[libcxx-commits] [libcxx] [libc++][pstl] Introduce a "statpart" backend (PR #230540)

[Louis Dionne via libcxx-commits]

================
@@ -0,0 +1,1221 @@
+//===----------------------------------------------------------------------===//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+//
+//===----------------------------------------------------------------------===//
+
+#ifndef _LIBCPP___PSTL_STATPART_H
+#define _LIBCPP___PSTL_STATPART_H
+
+#include <__algorithm/find_end.h>
+#include <__algorithm/find_if.h>
+#include <__algorithm/for_each.h>
+#include <__algorithm/is_heap_until.h>
+#include <__algorithm/lower_bound.h>
+#include <__algorithm/merge.h>
+#include <__algorithm/min.h>
+#include <__algorithm/min_element.h>
+#include <__algorithm/minmax_element.h>
+#include <__algorithm/mismatch.h>
+#include <__algorithm/move.h>
+#include <__algorithm/reverse.h>
+#include <__algorithm/search.h>
+#include <__algorithm/search_n.h>
+#include <__algorithm/stable_sort.h>
+#include <__algorithm/swap_ranges.h>
+#include <__algorithm/transform.h>
+#include <__algorithm/upper_bound.h>
+#include <__atomic/atomic.h>
+#include <__config>
+#include <__functional/identity.h>
+#include <__functional/operations.h>
+#include <__iterator/concepts.h>
+#include <__iterator/iterator_traits.h>
+#include <__iterator/move_iterator.h>
+#include <__iterator/next.h>
+#include <__iterator/reverse_iterator.h>
+#include <__memory/allocator.h>
+#include <__memory/construct_at.h>
+#include <__memory/destroy.h>
+#include <__memory/uninitialized_algorithms.h>
+#include <__memory/unique_ptr.h>
+#include <__numeric/reduce.h>
+#include <__numeric/transform_reduce.h>
+#include <__optional/nullopt_t.h>
+#include <__optional/optional.h>
+#include <__pstl/backend_fwd.h>
+#include <__pstl/cpu_algos/cpu_traits.h>
+#include <__pstl/cpu_algos/find_if.h>
+#include <__thread/thread.h>
+#include <__type_traits/is_assignable.h>
+#include <__type_traits/is_constructible.h>
+#include <__type_traits/is_execution_policy.h>
+#include <__type_traits/is_trivially_copyable.h>
+#include <__utility/convert_to_integral.h>
+#include <__utility/empty.h>
+#include <__utility/move.h>
+#include <__utility/pair.h>
+
+#if !defined(_LIBCPP_HAS_NO_PRAGMA_SYSTEM_HEADER)
+#  pragma GCC system_header
+#endif
+
+_LIBCPP_PUSH_MACROS
+#include <__undef_macros>
+
+#if _LIBCPP_STD_VER >= 17
+
+_LIBCPP_BEGIN_NAMESPACE_STD
+_LIBCPP_BEGIN_EXPLICIT_ABI_ANNOTATIONS
+namespace __pstl {
+namespace __statpart {
+
+// The backbone of the backend: applies the function to each of the indices [0, __chunk_count] in parallel.
+// The actual number of parallel workers and the order of application is unspecified.
+_LIBCPP_EXPORTED_FROM_ABI void
+__apply(size_t __chunk_count, void* __context, void (*__func)(void* __context, size_t __chunk)) noexcept;
+
+// The maximum number of partitions to be processed per each CPU
+constexpr inline size_t __max_partitions_per_cpu = 8;
+
+// The minimum number of elements to sort serially in a single partition
+constexpr inline size_t __min_sort_partition_size = 256;
+
+enum class __partitioning_result {
+  // returned when the partitioning result is not nullopt
+  __ok,
+  // returned when there is not enough element to produce a single partition or the final number of partitions is less
+  // than 2, i.e. there's no reason to process in parallel
+  __not_enough_elements,
+  // later: returned memory couldn't be allocated for the stashed forward iterators
+  __bad_alloc
+};
+
+template <class _Iterator>
+struct _LIBCPP_HIDE_FROM_ABI __iterator_range {
----------------
ldionne wrote:

Note that there are also range based versions of the PSTL since P3179R9. We should probably keep that in mind to avoid designing something that'll make this impossible or difficult to support.

https://github.com/llvm/llvm-project/pull/230540