Class to manage the scene object rendering queue. More...

#include <OgreRenderQueue.h>

Inheritance diagram for Ogre::RenderQueue:

Public Types
enum	Modes { V1_LEGACY , V1_FAST , FAST , PARTICLE_SYSTEM }

enum	RqSortMode { DisableSort , NormalSort , StableSort }

Public Member Functions
	RenderQueue (HlmsManager hlmsManager, SceneManager sceneManager, VaoManager *vaoManager)

	~RenderQueue ()

void	_compileShadersThread (size_t threadIdx)

void	_releaseManualHardwareResources ()

void	_warmUpShadersThread (size_t threadIdx)

void	addRenderableV1 (uint8 renderQueueId, bool casterPass, Renderable pRend, const MovableObject pMovableObject)
	Add a renderable (Ogre v1.x) object to the queue. More...

void	addRenderableV2 (size_t threadIdx, uint8 renderQueueId, bool casterPass, Renderable pRend, const MovableObject pMovableObject)
	Add a renderable (Ogre v2.0, i.e. More...

void	clear ()
	Empty the queue - should only be called by SceneManagers. More...

void	clearState ()
	The RenderQueue keeps track of API state to avoid redundant state change passes Calling this function forces the RenderQueue to re-set the Macro- & Blendblocks, shaders, and any other API dependendant calls on the next render. More...

void	frameEnded ()
	Called when the frame has fully ended (ALL passes have been executed to all RTTs) More...

RenderQueue::Modes	getRenderQueueMode (uint8 rqId) const

RqSortMode	getSortRenderQueue (uint8 rqId) const

void	render (RenderSystem *rs, uint8 firstRq, uint8 lastRq, bool casterPass, bool dualParaboloid)

void	renderPassPrepare (bool casterPass, bool dualParaboloid)
	If you need to call RenderQueue::render, then you must call this function. More...

void	renderSingleObject (Renderable pRend, const MovableObject pMovableObject, RenderSystem *rs, bool casterPass, bool dualParaboloid)
	Don't call this too often. Only renders v1 objects at the moment. More...

void	setRenderQueueMode (uint8 rqId, RenderQueue::Modes newMode)
	Sets the mode for the RenderQueue ID. More...

void	setSortRenderQueue (uint8 rqId, RqSortMode sortMode)
	Sets whether we should sort the render queue ID every frame. More...

void	warmUpShadersCollect (uint8 firstRq, uint8 lastRq, bool casterPass)
	Simulates what render() would do. More...

void	warmUpShadersTrigger (RenderSystem *rs)

Detailed Description

Class to manage the scene object rendering queue.

Remarks: Objects are grouped by material to minimise rendering state changes. The map from material to renderable object is wrapped in a class for ease of use.

: This class includes the concept of 'queue groups' which allows the application adding the renderable to specifically schedule it so that it is included in a discrete group. Good for separating renderables into the main scene, backgrounds and overlays, and also could be used in the future for more complex multipass routines like stenciling.

Remarks: By default, movables will be assigned the following render queue ID:
Decal: 0
Light: 0-5 (depends on light type, cannot be changed)
Item: 10
Rectangle2D: 10
v1::Entity: 110
v1::Rectangle2D: 110
ParticleSystem (v1):110
ParticleSystem (v2):kParticleSystemDefaultRenderQueueId
[.. more unlisted ..]; By default, the render queues have the following mode set:
0-99: FAST
100-199: V1_FAST
200-224: FAST
225-255: V1_FAST

Member Enumeration Documentation

◆ Modes

enum Ogre::RenderQueue::Modes

Enumerator
V1_LEGACY	This is the slowest mode. Renders the same or similar way to how Ogre 1.x rendered meshes. Only v1 entities can be put in this type of queue. Ideal for low level materials, InstancedEntity via InstanceManager, and any other weird custom object.
V1_FAST	Renders v1 entities using HLMS materials with some of the new benefits, but some deprecated features from Ogre 1.x might not be available or work properly (like global instancing buffer). Ideal for most v1 entities, particle effects and billboards using HLMS materials. Cannot be used by RenderSystems that don't support constant buffers (i.e. OpenGL ES 2)
FAST	Renders v2 items using HLMS materials with minimum driver overhead Recommended method for rendering extremely large amounts of objects of homogeneous and heterogenous meshes, with many different kinds of materials and textures. Only v2 Items can be put in this queue.
PARTICLE_SYSTEM	Renders ParticleSystemDef (i.e. ParticleFX2 plugin). Don't put v1 Particle Systems here.

◆ RqSortMode

enum Ogre::RenderQueue::RqSortMode

Enumerator
DisableSort
NormalSort
StableSort

Constructor & Destructor Documentation

◆ RenderQueue()

Ogre::RenderQueue::RenderQueue	(	HlmsManager *	hlmsManager,
		SceneManager *	sceneManager,
		VaoManager *	vaoManager
	)

◆ ~RenderQueue()

Ogre::RenderQueue::~RenderQueue ( )

Member Function Documentation

◆ _compileShadersThread()

void Ogre::RenderQueue::_compileShadersThread ( size_t threadIdx )

◆ _releaseManualHardwareResources()

void Ogre::RenderQueue::_releaseManualHardwareResources ( )

◆ _warmUpShadersThread()

void Ogre::RenderQueue::_warmUpShadersThread ( size_t threadIdx )

◆ addRenderableV1()

void Ogre::RenderQueue::addRenderableV1	(	uint8	renderQueueId,
		bool	casterPass,
		Renderable *	pRend,
		const MovableObject *	pMovableObject
	)

Add a renderable (Ogre v1.x) object to the queue.

