cuda_device.hpp

转到此文件的文档。

/*
 * Copyright (c) 2021-2025, NVIDIA CORPORATION.
 *
 * Licensed under the Apache License, Version 2.0 (the "License");
 * you may not use this file except in compliance with the License.
 * You may obtain a copy of the License at
 *
 * https://apache.ac.cn/licenses/LICENSE-2.0
 *
 * Unless required by applicable law or agreed to in writing, software
 * distributed under the License is distributed on an "AS IS" BASIS,
 * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 * See the License for the specific language governing permissions and
 * limitations under the License.
 */
#pragma once

#include <rmm/aligned.hpp>
#include <rmm/detail/error.hpp>
#include <rmm/detail/export.hpp>

#include <cuda_runtime_api.h>

#include <cstddef>
#include <utility>

namespace RMM_NAMESPACE {

struct cuda_device_id;
inline cuda_device_id get_current_cuda_device();

struct cuda_device_id {
 using value_type = int;

 cuda_device_id() noexcept : id_{get_current_cuda_device().value()} {}

 explicit constexpr cuda_device_id(value_type dev_id) noexcept : id_{dev_id} {}

 [[nodiscard]] constexpr value_type value() const noexcept { return id_; }

 // TODO re-add doxygen comment specifier /** for these hidden friend operators once this Breathe
 // bug is fixed: https://github.com/breathe-doc/breathe/issues/916

 [[nodiscard]] constexpr friend bool operator==(cuda_device_id const& lhs,
 cuda_device_id const& rhs) noexcept
 {
 return lhs.value() == rhs.value();
 }

 [[nodiscard]] constexpr friend bool operator!=(cuda_device_id const& lhs,
 cuda_device_id const& rhs) noexcept
 {
 return lhs.value() != rhs.value();
 }
 private
 value_type id_;
};

inline cuda_device_id get_current_cuda_device()
{
 cuda_device_id::value_type dev_id{-1};
 RMM_ASSERT_CUDA_SUCCESS(cudaGetDevice(&dev_id));
 return cuda_device_id{dev_id};
}

inline int get_num_cuda_devices()
{
 cuda_device_id::value_type num_dev{-1};
 RMM_ASSERT_CUDA_SUCCESS(cudaGetDeviceCount(&num_dev));
 return num_dev;
}

inline std::pair<std::size_t, std::size_t> available_device_memory()
{
 std::size_t free{};
 std::size_t total{};
 RMM_CUDA_TRY(cudaMemGetInfo(&free, &total));
 return {free, total};
}

inline std::size_t percent_of_free_device_memory(int percent)
{
 [[maybe_unused]] auto const [free, total] = rmm::available_device_memory();
 auto fraction = static_cast<double>(percent) / 100.0;
 return rmm::align_down(static_cast<std::size_t>(static_cast<double>(free) * fraction),
 rmm::CUDA_ALLOCATION_ALIGNMENT);
}

struct cuda_set_device_raii {
 explicit cuda_set_device_raii(cuda_device_id dev_id)
 : old_device_{get_current_cuda_device()},
 needs_reset_{dev_id.value() >= 0 && old_device_ != dev_id}
 {
 if (needs_reset_) { RMM_ASSERT_CUDA_SUCCESS(cudaSetDevice(dev_id.value())); }
 }
 ~cuda_set_device_raii() noexcept
 {
 if (needs_reset_) { RMM_ASSERT_CUDA_SUCCESS(cudaSetDevice(old_device_.value())); }
 }

 cuda_set_device_raii(cuda_set_device_raii const&) = delete;
 cuda_set_device_raii& operator=(cuda_set_device_raii const&) = delete;
 cuda_set_device_raii(cuda_set_device_raii&&) = delete;
 cuda_set_device_raii& operator=(cuda_set_device_raii&&) = delete;

 private
 cuda_device_id old_device_;
 bool needs_reset_;
};
 // end of group
} // namespace RMM_NAMESPACE