Problem Server Can't Keep Up. 32-bit System.

[Blanq820]

Hey guys, for the Mindcrack Pack FTB server I'm hosting on my computer. I keep getting the message, "[WARNING] Can't keep up! Did the system time change, or is the server overloaded?" spamming every 20-90 seconds. The server isn't lagging too bad most of the time, but there is never really a time where its not lagging at least a little bit; it made farming a creeper shard rather interesting.

My question is if there is any way to improve the server performance on a 32-bit system. What I'm using is:

Windows 7 x86
AMD Athlon II X4 645 Processor (3.1 GHz)
4 GB RAM
Java SE Runtime Environment (build 1.7.0_17-b02)
Java HotSpot Client VM (build 23.7-b01, mixed mode, sharing)

I haven't done anything to the server batch I got from downloading the mindcrack pack, so everything there is as-is.

Thanks in advance for the help!

 

[Francis Baster]

There's a new Java update already? Anyway if your server was not lagging to begin with, it's possible that your setup is not the issue. It could possibly be too many item entities in loaded chunks, chunkloaders, or players generating new terrain. Have you investigated these common causes of lag? The aPerf mod can give you detailed information on currently loaded entities, while the /chunkloaders command from ChickenChunks mod (which comes with MindCrack and Direwolf20) can give you information about the quantity and location of chunkloaders.

If not, I'm sorry I don't have any Windows performance tips to give you. Hopefully someone else can contribute

 

[glasspelican]

I dont know if this will help enough
open your startserver.bat file and add to the end of the first line "-server"

 

[Blanq820]

well, ive taken down all but one chunk loader and disassembled all the quarries. its helping a fair bit. as for the entities, theres nothing worth metnioning in that dept.
i also added the -server at the end of the line, took a while longer to bring up the server, but it seems like theres less of the cant keep up spams.

thank you both for your comments, they have at least helped out a little bit. i think the big problem may just be javax86's limitations.