Skip to content

[onedpl][ranges] copy_if implementation for sized output - rebase from fresh main state #2162

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Conversation

SergeyKopienko
Copy link
Contributor

This PR prepared for fixing CI errors in the #1898

dmitriy-sobolev and others added 30 commits January 17, 2025 22:08
Proposal for CPU backend implementation for histogram.
The ValueType should not be necessary for device copyability because it is not a member field of the helper struct. It is only the return value of the callable. Therefore, we should not base device copyability on it.

Signed-off-by: Dan Hoeflinger <[email protected]>
* [oneDPL][make] + usage ONEAPI_DEVICE_SELECTOR variable
Co-authored-by: Dmitriy Sobolev <[email protected]>
…en compiling with -O0 (#1997)

Signed-off-by: Matthew Michel <[email protected]>
Co-authored-by: Dmitriy Sobolev <[email protected]>
By adding `_ExecutionPolicy` into the kernel name, we can work around the duplicate kernel name issue in reduce-then-scan based algorithms. However, a library wide solution is still needed for #2041

---------

Signed-off-by: Matthew Michel <[email protected]>
This avoids the indirect windows.h inclusion with the tbb parallel backend which occurred with using tbb's enumerable_thread_specific class.

---------

Signed-off-by: Dan Hoeflinger <[email protected]>
Implement with CRTP.
Store construction arguments by values.
Replace unique_ptr with optional.

Co-authored-by: Dan Hoeflinger <[email protected]>
…iler optimization detection (#2046)

Reverts #2040 and instead uses the `__OPTIMIZE__` macro defined by clang-based compilers to detect -O0 compilation and compile reduce-then-scan paths with a sub-group size of 16 to work around hardware bugs on older integrated graphics architectures. This avoids the performance impact of the kernel bundle approach.
---------

Signed-off-by: Matthew Michel <[email protected]>
dmitriy-sobolev and others added 29 commits March 13, 2025 13:31
Updates to reflect status 

---------

Signed-off-by: Dan Hoeflinger <[email protected]>
Co-authored-by: Alexey Kukanov <[email protected]>
…mespace (#2135)

* Fix deprecation of oneapi::dpl::experimental::kt::esimd namespace. The deprecation attribute should be placed before the namespace's name.

---------

Signed-off-by: Matthew Michel <[email protected]>
* test/parallel_api/iterator/transform_iterator.pass.cpp - remove unused code in struct test_copy_if

* test/parallel_api/iterator/transform_iterator.pass.cpp - remove struct test_copy_if as unused
Provides a proposal for a public customization point for users to define to indicate if their types are passed directly to sycl kernels.

---------

Signed-off-by: Dan Hoeflinger <[email protected]>
Co-authored-by: Ruslan Arutyunyan <[email protected]>
Signed-off-by: Emmanuel Ferdman <[email protected]>
…rial backend)

# Conflicts:
#	include/oneapi/dpl/pstl/omp/parallel_scan.h
#	include/oneapi/dpl/pstl/parallel_backend_tbb.h
…fix warning in __result_and_scratch_storage_impl : declaration of '_T' shadows template parameter [-Wmicrosoft-template-shadow]
…fix C++ 17 compile error in struct __wait_and_get_value : 'auto' not allowed in function prototype
…fix C++17 compile error : 'auto' not allowed in function prototype
…fix C++17 compile error : expected member name or ';' after declaration specifiers
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

9 participants