Problems start with more than 22 players, Its pointless to use a full
dedicated for just 1 server.
As a side note, I performed an experiment where I put two 24 player
quickplay servers up on a P4 with 2 gigs of RAM, and there was no
problems running it and the gameplay didn't have any noticeable
latency. So, food for thought.
2011/7/19 Björn Rohlén<bjorn.roh...@gmail.com>:
And in this case, one really CAN question why this is so. I remember when
moving from Q3F/ETF (quake3/enemy territory based TeamFortress) to HL/SRC. I
had 2x24 servers on ONE p4-3ghz and 4gb ram and while both pubs were full
and packed with action, that ONE cpu would chew up ~85% cpu. Now, source is
based on the old quake-sources (they call it quake 1.5 since they got them
between ID did quake 1 and 2) and SURE, they've made all these nifty
improvments, such as directx support and what not ON THE CLIENT, but the
serverside seems to be from the RQ-era. There is NO justification for moving
around a bunch of clients on the same .bsp-format, doing the same work with
LESS entities compared to an hopelessly old infrastructure and eat 8 times
the amount of resources.
And getting worse with each release.
-The G
On Tue, Jul 19, 2011 at 11:54 PM, Kyle Sanderson<kyle.l...@gmail.com>wrote:
We've all in the same boat. There's not much you can do without
Valve's assistance, no one seems interested in helping. Both Movement
and Registration have gone to hell since the EP1 -> OB swap. I'm
constantly getting stuck in walls, Hitboxes are not lining up with
models. We're running on a E3-1270 (3.4ghz) with Turbo enabled. We can
get around 50 players before things get pretty sketchy. Performance
wise we're just about evened up with EP1 (Which was running on a
Q9550).
It's nice that it was ported to OB... It's absolute crap that the
useless cruft that existed in EP1 was ported to OB.
Kyle.
On Tue, Jul 19, 2011 at 2:36 PM, Evaldas<hlds_li...@evagames.eu> wrote:
X3220 is a 4 core CPU. First one of the quad core versions.
OrangeBox utilize only one core, so basically you are using
only 2.4GHz of all available. Try using the 100Hz kernel.
On Wed, 20 Jul 2011 00:27:07 +0300, Andres Pozos<javato...@yahoo.es>
wrote:
Hi:
Since the source engine migration to orangebox las year i noticed that
with each update the engine take more and more resources. A server with
more
than 20 slots take ALL resources availables.
My server:
Intel(R) Xeon(R) CPU X3220 @ 2.40GHz 8 cores
8GB RAM
Cpu(s): 90.4%us, 0.4%sy, 0.0%ni, 82.7%id, 1.3%wa, 0.0%hi, 0.2%si,
0.0%st
Mem: 8146960k total, 8034680k used, 112280k free, 143980k buffers
Swap: 6258680k total, 0k used, 6258680k free, 3473732k cached
Tried with Vanilla Centos 5,6, Debian 5 many different servers and all
have same problem. fps_max is limited to 70 but seems that isnt helping.
Is just me or theres any way to fix cpu usage?
Best regards
_______________________________________________
To unsubscribe, edit your list preferences, or view the list archives,
please visit:
http://list.valvesoftware.com/mailman/listinfo/hlds_linux
--
Evaldas,
EVAgames community
www.evagames.eu
_______________________________________________
To unsubscribe, edit your list preferences, or view the list archives,
please visit:
http://list.valvesoftware.com/mailman/listinfo/hlds_linux
_______________________________________________
To unsubscribe, edit your list preferences, or view the list archives,
please visit:
http://list.valvesoftware.com/mailman/listinfo/hlds_linux
_______________________________________________
To unsubscribe, edit your list preferences, or view the list archives, please
visit:
http://list.valvesoftware.com/mailman/listinfo/hlds_linux
_______________________________________________
To unsubscribe, edit your list preferences, or view the list archives, please
visit:
http://list.valvesoftware.com/mailman/listinfo/hlds_linux
_______________________________________________
To unsubscribe, edit your list preferences, or view the list archives, please
visit:
http://list.valvesoftware.com/mailman/listinfo/hlds_linux
