Performance tracking

[paroj]

https://github.com/paroj/OgreSceneNodeBenchmark

140*140 Nodes = 19600 Nodes

dispatches to:

SceneMgrQueuedRenderableVisitor::visit(const Pass* p, RenderableList& rs)

Frametime

Ogre	Configuration	Static (ms)	Animated (ms)
1.10.7	GL	42.64	65.60
1.10.7	GL3+	89.22	118.54
1.10.7	GL3+ (StateCache)	45.20	64.54
2.0	GL (SSE2, 4 Threads)	27.12	28.98
2.0	GL (C, 4 Threads)	26.97	27.56
2.0	GL (C, 1 Thread)	30.63	32.15
1.10.8	GL (vector storage)	33.10	48.88
1.11.dev	GL3+	38.16	53.16
1.11.dev	GL	33.23	45.75
1.11.dev	GL (StateCache)	29.25	41.60

Intel(R) Xeon(R) CPU E3-1231 v3 @ 3.40GHz

Memory

Ogre	SceneManager	MB
1.10.9	Octree	70
1.10.9	default	67
1.11.dev	default	62.5
2.0	default	66

[paroj]

the remaining bottleneck is likely in findVisibleNodes - more profiling needed.

[paroj]

Ogre	Configuration	Static (ms)
1.11.5dev	GL	39.42
1.11.5dev	GL + #985	36.00

AMD Ryzen 7 2700X Eight-Core Processor

[paroj]

the remaining bottleneck is likely in findVisibleNodes - more profiling needed.

yep.. Animated config:

Ogre 1.12	Ogre 2.0