Server stress test
Warboy
Join Date: 2012-09-05 Member: 157929Members, NS2 Playtester, Squad Five Blue, Subnautica Playtester
Join Date: 2012-09-05 Member: 157929Members, NS2 Playtester, Squad Five Blue, Subnautica Playtester
This discussion has been closed.
Comments
Clockspeed isn't important. But I did add location so everyone knows atleast. Classifiedcomputers.net [SW-Michigan]
Well, in fact it is. Trust us! :P
Considering my server gets the same ticks as 4.4Ghz servers. No, it is.
That would truly be revolutionary.
That would truly be revolutionary.<!--QuoteEnd--></div><!--QuoteEEnd-->
Seems to be the case, because it was full for 2 hours yesterday and it didn't go under 28 tickrate typically, It would drop below 28 when I raised the player cap.
When you say "like a 4.4GHz machine", you are assuming that a 4.4ghz machine will start to slow down at the same time a slower server would, but it would in fact handle higher players before it started to lose tickrate.
You can tell the tool to draw graph of a smaller period and then you can select specific ticks from which you'll see cost of each PROFILE group. Server owners and devs should focus on data from later parts of the round when there is more of everything and bottlenecks really show.
<img src="https://dl.dropbox.com/u/6615362/perfanalyzer.jpg" border="0" class="linked-image" />
It didn't appear costly on my client. Even if it'd cost 1 tick out of 30, it'd still be worth it to know when and why server is slowing down.
I'm basing mine off experience with hosting game servers in general. Because sorry but TF2 is more demanding on clock speed since it's mostly single threaded.
But I'm gonna do a test, I'm thinking about setting up 2 servers.
1 is a Core 2 Duo running at 3.33Ghz, The other is the main server which 2x Xeons running at 2.66Ghz, and we'll see what one lags first.
Which xeon are you running at 2.66Ghz? If you are running an older model X5355 which is comparable to the core 2 duo (65nm), then I can almost guarantee the core 2 duo will perform better.
If you are comparing one of the newer Xeon's running at 2.66GHz, then it will probably beat a core 2 duo, but this isn't the comparison you need to be making. When you said it performs the same as a server that is running at 4.4ghz, these are either on ivy bridge or on sandy bridge, which is much newer than a core 2 duo, and clock for clock, the ivy/sandy will perform better against your xeon running at 2.66.
If you compare an i5 2500K, running at stock speeds of 3.33GHz, and compare it to your xeons running at 2.66, then the 2500K will win.
I'm not saying that 2.66ghz is not enough, or your server won't perform. UWE have been making large improvements to the server code, and up until recently, anything below 4.0GHz would only run a 12 player server reliably, but it seems this has changed.
Which xeon are you running at 2.66Ghz? If you are running an older model X5355 which is comparable to the core 2 duo (65nm), then I can almost guarantee the core 2 duo will perform better.
If you are comparing one of the newer Xeon's running at 2.66GHz, then it will probably beat a core 2 duo, but this isn't the comparison you need to be making. When you said it performs the same as a server that is running at 4.4ghz, these are either on ivy bridge or on sandy bridge, which is much newer than a core 2 duo, and clock for clock, the ivy/sandy will perform better against your xeon running at 2.66.
If you compare an i5 2500K, running at stock speeds of 3.33GHz, and compare it to your xeons running at 2.66, then the 2500K will win.
I'm not saying that 2.66ghz is not enough, or your server won't perform. UWE have been making large improvements to the server code, and up until recently, anything below 4.0GHz would only run a 12 player server reliably, but it seems this has changed.<!--QuoteEnd--></div><!--QuoteEEnd-->
Core 2 Duo E6600 @ 3.33 vs 2x Xeon E5430 @ 2.66
Sorry, but this is pure nonsense @ the current state. (218)
<!--quoteo(post=1972796:date=Sep 6 2012, 04:09 PM:name=Warboy)--><div class='quotetop'>QUOTE (Warboy @ Sep 6 2012, 04:09 PM) <a href="index.php?act=findpost&pid=1972796"><{POST_SNAPBACK}></a></div><div class='quotemain'><!--quotec-->Clockspeed isn't important.<!--QuoteEnd--></div><!--QuoteEEnd-->
This tells me, you have no clue what you are talking about.
