TEST: 128+ players, with some changes, Saturday 10th August (10.8.13) 7PM BST (GMT+1)

[comp_uter15776 1]

Hello all,

Let me just say I'm not claiming this is an uninspired idea, I think Blue1 hit something on the nail with the experimenting. I'd like to do something similar, but with a twist.

Firstly, the server we use is a linux via WINE emulator, so it will be a nice comparison to view how a sub-linux environment compares with a windows DS, considering it's not technically native.

I'd like to try push the boat out and see if we could break 128 players. The gamemodes would be a Sector Control (upscaled with say, 75 on a side, so in fact space for 150), and an AAD where BLUFOR protects the Independent forces for 2 hours from OPFOR. St0rm's servers should be able to cope, they are also able to cover any other demanding game of today, and again it would be interesting to compare specification to performance and then match it up to Blue1's readings. Both gamemodes features vehicles and Tonic's VAS, so we would gain a greater understanding of the playability when vehicles are used heavily.

St0rm.net's server specs:

OS: Debian 7 (Wheezy)

CPU: Intel Xeon E3-1245 (3.4GHz), 8 cores

RAM: 32GB

HDD: 4TB

Connection: Gigabit line

I believe the maximum AI would be around 45-55, as AI is removed for all but the AAD gamemode (where you have to defend them, which is what gives BLUFOR a challenge! >:D)

Each gamemode would test a different area of ArmA conflict, as the AAD mode is heavily centred around one area and thus is for a close battle with a lot of action in one spot, whereas the Sector Control mission spans the whole island and should give us an idea of a playthrough on a wider scale but with less action at once, more mini-battles instead.

So, the event details:

Saturday

7PM BST (British Summer Time, which is GMT+1)

3rd August 2013 10th August 2013

Server name: [st0rm.net][uS] - ArmA3 Beta STABLE 150 Player Test

STABLE BUILD SERVER, please change your builds over to Stable if you are on Dev via Steam->ArmA 3 Beta->Right click->Properties->Beta->no beta OWTTE->Close

This will be an open server for testing, there will be a couple of admins to check everything is ticking over. Please post suggestions and/or feedback in this thread, I will be monitoring server stats closely throughout, as well as chat.

Regards,

Harry

P.S. We run a teamspeak server (512 slot) at ts.st0rm.net so please feel free to join that and express your emotions. It is hosted on a different server so it will not be affected by and untoward lag or problematic server behaviour.

Results are in - Although we didn't reach the same numbers as Blue1, the event was fairly successful.

We peaked at 20 players and the server coped extremely well with all, SFPS never dropped below 100 and ping was roughly 100 for all members (below that in some cases), which isn't bad for the server being in canada. Whilst this test went well, I can't declare it as conclusive purely due to the fact that the number of players was not high enough to pose a stress test on the server.

However: If setup correctly, it would seem that Linux + WINE is a very efficient combo.

Edited August 11, 2013 by Comp_uter15776

[Blue1 10]

Hi Harry,

nice one =)

Suggestion: to really compare, i'd like to see our actual Blue1CTI Mission runinng a while as well?

If you want i can provide you with everything you need to do that.

Also i think it would be a good idea to use @ASM as i did.

But i am not sure if it is running at WINE? But i will ask Fred41 ASAP.

Let me know here or via PM if i can support you anyhow!

The more Results we can provide to "open up the eyes" of BIS and their MP issues - the better =)

Edit* and this is a good chance to proof what Dwarden seemed to implicate ("DEV-branch is x-times faster...") ;) sorry -could not resist, again^^

Best regards,

Blue1

[comp_uter15776 1]

Hey Blue1, thanks for the reply.

I'll happily stick it on for easier comparison. Just send me what I need and we'll implement it into the cycle.

We have initiated use of ASM on it, however, as expected, we have ran into trouble. Currently I have a friend of mine looking into it but in any case hopefully we can get some metrics out of it.

And indeed, more results gives a greater reliability on these tests :D Including checking the dev branch, it is partly on that to test it vs stable, but also because the gamemodes work better on it atm >:D

Regards,

Harry

[Kaijev 1]

I'll try and make it if I'm available then

[stefeman 1]

I'll definitely join.

I try to bring my friends too :)

Thank you for trying this thing out!

[comp_uter15776 1]

Great, thanks guys. The more people that are present the better it is for testing!

[maddogx 13]

Thread title edited to reflect the changed date.

