Mcpc version B269
mekanism version 6.0.5

The server has been having TPS drops ranging from 18TPS-9TPS and the only thing that has a update time higher then 100u/s is mekanism's handler at about 200-7000 u/s so I'm assuming the TPS drop is due to that.

I may not have any pictures of mekanism reaching 7000 however I've seen it steady at 6400 and jumping to 7000 at times, but here is one of it at nearly 400u/s http://puu.sh/9pBlV/fd42653238.png

Please tell me what else you would like to diagnose this and come up with a resolution.

I was using the mod opis to track that at the time, just tell me what else you want me to use for you and I'll do it. As for TE it's handler is pretty high as well however when mekanism was spiking over 2000u/s last night TE remained at about 100-300 so not in the same range.

What software are you using to track that? Also, you said that the only thread that has an update time higher than 100 u/s is Mekanism, but it looks like TE's handler is reaching 180. Did you mean 200?

I just checked for my Server, Opis is saying the same thing for me. 7000+.us For Mek Handler.

Also Aidan, that update really did help with Packets :)

If you wouldn't mind tracing the source of the issue consider downloading visualvm (http://visualvm.java.net/download.html) and taking a snapshot of the CPU usage. This will help me quite a bit :)

Hi Aidan,

Using Meka V6.0.5.46 this appears to be related to this problem.

Here is some warmroast output with mekanism's handler consuming way more CPU than it used to. This appears to be the cause of my server locking up a few times per minute. I've expanded what appear to be the problem areas.

https://dl.dropboxusercontent.com/u/82904851/Warmroast_AE/WarmRoast.htm

VisualVM didn't give us much useful information.

Another warmroast taken.

https://dl.dropboxusercontent.com/u/82904851/MekanismOMG/WarmRoast.htm

#1 looks like OK behavior - the CPU usage is a bit higher because it's creating networks after chunk load.
As for the second...yeah, I'll see what I can do. That's not normal.

Hi Aidan,

Thanks for the chat in IRC.

Here's warmroast output with 11 Players online, problem area's expanded. It appears the first area is due to chunk churning. We are still looking for the In-Game area causing this. Scroll down and you can see the energy network as well.

https://dl.dropboxusercontent.com/u/82904851/Mekanism1404-3/WarmRoast.htm

Here is Opis output for the Network Handler at the same time (Running @ 16k): https://i.imgur.com/OMRWQqp.png

@Cozza38, would you mind running a few more tests on this when the server load is extremely high?

Alright! I have now done a few major refactoring that should cut transmitter usage by at least 50%. Will be present in v7.