TheBuzzSaw
Comments
They stopped publicizing it? I grabbed it straight off my download page just like every other server release.
https://s3.amazonaws.com/Minecraft.Download/versions/1.7.2/minecraft_server.1.7.2.exe
This is still a problem in release 1.7.2. The ONLY text I get (Windows 7 64-bit running Oracle Java 64-bit) is the "Memory use" and the "Avg tick" (and their respective values). Everything else is solid white.
The problem seems gone for me. I'm still running the same hardware with Mint 14 MATE and OpenJDK (64-bit). Again, I have no idea if it was a kernel update, a Java update, or a Minecraft update (or combination of those), but the CPU only spikes briefly when loading world chunks. General gameplay falls below 100%. So, for what it's worth, I have not observed this problem in 1.5.1.
I hope it is fixed for everyone. It is an issue worth monitoring. When the problem was happening, my system became unresponsive. It really was using 100% at the expense of everything else. I couldn't even drag a Nautilus/Caja window across my desktop; it was so bad.
I just installed Mint 14 64-bit. (I abhor Unity, but I love Ubuntu repos.)
I am running Mint 14 MATE with no compiz. So far, no 100% CPU usage issues. It spikes when (presumably) it has to load new areas, but general gameplay is fine. I will begin enabling certain desktop effects/compositors and see if that affects the CPU.
Good news. The update 1.4.5 seems to have fixed it. Granted, I don't know whether it was Minecraft's update or one of my system updates to Ubuntu/NVIDIA that fixed it. It patches so often. I will keep an eye on it. I think the real test will be in making a new world and playing some heavy duty multiplayer.
I did manage to attach VisualVM (profiler) to it. Sadly, it made the game unplayable. XD But if you ever want a snapshot from that, I can post it here.
It is using 100% all the time. I can stand still, and it will stay at 100%. I can walk around, and it will stay at 100%. It will periodically shift which core is maxed, but one of them is always maxed. And yeah, as noted in my specs above, my computer is no slouch. If I run the game in Windows, this problem does not exist.
I updated the environment to clarify: I am indeed running 64-bit.
And yeah, I figured you had sleeps in the code. (If you didn't, the Windows and Mac versions would be crying out loud right now.) This makes me wonder if there is something wrong with the Linux JVM and/or its handling of timers/sleeps.
And yes, I've updated my LWJGL. Are you referring to one beside the one that the Minecraft client manages/updates?
I've allocated 2 GB and 4GB to the client with no change. One core hits 100% and stays there. Even if I stand still and do nothing, it stays at 100%.
I am a game developer myself.
<StabInTheDark> This behavior presents itself when there is a loop in the code with absolutely no sleep calls. Oftentimes, when developers build their first game loop, the game eats 100% needlessly because the application never relinquishes control of the CPU. So, there is likely a loop or two in the code that would really benefit from a 1 millisecond sleep (or even 0 millisecond sleep). </StabInTheDark>
Good to know. I'll give the raw JAR file a try tonight.
And yes, the server works fine. I just miss being able to interact with that console and see log entries. Others connect just fine, and I see their names in the lower left box.