. Cleanup KOKKOS_CONFIGURE_CORE (details / githubweb)
  8. [4.2.01]: changelog update  (#6656) (details / githubweb)
  9. Clean up shift_{right, left}_team_impl (#6821) (details / githubweb)
  10. Avoid detecting unwanted fences in the parallel_scan_no_fence test (details / githubweb)
  11. Add test for current CTAD support with RangePolicy (#6803) (details / githubweb)
  12. multi-GPU support: Add test for all policies (#6782) (details / githubweb)
  13. Implement sort_by_key (#6801) (details / githubweb)
  14. Fix scorecard workflow (#6831) (details / githubweb)
  15. Remove redundant RangePolicy constructor (#6841) (details / githubweb)
  16. Bump Google Benchmark version v1.{6.2 -> 7.1} in CMake FetchContent (details / githubweb)
  17. Remove variadic range policy constructor (#6845) (details / githubweb)
  18. Avoid unused variable warning in TestRangePolicyCTAD.cpp (details / githubweb)
  19. bytes_and_flops with CMake (details / githubweb)
  20. bytes_and_flops fix a small bug in command line argument (details / githubweb)
  21. add gather benchmark to CMake (details / githubweb)
  22. add atomic benchmark to CMake (details / githubweb)
  23. add policy_performance benchmark to CMake (details / githubweb)
  24. fix some warnings in policy_performance benchmark (details / githubweb)
  25. policy_benchmark: fix indentation (details / githubweb)
  26. Benchmarks: disable 2 benchmarks for OpenMPTarget (details / githubweb)
  27. bytes_and_flops: fix a counter name (details / githubweb)
  28. Fix C-style cast (details / githubweb)
  29. Fix amdclang++ compilation (#6857) (details / githubweb)
  30. Fix fallback implementation for sort_by_key (#6856) (details / githubweb)
  31. Update the unsafe implicit conversion error message in MDRangePolicy (#6855) (details / githubweb)
  32. Split some classes from Kokkos_ViewMapping (#6859) (details / githubweb)
  33. update mdspan tpl (details / githubweb)
  34. CTAD (deduction guides) for RangePolicy (#6850) (details / githubweb)
  35. CTAD (deduction guides) for MDRangePolicy (#5516) (details / githubweb)
  36. Add support for rocThrust in sort when using HIP (#6793) (details / githubweb)
  37. Add HIP specialization for sort-by-key (details / githubweb)
  38. Fix unused variable warning in TestMDRangePolicyCTAD.cpp (details / githubweb)
  39. Improve message on view out of bounds access and always abort (#6861) (details / githubweb)
  40. Fix linking with rothrust in downstream applications (details / githubweb)
  41. SYCL: Make sure to call find_dependency for oneDPL if necessary  (#6870) (details / githubweb)
  42. Don't use Fedora development version in GitHub CI (details / githubweb)
  43. kokkos_tpls.cmake: update default option to enable rocthrust (details / githubweb)
  44. Scratch space fix for MultiGPU (#6866) (details / githubweb)
  45. [ci skip] Bump version number to 4.3.99 (details / githubweb)
  46. Fix deep copy when filling Rank-7 views (details / githubweb)
  47. TestViewCopy_c.hpp: add new unit test for deep copy (ViewFill) (details / githubweb)
  48. TestViewCopy_c.hpp: better handling for OpenMPTarget (details / githubweb)
  49. Use builtin for atomic_fetch in the HIP backend (details / githubweb)
  50. Fix Makefile.kokkos for Threads (details / githubweb)
  51. Fix a bug in Makefile when using AMD GPU architectures (#6892) (details / githubweb)
  52. Cuda: Fix configuring with CMake 3.28.4 (#6898) (details / githubweb)
  53. Update Intel GPU architectures in Makefile (#6895) (details / githubweb)
  54. Drop Experimental::LayoutTiled class template (details / githubweb)
  55. Get rid of now unnecessary use of is_layouttiled trait (details / githubweb)
