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
97 lines
3.5 KiB
C++
97 lines
3.5 KiB
C++
//===----------------------------------------------------------------------===//
|
|
//
|
|
// Part of the libcu++ 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) 2025 NVIDIA CORPORATION & AFFILIATES.
|
|
//
|
|
//===----------------------------------------------------------------------===//
|
|
|
|
#ifndef _CUDA___NUMERIC_NARROW_H
|
|
#define _CUDA___NUMERIC_NARROW_H
|
|
|
|
#include <cuda/std/detail/__config>
|
|
|
|
#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
|
|
|
|
#include <cuda/std/__exception/exception_macros.h>
|
|
#include <cuda/std/__host_stdlib/stdexcept>
|
|
#include <cuda/std/__type_traits/is_arithmetic.h>
|
|
#include <cuda/std/__type_traits/is_constructible.h>
|
|
#include <cuda/std/__type_traits/is_signed.h>
|
|
#include <cuda/std/__utility/forward.h>
|
|
|
|
#include <cuda/std/__cccl/prologue.h>
|
|
|
|
_CCCL_BEGIN_NAMESPACE_CUDA
|
|
|
|
//! Uses static_cast to cast a value \p __from to type \p _To. \p _To needs to be constructible from \p _From, and \p
|
|
//! implement operator!=. This function is intended to show that narrowing and a potential change of the value is
|
|
//! intended. Modelled after `gsl::narrow_cast`. See also the C++ Core Guidelines <a
|
|
//! href="https://isocpp.github.io/CppCoreGuidelines/CppCoreGuidelines#Res-narrowing">ES.46</a> and <a
|
|
//! href="https://isocpp.github.io/CppCoreGuidelines/CppCoreGuidelines#Res-casts-named">ES.49</a>.
|
|
template <class _To, class _From>
|
|
[[nodiscard]] _CCCL_API constexpr _To
|
|
narrow_cast(_From&& __from) noexcept(noexcept(static_cast<_To>(::cuda::std::forward<_From>(__from))))
|
|
{
|
|
return static_cast<_To>(::cuda::std::forward<_From>(__from));
|
|
}
|
|
|
|
#if _CCCL_HAS_EXCEPTIONS()
|
|
struct narrowing_error : ::std::runtime_error
|
|
{
|
|
_CCCL_HOST_API narrowing_error()
|
|
: ::std::runtime_error("Narrowing error")
|
|
{}
|
|
};
|
|
#endif // _CCCL_HAS_EXCEPTIONS()
|
|
|
|
//! Uses static_cast to cast a value \p __from to type \p _To and checks whether the value has changed. \p _To needs
|
|
//! to be constructible from \p _From and vice versa, and \p implement operator!=. Throws \ref narrowing_error in host
|
|
//! code and traps in device code if the value has changed. Modelled after `gsl::narrow`. See also the C++ Core
|
|
//! Guidelines <a href="https://isocpp.github.io/CppCoreGuidelines/CppCoreGuidelines#Res-narrowing">ES.46</a> and <a
|
|
//! href="https://isocpp.github.io/CppCoreGuidelines/CppCoreGuidelines#Res-casts-named">ES.49</a>.
|
|
template <class _To, class _From>
|
|
[[nodiscard]] _CCCL_API constexpr _To narrow(_From __from)
|
|
{
|
|
static_assert(::cuda::std::is_constructible_v<_From, _To>);
|
|
static_assert(::cuda::std::is_constructible_v<_To, _From>);
|
|
|
|
const auto __converted = static_cast<_To>(__from);
|
|
if (static_cast<_From>(__converted) != __from)
|
|
{
|
|
_CCCL_THROW(::cuda::narrowing_error);
|
|
}
|
|
|
|
if constexpr (::cuda::std::is_arithmetic_v<_From>)
|
|
{
|
|
if constexpr (::cuda::std::is_signed_v<_From> && !::cuda::std::is_signed_v<_To>)
|
|
{
|
|
if (__from < _From{})
|
|
{
|
|
_CCCL_THROW(::cuda::narrowing_error);
|
|
}
|
|
}
|
|
if constexpr (!::cuda::std::is_signed_v<_From> && ::cuda::std::is_signed_v<_To>)
|
|
{
|
|
if (__converted < _To{})
|
|
{
|
|
_CCCL_THROW(::cuda::narrowing_error);
|
|
}
|
|
}
|
|
}
|
|
return __converted;
|
|
}
|
|
|
|
_CCCL_END_NAMESPACE_CUDA
|
|
|
|
#include <cuda/std/__cccl/epilogue.h>
|
|
|
|
#endif // _CUDA___NUMERIC_NARROW_H
|