hpx/async_cuda/cuda_executor.hpp

Defined in header hpx/async_cuda/cuda_executor.hpp.

See Public API for a list of names and headers that are part of the public HPX API.

namespace hpx

namespace cuda

namespace experimental

struct cuda_executor : public hpx::cuda::experimental::cuda_executor_base

Public Functions

inline explicit cuda_executor(std::size_t device, bool event_mode = true)

inline ~cuda_executor()

template<typename F, typename ...Ts>
inline decltype(auto) friend tag_invoke(hpx::parallel::execution::post_t, cuda_executor const &exec, F &&f, Ts&&... ts)

template<typename F, typename ...Ts>
inline decltype(auto) friend tag_invoke(hpx::parallel::execution::async_execute_t, cuda_executor const &exec, F &&f, Ts&&... ts)

Protected Functions

template<typename R, typename ...Params, typename ...Args>
inline void post(R (*cuda_function)(Params...), Args&&... args) const

template<typename R, typename ...Params, typename ...Args>
inline hpx::future<void> async(R (*cuda_kernel)(Params...), Args&&... args) const

struct cuda_executor_base

Subclassed by hpx::cuda::experimental::cuda_executor

Public Types

using future_type = hpx::future<void>

Public Functions

inline cuda_executor_base(std::size_t device, bool event_mode)

inline future_type get_future() const

Protected Attributes

int device_

bool event_mode_

cudaStream_t stream_

std::shared_ptr<hpx::cuda::experimental::target> target_

namespace parallel

namespace execution