  56. Deprecate is_layouttiled trait (details / githubweb)
  57. Move `Kokkos::Array` tests to a more suitable place (#6905) (details / githubweb)
  58. Do not return a copy of the input functor for Kokkos::Experimental::for_each (details / githubweb)
  59. Drop specialization of ViewMapping for Kokkos::Array (details / githubweb)
  60. Accommodate users that depend on a code that define silly macros (#6909) (details / githubweb)
  61. SYCL: Fix multi-GPU support and add test (#6887) (details / githubweb)
  62. SYCL: Unroll shuffle loops for top-level parallel_reduce and parallel_scan (#6750) (details / githubweb)
  63. Preparing readme for develop as the default branch (#6796) (details / githubweb)
  64. CHANGELOG.md: 4.3.00 update (details / githubweb)
  65. Refactor: Uniformize `create_mirror*` parameter name for views (#6917) (details / githubweb)
  66. SYCL: Improve team_reduce implementation (#6562) (details / githubweb)
  67. Use recommended/max team size functions in Cuda ParallelFor and Reduce constructors (#6891) (details / githubweb)
  68. SYCL: Fix range in subgroup scan for workgroup_scan (details / githubweb)
  69. Add ROCm 6.0 in the nightly CI (details / githubweb)
  70. MI300 support unified memory support (#6877) (details / githubweb)
  71. cuda: reduction with `RangePolicy`: fix grid dimensions to work for large values and avoid overflow (#6578) (details / githubweb)
  72. Fix nightly CI (details / githubweb)
  73. Deprecate trailing Proxy template argument in Kokkos::Array (details / githubweb)
  74. CUDA: Update nvcc_wrapper (details / githubweb)
  75. bring back --fmad option to nvcc_wrapper (#6931) (details / githubweb)
  76. Remove unnecessary header include (details / githubweb)
  77. Adding converting constructor in Kokkos::RandomAccessIterator (#6929) (details / githubweb)
  78. Prefer standard C++ feature testing to guard the C++20 requires expression (details / githubweb)
  79. Add `kokkos_swap(Array<T, N>)` sepcialization (details / githubweb)
  80. Deprecate specialization of Kokkos::pair for a single element (details / githubweb)
  81. Fix deprecated warning from `Kokkos::Array` specialization (#6945) (details / githubweb)
  82. [ci skip] Enable deprecated code and deprecated warnings in nightly CI (details / githubweb)
  83. Serial: Use the provided execution space instance in TeamPolicy (details / githubweb)
  84. Fixed the link for P6601 (Threads backend change) (details / githubweb)
  85. Add test (details / githubweb)
  86. Use Kokkos::AUTO for OpenMPTarget (details / githubweb)
  87. kokkos_swap(Array) member friend should not be templated on some other type U (details / githubweb)
  88. Fix noexcept specification for kokkos_swap on zero-sized arrays (details / githubweb)
  89. Replace deprecated sycl::device_ptr/sycl::host_ptr (details / githubweb)
  90. Introduce alias based on feature macro (details / githubweb)
  91. Kokkos::Impl::SYCLTypes:: -> Kokkos::Impl::sycl_ (details / githubweb)
  92. Fix wrong macro guards for deprecated Kokkos::pair<T1,void> specialization (details / githubweb)
  93. Fix support for Kokkos::Array of const-qualified element type (details / githubweb)
  94. Try to fix the CUDA 11.0 build (details / githubweb)
  95. Revert "Try to fix the CUDA 11.0 build" (details / githubweb)
  96. Alternate definition of Impl::is_nothrow_swappable_v for NVCC version less than 11.4 (details / githubweb)
  97. Avoid introducing a 2nd definition of the Impl::swappable trait (details / githubweb)
  98. Do not require OpenMP support for languages other than CXX (details / githubweb)
  99. Update version (details / githubweb)
