CPPScheduler Class Reference

C++11 implementation of a pool of threads to automatically split a kernel's execution among several threads. More...

#include <CPPScheduler.h>

Collaboration diagram for CPPScheduler:

Public Member Functions
	CPPScheduler ()
	Constructor: create a pool of threads. More...
	~CPPScheduler ()
	Default destructor. More...
void	set_num_threads (unsigned int num_threads) override
	Sets the number of threads the scheduler will use to run the kernels. More...
void	set_num_threads_with_affinity (unsigned int num_threads, BindFunc func) override
	Sets the number of threads the scheduler will use to run the kernels but also using a binding function to pin the threads to given logical cores. More...
unsigned int	num_threads () const override
	Returns the number of threads that the SingleThreadScheduler has in its pool. More...
void	schedule (ICPPKernel *kernel, const Hints &hints) override
	Runs the kernel in the same thread as the caller synchronously. More...
void	schedule_op (ICPPKernel *kernel, const Hints &hints, const Window &window, ITensorPack &tensors) override
	Runs the kernel in the same thread as the caller synchronously. More...
Public Member Functions inherited from IScheduler
	IScheduler ()
	Default constructor. More...
virtual	~IScheduler ()=default
	Destructor. More...
virtual void	run_tagged_workloads (std::vector< Workload > &workloads, const char *tag)
	Execute all the passed workloads. More...
CPUInfo &	cpu_info ()
	Get CPU info. More...
unsigned int	num_threads_hint () const
	Get a hint for the best possible number of execution threads. More...

Static Public Member Functions
static CPPScheduler &	get ()
	Access the scheduler singleton. More...

Additional Inherited Members
Public Types inherited from IScheduler
enum	StrategyHint { STATIC, DYNAMIC }
	Strategies available to split a workload. More...
using	BindFunc = std::function< int(int, int)>
	Function to be used and map a given thread id to a logical core id. More...
using	Workload = std::function< void(const ThreadInfo &)>
	Signature for the workloads to execute. More...
Static Public Attributes inherited from IScheduler
static constexpr unsigned int	split_dimensions_all = std::numeric_limits<unsigned>::max()
	When arm_compute::ISchedular::Hints::_split_dimension is initialized with this value then the schedular is free to break down the problem space over as many dimensions as it wishes. More...

Detailed Description

C++11 implementation of a pool of threads to automatically split a kernel's execution among several threads.

It has 2 scheduling modes: Linear or Fanout (please refer to the implementation for details) The mode is selected automatically based on the runtime environment. However it can be forced via an environment variable ARM_COMPUTE_CPP_SCHEDULER_MODE. e.g.: ARM_COMPUTE_CPP_SCHEDULER_MODE=linear # Force select the linear scheduling mode ARM_COMPUTE_CPP_SCHEDULER_MODE=fanout # Force select the fanout scheduling mode

Definition at line 42 of file CPPScheduler.h.

Constructor & Destructor Documentation

CPPScheduler()

CPPScheduler

(

)

Constructor: create a pool of threads.

Definition at line 435 of file CPPScheduler.cpp.

                           : _impl(std::make_unique<Impl>(num_threads_hint()))
{
}

~CPPScheduler()

~CPPScheduler ( )

default

Default destructor.

Member Function Documentation

get()

CPPScheduler & get ( )

static

Access the scheduler singleton.

Note

this method has been deprecated and will be remover in future releases

Returns

The scheduler

Definition at line 429 of file CPPScheduler.cpp.

{
    static CPPScheduler scheduler;
    return scheduler;
}

num_threads()

unsigned int num_threads ( ) const

overridevirtual

Returns the number of threads that the SingleThreadScheduler has in its pool.

Returns

Number of threads available in SingleThreadScheduler.

Implements IScheduler.

Definition at line 455 of file CPPScheduler.cpp.

{
    return _impl->num_threads();
}

Referenced by CPPScheduler::set_num_threads(), and CPPScheduler::set_num_threads_with_affinity().

schedule()

void schedule ( ICPPKernel *  kernel, const Hints &  hints  )

overridevirtual

Runs the kernel in the same thread as the caller synchronously.

Parameters

[in]	kernel	Kernel to execute.
[in]	hints	Hints for the scheduler.

Implements IScheduler.

Definition at line 552 of file CPPScheduler.cpp.

{
    ITensorPack tensors;
    schedule_common(kernel, hints, kernel->window(), tensors);
}

References IKernel::window().

Referenced by TEST_CASE().

schedule_op()

void schedule_op ( ICPPKernel *  kernel, const Hints &  hints, const Window &  window, ITensorPack &  tensors  )

overridevirtual

Runs the kernel in the same thread as the caller synchronously.

Parameters

[in]	kernel	Kernel to execute.
[in]	hints	Hints for the scheduler.
[in]	window	Window to use for kernel execution.
[in]	tensors	Vector containing the tensors to operate on.

Implements IScheduler.

Definition at line 547 of file CPPScheduler.cpp.

{
    schedule_common(kernel, hints, window, tensors);
}

set_num_threads()

void set_num_threads ( unsigned int  num_threads )

overridevirtual

Sets the number of threads the scheduler will use to run the kernels.

Parameters

[in]

num_threads

If set to 0, then one thread per CPU core available on the system will be used, otherwise the number of threads specified.

Implements IScheduler.

Definition at line 441 of file CPPScheduler.cpp.

{
    // No changes in the number of threads while current workloads are running
    arm_compute::lock_guard<std::mutex> lock(_impl->_run_workloads_mutex);
    _impl->set_num_threads(num_threads, num_threads_hint());
}

References CPPScheduler::num_threads(), and IScheduler::num_threads_hint().

Referenced by TEST_CASE().

set_num_threads_with_affinity()

void set_num_threads_with_affinity ( unsigned int  num_threads, BindFunc  func  )

overridevirtual

Sets the number of threads the scheduler will use to run the kernels but also using a binding function to pin the threads to given logical cores.

Parameters

[in]	num_threads	If set to 0, then one thread per CPU core available on the system will be used, otherwise the number of threads specified.
[in]	func	Binding function to use.

Reimplemented from IScheduler.

Definition at line 448 of file CPPScheduler.cpp.

{
    // No changes in the number of threads while current workloads are running
    arm_compute::lock_guard<std::mutex> lock(_impl->_run_workloads_mutex);
    _impl->set_num_threads_with_affinity(num_threads, num_threads_hint(), func);
}

References CPPScheduler::num_threads(), and IScheduler::num_threads_hint().

---

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

• arm_compute/runtime/CPP/CPPScheduler.h
• src/runtime/CPP/CPPScheduler.cpp