class to create an executable CUDA graph with unique ownership More...

#include <taskflow/cuda/cuda_graph_exec.hpp>

Inheritance diagram for tf::cudaGraphExecBase< Creator, Deleter >:

Collaboration diagram for tf::cudaGraphExecBase< Creator, Deleter >:

Public Types
using	base_type = std::unique_ptr<std::remove_pointer_t<cudaGraphExec_t>, Deleter>
	base std::unique_ptr type

Public Member Functions
template<typename... ArgsT>
	cudaGraphExecBase (ArgsT &&... args)
	constructs a `cudaGraphExec` object by passing the given arguments to the executable CUDA graph creator

	cudaGraphExecBase (cudaGraphExecBase &&)=default
	constructs a `cudaGraphExec` from the given rhs using move semantics

cudaGraphExecBase &	operator= (cudaGraphExecBase &&)=default
	assign the rhs to `*this` using move semantics

template<typename C>
void	host (cudaTask task, C &&callable, void *user_data)
	updates parameters of a host task

template<typename F, typename... ArgsT>
void	kernel (cudaTask task, dim3 g, dim3 b, size_t shm, F f, ArgsT... args)
	updates parameters of a kernel task

void	memset (cudaTask task, void *dst, int ch, size_t count)
	updates parameters of a memset task

void	memcpy (cudaTask task, void tgt, const void src, size_t bytes)
	updates parameters of a memcpy task

template<typename T, std::enable_if_t< is_pod_v< T > &&(sizeof(T)==1\|\|sizeof(T)==2\|\|sizeof(T)==4), void > * = nullptr>
void	zero (cudaTask task, T *dst, size_t count)
	updates parameters of a memset task to a zero task

template<typename T, std::enable_if_t< is_pod_v< T > &&(sizeof(T)==1\|\|sizeof(T)==2\|\|sizeof(T)==4), void > * = nullptr>
void	fill (cudaTask task, T *dst, T value, size_t count)
	updates parameters of a memset task to a fill task

template<typename T, std::enable_if_t<!std::is_same_v< T, void >, void > * = nullptr>
void	copy (cudaTask task, T tgt, const T src, size_t num)
	updates parameters of a memcpy task to a copy task

template<typename C>
void	single_task (cudaTask task, C c)
	updates a single-threaded kernel task

template<typename I, typename C, typename E = cudaDefaultExecutionPolicy>
void	for_each (cudaTask task, I first, I last, C callable)
	updates parameters of a `for_each` kernel task created from the CUDA graph of `*this`

template<typename I, typename C, typename E = cudaDefaultExecutionPolicy>
void	for_each_index (cudaTask task, I first, I last, I step, C callable)
	updates parameters of a `for_each_index` kernel task created from the CUDA graph of `*this`

template<typename I, typename O, typename C, typename E = cudaDefaultExecutionPolicy>
void	transform (cudaTask task, I first, I last, O output, C c)
	updates parameters of a `transform` kernel task created from the CUDA graph of `*this`

template<typename I1, typename I2, typename O, typename C, typename E = cudaDefaultExecutionPolicy>
void	transform (cudaTask task, I1 first1, I1 last1, I2 first2, O output, C c)
	updates parameters of a `transform` kernel task created from the CUDA graph of `*this`

Detailed Description

template<typename Creator, typename Deleter>
class tf::cudaGraphExecBase< Creator, Deleter >

class to create an executable CUDA graph with unique ownership

Template Parameters

Creator	functor to create the stream (used in constructor)
Deleter	functor to delete the stream (used in destructor)

This class wraps a cudaGraphExec_t handle with std::unique_ptr to ensure proper resource management and automatic cleanup.

Constructor & Destructor Documentation

◆ cudaGraphExecBase()

template<typename Creator, typename Deleter>

template<typename... ArgsT>

tf::cudaGraphExecBase< Creator, Deleter >::cudaGraphExecBase ( ArgsT &&... args )

inlineexplicit

constructs a cudaGraphExec object by passing the given arguments to the executable CUDA graph creator

Constructs a cudaGraphExec object by passing the given arguments to the executable CUDA graph creator

Parameters

args	arguments to pass to the executable CUDA graph creator

Member Function Documentation

◆ copy()

template<typename Creator, typename Deleter>

template<typename T, std::enable_if_t<!std::is_same_v< T, void >, void > *>

void tf::cudaGraphExecBase< Creator, Deleter >::copy	(	cudaTask	task,
		T *	tgt,
		const T *	src,
		size_t	num )

updates parameters of a memcpy task to a copy task

The method is similar to tf::cudaFlow::copy but operates on a task of type tf::cudaTaskType::MEMCPY. The source/destination memory may have different address values but must be allocated from the same contexts as the original source/destination memory.

