Jorbtar - THEIR ALIIIVVVEEE!!!!!

[dagrimreefah]

I believe this article explains it all; very informative.

Had you already encountered this article: Jonathan Corbet, The trouble
with stable pages, March 13, 2012 (http://lwn.net/Articles/486311/) ?

It might be the "newly added" stable pages feature added to the kernel Loftar upgraded to.

[Grable]

Thanks for still trying to fix this, loftar.

[Ogrim]

Been lurking, signing on to the forums about once a week...

Just logged on today and crapped my pants.

Loftar... Jorb...


THANK YOU.

[gummi]

Yay! Hope!

Thank you

[bot1]

lags are fixed and w7 pls

[Wipeout]

Game wasn't intended to have world resets, for example W5 ended because of loss of world data or something like that.

[overtyped]

He won't reset yet, theres too many things he needs to fix, for example bug ram, and making claims despawn after a certain period of time. otherwise the next world will turn into the same cesspool of random villages that are 100% abandoned but still claimed.

[dagrimreefah]

I don't know if anyone else noticed but at least so far today seems a SHITLOAD less laggy.

[loftar]

For those who are interested, the last lagfix consisted of a patch I wrote for BDB to enable me to put its region files in a different directory, so I put them on tmpfs. That avoids pdflush going in and flushing dirty region data to disk, and therefore avoids the long disk queues that it caused, and also prevents those pages at all being under writeback and therefore subject to stable-page-wait. I'm still strongly considering patching away stable pages in the kernel, as well, but as it looks thus far, this seems quite good for what it is. Also, I hear that Linux 3.9 is supposed to do away with I/O-waiting for stable pages, so I might just wait for that instead of patching. It should be coming out rather soon anyway.

[Ninijutsu]

Game wasn't intended to have world resets, for example W5 ended because of loss of world data or something like that.

It is intended to reset with huge game updates.