TEST: 128 players? possible at Arma3? Sunday, 28.7.13

Yeah I think 64 players is pushing it for most servers. In order to host a 128 player server youd have to have a beast of a server with lighning-fast internet. I'm talking like dedicated server CPUs (Intel Xeon) and 1-2GBps internet.

It also depending on the mission and custom addons.

In A2, I played in a Berzerk server with 90+ players on it without any issue.

I tried the test. I joined when there was about 77 players on and it seemed totally fine for about an hour. I logged off and came back on 1/2 hour later when it was at about 122 I believe (or at least over 100) and there were vehicle lags/warps but the infantry seemed ok. I got bounced soon after I think the server restarted.

Anyway there is hope!! Good luck!

pls read this first

@Papanowel: thats exactly what i say! A well optimized Mission, running on a Server with the correct settings - can make A3 also a 128player expirience.

Biggest problem is - and stays -BIS is not going to optimize Multicores or Threading ...i mean... you will see after i posted my results, ... it is so sad :( u have a bigfcking machine of a server and the game itself isn't using its capabilitys.

Ofcourse, regarding to professional hosters it is way better to use only 20-30% of CPU-capabilitys. ...but i guess, it is the community bringing up the game! ...and buying/renting Rootservers and getting "rewarded" with the crappy usage of servers performance.

Since now, no DEV answered to real solutions (oh, do you know you have to buy a SECOND COPY of A3 to get a HeadlessClient to run?, Funny, eh?....) if they are going to listen to hosters/editors... or not or just tell them to ...err..okay,..you know...

Would make the final difference.

I understand Xeno for leaving. But i am not giving up for the moment.

(no, i won't explain, google for it, if you want - but he is right in everything he mentioned, for years and BIS had time enough to fix it.)

A3 will never get that popular as, for best example BF3 - if they make it so much fcking complicated / or lets say, if they don't share knowledge to modders/editors who they, at least they say - so much want / need!

and yes i will prove this message (ask RWJ ;) haha)

Testresults added - see first Post.

cya

Yes thanks for sharing the data very helpful.

Hi Dwarden,

yes - of course! I want to supply a Stable Map / mission to the biggest number of players possible. And they aren't mostly on DEV, sadly.

So i decided to take my Stable-Server-branch for it, as we got provided by a nice Beta update lately- which you state in your REPs - won't be changed so much at the Core-Engine anymore, ...

But i would like to see some more Feedback or maybe even Hints (especially from you) to the results and conclusions (or lets say, some more replying to the now given/proven facts of things that you left us in "guesswork" before)

instead of "a one-lined sentence, trying to reduce the efforts we took - just by saying it could be the "wrong" branch and with that implicating all is 500%better/faster MP at DEV,........."

your turn.

*popcorn*

Thanks and best regards,

Blue1

Hi Blue,

thanks for making this large test possible and for sharing the results.

Such a large scale test is very helpful for understanding the mechanic and to come to know the limits of the current engine.

Especially the limiting impact of the script processor is well shown by the results.

For me, the performance of the current engine (stable branch) is already very impressiv and i am hopeful that it can be improved a bit more in the release.

The problems we currently observe on very many servers are mainly the result of not well optimized mission script code and/or bad server configuration.

I think what arma 3 really needs, to have a better performing server/mission landscape, is:

1. better documentation (especially engine internal details are very bad or not documented), detailed information/warnings for performance hungry script commands

2. better development tools (what about a profiler for sqf?)

3. better support for monitoring tools (like ASM:)), i would recommend a simple shared memory interface, filled periodical with engine states currently not available or difficult to monitor,

like network bandwith, number of running VM scripts e.t.c.

But i fully understand, that some things simple need time and man power.

So i am hopefully that BI-developers are already aware of the problems and maybe something is already WIP.

Best regards,

Fred41

Hi Fred,

at first, again a very big THANK YOU, for helping me withthe Tests,

providing me with those Scriptchanges and Optimizations of my - now our - Mission and,

last but not least, the very helpful ASM!

I totally agree with your statements/suggestions above, and i think - most other Missionmakers etc. do as well! Or?

Especially the limiting impact of the script processor is well shown by the results.

THIS is what i wanted to approve and i am (still) very disappointed as none of the Developers made a usefull/helpful/mysterious hint what we maybe could expect in the near future.

But i think, as its already stated in the actual Sit/SpotReps - we can't expect anymore major changes to the Main-Engine.

