uni20::async::DebugCudaScheduler Class Reference

Configurable debug scheduler that admits ordinary and CUDA tasks. More...

#include <debug_cuda_scheduler.hpp>

Inheritance diagram for uni20::async::DebugCudaScheduler:

Collaboration diagram for uni20::async::DebugCudaScheduler:

Public Member Functions
	DebugCudaScheduler (cuda::Device default_device, DebugSchedulerOptions options={})
	Construct with an explicit default CUDA activation device.
	DebugCudaScheduler (DebugSchedulerOptions options={})
	Construct with device zero as the default CUDA activation device.
void	schedule (AsyncTask &&task) override
	Enqueue a task for later run.
void	schedule (CudaTask &&task) override
	Enqueue a CUDA task without establishing device affinity.
void	schedule (CudaTask &&task, int device) override
	Bind and enqueue a CUDA task for initial execution on a device.
Public Member Functions inherited from uni20::async::DebugScheduler
bool	can_run () const noexcept
	Reports whether at least one runnable task exists.
	DebugScheduler (DebugSchedulerOptions options={})
	Default-construct an empty scheduler.
bool	done () const noexcept
	Check whether the runnable queue is empty.
void	help_while_waiting (const WaitPredicate &is_ready) override
	Drives one runnable task while waiting for readiness.
DebugSchedulerOptions const &	options () const noexcept
	Return the scheduler's immutable ordering configuration.
void	pause () override
	Block the scheduler from running.
void	resume () override
	Unblock the scheduler.
void	run ()
	Run one batch of scheduled coroutines in the configured order.
void	run_all ()
	Run until no runnable tasks remain.
void	schedule (AsyncTask &&task) override
	Enqueue a task for later run.
	~DebugScheduler ()
	Destructor.
Public Member Functions inherited from uni20::async::IAsyncScheduler
template<class Function > requires std::invocable<Function&, std::size_t>
void	execute_batch (std::size_t size, Function &&function)
	Execute a synchronous batch of independent lightweight work items.
virtual void	wait_for (const WaitPredicate &is_ready)
	Block the calling thread until is_ready returns true.
virtual void	wait_for (WaitRequest const &request)
	Block until a structured wait request becomes ready.
	~IAsyncScheduler () override=default
	Virtual destructor.
Public Member Functions inherited from uni20::async::IScheduler
virtual	~IScheduler ()=default
	Virtual destructor.
Public Member Functions inherited from uni20::async::ICudaScheduler
	~ICudaScheduler () override=default

Additional Inherited Members
Public Types inherited from uni20::async::IAsyncScheduler
using	WaitPredicate = std::function< bool()>
using	WaitWakeup = std::function< void()>
using	WaitWakeupRegistration = std::function< void(WaitWakeup)>
Protected Member Functions inherited from uni20::async::DebugScheduler
void	enqueue_task (BasicTask &&task)
	Add one already-bound task to the runnable queue.

Detailed Description

Configurable debug scheduler that admits ordinary and CUDA tasks.

All tasks share one runnable queue and the ordering policy inherited from DebugScheduler. Each CUDA activation selects the device stored in its promise and restores the calling thread's previous device when it suspends or completes.

Constructor & Destructor Documentation

DebugCudaScheduler() [1/2]

uni20::async::DebugCudaScheduler::DebugCudaScheduler ( DebugSchedulerOptions  options = {} )

inlineexplicit

Construct with device zero as the default CUDA activation device.

DebugCudaScheduler() [2/2]

uni20::async::DebugCudaScheduler::DebugCudaScheduler ( cuda::Device  default_device, DebugSchedulerOptions  options = {}  )

inlineexplicit

Construct with an explicit default CUDA activation device.

Parameters

default_device	Device used while a CUDA task has no affinity.
options	Deterministic runnable ordering configuration.

Member Function Documentation

schedule() [1/3]

void uni20::async::DebugScheduler::schedule ( AsyncTask &&  task )

inlineoverridevirtual

Enqueue a task for later run.

Parameters

task	An AsyncTask bound to this scheduler.

Implements uni20::async::IAsyncScheduler.

schedule() [2/3]

void uni20::async::DebugCudaScheduler::schedule ( CudaTask &&  task )

inlineoverridevirtual

Enqueue a CUDA task without establishing device affinity.

Parameters

task	CUDA task to admit through the default device context.

Implements uni20::async::ICudaScheduler.

schedule() [3/3]

void uni20::async::DebugCudaScheduler::schedule ( CudaTask &&  task, int  device  )

inlineoverridevirtual

Bind and enqueue a CUDA task for initial execution on a device.

Parameters

task	CUDA task to admit.
device	CUDA runtime device ordinal.

Implements uni20::async::ICudaScheduler.

---

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

• debug_cuda_scheduler.hpp