The elastic data memory request generator to read protobuf trace containing execution trace annotated with data and ordering dependencies. More...

#include <trace_cpu.hh>

Classes
class	GraphNode
	The struct GraphNode stores an instruction in the trace file. More...

class	HardwareResource
	The HardwareResource class models structures that hold the in-flight nodes. More...

class	InputStream
	The InputStream encapsulates a trace file and the internal buffers and populates GraphNodes based on the input. More...

struct	ReadyNode
	Struct to store a ready-to-execute node and its execution tick. More...

Public Member Functions
	ElasticDataGen (TraceCPU &_owner, const std::string &_name, MasterPort &_port, MasterID master_id, const std::string &trace_file, TraceCPUParams *params)

Tick	init ()
	Called from TraceCPU init(). More...

void	adjustInitTraceOffset (Tick &offset)
	Adjust traceOffset based on what TraceCPU init() determines on comparing the offsets in the fetch request and elastic traces. More...

const std::string &	name () const
	Returns name of the ElasticDataGen instance. More...

void	exit ()
	Exit the ElasticDataGen. More...

bool	readNextWindow ()
	Reads a line of the trace file. More...

template<typename T >
void	addDepsOnParent (GraphNode *new_node, T &dep_array, uint8_t &num_dep)
	Iterate over the dependencies of a new node and add the new node to the list of dependents of the parent node. More...

void	execute ()
	This is the main execute function which consumes nodes from the sorted readyList. More...

PacketPtr	executeMemReq (GraphNode *node_ptr)
	Creates a new request for a load or store assigning the request parameters. More...

void	addToSortedReadyList (NodeSeqNum seq_num, Tick exec_tick)
	Add a ready node to the readyList. More...

void	printReadyList ()
	Print readyList for debugging using debug flag TraceCPUData. More...

void	completeMemAccess (PacketPtr pkt)
	When a load writeback is received, that is when the load completes, release the dependents on it. More...

bool	isExecComplete () const
	Returns the execComplete variable which is set when the last node is executed. More...

bool	checkAndIssue (const GraphNode *node_ptr, bool first=true)
	Attempts to issue a node once the node's source dependencies are complete. More...

uint64_t	getMicroOpCount () const
	Get number of micro-ops modelled in the TraceCPU replay. More...

void	regStats ()

Private Types
typedef uint64_t	NodeSeqNum
	Node sequence number type. More...

typedef uint64_t	NodeRobNum
	Node ROB number type. More...

typedef ProtoMessage::InstDepRecord::RecordType	RecordType

typedef ProtoMessage::InstDepRecord	Record

Private Attributes
TraceCPU &	owner
	Reference of the TraceCPU. More...

MasterPort &	port
	Reference of the port to be used to issue memory requests. More...

const MasterID	masterID
	MasterID used for the requests being sent. More...

InputStream	trace
	Input stream used for reading the input trace file. More...

std::string	genName
	String to store the name of the FixedRetryGen. More...

PacketPtr	retryPkt
	PacketPtr used to store the packet to retry. More...

bool	traceComplete
	Set to true when end of trace is reached. More...

bool	nextRead
	Set to true when the next window of instructions need to be read. More...

bool	execComplete
	Set true when execution of trace is complete. More...

const uint32_t	windowSize
	Window size within which to check for dependencies. More...

HardwareResource	hwResource
	Hardware resources required to contain in-flight nodes and to throttle issuing of new nodes when resources are not available. More...

std::unordered_map< NodeSeqNum, GraphNode * >	depGraph
	Store the depGraph of GraphNodes. More...

std::queue< const GraphNode * >	depFreeQueue
	Queue of dependency-free nodes that are pending issue because resources are not available. More...

std::list< ReadyNode >	readyList
	List of nodes that are ready to execute. More...

Stats::Scalar	maxDependents
	Stats for data memory accesses replayed. More...

Stats::Scalar	maxReadyListSize

Stats::Scalar	numSendAttempted

Stats::Scalar	numSendSucceeded

Stats::Scalar	numSendFailed

Stats::Scalar	numRetrySucceeded

Stats::Scalar	numSplitReqs

Stats::Scalar	numSOLoads

Stats::Scalar	numSOStores

Stats::Scalar	dataLastTick
	Tick when ElasticDataGen completes execution. More...