  100. Add Linux Foundation notice and fix C++ standard (details / githubweb)
  101. SYCL: Avoid deprecated floating-point number abs overloads (#6959) (details / githubweb)
  102. OpenMPTarget: Update loop order in MDRange (#6925) (details / githubweb)
  103. Threads: Don't silently allow m_instance to be a nullptr (#6969) (details / githubweb)
  104. Fix enabling OpenMP with HIP and "compile as CMake language" (details / githubweb)
  105. Fix OpenMP+CUDA when `Kokkos_ENABLE_COMPILE_AS_CMAKE_LANGUAGE` is `ON` (details / githubweb)
  106. Fix fedora CI builds with flang-new (details / githubweb)
  107. Fix Copyright file (details / githubweb)
  108. Also use is_nothrow_swappable workaround for Intel Classic Compilers (#6983) (details / githubweb)
  109. Add thread-safety tests (#6938) (details / githubweb)
  110. Fix deprecation warnings with GCC for pair<T1,void> comparison operators (details / githubweb)
  111. Fix TPL_LIBRARY_SUFFIXES for 32-bit build (details / githubweb)
  112. Avoid duplicated definition of KOKKOS_IMPL_32BIT (details / githubweb)
  113. Disable failing parallel_scan_with_reducers test (details / githubweb)
  114. Fix a bug when using realloc on views of non-default constructible element types (#6993) (details / githubweb)
  115. SYCL: Prepare Parallel* for Graphs (#6988) (details / githubweb)
  116. SYCL: Fix deprecation in custom parallel_for RangePolicy implementation (details / githubweb)
  117. [ci skip] update changelog for 4.3.1 (#6995) (details / githubweb)
  118. OpenMPTarget: Use mutex lock for parallel scan. (details / githubweb)
  119. Remove cuSPARSE TPL (details / githubweb)
  120. SYCL: Print submission command queue property (#7004) (details / githubweb)
  121. Test DualView resize/realloc for types without default constructor (details / githubweb)
  122. Suppress deprecated warnings via pragma push/pop in the tests (#6999) (details / githubweb)
  123. This PR changes the default execution behavior of the parallel_for(team-policy) constructs in the OpenACC backend. (details / githubweb)
  124. Use parallel_for instead of parallel_reduce for check (details / githubweb)
  125. Fix unique_any_senders nvcc template deduction (details / githubweb)
  126. Properly guard deprecated `Kokkos_Vector.hpp` header self contained test (#7016) (details / githubweb)
  127. mdspan converting constructors (#6830) (details / githubweb)
  128. SYCL: Don't use shuffles for top-level reductions (#7009) (details / githubweb)
  129. Add a CITATION.cff file (#7008) (details / githubweb)
  130. Always run Graph tests (#7011) (details / githubweb)
  131. Fix undefined behavior in is_zero_byte (#7014) (details / githubweb)
  132. Disable OpenMPTarget Kokkos::Graph test (does not compile) (details / githubweb)
  133. Fixup for disabling deprecation warnings with NVC++ (details / githubweb)
  134. Remove unused CudaInternal::cuda_{malloc,free}_async_wrapper (details / githubweb)
  135. Improve `Impl::is_zero_byte()` (#7017) (details / githubweb)
  136. OpenMPTarget: Fix compiling Graph tests (#7020) (details / githubweb)
  137. SYCL: Skip launch_six Graph test (details / githubweb)
  138. HIP: Use builtin atomic for compare_exchange (#7000) (details / githubweb)
  139. Fix Kokkos README's FENL link (details / githubweb)
  140. Copied the deduction guides and test cases over from branch (details / githubweb)
  141. Refactor: Replace SFINAE by `if constexpr` for `create_mirror*` functions (#6955) (details / githubweb)
  142. Workaround icpc warnings (details / githubweb)
  143. Adding `Kokkos::to_array` (#6375) (details / githubweb)
  144. Fix adjacent difference (#6922) (details / githubweb)
