OK, so I'm wondering if some people could share their experiences with me
on this issue. I'm running a TF2 srcds server... and I've tried it on
Ubuntu 10.10, Gentoo, Windows Server 2008, and Windows Server 2003, yet I
get the same results in FPS. I've been noticing on my server that FPS can
Does anybody know about what type of hardware it'd actually take to run a
24 person TF2 and a 12 person TF2, with 5 sourcetv slots (including
constant 2 recording slots), and to run hlstatsx:ce for the 24 person TF2
on the same machine, but have the web server and database on a different
Does anybody know if these problems occur for the old engine, like with
CS:S (non-beta)? I'm also thinking if this is an issue with the
TF2/orangebox engine, the new CS:S will probably be affected in the same
way.
___
To unsubscribe, edit your list
That's the problem though... We can have a freshly formatted machine,
install practically any OS to it... nothing else but the OS and the srcds
stuff... and load up 24 slots, fill them, and that fps does drop below 66
at times; with no addons, plugins, stats, or anything else going on.
We're renting a VDS from here:
https://www.nfoservers.com/order-virtual-dedicated-server.php
It's the first plan, so... 1 GB RAM, plus a HT'ed Nehalem core.. probably
at least 2 GHz. They also moved us to different hardware and it still does
the same thing.
You think there's too much scheduler latency even on a Xen? I'm pretty
sure they're using Xen. I use either 'rcon stats' or fpsmeter.org for the
main part, which uses the same technique. We tried putting a 24 slot pub
TF2 and a 12 slot private TF2 on it and it pretty much choked...
Yeah, we went with this host on VDS expecting to be able to run one 24
slot pub and a 12 slot private and even when they were full have no
problems... which is clearly not the case with the 1 core VDS package.
On Wed, 10 Nov 2010 15:39:03 -0600, ja...@pctoolbin.com wrote:
Well, I had three
Aaron, you get that on the single core plan? Because with just a single 24
slot running on ours, with no non-OS processes, on Win 2003, here are
actual readings from 'rcon stats':
514.59, 85.32, 504.91, 54.04, 63.97, 42.70, 56.95, 60.04
The host temporarily gave us another core, but it
Yes; SourceTV is a requirement for us. But upon disabling it, I don't see
much, if any, difference.
We will play with the threads and cvars and reply back.
On Wed, 10 Nov 2010 15:05:28 -0800, John lists.va...@nuclearfallout.net
wrote:
Also, you're not running a SourceTV in there at the
Are you saying its going to always drop below the tick rate on a VDS,
direct hardware, or both?
On Wed, 10 Nov 2010 18:16:49 -0500, Gary Stanley
FPS is measured by syscalls. Because of context switching, interrupt
overhead, syscall latency, and scheduler latency, you'll always have
10 matches
Mail list logo