Detailed Description

The elastic data memory request generator to read protobuf trace containing execution trace annotated with data and ordering dependencies.

It deduces the time at which to send a load/store request by tracking the dependencies. It attempts to send a memory request for a load/store without performing real execution of micro-ops. If L1 cache port sends packet succesfully, the generator checks which instructions became dependency free as a result of this and schedules an event accordingly. If it fails to send the packet, it waits for a retry from the cache.

Definition at line 554 of file trace_cpu.hh.

Member Typedef Documentation

◆ NodeRobNum

typedef uint64_t TraceCPU::ElasticDataGen::NodeRobNum

private

Node ROB number type.

Definition at line 563 of file trace_cpu.hh.

◆ NodeSeqNum

typedef uint64_t TraceCPU::ElasticDataGen::NodeSeqNum

private

Node sequence number type.

Definition at line 560 of file trace_cpu.hh.

◆ Record

typedef ProtoMessage::InstDepRecord TraceCPU::ElasticDataGen::Record

private

Definition at line 566 of file trace_cpu.hh.

◆ RecordType

typedef ProtoMessage::InstDepRecord::RecordType TraceCPU::ElasticDataGen::RecordType

private

Definition at line 565 of file trace_cpu.hh.

Constructor & Destructor Documentation

◆ ElasticDataGen()

TraceCPU::ElasticDataGen::ElasticDataGen	(	TraceCPU &	_owner,
		const std::string &	_name,
		MasterPort &	_port,
		MasterID	master_id,
		const std::string &	trace_file,
		TraceCPUParams *	params
	)

inline

Definition at line 849 of file trace_cpu.hh.

References DPRINTF, TraceCPU::init(), and ArmISA::offset.

Member Function Documentation

◆ addDepsOnParent()

template<typename T >

void TraceCPU::ElasticDataGen::addDepsOnParent	(	GraphNode *	new_node,
		T &	dep_array,
		uint8_t &	num_dep
	)

Iterate over the dependencies of a new node and add the new node to the list of dependents of the parent node.

Parameters

new_node new node to add to the graph

Template Parameters

dep_array the dependency array of type rob or register, that is to be iterated, and may get modified

Parameters

num_dep the number of dependencies set in the array which may get modified during iteration

Definition at line 397 of file trace_cpu.cc.

◆ addToSortedReadyList()

void TraceCPU::ElasticDataGen::addToSortedReadyList	(	NodeSeqNum	seq_num,
		Tick	exec_tick
	)

Add a ready node to the readyList.

When inserting, ensure the nodes are sorted in ascending order of their execute ticks.

Parameters

seq_num	seq. num of ready node
exec_tick	the execute tick of the ready node

Definition at line 801 of file trace_cpu.cc.

References TraceCPU::ElasticDataGen::ReadyNode::execTick, and TraceCPU::ElasticDataGen::ReadyNode::seqNum.

◆ adjustInitTraceOffset()

void TraceCPU::ElasticDataGen::adjustInitTraceOffset ( Tick & offset )

Adjust traceOffset based on what TraceCPU init() determines on comparing the offsets in the fetch request and elastic traces.

Parameters

trace_offset trace offset set by comparing both traces

Definition at line 331 of file trace_cpu.cc.

References ArmISA::offset.

Referenced by TraceCPU::init().

◆ checkAndIssue()

bool TraceCPU::ElasticDataGen::checkAndIssue	(	const GraphNode *	node_ptr,
		bool	first = `true`
	)

Attempts to issue a node once the node's source dependencies are complete.

If resources are available then add it to the readyList, otherwise the node is not issued and is stored in depFreeQueue until resources become available.

Parameters

node_ptr	pointer to node to be issued
first	true if this is the first attempt to issue this node

Returns: true if node was added to readyList

Definition at line 694 of file trace_cpu.cc.