  145. TestArray: add intel guard to to_array implicit conversion test (details / githubweb)
  146. Avoid using "#if not defined" (details / githubweb)
  147. (Rebase) Partial fix to compile time issues w/nvcc + Kokkos_ENABLE_DEBUG_BOUNDS_CHECK (#7013) (details / githubweb)
  148. Add dependabot for GitHub Actions (details / githubweb)
  149. Added static_asserts for out of range tuple_element and get (details / githubweb)
  150. Add tests to `ScopeGuard` (#7028) (details / githubweb)
  151. Improve Kokkos Graphs (#7039) (details / githubweb)
  152. Introduce `KOKKOS_DEDUCTION_GUIDE` macro to allow user-defined deduction guide in device code for clang compiler (#6954) (details / githubweb)
  153. Add nightly CI on Frontier (#7048) (details / githubweb)
  154. Update scorecard GitHub workflow (details / githubweb)
  155. Fix unused symbols in CTAD tests (details / githubweb)
  156. Cuda: Fix nvcc warnings (details / githubweb)
  157. Fix quotation marks in CXX flags (details / githubweb)
  158. Fix kokkos_swap (details / githubweb)
  159. Fix array size (details / githubweb)
  160. Fix gtest (details / githubweb)
  161. Fix .jenkins whitespce (details / githubweb)
  162. Only use -Werror all-warnings with explicit nvcc_wrapper (details / githubweb)
  163. Fix dangling reference (details / githubweb)
  164. Suppress 'long double' is treated as 'double' in device code (details / githubweb)
  165. Use -Xcudafe --diag_suppress=20208 for 11.6 build; nothing else seems to help (details / githubweb)
  166. Try moving pragma suppress to tests (details / githubweb)
  167. Use -Xcudafe --diag_suppress=20208 in Makefile build (details / githubweb)
  168. Move dependabot to .github (details / githubweb)
  169. Bump Jimver/cuda-toolkit from 0.2.14 to 0.2.15 (details / githubweb)
  170. Bump DoozyX/clang-format-lint-action from 0.16.2 to 0.17 (details / githubweb)
  171. Bump version in the readme (details / githubweb)
  172. Restrict to_array subtest to NVCC >= 11.4.0 (details / githubweb)
  173. [ci skip] Use ROCM 6.1 in the nightly CI and disable one test (details / githubweb)
  174. Switch to using functors in sort_by_key_via_sort (#7059) (details / githubweb)
  175. Fix mdspan test (details / githubweb)
  176. Update to CUDA 12.4.1 in MSVC CI (details / githubweb)
  177. Workaround MSVC compiler issues in Views (details / githubweb)
  178. Avoid lambda in sort_by_key_via_sort (details / githubweb)
  179. Fix using CUDAToolkit for CMake 3.28.4 and higher (details / githubweb)
  180. Move check CMake 3.20.1 with nvhpc (details / githubweb)
  181. Fix closing brackets (details / githubweb)
  182. KOKKOS_CUDA_ERROR->DEFAULT_MSG (details / githubweb)
  183. Workaround icpc "missing return statement at end of non-void function" (details / githubweb)
  184. Enable MDSPAN support by default (#7069) (details / githubweb)
  185. Add SpaceAwareAccessor (details / githubweb)
  186. Add basic tests for SpaceAwareAccessor (details / githubweb)
  187. Use SpaceAwareAccessor in View mdspan-interop (details / githubweb)
  188. Add specialization of SpaceAwareAcc for AnonymousSpace (details / githubweb)
  189. Address review comments (details / githubweb)
  190. Fix incompatible dynamic_extent definition in Kokkos (details / githubweb)
  191. SpaceAwareAccessor: fix issues (no-unique-address, is_empty) (details / githubweb)
  192. Fix using shared libraries and -fvisibility=hidden (#7065) (details / githubweb)
  193. OpenACC: Skip exec_space_thread_safety_range_scan (#7022) (details / githubweb)
  194. Structured binding support for Kokkos::complex (#7040) (details / githubweb)
