00:34 < bridge> [teeworlds] <Dune> hey, I don't have this issue
11:22 < bridge> [teeworlds] <Learath2> the traffic isn't encrypted, should be trivial to diagnose this with wireshark
19:39 < bridge> [teeworlds] <fokkonaut> https://cdn.discordapp.com/attachments/490150878934990850/712358850849800222/Screenshot_20200519-193822_Chrome.jpg
19:39 < bridge> [teeworlds] <fokkonaut> 101 players! record :)
20:58 < bridge> [teeworlds] <Dune> didn't you turn it off because it wasn't working well past 72?
21:18 < bridge> [teeworlds] <fokkonaut> I turned it back on
21:18 < bridge> [teeworlds] <fokkonaut> Its pretty much just the inefficient algorithm
21:18 < bridge> [teeworlds] <fokkonaut> i think it can be improved
21:21 < bridge> [teeworlds] <redix> wow 25 ping -> > 100 prediction time... seems like 128 players is really too much for the algorithms =\
21:57 < bridge> [teeworlds] <fokkonaut> :D
21:58 < bridge> [teeworlds] <fokkonaut> Do you have any ideas to improve it?
22:30 < bridge> [teeworlds] <redix> Depends on what the limiting factor is. If it's the physics code, dividing the world into several chunks might help. So you are only checking the collision of tees and other entities which are inside the same chunk. Might aswell be the netcode, snap creation, 64 <-> 128 mapping code or whatever but i rather suspect the physics code
22:31 < bridge> [teeworlds] <redix> some profiling might be good
22:37 < bridge> [teeworlds] <redix> pred time seems better now but the pred margin graph has frequent spikes
22:37 < bridge> [teeworlds] <redix> https://cdn.discordapp.com/attachments/490150878934990850/712403510121070612/screenshot_2020-05-19_22-35-56.png
22:39 < bridge> [teeworlds] <redix> https://cdn.discordapp.com/attachments/490150878934990850/712404009398567022/screenshot_2020-05-19_22-37-57.png
22:39 < bridge> [teeworlds] <redix> vanilla server on the same machine as comparison