References TraceCPU::ElasticDataGen::GraphNode::compDelay, DPRINTFR, TraceCPU::ElasticDataGen::GraphNode::numRegDep, TraceCPU::ElasticDataGen::GraphNode::numRobDep, TraceCPU::ElasticDataGen::GraphNode::robNum, TraceCPU::ElasticDataGen::GraphNode::seqNum, and TraceCPU::ElasticDataGen::GraphNode::typeToStr().

◆ completeMemAccess()

void TraceCPU::ElasticDataGen::completeMemAccess ( PacketPtr pkt )

When a load writeback is received, that is when the load completes, release the dependents on it.

This is called from the dcache port recvTimingResp().

Definition at line 734 of file trace_cpu.cc.

References DPRINTF, DTRACE, Packet::isWrite(), and Packet::req.

Referenced by TraceCPU::dcacheRecvTimingResp().

◆ execute()

void TraceCPU::ElasticDataGen::execute ( )

This is the main execute function which consumes nodes from the sorted readyList.

First attempt to issue the pending dependency-free nodes held in the depFreeQueue. Insert the ready-to-issue nodes into the readyList. Then iterate through the readyList and when a node has its execute tick equal to curTick(), execute it. If the node is a load or a store call executeMemReq() and if it is neither, simply mark it complete.

Definition at line 427 of file trace_cpu.cc.

References curTick(), DPRINTF, DPRINTFR, DTRACE, and panic.

Referenced by TraceCPU::schedDcacheNext().

◆ executeMemReq()

PacketPtr TraceCPU::ElasticDataGen::executeMemReq ( GraphNode * node_ptr )

Creates a new request for a load or store assigning the request parameters.

Calls the port's sendTimingReq() and returns a packet if the send failed so that it can be saved for a retry.

Parameters

node_ptr pointer to the load or store node to be executed

Returns: packet pointer if the request failed and nullptr if it was sent successfully

Definition at line 621 of file trace_cpu.cc.

References Packet::createRead(), Packet::createWrite(), Packet::dataDynamic(), DPRINTF, TraceCPU::ElasticDataGen::GraphNode::flags, TraceCPU::ElasticDataGen::GraphNode::isLoad(), TraceCPU::ElasticDataGen::GraphNode::isStrictlyOrdered(), TraceCPU::ElasticDataGen::GraphNode::pc, TraceCPU::ElasticDataGen::GraphNode::physAddr, TraceCPU::ElasticDataGen::GraphNode::seqNum, TraceCPU::ElasticDataGen::GraphNode::size, and TraceCPU::ElasticDataGen::GraphNode::virtAddr.

◆ exit()

void TraceCPU::ElasticDataGen::exit ( )

Exit the ElasticDataGen.

Definition at line 338 of file trace_cpu.cc.

◆ getMicroOpCount()

uint64_t TraceCPU::ElasticDataGen::getMicroOpCount ( ) const

inline

Get number of micro-ops modelled in the TraceCPU replay.

Definition at line 977 of file trace_cpu.hh.

References TraceCPU::regStats().

◆ init()

Tick TraceCPU::ElasticDataGen::init ( )

Called from TraceCPU init().

Reads the first message from the input trace file and returns the send tick.

Returns: Tick when first packet must be sent

Definition at line 301 of file trace_cpu.cc.

References DPRINTF, DTRACE, and panic.

Referenced by TraceCPU::init().

◆ isExecComplete()

bool TraceCPU::ElasticDataGen::isExecComplete ( ) const

inline

Returns the execComplete variable which is set when the last node is executed.

Returns: bool true if execComplete is set, false otherwise.

Definition at line 962 of file trace_cpu.hh.

Referenced by TraceCPU::schedDcacheNext().

◆ name()

const std::string& TraceCPU::ElasticDataGen::name ( ) const

inline

Returns name of the ElasticDataGen instance.

Definition at line 886 of file trace_cpu.hh.

References X86ISA::exit.

Referenced by TraceCPU::FixedRetryGen::regStats().

◆ printReadyList()

void TraceCPU::ElasticDataGen::printReadyList ( )

Print readyList for debugging using debug flag TraceCPUData.

Definition at line 858 of file trace_cpu.cc.

References DPRINTF, and DPRINTFR.

