CCCL (CUDA C++ Core Libraries) provides: - CUB: device/block/warp-level GPU primitives (reduce, scan, sort, topk) - Thrust: high-level parallel algorithms (transform_reduce, sort, scan) - libcudacxx: CUDA C++ standard library (atomics, barriers, memory) - cudax: experimental features (memory resources, allocators) - Tuning policies: per-SM hardware-specific algorithm parameters Competition optimization vectors mapped to CCCL: - Output TPS (83% weight): warp_reduce, block_reduce, device_topk - Input TPS (14% weight): device_scan, block_load, prefetch - Cache TPS (3% weight): prefix caching strategy patterns - Memory (0.9 util): pooled/cached/buddy allocators Source: https://github.com/NVIDIA/cccl (shallow clone, HEAD only) License: Apache-2.0
167 lines
4.3 KiB
C++
167 lines
4.3 KiB
C++
//===----------------------------------------------------------------------===//
|
|
//
|
|
// 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
|
|
// SPDX-FileCopyrightText: Copyright (c) 2024 NVIDIA CORPORATION & AFFILIATES.
|
|
//
|
|
//===----------------------------------------------------------------------===//
|
|
#ifndef LIBCXX_TEST_SUPPORT_TEST_RANGE_H
|
|
#define LIBCXX_TEST_SUPPORT_TEST_RANGE_H
|
|
|
|
#include <cuda/std/iterator>
|
|
#include <cuda/std/ranges>
|
|
|
|
#include "test_iterators.h"
|
|
|
|
struct sentinel
|
|
{
|
|
template <class I, cuda::std::enable_if_t<cuda::std::input_or_output_iterator<I>, int> = 0>
|
|
TEST_FUNC friend bool operator==(sentinel const&, I const&)
|
|
{
|
|
return true;
|
|
}
|
|
template <class I, cuda::std::enable_if_t<cuda::std::input_or_output_iterator<I>, int> = 0>
|
|
TEST_FUNC friend bool operator==(I const&, sentinel const&)
|
|
{
|
|
return true;
|
|
}
|
|
template <class I, cuda::std::enable_if_t<cuda::std::input_or_output_iterator<I>, int> = 0>
|
|
TEST_FUNC friend bool operator!=(sentinel const&, I const&)
|
|
{
|
|
return false;
|
|
}
|
|
template <class I, cuda::std::enable_if_t<cuda::std::input_or_output_iterator<I>, int> = 0>
|
|
TEST_FUNC friend bool operator!=(I const&, sentinel const&)
|
|
{
|
|
return false;
|
|
}
|
|
};
|
|
|
|
template <template <class...> class I, cuda::std::enable_if_t<cuda::std::input_or_output_iterator<I<int*>>, int> = 0>
|
|
struct test_range
|
|
{
|
|
TEST_FUNC I<int*> begin()
|
|
{
|
|
return I<int*>{nullptr};
|
|
}
|
|
TEST_FUNC I<int const*> begin() const
|
|
{
|
|
return I<int const*>{nullptr};
|
|
}
|
|
TEST_FUNC sentinel end()
|
|
{
|
|
return sentinel{};
|
|
}
|
|
TEST_FUNC sentinel end() const
|
|
{
|
|
return sentinel{};
|
|
}
|
|
};
|
|
|
|
template <template <class...> class I, cuda::std::enable_if_t<cuda::std::input_or_output_iterator<I<int*>>, int> = 0>
|
|
struct test_non_const_range
|
|
{
|
|
TEST_FUNC I<int*> begin()
|
|
{
|
|
return I<int*>{nullptr};
|
|
}
|
|
TEST_FUNC sentinel end()
|
|
{
|
|
return sentinel{};
|
|
}
|
|
};
|
|
|
|
template <template <class...> class I, cuda::std::enable_if_t<cuda::std::input_or_output_iterator<I<int*>>, int> = 0>
|
|
struct test_common_range
|
|
{
|
|
TEST_FUNC I<int*> begin()
|
|
{
|
|
return I<int*>{nullptr};
|
|
}
|
|
TEST_FUNC I<int const*> begin() const
|
|
{
|
|
return I<int const*>{nullptr};
|
|
}
|
|
TEST_FUNC I<int*> end()
|
|
{
|
|
return I<int*>{nullptr};
|
|
}
|
|
TEST_FUNC I<int const*> end() const
|
|
{
|
|
return I<int const*>{nullptr};
|
|
}
|
|
};
|
|
|
|
template <template <class...> class I, cuda::std::enable_if_t<cuda::std::input_or_output_iterator<I<int*>>, int> = 0>
|
|
struct test_non_const_common_range
|
|
{
|
|
TEST_FUNC I<int*> begin()
|
|
{
|
|
return I<int*>{nullptr};
|
|
}
|
|
TEST_FUNC I<int*> end()
|
|
{
|
|
return I<int*>{nullptr};
|
|
}
|
|
};
|
|
|
|
template <template <class...> class I, cuda::std::enable_if_t<cuda::std::input_or_output_iterator<I<int*>>, int> = 0>
|
|
struct test_view : cuda::std::ranges::view_base
|
|
{
|
|
TEST_FUNC I<int*> begin()
|
|
{
|
|
return I<int*>{nullptr};
|
|
}
|
|
TEST_FUNC I<int const*> begin() const
|
|
{
|
|
return I<int const*>{nullptr};
|
|
}
|
|
TEST_FUNC sentinel end()
|
|
{
|
|
return sentinel{};
|
|
}
|
|
TEST_FUNC sentinel end() const
|
|
{
|
|
return sentinel{};
|
|
}
|
|
};
|
|
|
|
struct BorrowedRange
|
|
{
|
|
TEST_FUNC int* begin() const
|
|
{
|
|
return nullptr;
|
|
}
|
|
TEST_FUNC int* end() const
|
|
{
|
|
return nullptr;
|
|
}
|
|
TEST_FUNC BorrowedRange(BorrowedRange&&) = delete;
|
|
};
|
|
|
|
namespace cuda::std::ranges
|
|
{
|
|
template <>
|
|
inline constexpr bool enable_borrowed_range<BorrowedRange> = true;
|
|
} // namespace cuda::std::ranges
|
|
|
|
static_assert(!cuda::std::ranges::view<BorrowedRange>);
|
|
static_assert(cuda::std::ranges::borrowed_range<BorrowedRange>);
|
|
|
|
using BorrowedView = cuda::std::ranges::empty_view<int>;
|
|
static_assert(cuda::std::ranges::view<BorrowedView>);
|
|
static_assert(cuda::std::ranges::borrowed_range<BorrowedView>);
|
|
|
|
using NonBorrowedView = cuda::std::ranges::single_view<int>;
|
|
static_assert(cuda::std::ranges::view<NonBorrowedView>);
|
|
static_assert(!cuda::std::ranges::borrowed_range<NonBorrowedView>);
|
|
|
|
template <class Range>
|
|
inline constexpr bool simple_view =
|
|
cuda::std::ranges::view<Range> && cuda::std::ranges::range<const Range>
|
|
&& cuda::std::same_as<cuda::std::ranges::iterator_t<Range>, cuda::std::ranges::iterator_t<const Range>>
|
|
&& cuda::std::same_as<cuda::std::ranges::sentinel_t<Range>, cuda::std::ranges::sentinel_t<const Range>>;
|
|
|
|
#endif // LIBCXX_TEST_SUPPORT_TEST_RANGE_H
|