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
35 lines
1.0 KiB
Plaintext
35 lines
1.0 KiB
Plaintext
// SPDX-FileCopyrightText: Copyright (c) 2011-2025, NVIDIA CORPORATION. All rights reserved.
|
|
// SPDX-License-Identifier: BSD-3
|
|
|
|
#pragma once
|
|
|
|
#include <cub/config.cuh>
|
|
|
|
#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
|
|
|
|
CUB_NAMESPACE_BEGIN
|
|
namespace detail
|
|
{
|
|
#ifndef _CCCL_DOXYGEN_INVOKED // Do not document
|
|
|
|
// NOTE: bit_cast cannot be always used because __half, __nv_bfloat16, etc. are not trivially copyable
|
|
template <typename Output, typename Input>
|
|
[[nodiscard]] _CCCL_DEVICE _CCCL_FORCEINLINE Output unsafe_bitcast(const Input& input)
|
|
{
|
|
Output output;
|
|
static_assert(sizeof(input) == sizeof(output), "wrong size");
|
|
// NOLINTNEXTLINE(bugprone-undefined-memory-manipulation)
|
|
::memcpy(&output, &input, sizeof(input));
|
|
return output;
|
|
}
|
|
|
|
#endif // !_CCCL_DOXYGEN_INVOKED
|
|
} // namespace detail
|
|
CUB_NAMESPACE_END
|