NPL::CNeuronFileState Class Reference

each neuron file that is activated will have a CNeuronFileState structure kept. More...

#include <NeuronFileState.h>

Public Member Functions
	CNeuronFileState (const std::string &filename)
bool	IsProcessing () const
	whether the message is still being processed in the activation function. More...
void	SetProcessing (bool val)
bool	PushMessage (NPLMessage_ptr &msg)
bool	PopMessage (NPLMessage_ptr &msg)
	Pop a message. More...
bool	GetMessage (NPLMessage_ptr &msg)
	get the message without removing it from the stack
bool	IsEmpty ()
	if message queue is empty. More...
void	ClearMessage ()
	clear all messages
bool	PopClearState ()
void	Tick (int nFrameMoveId)
	increase internal tick counter for activation. More...
int	GetTick (int nFrameMoveId)
	get tick count for current frame. More...
int32	GetMaxQueueSize () const
	max number of messages in the pending message queue. More...
void	SetMaxQueueSize (int32 val)
const std::string &	GetFilename () const
	neuron file name
bool	IsPreemptive ()
	Preemptive multi-tasking is simulated by counting instructions. More...
int32	GetPreemptiveInstructionCount () const
	if not 0, we will use coroutine and debug hook to simulate preemptive scheduling in user mode. More...
void	SetPreemptiveInstructionCount (int32 val)

Protected Member Functions
void	ClearMessageImp ()

Protected Attributes
int32	m_nActivationThisTick
	total number of activation calls in this/last tick. More...
int32	m_nFrameMoveId
	different for each tick
int32	m_nTotalActivations
	total number of activations since program starts. More...
int32	m_nPreemptiveInstructionCount
	if not 0, we will use coroutine and debug hook to simulate preemptive scheduling in user mode. More...
DWORD	m_nActionState
	action state
bool	m_isProcessing
	whether the message is still being processed in the activation function. More...
int32	m_nMaxQueueSize
	max number of messages in the pending message queue. More...
std::deque< NPLMessage_ptr >	m_queue
	the pending message queue
std::string	m_filename
	neuron file name

Friends
class	CNPLRuntimeState

Detailed Description

each neuron file that is activated will have a CNeuronFileState structure kept.

This class is only used by the local NPL thread, so no lock is needed.

Member Function Documentation

GetMaxQueueSize()

int32 NPL::CNeuronFileState::GetMaxQueueSize

(

)

const

max number of messages in the pending message queue.

Any new incoming messages will be dropped. By default its size is same as parent NPL thread's message queue size

GetPreemptiveInstructionCount()

int32 NPL::CNeuronFileState::GetPreemptiveInstructionCount

(

)

const

if not 0, we will use coroutine and debug hook to simulate preemptive scheduling in user mode.

GetTick()

int NPL::CNeuronFileState::GetTick

(

int

nFrameMoveId

)

get tick count for current frame.

Parameters

nFrameMoveId

parent NPL thread's frame move id

IsEmpty()

bool NPL::CNeuronFileState::IsEmpty

(

)

if message queue is empty.

IsPreemptive()

bool NPL::CNeuronFileState::IsPreemptive

(

)

Preemptive multi-tasking is simulated by counting instructions.

IsProcessing()

bool NPL::CNeuronFileState::IsProcessing

(

)

const

whether the message is still being processed in the activation function.

some activation will set this flag to true in case their processing is cross multiple NPL time slice (tick).

PopMessage()

bool NPL::CNeuronFileState::PopMessage

(

NPLMessage_ptr &

msg

)

Pop a message.

Returns

true if succeed or false if queue is empty.

PushMessage()

bool NPL::CNeuronFileState::PushMessage

(

NPLMessage_ptr &

msg

)

Returns

true if succeed, or false if queue is full.

Tick()

void NPL::CNeuronFileState::Tick

(

int

nFrameMoveId

)

increase internal tick counter for activation.

Parameters

nFrameMoveId

parent NPL thread's frame move id

Member Data Documentation

m_isProcessing

bool NPL::CNeuronFileState::m_isProcessing

protected

whether the message is still being processed in the activation function.

some activation will set this flag to true in case their processing is cross multiple NPL time slice (tick).

m_nActivationThisTick

int32 NPL::CNeuronFileState::m_nActivationThisTick

protected

total number of activation calls in this/last tick.

m_nMaxQueueSize

int32 NPL::CNeuronFileState::m_nMaxQueueSize

protected

max number of messages in the pending message queue.

Any new incoming messages will be dropped. By default its size is same as parent NPL thread's message queue size

m_nPreemptiveInstructionCount

int32 NPL::CNeuronFileState::m_nPreemptiveInstructionCount

protected

if not 0, we will use coroutine and debug hook to simulate preemptive scheduling in user mode.

m_nTotalActivations

int32 NPL::CNeuronFileState::m_nTotalActivations

protected

total number of activations since program starts.

---

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

• Client/trunk/ParaEngineClient/NPL/NeuronFileState.h
• Client/trunk/ParaEngineClient/NPL/NeuronFileState.cpp