  195. Fix thread-safety for the Serial backend (details / githubweb)
  196. Update from GCC 13 to 14 and use C++ 26 in Jenkins nightly (details / githubweb)
  197. Move `layout_iterate_type_selector` into Impl namespace (#7076) (details / githubweb)
  198. Use correct includes for spaceawareaccessor (details / githubweb)
  199. Improve performance for deleting an instance. (details / githubweb)
  200. fix refcount exception safety (#6289) (details / githubweb)
  201. Don't delete special member functions explicitly (details / githubweb)
  202. Complex needs a tuple include (details / githubweb)
  203. Update internal desul file copies to 60c1115 (details / githubweb)
  204. Update desul version in github workflow (details / githubweb)
  205. Resolve various bogus icpc -Werror (#7079) (details / githubweb)
  206. Remove support for NVHPC as CUDA device compiler (details / githubweb)
  207. remove cmake options to change default spaces (details / githubweb)
  208. Temporarily disable failing parts in the TestMDSpan.hpp for the OpenACC (details / githubweb)
  209. Add missing space (details / githubweb)
  210. Fix more icpc issues (details / githubweb)
  211. Bump Jimver/cuda-toolkit from 0.2.15 to 0.2.16 (details / githubweb)
  212. Copy print_configuration setting in combination of kokkos settings. (details / githubweb)
  213. Couple more icpc -Werror fixes (details / githubweb)
  214. Cleanup KokkosP hooks in `Profiling::` (#7096) (details / githubweb)
  215. Remove FIXME_NVHPC 23.7 guards (details / githubweb)
  216. Fix SpaceAwareAccessor based on usage experiment in View (#7088) (details / githubweb)
  217. Added width 8 abi for avx2 (details / githubweb)
  218. Added for AVX512 (details / githubweb)
  219. Added for width 4 for NEON (details / githubweb)
  220. clang-formatted (details / githubweb)
  221. Disabling simd unit tests from building for Windows+CUDA build (details / githubweb)
  222. Workaround for the compilation failure for rocm 5.6-6.0 (details / githubweb)
  223. Added a comment about is_type structs (details / githubweb)
  224. clang formating (details / githubweb)
  225. Restrict jenkins CI not to run on hopper for nvcc < 11.8 (details / githubweb)
  226. cmake: fix relative to find kokkos_compiler_launcher (details / githubweb)
  227. Clean up KOKKOS_LIB_INCLUDE_DIRECTORIES, append include directories to associated targets in Trilinos builds (#7103) (details / githubweb)
  228. Workaround to ice with icpc when using -no-ip (#7106) (details / githubweb)
  229. move view allocation related functionality to a new header (#7110) (details / githubweb)
  230. SYCL CI: Manually build oneDPL (details / githubweb)
  231. Add missing concepts include in test (details / githubweb)
  232. Github CI: Test with C++17, C++20, and C++23 (#7082) (details / githubweb)
  233. Update base image for ROCm 5.6 (details / githubweb)
  234. Update mdspan to 98a12b01b51b2 (details / githubweb)
  235. Add AtomicAccessorRelaxed (#7089) (details / githubweb)
  236. Fix div by zero in math ops testing (details / githubweb)
  237. Fix and test with -fsanitize=undefined in GitHub CI (#7104) (details / githubweb)
  238. Bump actions/upload-artifact from 4.3.3 to 4.3.4 (details / githubweb)
  239. Update HIP nightly build base image Ubuntu 20.04 -> 22.04 (details / githubweb)
  240. Deprecate `RawMemoryAllocationFailure::FailureMode::MaximumCudaUVMAllocationsExceeded` (#7120) (details / githubweb)
  241. Refactor: Move logic of `create_mirror*` to `Impl::create_mirror*` (#7061) (details / githubweb)
  242. Define minimum compiler versions for C++20 support (details / githubweb)
  243. Drop (unused) cuda uvm allocations counter (details / githubweb)
