Skip to content

[v2] Optimisation candidates #49

@Cubicake

Description

@Cubicake

v2 is already performing incredibly well, but the following places are candidates for optimisation.

AbstractBlockView#isBlockOccluding - 82% of RaycastUtil#raycast is from this method. RaESP can currently do 1000 recasts in ~ 0.5ms, but chunk section occlusion would benefit from increasing this number. The ConcurrentHashMap lookup would need to be replaced, maybe with fastutil?

AbstractBlockView#getNeedingRecheck - Relatively expensive iteration of a ConcurrentHashMap, not sure if it is possible to optimise this at all.

PacketEventsBlockViewController#ingestChunkAndSetTileEntitiesToHiddenBlocks - This is by far the least performant method currently, however unfortunately the bottleneck is in PacketEvent's Chunk_v1_18#getBlockID, which I doubt can be optimised from our side.

Metadata

Metadata

Labels

No labels
No labels

Projects

No projects

Milestone

No milestone

Relationships

None yet

Development

No branches or pull requests

Issue actions