Server lag related to block update
LemADEC opened this issue · 8 comments
As found on our public server using FTB Inventions 1.0.1, there's 10 to 40% load on an event handler related to ProjectRed. See below screenshot at 17% from a 626 seconds profiling:
On shorter period, the same event handler can take 40 % of the load by itself.
Related mods version are:
Forge 1614
ProjectRed-1.7.10-4.7.0pre12.95
MrTJPCore-1.1.0.33
I would like to know if this issue has been fixed as i'm running a modded server with PR on it and one of my IC chips has a traffic light sequence built in and I want to make sure that it won't lag if too many are placed within one world.
@LemADEC Did you find a solution to this? Seems like some sort of regeneration.
@MrTJP please.
Note: retro_gen is false on config, though it's definitely something with moss feature.
What do you mean by regeneration?
The issue is related to the core mod reproducing vanilla 'random ticks' to apply custom effects on every chunks. There's apparently a memory leak as it reload chunks (about 2% direct impact, way more in practice). More importantly, the core of the function takes a lot of time to scan each chunk, so there might be too many handlers registered or some filtering issue involved, etc.
I get this issue too; we resorted to ASMing out the ticker entirely. We have retrogen and mossy spread off. In fact, we have all PR worldgen off.