Theory Based Hosting [Windows Server 2008 64-bit]

[BehaartesEtwas]

my servers are running with 66 fps for a couple of month now, nobody noticed a difference when I changed it.

[Zero1028]

I just noticed this, I revamped a couple things and i am running at 33 tic / 66 FPS and I get 250-300 FPS average, the hitbox is amazing and reg as well.

- Travis

[Hazz]

I'm interested about how the FPS doesn't make a difference any more, is it since the move over to the OB engine?

[Zero1028]

I just wanted to update you guys ...

L4D Server is using 0.02-0.04% CPU Usage... Thats with the settings maxed and on versus.
CSS - 64 Slot Server 500 FPS, 100 TIck, 0.30 - 1.33% CPU Usage at 50% Full Capacity. 0.33 - 0.74% with 63 Bots.
TF2 - 20 Slot Server 500 FPS, 100 Tick, - 0.74-1.20% CPU Usage at 100% Full Capacity.
L4D2 - Sames as L4D but with .5% Increase

.. So I can honestly say this machine is probably overpowered considering that all servers were running at the same time and at near full capcity and I was not using over 10% Total CPU Usage. So I will be with 40 Gameservers under 50% Total CPU Usage.

[Zero1028]

Btw, the monitoring has been done on extensive number of monitoring software programs both provided by OS and SuperMicro. From what I am getting from these readings is,

1 CSS Server 20 Slots 500 FPS 100 Tick averages

And this averages 0.0007 Ghz CPU Consumption.

0.0007 x 32 Game Servers = 0.0244 Ghz CPU Consumption ...

And I have been testing this for 4 days now and I keep getting the same readings even with the servers live. So either I am doing something wrong or I am correct. Either way if this is correct I will be able to host more than 64 CSS 20 Slot 500 FPS 100 Tick game servers without noticing any big CPU hog...

The total CPU consumption for 64 game servers would in fact be,

0.045 Ghz Total ...

[BehaartesEtwas]

(01-13-2011, 01:23 AM)Hazz Wrote:  I'm interested about how the FPS doesn't make a difference any more, is it since the move over to the OB engine?

http://wiki.fragaholics.de/index.php/EN:Server_FAQ#How_important_is_it_to_run_a_srcds-based_server_with_high_fps.3F

[Hazz]

Thanks for that, still running my CSS servers at 500FPS! Hopefully lowering it will save me some resources.

[joe.powell]

Just changed all my servers down to 66FPS, and didn't mention it to anyone.

So far nobody has noticed a single thing has changed. Success tbh.

[realchamp]

(01-14-2011, 07:24 AM)joe.powell Wrote:  Just changed all my servers down to 66FPS, and didn't mention it to anyone.

So far nobody has noticed a single thing has changed. Success tbh.

Theoreticly it shouldn't affect anything. As long as it doesn't go below 66fps
The tickrate depends on it being atleast the same as fps, so incase the fps is 40 - the tickrate would be 40 too

[Hazz]

Set all our servers to 100FPS, all servers running 66 tick. Going to see how it goes.

[Zylius]

Thank you guys. I was at my ISP where DataCenter is and my server is. Changing the network card of my humble server athlon x2 2,5GHZ and 2GB of ram Running windows 2008 R2 and 3 srcds boxes. Then i remembered this thread, and that fps doesnt matter any more. So i just stopped the fps boost. ( I could have done it remotely but... i don't know ) Then i came home and looked at my CPU graph i was like "Holy sh1t my server is probably completely EMPTY! what happened" And then i looked and were was actually ~45 people playing on the server. I was stunned . Now i can host more servers. Fps boost is bullsh1t. (the spike at the very end of graph is psychostats calculating)

[Monk]

It's because when you boost a server, Sleep(1) on windows gets expensive because you need more CPU time to make sure your clocks wake up.

FPS boosting was good at one time, but not any longer. It doesn't do shit except suck up CPU, and causes people to cry, piss and moan when they cannot kill anyone.