Small but notable lag with enabled Radiation
Krutoy242 opened this issue ยท 2 comments
Overhauled, radiation_enabled=true
.
I'm maintaining E2E-E server. I found that NC:O causing about 2% of TPS lag.
By looking through source code, i found that most heavy calls is loadedChunks.toArray and transferRadiationFromProviderToChunkBuffer.
- How can I improve TPS on my server without completely disabling radiation? My
radiation_world_chunks_per_tick
set to5
, but seems likeloadedChunks.toArray
would constantly consume same amount of CPU, no matter of config. - How I could optimize
transferRadiationFromProviderToChunkBuffer
or make it not called at all? I cant find a config for disabling TEs to contribute chunk radiation.
Profile: spark.lucko.me/dgYOuB47ke
Logs: debug.zip
Hmm, this seems like a tricky one to get around totally, but I have a potential idea to avoid the conversion of the collection of chunks into an array. I'll add it to the next version of NCO, but it might be a bit of time before NCO as a whole is stable again.
I can also add the config option you describe to disable TEs contributing to chunk radiation :)