Integrate CUDASTF -> CudaX #2572
Commits on Oct 16, 2024
-
-
-
-
-
-
-
-
Revert "Try to avoid ambiguous operators"
This reverts commit df1777f.
-
-
-
-
-
-
-
-
-
Rename cudax test targets to match CCCL conventions. (NVIDIA#2568)
<project>.<dialect>.test.<testname> instead of <project>.<dialect>.<testname>_tests
-
Update project list in issue template (NVIDIA#2532)
* Update project list in issue template * Update .github/ISSUE_TEMPLATE/feature_request.yml Co-authored-by: Jake Hemstad <[email protected]> --------- Co-authored-by: Jake Hemstad <[email protected]>
-
-
Fix an error in the documentation of the ci/build_common.sh script wh…
…ich mentions -arch=80 instead of -arch 80 (NVIDIA#2574)
-
gcc-14 gained the ability to mangle
noexceptexpressions (NVIDIA#2565)* gcc-14 gained the ability to mangle `noexcept` expressions * Update libcudacxx/include/cuda/std/__concepts/__concept_macros.h --------- Co-authored-by: Michael Schellenberger Costa <[email protected]>
-
Miscellaneous simple fixes (NVIDIA#2575)
* Do not include our own string.h file (NVIDIA#2444) That might conflict with the host standard library * Add missing `` in docs/cub/index.rst * Very minor grammar fix in docs/cub/index.rst * Remove outdated "Getting The Thrust Source Code" section. * Fix TEST_THRUST_PATH in c/parallel/test/CMakeLists.txt (looks like a copy-paste-forgot-to-modify-oversight) * Remove a stray semicolon in c/parallel/src/reduce.cu (functionally inconsequential). * Fix a memory leak in c/parallel/test/c2h.h (this leak is only in a test and most likely not critical at all). --------- Co-authored-by: Michael Schellenberger Costa <[email protected]>
-
Avoid including
yvals.hwhen the compiler is not MSVC. (NVIDIA#2545)* Only include yvals when the compiler is MSVC * Use the right check --------- Co-authored-by: Michael Schellenberger Costa <[email protected]>
-
Fix popc.h when architecture is not x86 on MSVC. (NVIDIA#2524)
* Fix popc when architecture is not x86 * Update libcudacxx/include/cuda/std/__bit/popc.h --------- Co-authored-by: Michael Schellenberger Costa <[email protected]>
-
fix the forwarding of the receiver in the
just_fromalgorithm (NVID……IA#2569) * fix the forwarding of the receiver in the `just_from` algorithm * suppress clang deprecation warning about `prop` and `env` * suppress clang warning about unknown diagnostic
-
Block type pack indexing on NVCC (NVIDIA#2563)
* Block type indexing if the compiler is NVCC. Co-authored-by: Michael Schellenberger Costa <[email protected]>
-
Cleanup the semaphore headers (NVIDIA#2441)
* Move `__atomic_semaphore_base` to its own file * Move `counting_semaphore` to its own file * Move `cuda::counting_semaphore` to its own file * Cleanup the semaphore headers
-
Add
_CCCL_GRID_CONSTANTmacro (NVIDIA#2530)* add _CCCL_GRID_CONSTANT__
-
Add
_CCCL_RESTRICTmacro (NVIDIA#2529)* add _CCCL_RESTRICT macro
-
Try to use the same redefinition of
__assert_failas pytorch has (N……VIDIA#2577) * Try to use the same redefinition of `__assert_fail` as pytorch has * Also dont use the `__THROW` macro as that might be undefined
-
-
Expose parts of
<cuda/std/memory>(NVIDIA#2502)This exposes some parts of `<memory>` that are frequently used and safe to use everywhere. We do not expose some features like allocators and smart pointers until we are sure that they are usefull and properly implemented. Co-authored-by: Michael Schellenberger Costa <[email protected]>
-
add a config macro for testing support for inline variables (NVIDIA#2581
) * add a config macro for testing support for inline variables * Drop `_LIBCUDACXX_` macros * Use a consistent macro for enabling variable templates * Move it in repo.toml --------- Co-authored-by: Michael Schellenberger Costa <[email protected]>
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
Commits on Oct 17, 2024
Commits on Oct 18, 2024
Commits on Oct 19, 2024
Commits on Oct 21, 2024
Commits on Oct 22, 2024
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
Revert "use a fallback source_location for old CUDA versions on MSVC"
This reverts commit 98014f9.
-
-
-
-
-
-
-
Co-authored-by: Michael Schellenberger Costa <[email protected]>
