56fd68e7dd
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
36 lines
1.2 KiB
C++
36 lines
1.2 KiB
C++
// SPDX-FileCopyrightText: Copyright (c) 2024, NVIDIA Corporation. All rights reserved.
|
|
// SPDX-License-Identifier: Apache-2.0
|
|
|
|
/// \file alignment.h
|
|
/// \brief Type-alignment utilities.
|
|
|
|
#pragma once
|
|
|
|
#include <thrust/detail/config.h>
|
|
|
|
#if defined(_CCCL_IMPLICIT_SYSTEM_HEADER_GCC)
|
|
# pragma GCC system_header
|
|
#elif defined(_CCCL_IMPLICIT_SYSTEM_HEADER_CLANG)
|
|
# pragma clang system_header
|
|
#elif defined(_CCCL_IMPLICIT_SYSTEM_HEADER_MSVC)
|
|
# pragma system_header
|
|
#endif // no system header
|
|
|
|
THRUST_NAMESPACE_BEGIN
|
|
namespace detail
|
|
{
|
|
/// \p aligned_reinterpret_cast `reinterpret_cast`s \p u of type \p U to `void*`
|
|
/// and then `reinterpret_cast`s the result to \p T. The indirection through
|
|
/// `void*` suppresses compiler warnings when the alignment requirement of \p *u
|
|
/// is less than the alignment requirement of \p *t. The caller of
|
|
/// \p aligned_reinterpret_cast is responsible for ensuring that the alignment
|
|
/// requirements are actually satisfied.
|
|
template <typename T, typename U>
|
|
_CCCL_HOST_DEVICE T aligned_reinterpret_cast(U u)
|
|
{
|
|
// NOLINTNEXTLINE(bugprone-casting-through-void)
|
|
return reinterpret_cast<T>(reinterpret_cast<void*>(u));
|
|
}
|
|
} // end namespace detail
|
|
THRUST_NAMESPACE_END
|