[comp_uter15776 1]

Thanks MadDog, I was wondering if that'd happen or not xD

[comp_uter15776 1]

Update: This should still be going ahead and I've reformatted some missions to 75vs75 so despite whatever the ingame notification says (I think for some reason it limits to 64 max even if you have more slots?) We can actually fit over double that on :P

[SnR 1]

Good luck we are all watching.:cool:

[comp_uter15776 1]

Thanks SnR :D

Last post before the event (hopefully): As far as timing goes, I plan to officially begin the test as soon as the livestream from BIS is over :D <Stops more timezone confusion>

[stefeman 1]

Can we get IP for remote/direct connection?

[comp_uter15776 1]

arma3.st0rm.net:2302

UPDATE: ArmA 3's Livestream has been postponed by roughly 1 hour so we will have to by 2 unfortunately. That's just the way it is :( So, you're looking at about 9PM if you're in the UK to start as the livestream is now predicted at 7PM

[stefeman 1]

How many hours from now the server starts up?

I'm running stable version, yet im unable to see the server in listing xD

[comp_uter15776 1]

EDIT: We are go!

---------- Post added at 20:20 ---------- Previous post was at 18:36 ----------

Oh, the address is arma3.st0rm.net:2302

Edited August 10, 2013 by Comp_uter15776

[SnR 1]

Who joined and how did it go ?

[comp_uter15776 1]

Numerous groups and people, I shan't go and post them all here but the original post was updated with a bit of info. The event went really well server-wise and linux/wine didn't show to have any problems, but then the test wasn't as straining as it could be.

[Blue1 10]

Hi again,

sorry i couldn't take part or help that much due to very much RL stuff the past days.

But, please could you explain this one (from your opening Post/results) a little closer?

the server coped extremely well with all, SFPS never dropped below 100 and....

AFAIK maximum ServerFPS-Value is capped at a maximum of 50?

What measurements / tools etc did you use? Got any @ASM Screenshots? CPUs usage-infos?

So if i get it right, you had only 20 as the max playercount online and playing?

I think, as you already mentioned, this isn't enough to "guess" about a proper result on higher loads (i.e. +60 players + AI etc)

EDIT: why did you decide to run this on the Stable Branch instead of the previously announced DEV?

Maybe you could set those maps with 100+ Players up again and let all measurements / statistics run a while, until you get the server full enough?

best regards,

Blue1

Edited August 12, 2013 by Blue1

[comp_uter15776 1]

Hi again,

sorry i couldn't take part or help that much due to very much RL stuff the past days.

But, please could you explain this one (from your opening Post/results) a little closer?

AFAIK maximum ServerFPS-Value is capped at a maximum of 50?

What measurements / tools etc did you use? Got any @ASM Screenshots? CPUs usage-infos?

So if i get it right, you had only 20 as the max playercount online and playing?

I think, as you already mentioned, this isn't enough to "guess" about a proper result on higher loads (i.e. +60 players + AI etc)

EDIT: why did you decide to run this on the Stable Branch instead of the previously announced DEV?

Maybe you could set those maps with 100+ Players up again and let all measurements / statistics run a while, until you get the server full enough?

best regards,

Blue1

You are correct. I couldn't get any screenshots however as WINE is a bitch like that xD (I was able to see them though). As for the SFPS - I accidentally put 100 in twice ;s 100 should be ping, SFPS was never below 45 ish except for one time when I had the map reload and it dropped a tad more.

There were a maximum of 45 AI and 20 players at the highest player count, but the numbers never increased above that (BIS just so decided to make their livestream and then their not-so-livestream coincide with my event :( )

For the branches; it was purely because more people were on it and it'd be a better representative of what the final build would be like (as dev rapidly changes). I think also your server's was on stable too.

I could set the maps up and everything again but it's just a matter of getting people in there ;)

[Blue1 10]

Screenshots: uhm,..okay :confused:

Pardon, ..but, yes as you say ...45 AI...and 20 players nothing to with loadtest, to be honest.

Of course SFPS wouldn't ever drop at that values.

Make it 450 AI or something, for example, if you don't get more Players in.

Yes -your timing was bad luck i guess.

I think, the current state of DEVbranch is more likely to be representative to the final, than the "actual"( not to say old) Stablebranch.

At my Test, the big-Beta-Update was 2 days young, if i remember right...

Thats why i chose it (and therefore you are right: more players).

My expirience is, if there is a server up with 100+slots shown at the gamebrowser (yes it does, after the map loaded, if it is set correctly!)

and the first people are online, more will follow automatically - just to find out whats happening there :)

goodluck!