Sorry, but this is pure nonsense @ the current state. (218)
This tells me, you have no clue what you are talking about.<!--QuoteEnd--></div><!--QuoteEEnd-->
Yes I am, It only dips below 28 when there is lots of whips/hydras/turrets on the map. I've had several people stress test it the other day. But I'm gonna try the 3.33Ghz Dual-Core 2nd server for the time being.
You sir need to stop assuming. I've been hosting servers since HL1.
<!--quoteo(post=1972922:date=Sep 6 2012, 06:31 PM:name=Warboy)--><div class='quotetop'>QUOTE (Warboy @ Sep 6 2012, 06:31 PM) <a href="index.php?act=findpost&pid=1972922"><{POST_SNAPBACK}></a></div><div class='quotemain'><!--quotec-->Seems to be the case, because it was full for 2 hours yesterday and it didn't go under 28 tickrate typically, It would drop below 28 when I raised the player cap.<!--QuoteEnd--></div><!--QuoteEEnd-->
to:
<!--quoteo(post=1974109:date=Sep 8 2012, 08:50 PM:name=Warboy)--><div class='quotetop'>QUOTE (Warboy @ Sep 8 2012, 08:50 PM) <a href="index.php?act=findpost&pid=1974109"><{POST_SNAPBACK}></a></div><div class='quotemain'><!--quotec-->Yes I am, It only dips below 28 when there is lots of whips/hydras/turrets on the map. I've had several people stress test it the other day. But I'm gonna try the 3.33Ghz Dual-Core 2nd server for the time being.<!--QuoteEnd--></div><!--QuoteEEnd-->
So your miracle 2,6 ghz server can handle the game? fine
Why you switch to the 3,3ghz server then?
Btw. the 3,3 ghz server cant handle 18 slots aswell.
I know one player with an miracle client. His cpu @stockspeed (3,4ghz) never goes down under 60fps.
Its amazing, cause every client, really EVERY client goes down to 30-40 fps in lategame. No matter what ghz they have.
to:
So your miracle 2,6 ghz server can handle the game? fine
Why you switch to the 3,3ghz server then?
Btw. the 3,3 ghz server cant handle 18 slots aswell.
I know one player with an miracle client. His cpu @stockspeed (3,4ghz) never goes down under 60fps.
Its amazing, cause every client, really EVERY client goes down to 30-40 fps in lategame. No matter what ghz they have.<!--QuoteEnd--></div><!--QuoteEEnd-->
No, Because if you know anything about design, adding lots whips, hydras and turrets causes the server to do MUCH more calculations then in normal gameplay.
Because most endgame cases might have a 1/2 dozen whips/hydras/turrets, during the stress test, there were alot.
Sorry, but things like that pissing me off.
Sorry, but things like that pissing me off.<!--QuoteEnd--></div><!--QuoteEEnd-->
what?
That didn't make any sense....No offense.
Show me and the rest of the serveradmins, that an 2,6 ghz server can hold around 28 ticks with 18 slots. (btw. the 28 ticks are an bug)
I'll be very interested in seeing this. It would also be nice having a replay/stress test tool for servers.
<!--quoteo(post=1974156:date=Sep 8 2012, 05:59 PM:name=dePARA)--><div class='quotetop'>QUOTE (dePARA @ Sep 8 2012, 05:59 PM) <a href="index.php?act=findpost&pid=1974156"><{POST_SNAPBACK}></a></div><div class='quotemain'><!--quotec-->I want to play on your miracle server.
Show me and the rest of the serveradmins, that an 2,6 ghz server can hold around 28 ticks with 18 slots. (btw. the 28 ticks are an bug)<!--QuoteEnd--></div><!--QuoteEEnd-->
Yea, Good job not proving your point and once you got jealous, You decided to DDoS my server.
I doubt that dePARA has a reason to be jealous, since he runs the most populated servers in the EU.
And if you cant provide any proof for your accusation, i would also be very careful making a statement like this in a public forum.
And if you cant provide any proof for your accusation, i would also be very careful making a statement like this in a public forum.<!--QuoteEnd--></div><!--QuoteEEnd-->
I can provide proof it was indeed DDoS'ed, and not just laggy or whatever.
Server tick stay high, while players were getting 999 ping, then when I noticed the DDoS, I deployed counter-measures and have a firewall log. and a chat log of what he was saying in the server.
Server tick stay high, while players were getting 999 ping, then when I noticed the DDoS, I deployed counter-measures and have a firewall log. and a chat log of what he was saying in the server.<!--QuoteEnd--></div><!--QuoteEEnd-->
That is still no proof that he has done anything about the DDoS...