Each Scheduling boundary is associated with ready queues. More...

#include "llvm/CodeGen/VLIWMachineScheduler.h"

Public Member Functions
	VLIWSchedBoundary (unsigned ID, const Twine &Name)
	Pending queues extend the ready queues with the same ID and the PendingFlag set.
	~VLIWSchedBoundary ()
VLIWSchedBoundary &	operator= (const VLIWSchedBoundary &other)=delete
	VLIWSchedBoundary (const VLIWSchedBoundary &other)=delete
void	init (VLIWMachineScheduler dag, const TargetSchedModel smodel)
bool	isTop () const
bool	checkHazard (SUnit *SU)
	Does this SU have a hazard within the current instruction group.
void	releaseNode (SUnit *SU, unsigned ReadyCycle)
void	bumpCycle ()
	Move the boundary of scheduled code by one cycle.
void	bumpNode (SUnit *SU)
	Move the boundary of scheduled code by one SUnit.
void	releasePending ()
	Release pending ready nodes in to the available queue.
void	removeReady (SUnit *SU)
	Remove SU from the ready set for this boundary.
SUnit *	pickOnlyChoice ()
	If this queue only has one ready candidate, return it.
bool	isLatencyBound (SUnit *SU)

Public Attributes
VLIWMachineScheduler *	DAG = nullptr
const TargetSchedModel *	SchedModel = nullptr
ReadyQueue	Available
ReadyQueue	Pending
bool	CheckPending = false
ScheduleHazardRecognizer *	HazardRec = nullptr
VLIWResourceModel *	ResourceModel = nullptr
unsigned	CurrCycle = 0
unsigned	IssueCount = 0
unsigned	CriticalPathLength = 0
unsigned	MinReadyCycle = std::numeric_limits<unsigned>::max()
	MinReadyCycle - Cycle of the soonest available instruction.
unsigned	MaxMinLatency = 0

Detailed Description

Each Scheduling boundary is associated with ready queues.

It tracks the current cycle in whichever direction at has moved, and maintains the state of "hazards" and other interlocks at the current cycle.

Definition at line 127 of file VLIWMachineScheduler.h.

Constructor & Destructor Documentation

◆ VLIWSchedBoundary() [1/2]

llvm::ConvergingVLIWScheduler::VLIWSchedBoundary::VLIWSchedBoundary	(	unsigned	ID,
		const Twine &	Name )

inline

Pending queues extend the ready queues with the same ID and the PendingFlag set.

Definition at line 150 of file VLIWMachineScheduler.h.

References Available, llvm::ConvergingVLIWScheduler::ConvergingVLIWScheduler(), llvm::ConvergingVLIWScheduler::LogMaxQID, and Pending.

Referenced by operator=(), and VLIWSchedBoundary().

◆ ~VLIWSchedBoundary()

ConvergingVLIWScheduler::VLIWSchedBoundary::~VLIWSchedBoundary ( )

Definition at line 340 of file VLIWMachineScheduler.cpp.

References HazardRec, and ResourceModel.

◆ VLIWSchedBoundary() [2/2]

llvm::ConvergingVLIWScheduler::VLIWSchedBoundary::VLIWSchedBoundary ( const VLIWSchedBoundary & other )

delete

References VLIWSchedBoundary().

Member Function Documentation

◆ bumpCycle()

void ConvergingVLIWScheduler::VLIWSchedBoundary::bumpCycle ( )

Move the boundary of scheduled code by one cycle.

Definition at line 384 of file VLIWMachineScheduler.cpp.

References assert(), Available, CheckPending, CurrCycle, llvm::dbgs(), HazardRec, IssueCount, isTop(), LLVM_DEBUG, MinReadyCycle, and SchedModel.

Referenced by bumpNode(), and pickOnlyChoice().

◆ bumpNode()

void ConvergingVLIWScheduler::VLIWSchedBoundary::bumpNode ( SUnit * SU )

Move the boundary of scheduled code by one SUnit.

Definition at line 411 of file VLIWMachineScheduler.cpp.

References bumpCycle(), CurrCycle, llvm::dbgs(), llvm::SUnit::getInstr(), HazardRec, llvm::SUnit::isCall, IssueCount, isTop(), LLVM_DEBUG, ResourceModel, and SchedModel.

◆ checkHazard()

bool ConvergingVLIWScheduler::VLIWSchedBoundary::checkHazard ( SUnit * SU )

Does this SU have a hazard within the current instruction group.

The scheduler supports two modes of hazard recognition. The first is the ScheduleHazardRecognizer API. It is a fully general hazard recognizer that supports highly complicated in-order reservation tables (ScoreboardHazardRecognizer) and arbitrary target-specific logic.

The second is a streamlined mechanism that checks for hazards based on simple counters that the scheduler itself maintains. It explicitly checks for instruction dispatch limitations, including the number of micro-ops that can dispatch per cycle.

TODO: Also check whether the SU must start a new group.

Definition at line 358 of file VLIWMachineScheduler.cpp.

References llvm::SUnit::getInstr(), HazardRec, IssueCount, llvm::ScheduleHazardRecognizer::NoHazard, and SchedModel.

Referenced by releaseNode(), and releasePending().