  244. Do not bother catching memory allocation failure and rethrow (details / githubweb)
  245. Fix using and, or, xor in desul with MSVC (#7124) (details / githubweb)
  246. Do not swallow errors when deallocating memory with CUDA (details / githubweb)
  247. Let the throwing push finalize hook calls terminate test actually run (details / githubweb)
  248. Drop (unused) KOKKOS_ADD_ADVANCED_TEST TriBITS function (details / githubweb)
  249. Simplify the logic when finalizing and calling the registered functions (details / githubweb)
  250. [ci skip] Fix ROCm version to 6.1.2 in nightly CI (details / githubweb)
  251. Disable the PushFinalizeHookTerminate test on Windows (details / githubweb)
  252. OpenMP: Ensure kernels submitted by multiple threads to the same instance don't run concurrently (#6151) (details / githubweb)
  253. SYCL: Add missing include for std::stringstream (details / githubweb)
  254. SYCL: Add support for Graphs (#6912) (details / githubweb)
  255. [ci skip] rename jenkins build (details / githubweb)
  256. NVHPC: Disable failing NaN tests (details / githubweb)
  257. no_device_stack is unknown (details / githubweb)
  258. Check for LIBCXX 10 or later for C++20 and later (#7123) (details / githubweb)
  259. Get rid of `RawMemoryAllocationFailure::AllocationMechanism` and derived backend-specific exceptions (#7139) (details / githubweb)
  260. Add `likely` and `unlikely` attribute from C++20 to ref counting in views (#6730) (details / githubweb)
  261. Throw bad alloc if omp_target_alloc() returns nullptr (details / githubweb)
  262. Enable test_view_allocation_error with OpenMPTarget (details / githubweb)
  263. Fix gcc-14 C++26 nightly jenkins build (#7137) (details / githubweb)
  264. Fix Trilinos nightly failure due to `create_mirror*` refactor (#7126) (details / githubweb)
  265. fixup! Throw bad alloc if omp_target_alloc() returns nullptr (details / githubweb)
  266. Add CMake options to control compilation flags for AMD GPUs (#7127) (details / githubweb)
  267. Fix using View without corresponding mdspan-type (#7140) (details / githubweb)

Pull request #6828 event at 9:55:49 AM on Jul 19, 2024

This run spent:

  • 3 hr 17 min waiting;
  • 2 hr 0 min build duration;
  • 2 hr 0 min total from scheduled to completion.
Revision: c847e6c9923daf1496e0e0bc36b02de8bbb29c5f
Repository: https://github.com/kokkos/kokkos.git
  • PR-6828
Revision: b23055b52e46b46ae29e4a84aeaecbcb3f21bce5
Repository: https://github.com/kokkos/kokkos.git
  • PR-6828
Revision: 894328011128462d781f9478b1411cc3d031ba21
Repository: https://github.com/kokkos/kokkos.git
  • PR-6828
Revision: 862f17f9ed27455a2933d89d680f50f695dbae7f
Repository: https://github.com/kokkos/kokkos.git
  • PR-6828
Revision: 1ff46c5aebf96fa8eb79ea1bbae27811151f0505
Repository: https://github.com/kokkos/kokkos.git
  • PR-6828
Revision: c344d3a4b992e293f700a9d31343adf25077db9e
Repository: https://github.com/kokkos/kokkos.git
  • PR-6828
Revision: a24e5187394427a0412e990c990c3e7ddeb0101f
Repository: https://github.com/kokkos/kokkos.git
  • PR-6828
Revision: 58108c1c82ec01db0ed457bf1d22714e89dd1aee
Repository: https://github.com/kokkos/kokkos.git
  • PR-6828
Revision: 5c735265452c837caaeca4077c9c232df9fbebe4
Repository: https://github.com/kokkos/kokkos.git
  • PR-6828
Revision: ec480f4b1dd315814bbf24c4cbeea528bde975c0
Repository: https://github.com/kokkos/kokkos.git
  • PR-6828
Revision: c864237fb39c65711325a327ae11b472e2d57a67
Repository: https://github.com/kokkos/kokkos.git
  • PR-6828
Revision: 654bfde0f848b88a843c9381056de1dfb4b458aa
Repository: https://github.com/kokkos/kokkos.git
  • PR-6828