◆ readNextWindow()

bool TraceCPU::ElasticDataGen::readNextWindow ( )

Reads a line of the trace file.

Returns the tick when the next request should be generated. If the end of the file has been reached, it returns false.

Returns: bool false if end of file has been reached else true

Definition at line 344 of file trace_cpu.cc.

References DPRINTF, TraceCPU::ElasticDataGen::GraphNode::numRegDep, TraceCPU::ElasticDataGen::GraphNode::numRobDep, TraceCPU::ElasticDataGen::GraphNode::regDep, TraceCPU::ElasticDataGen::GraphNode::robDep, and TraceCPU::ElasticDataGen::GraphNode::seqNum.

◆ regStats()

void TraceCPU::ElasticDataGen::regStats ( )

Definition at line 245 of file trace_cpu.cc.

References SimObject::name().

Referenced by TraceCPU::regStats().

Member Data Documentation

◆ dataLastTick

Stats::Scalar TraceCPU::ElasticDataGen::dataLastTick

private

Tick when ElasticDataGen completes execution.

Definition at line 1053 of file trace_cpu.hh.

◆ depFreeQueue

std::queue<const GraphNode*> TraceCPU::ElasticDataGen::depFreeQueue

private

Queue of dependency-free nodes that are pending issue because resources are not available.

This is chosen to be FIFO so that dependent nodes which become free in program order get pushed into the queue in that order. Thus nodes are more likely to issue in program order.

Definition at line 1037 of file trace_cpu.hh.

◆ depGraph

std::unordered_map<NodeSeqNum, GraphNode*> TraceCPU::ElasticDataGen::depGraph

private

Store the depGraph of GraphNodes.

Definition at line 1028 of file trace_cpu.hh.

◆ execComplete

bool TraceCPU::ElasticDataGen::execComplete

private

Set true when execution of trace is complete.

Definition at line 1008 of file trace_cpu.hh.

◆ genName

std::string TraceCPU::ElasticDataGen::genName

private

String to store the name of the FixedRetryGen.

Definition at line 996 of file trace_cpu.hh.

◆ hwResource

HardwareResource TraceCPU::ElasticDataGen::hwResource

private

Hardware resources required to contain in-flight nodes and to throttle issuing of new nodes when resources are not available.

Definition at line 1025 of file trace_cpu.hh.

◆ masterID

const MasterID TraceCPU::ElasticDataGen::masterID

private

MasterID used for the requests being sent.

Definition at line 990 of file trace_cpu.hh.

Referenced by TraceCPU::FixedRetryGen::send().

◆ maxDependents

Stats::Scalar TraceCPU::ElasticDataGen::maxDependents

private

Stats for data memory accesses replayed.

Definition at line 1043 of file trace_cpu.hh.

◆ maxReadyListSize

Stats::Scalar TraceCPU::ElasticDataGen::maxReadyListSize

private

Definition at line 1044 of file trace_cpu.hh.

◆ nextRead

bool TraceCPU::ElasticDataGen::nextRead

private

Set to true when the next window of instructions need to be read.

Definition at line 1005 of file trace_cpu.hh.

◆ numRetrySucceeded

Stats::Scalar TraceCPU::ElasticDataGen::numRetrySucceeded

private

Definition at line 1048 of file trace_cpu.hh.

Referenced by TraceCPU::FixedRetryGen::regStats(), and TraceCPU::FixedRetryGen::tryNext().

◆ numSendAttempted

Stats::Scalar TraceCPU::ElasticDataGen::numSendAttempted

private

Definition at line 1045 of file trace_cpu.hh.

Referenced by TraceCPU::FixedRetryGen::regStats(), and TraceCPU::FixedRetryGen::tryNext().

◆ numSendFailed

Stats::Scalar TraceCPU::ElasticDataGen::numSendFailed

private

Definition at line 1047 of file trace_cpu.hh.

Referenced by TraceCPU::FixedRetryGen::regStats(), and TraceCPU::FixedRetryGen::tryNext().

◆ numSendSucceeded

Stats::Scalar TraceCPU::ElasticDataGen::numSendSucceeded

private

Definition at line 1046 of file trace_cpu.hh.