As such things usually have to be fixed (pre-)Alpha ...and now in late Beta it could maybe lead to other problems - which would maybe prevent a timely Release.

this makes your statement:

1. better documentation (especially engine internal details are very bad or not documented), detailed information/warnings for performance hungry script commands

even more important.

So, why the hell - if BIS says "modders&mission makers are so important to us"...

why isn't there a lot more "do's and don'ts" regarding to avoid heavy errors/faults in scripting, leading to better, faster codes / better, more playable MP missions...

I know, keeping a good and usefull Documentation is a lot of work.

But compared to "make a Documentation after you are done" .... it is only "half the work", if you keep it up right at the start and while the whole progress of a Project!

3. better support for monitoring tools ...

Yes of course this would be helpful!

But if all important informations were available in a Documentation - Contentcreators wouldn't have to measure/find out "what is happening when and why or why not"...

But i fully understand, that some things simple need time and man power.

So i am hopefully that BI-developers are already aware of the problems and maybe something is already WIP.

Yes, sure i am aware about that too.

*And no, as i already said before, i don't believe that BIS can/will change anything of all those things and wellknown "problems".

Why? Because they haven't done it since Arma2.

*in this special case i'd love to be proven wrong

best regards,

Blue1

New test proposed, 128+ this sat.

Link: http://forums.bistudio.com/showthread.php?160391-TEST-128-players-with-some-changes-Saturday-3rd-August-(3-8-13)-7PM-BST-(GMT-1)

Perhaps we can compare the stats aftwards... Linux+WINE might even be on par with original Windows!

a few hundred "beta-testers" joined the test. (around 500 different GUIDs joined!)

thank you very much for that.

but to be honest - i am quite disappointed.

i am, one the one side, more disappointed about that BIS developers ... not taking care or whatever community is trying or communicating / reveal / ....whatever.

and on the other side - of this community. no feedbacks? no sharing of of hardware-specs and personal opinions of this test?

Are there so little ppl on this forum? i mean, the usual steam-customers aren't really interested that this is BETA and some feedback could be provided for the much more cheaper price of a game.

but i guess - they care as much as BIS does for MP community.

whatever - after around 36hours i got the first messages where my servers are - why they aren't online and so on,....

answered in short: i am pissed off.

ask me why - if you want to.

best regards,

Blue1

---------- Post added at 01:29 ---------- Previous post was at 01:03 ----------

and yes - i am aware that BIS won't even notice that my servers are down - or what i have / we tried to provide.

But - i am not sure if they are going to be a serious company, $driven - hunting for success, having a management at all.

for the moment, it looks to me as they are a garage-band, trying and trying to get professional.

but if i am allowed to make an advice: get a communitymanager if you rely on a MP or contentcreator-friendly expirience (as you say).

for the the moment - you aren't communityfriendly, as you don't provide us with informations or hints or documentations that are usefull to raise the immersion of your work.

thanks - and i will take the forum-ban as an (cheap and easy) answer too.

*still popcorn*

still your turn - regarding an even more ambitious customership than 5-10 years ago.

best wishes

Grab a Snickers.

Seriously I don't know what you want. I read you post a second time, but I just don't get it.

You say BIS isn't community friendly? Really?

You say there is no documentation? Well, there is. I just assume you mean for scripting, but you could be talking about anything.

And BIS isn't community friendly? Show me a developer who is more in touch with its community than BIS. This forum is the best example.

I don't know what you want to accomplish with this post, but you come off as a whiny child.

So I guess single core processing is where it's at now-a-days? Back to the ol' Pentium 4!

The FPS on the server was fine all the time, but it started lagging when there were a lot of players.

Grab a Snickers.

Seriously I don't know what you want. I read you post a second time, but I just don't get it.

You say BIS isn't community friendly? Really?

You say there is no documentation? Well, there is. I just assume you mean for scripting, but you could be talking about anything.

And BIS isn't community friendly? Show me a developer who is more in touch with its community than BIS. This forum is the best example.

I don't know what you want to accomplish with this post, but you come off as a whiny child.

I grabbed a Snickers, thx,

i love answers like that... seriously!

maybe you read it again and this time also try to understand it.

I really love to be called a whiny child by people not understanding whats up, especially if they didn't even try to change something by making some efforts. did you? for the community?

You "assume" right - as this whole post (i assume you din't read it at all?) is around scripting /scripts-performance and server-performance.