See also: addRenderable

◆ addRenderableV2()

void Ogre::RenderQueue::addRenderableV2	(	size_t	threadIdx,
		uint8	renderQueueId,
		bool	casterPass,
		Renderable *	pRend,
		const MovableObject *	pMovableObject
	)

Add a renderable (Ogre v2.0, i.e.

Items; they use VAOs) object to the queue.

Remarks: If sorting mode is set to DisableSort (RenderQueue::setSortRenderQueue) for the given renderQueueId, then the order in which renderables are added via RenderQueue::addRenderableV2 (and V1) determines the render order (the first added get rendered first) within that renderQueueId.

Parameters

threadIdx	The unique index of the thread from which this function is called from. Valid range is [0; SceneManager::mNumWorkerThreads)
renderQueueId	The ID of the render queue. Must be the same as the ID in pMovableObject->getRenderQueueGroup()
casterPass	Whether we're performing the shadow mapping pass.
pRend	Pointer to the Renderable to be added to the queue.
pMovableObject	Pointer to the MovableObject linked to the Renderable.

◆ clear()

void Ogre::RenderQueue::clear ( )

Empty the queue - should only be called by SceneManagers.

◆ clearState()

void Ogre::RenderQueue::clearState ( )

The RenderQueue keeps track of API state to avoid redundant state change passes Calling this function forces the RenderQueue to re-set the Macro- & Blendblocks, shaders, and any other API dependendant calls on the next render.

Remarks: Calling this function inside render or renderES2 won't have any effect.

◆ frameEnded()

void Ogre::RenderQueue::frameEnded ( )

Called when the frame has fully ended (ALL passes have been executed to all RTTs)

◆ getRenderQueueMode()

RenderQueue::Modes Ogre::RenderQueue::getRenderQueueMode ( uint8 rqId ) const

◆ getSortRenderQueue()

RqSortMode Ogre::RenderQueue::getSortRenderQueue ( uint8 rqId ) const

◆ render()

void Ogre::RenderQueue::render	(	RenderSystem *	rs,
		uint8	firstRq,
		uint8	lastRq,
		bool	casterPass,
		bool	dualParaboloid
	)

◆ renderPassPrepare()

void Ogre::RenderQueue::renderPassPrepare	(	bool	casterPass,
		bool	dualParaboloid
	)

If you need to call RenderQueue::render, then you must call this function.

This function MUST be called (all listed functions are called in this order):

After RenderSystem::beginRenderPassDescriptor
After SceneManager::fireRenderQueueStarted
Before RenderSystem::executeRenderPassDescriptorDelayedActions
Before RenderQueue::render

Note that fireRenderQueueStarted just fires arbitrary listeners. You don't have to call that function if you are sure you don't need it.

To clarify functions are called in this order:

mRenderSystem->beginRenderPassDescriptor();
mSceneManager->fireRenderQueueStarted();
mRenderQueue->renderPassPrepare();
mRenderSystem->executeRenderPassDescriptorDelayedActions();
mRenderQueue->render();

Calling these functions in proper order is needed for best compatibility with Metal

Parameters

casterPass
dualParaboloid

◆ renderSingleObject()

void Ogre::RenderQueue::renderSingleObject	(	Renderable *	pRend,
		const MovableObject *	pMovableObject,
		RenderSystem *	rs,
		bool	casterPass,
		bool	dualParaboloid
	)

Don't call this too often. Only renders v1 objects at the moment.

◆ setRenderQueueMode()

void Ogre::RenderQueue::setRenderQueueMode	(	uint8	rqId,
		RenderQueue::Modes	newMode
	)

Sets the mode for the RenderQueue ID.

See also: RenderQueue::Modes

Parameters

rqId	ID of the render queue
newMode	The new mode to use.

◆ setSortRenderQueue()

void Ogre::RenderQueue::setSortRenderQueue	(	uint8	rqId,
		RqSortMode	sortMode
	)

Sets whether we should sort the render queue ID every frame.

Parameters

rqId	ID of the render queue
bSort	When false, the render queue group won't be sorted. Useful when the RQ needs to be drawn exactly in the order that Renderables were added, or when you have a deep CPU bottleneck where the time taken to sort hurts more than it is supposed to help.

◆ warmUpShadersCollect()

void Ogre::RenderQueue::warmUpShadersCollect	(	uint8	firstRq,
		uint8	lastRq,
		bool	casterPass
	)

Simulates what render() would do.

But doesn't actually render and skips as much as possible We are only interested in generating the PSOs for compiling them early.

Parameters

firstRq
lastRq
casterPass

◆ warmUpShadersTrigger()

void Ogre::RenderQueue::warmUpShadersTrigger ( RenderSystem * rs )

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

OgreRenderQueue.h

Public Types

Public Member Functions

Detailed Description

Member Enumeration Documentation

◆ Modes

◆ RqSortMode

Constructor & Destructor Documentation

◆ RenderQueue()

◆ ~RenderQueue()

Member Function Documentation

◆ _compileShadersThread()

◆ _releaseManualHardwareResources()

◆ _warmUpShadersThread()

◆ addRenderableV1()

◆ addRenderableV2()

◆ clear()

◆ clearState()

◆ frameEnded()

◆ getRenderQueueMode()

◆ getSortRenderQueue()

◆ render()

◆ renderPassPrepare()

◆ renderSingleObject()

◆ setRenderQueueMode()

◆ setSortRenderQueue()

◆ warmUpShadersCollect()

◆ warmUpShadersTrigger()