Referenced by TraceCPU::FixedRetryGen::regStats(), and TraceCPU::FixedRetryGen::tryNext().

◆ numSOLoads

Stats::Scalar TraceCPU::ElasticDataGen::numSOLoads

private

Definition at line 1050 of file trace_cpu.hh.

◆ numSOStores

Stats::Scalar TraceCPU::ElasticDataGen::numSOStores

private

Definition at line 1051 of file trace_cpu.hh.

◆ numSplitReqs

Stats::Scalar TraceCPU::ElasticDataGen::numSplitReqs

private

Definition at line 1049 of file trace_cpu.hh.

◆ owner

TraceCPU& TraceCPU::ElasticDataGen::owner

private

Reference of the TraceCPU.

Definition at line 984 of file trace_cpu.hh.

Referenced by TraceCPU::IcachePort::recvReqRetry(), TraceCPU::DcachePort::recvReqRetry(), and TraceCPU::DcachePort::recvTimingResp().

◆ port

MasterPort& TraceCPU::ElasticDataGen::port

private

Reference of the port to be used to issue memory requests.

Definition at line 987 of file trace_cpu.hh.

Referenced by TraceCPU::FixedRetryGen::send(), and TraceCPU::FixedRetryGen::tryNext().

◆ readyList

std::list<ReadyNode> TraceCPU::ElasticDataGen::readyList

private

List of nodes that are ready to execute.

Definition at line 1040 of file trace_cpu.hh.

◆ retryPkt

PacketPtr TraceCPU::ElasticDataGen::retryPkt

private

PacketPtr used to store the packet to retry.

Definition at line 999 of file trace_cpu.hh.

Referenced by TraceCPU::FixedRetryGen::send(), and TraceCPU::FixedRetryGen::tryNext().

◆ trace

InputStream TraceCPU::ElasticDataGen::trace

private

Input stream used for reading the input trace file.

Definition at line 993 of file trace_cpu.hh.

Referenced by TraceCPU::FixedRetryGen::exit(), and TraceCPU::FixedRetryGen::nextExecute().

◆ traceComplete

bool TraceCPU::ElasticDataGen::traceComplete

private

Set to true when end of trace is reached.

Definition at line 1002 of file trace_cpu.hh.

Referenced by TraceCPU::FixedRetryGen::nextExecute(), and TraceCPU::FixedRetryGen::tryNext().

◆ windowSize

const uint32_t TraceCPU::ElasticDataGen::windowSize

private

Window size within which to check for dependencies.

Its value is made equal to the window size used to generate the trace which is recorded in the trace header. The dependency graph must be populated enough such that when a node completes, its potential child node must be found and the dependency removed before the completed node itself is removed. Thus as soon as the graph shrinks to become smaller than this window, we read in the next window.

Definition at line 1019 of file trace_cpu.hh.

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

cpu/trace/trace_cpu.hh
cpu/trace/trace_cpu.cc

Classes

Public Member Functions

Private Types

Private Attributes

Detailed Description

Member Typedef Documentation

◆ NodeRobNum

◆ NodeSeqNum

◆ Record

◆ RecordType

Constructor & Destructor Documentation

◆ ElasticDataGen()

Member Function Documentation

◆ addDepsOnParent()

◆ addToSortedReadyList()

◆ adjustInitTraceOffset()

◆ checkAndIssue()

◆ completeMemAccess()

◆ execute()

◆ executeMemReq()

◆ exit()

◆ getMicroOpCount()

◆ init()

◆ isExecComplete()

◆ name()

◆ printReadyList()

◆ readNextWindow()

◆ regStats()

Member Data Documentation

◆ dataLastTick

◆ depFreeQueue

◆ depGraph

◆ execComplete

◆ genName

◆ hwResource

◆ masterID

◆ maxDependents

◆ maxReadyListSize

◆ nextRead

◆ numRetrySucceeded

◆ numSendAttempted

◆ numSendFailed

◆ numSendSucceeded

◆ numSOLoads

◆ numSOStores

◆ numSplitReqs

◆ owner

◆ port

◆ readyList

◆ retryPkt

◆ trace

◆ traceComplete

◆ windowSize