As you can read above, i wrote "for the the moment - you aren't communityfriendly, as you don't provide us with informations or hints or documentations that are usefull to raise the immersion of your work."

i thought it wouldn't be necessary, but it seems, it is...so just for you - try to read it like that:

"for the the moment - TO ME (as an hoster/editor/contentcreator and at last a player&customer) - BIS aren't SO MUCH communityfriendly (meaning the contentcreators/hosterscommunity, not the community at all...), as they don't provide us with as much as informations or hints or documentations that are usefull-as they SHOULD - to raise the immersion of their work and through that enable the possibilitys to a better/faster MP expirience and Mission-landscape".

got it now? geez.

@Zooloo75

So I guess single core processing is where it's at now-a-days? Back to the ol' Pentium 4!

yes, it seems like :(

@TSAndrey

The FPS on the server was fine all the time, but it started lagging when there were a lot of players.

thank you for feedback. i think, at the actual state of the mission (v7.2) its playable at around 80 to 90 players, then it is getting too much for the engine.

Do you mean "lot of players" at the server at all or just if they are around you (like in the mainbases)?

---------- Post added at 14:49 ---------- Previous post was at 14:17 ----------

@blue1 my post in thread isn't indication of being 'aware' ? what else you want me / us to do ?

documentation? for BE? for dedicated server? in the BIKI (Community Wiki) and here on BIForums ...

Dear Dwarden,

no it isn't at all.

To me (!) it "sounded" like a "Geezus why dafuq you are doing such a Test on Stable-Branch" or something like that... instead of the expected real feedback/answer to the shown/proven facts.

If you ask me like that, what i want you to do:

As you are the Communitymanager,

I'd like if you read the Thread again.

I ask you kindly to answer to these Results.

I ask you kindly to tell us if we can expect any changes to the obviously and wellknown MP-limitations (do i need to state them all again?).

I ask you kindly to provide more Documentation to anything that is related on Missions-/Scripts-/Servers-Performance -FROM FIRST HAND,

not from Community-collected Guessworks coming from "try&error-sessions" and then maybe confirmed by you - or mostly not.

For BE? I never mentioned BE in this Thread at all - wasn't part of measurements/questions.

And of course i know BIKI and so on... :j:

And before anyone may forget what this is all about:

i love this game and his possibilitys - and my opinion is, that it should be possible to make immersive Missions with 128 (or more) Players for "large scale" PVP, which is somewhat "Standard" already in 2013 (like good Multicoreusage on Servers too^^).

Thanks and best regards,

Blue1

you failed to realized i edited my 1st post where i was asking what branch it was done on, into 'that i figued out it was on stable branch' ... sigh

Oh crap :eek: Indeed i haven't seen that unedited post before.

So i want to apologize for the assumption and my misapprehension of the edited post afterwards. Sorry!

and i'm sorry that i can't offer you more that was already revealed in relation to the dedicated severs, in the BIKI/BIFORUM etc., for the moment ...

Okay, as i already wrote - i know how much work it is to create a Documentation and keep it up2date.

But i really wish - you guys will find time and manpower to further optimize the stated MP/Server issues and Bottlenecks,

as this is nothing we could do on our side, even with the BIKI :(

Thanks and sorry again.

best regards,

Blue1

The maximum number of players for a smooth gameplay though in all ArmA games IS small.

For such available player assets and ArmA's usual Xlarge terrains..a number of 200+ players should have been feasible imo.

The only limitation i can easily understand it's.. players hardware-when they (occasionally) have to render the units+weapons high poly models-

but this should be restriction belonging to client only..

thank you for feedback. i think, at the actual state of the mission (v7.2) its playable at around 80 to 90 players, then it is getting too much for the engine.

Do you mean "lot of players" at the server at all or just if they are around you (like in the mainbases)?

Everyone started lagging when then were a lot of players online, I don't think it was only around populated places.

@Dwarden

Ok..thanks.

I really hope a higher player count with smooth gameplay is among goals planned.

Everybody knows everything is W.I.P. for now

@TSAndrey

Yes..it's probably the player number present on server.

Although i was not present at the event..i m sure admin had locked server and mission/test started with all players present to avoid connects/disconnects that dramatically burdens servers stability..

the more people are in one place doing relevant stuff the data transferred should increase factorial

and the nearer players are too each other the more data gets relevant.

like:

2 players. distance 10 km. Mouse movement: irrelevant => no transfer

2 players. distance 10 m. Mouse movement: relevant => transfer

questions: how much data is range relevant and how much is always sending. Whats the barrier. Can attribute of datas relevance be changed. can data send edited by missions (indirect like disabling inventories of cars, lower number of placing objects, lower resolution of mouse movement tracking lower polling rates for movable objects etc etc). what are the biggest data packets.

hey blue, your inbox is full, please delete some messages, i cant reply to you.