UPBGE: Optimize process on culled objects.

Previously the process does in the render for culled objects was too slow due to some reasons:
- The matrix was updated for culled objects.
- All the mesh user infos was updated for culled objects.

To solve all of this issues the culled boolean is removed in RAS_MeshUser which assume that
ActivateMeshSlots always add the mesh slots in the display array bucket.
By doing this we can move the culled check in KX_GameObject::UpdateBuckets and avoid
update all mesh user too.

Also the value checked to update in UpdateBuckets are reordered:
!m_bCulled && m_bVisible && m_meshUser

The UPBGE was slower than the BF BGE for culled objects. For example in a scene of 10000 objects
culled:
UPBGE: 2.32ms
BF BGE: 1.60ms

With this patch the istuation is reverted:
UPBGE: 0.51ms
BF BGE: 1.60ms

32% faster.

source/gameengine/Ketsji/KX_GameObject.cpp

@@ -771,12 +771,13 @@ void KX_GameObject::AddMeshUser()
 
 void KX_GameObject::UpdateBuckets()
 {
-	if (GetSGNode() && m_meshUser) {
-		if (GetSGNode()->IsDirty())
+	// Update datas and add mesh slot to be rendered only if the object is not culled.
+	if (!m_bCulled && m_bVisible && m_meshUser) {
+		if (m_pSGNode->IsDirty()) {
 			GetOpenGLMatrix();
+		}
 
 		m_meshUser->SetColor(m_objectColor);
-		m_meshUser->SetCulled(m_bCulled || !m_bVisible);
 		m_meshUser->SetFrontFace(!m_bIsNegativeScaling);
 		m_meshUser->ActivateMeshSlots();
 	}

source/gameengine/Rasterizer/RAS_MeshUser.cpp

@@ -31,12 +31,10 @@
 
 RAS_MeshUser::RAS_MeshUser(void *clientobj)
 	:m_frontFace(true),
-	m_culled(true),
 	m_color(MT_Vector4(0.0f, 0.0f, 0.0f, 0.0f)),
 	m_matrix(NULL),
 	m_clientObject(clientobj)
 {
-
 }
 
 RAS_MeshUser::~RAS_MeshUser()
@@ -54,11 +52,6 @@ bool RAS_MeshUser::GetFrontFace() const
 	return m_frontFace;
 }
 
-bool RAS_MeshUser::GetCulled() const
-{
-	return m_culled;
-}
-
 const MT_Vector4& RAS_MeshUser::GetColor() const
 {
 	return m_color;
@@ -84,11 +77,6 @@ void RAS_MeshUser::SetFrontFace(bool frontFace)
 	m_frontFace = frontFace;
 }
 
-void RAS_MeshUser::SetCulled(bool culled)
-{
-	m_culled = culled;
-}
-
 void RAS_MeshUser::SetColor(const MT_Vector4& color)
 {
 	m_color = color;
@@ -101,10 +89,8 @@ void RAS_MeshUser::SetMatrix(float *matrix)
 
 void RAS_MeshUser::ActivateMeshSlots()
 {
-	if (!m_culled) {
-		for (RAS_MeshSlotList::iterator it = m_meshSlots.begin(), end = m_meshSlots.end(); it != end; ++it) {
-			RAS_MeshSlot *ms = *it;
-			ms->m_displayArrayBucket->ActivateMesh(ms);
-		}
+	for (RAS_MeshSlotList::iterator it = m_meshSlots.begin(), end = m_meshSlots.end(); it != end; ++it) {
+		RAS_MeshSlot *ms = *it;
+		ms->m_displayArrayBucket->ActivateMesh(ms);
 	}
 }

source/gameengine/Rasterizer/RAS_MeshUser.h

@@ -37,8 +37,6 @@ class RAS_MeshUser
 private:
 	/// OpenGL face wise.
 	bool m_frontFace;
-	/// Object culling flag.
-	bool m_culled;
 	/// Object color.
 	MT_Vector4 m_color;
 	/// Object transformation matrix.
@@ -54,14 +52,12 @@ class RAS_MeshUser
 
 	void AddMeshSlot(RAS_MeshSlot *meshSlot);
 	bool GetFrontFace() const;
-	bool GetCulled() const;
 	const MT_Vector4& GetColor() const;
 	float *GetMatrix() const;
 	void *GetClientObject() const;
 	RAS_MeshSlotList& GetMeshSlots();
 
 	void SetFrontFace(bool frontFace);
-	void SetCulled(bool culled);
 	void SetColor(const MT_Vector4& color);
 	void SetMatrix(float *matrix);