◆ fill()

template<typename Creator, typename Deleter>

template<typename T, std::enable_if_t< is_pod_v< T > &&(sizeof(T)==1||sizeof(T)==2||sizeof(T)==4), void > *>

void tf::cudaGraphExecBase< Creator, Deleter >::fill	(	cudaTask	task,
		T *	dst,
		T	value,
		size_t	count )

updates parameters of a memset task to a fill task

The method is similar to tf::cudaFlow::fill but operates on a task of type tf::cudaTaskType::MEMSET.

The source/destination memory may have different address values but must be allocated from the same contexts as the original source/destination memory.

◆ host()

template<typename Creator, typename Deleter>

template<typename C>

void tf::cudaGraphExecBase< Creator, Deleter >::host	(	cudaTask	task,
		C &&	callable,
		void *	user_data )

updates parameters of a host task

This method updates the parameter of the given host task (similar to tf::cudaFlow::host).

◆ kernel()

template<typename Creator, typename Deleter>

template<typename F, typename... ArgsT>

void tf::cudaGraphExecBase< Creator, Deleter >::kernel	(	cudaTask	task,
		dim3	g,
		dim3	b,
		size_t	shm,
		F	f,
		ArgsT...	args )

updates parameters of a kernel task

The method is similar to tf::cudaFlow::kernel but operates on a task of type tf::cudaTaskType::KERNEL. The kernel function name must NOT change.

◆ memcpy()

template<typename Creator, typename Deleter>

void tf::cudaGraphExecBase< Creator, Deleter >::memcpy	(	cudaTask	task,
		void *	tgt,
		const void *	src,
		size_t	bytes )

updates parameters of a memcpy task

The method is similar to tf::cudaFlow::memcpy but operates on a task of type tf::cudaTaskType::MEMCPY. The source/destination memory may have different address values but must be allocated from the same contexts as the original source/destination memory.

◆ memset()

template<typename Creator, typename Deleter>

void tf::cudaGraphExecBase< Creator, Deleter >::memset	(	cudaTask	task,
		void *	dst,
		int	ch,
		size_t	count )

updates parameters of a memset task

The method is similar to tf::cudaFlow::memset but operates on a task of type tf::cudaTaskType::MEMSET. The source/destination memory may have different address values but must be allocated from the same contexts as the original source/destination memory.

◆ single_task()

template<typename Creator, typename Deleter>

template<typename C>

void tf::cudaGraphExecBase< Creator, Deleter >::single_task	(	cudaTask	task,
		C	c )

updates a single-threaded kernel task

This method is similar to cudaFlow::single_task but operates on an existing task.

◆ zero()

template<typename Creator, typename Deleter>

template<typename T, std::enable_if_t< is_pod_v< T > &&(sizeof(T)==1||sizeof(T)==2||sizeof(T)==4), void > *>

void tf::cudaGraphExecBase< Creator, Deleter >::zero	(	cudaTask	task,
		T *	dst,
		size_t	count )

updates parameters of a memset task to a zero task

The method is similar to tf::cudaFlow::zero but operates on a task of type tf::cudaTaskType::MEMSET.

The source/destination memory may have different address values but must be allocated from the same contexts as the original source/destination memory.

The documentation for this class was generated from the following files:

taskflow/cuda/cuda_graph_exec.hpp
taskflow/cuda/algorithm/for_each.hpp
taskflow/cuda/algorithm/transform.hpp

Public Types

Public Member Functions

Detailed Description

Constructor & Destructor Documentation

◆ cudaGraphExecBase()

Member Function Documentation

◆ copy()

◆ fill()

◆ host()

◆ kernel()

◆ memcpy()

◆ memset()

◆ single_task()

◆ zero()