Maximum Number of Players

[CouchMonkey 0]

After some more investigation we still seem not to find any such hard coded limit. However, please note that there are two things to do in order to allow high number of players:

1. the mission should have enough playable slots (e.g. 100)

2. the config of the dedicated host should set "maxPlayers" to desired value (e.g. 100)

Maruk

1. The server had more than enough playable slots. In fact we had over 120.

2. Please read the original post:

...I set the maxplayers in the config to, say, 99 and it bumps it down to 65 in gamespy. In addition if you put in, say, 200, you can only have 20...

I dont know if the problem is with the game or gamespy, but there is indeed a problem. At the very least it wont recognize 3 digit numbers for the maxplayers.

You still have not clearly stated that you actually had more than 65 human players try to connect to the server.

Please understand that when it comes to reporting bugs, being able to answer the critical questions is very, very, very important, and thus far you have NOT given the proper confirmation. What it SAYS and what it DOES are two different things, and you have NOT made it clear that you have actually had 66 live human players try to connect to the server.

And for the second time. Yes, we have tried it. The 66th person cannot join. I don't know how I can be more clear.

[CouchMonkey 0]

I hate to be a bother, but is there any update on this?

Thanks for your help

[ade_mcc 0]

Quote[/b] ]Sorry if this has been brought up before. I did a search and couldn’t find anything (BTW BIS figure out a way to go to the specific page of the post of the search result and not the whole post, those things get to like 190 pages).

(On topic, it was asked in the initial post and never answered.)

In the 190 page topic, click 'Print this topic', then use CTR-F to find what you need. Voila.

Can't help with the maximum number of players mind.

[CouchMonkey 0]

That's a pretty good idea. Thank goodness I am not on a 56k connection

Alrighty, back on topic. I am installing the beta on our server now so see if the new beta MP interface resolves the issue.

[CouchMonkey 0]

Any notes on this from the programmers? I noticed the demo server goes up to 68 players so we are getting closer .

Thanks

[MidTower 0]

Sys.specs are a not true..Even on a high End PC you cant run this game smooth...

i'm getting sick of wankers like you

the game runs fine on a high spec system

e6600 @ 3.2Ghz

2GB ram

8800GTX 768Mb video ram

raptor HD

X-Fi

all in a nice case with good cooling, and a decent 650w PSU

with no driver tweaks or OCing the GPU, i have a perfectly smooth gameplay with every setting maxed out on 3500m VD fraps tells me i sit at the refresh rate of my monitor for just about every part of the game at 1600x1200, it dips in woods to about 50FPS

so don't come saying that it doesn't run on a high spec system.

get a life, get a real PC and stop posting when u clearly know nothing

Wow! I thought he had a right to post here, just like you do.

By the way he's right!

[Johnnie_Walker 0]

Can someone tell us if there is a max players limit for 1.04 version?

[walker 0]

Hi all

I guess the things to do are the maths and the tests.

With the new 1.04 servers now out and the 1.05 in a week or so we need to start working out a plan for some scientific testing to see what we can get out of ArmA.

Some funding would be nice  

Suggested form for some scientific testing

Obviously trying servers with 20 people 30 people, 40 people, 50 people, 76 people and 100 people on them are needed.

Baseline Testing on a LAN

It would be nice to do some LAN testing to establish a baseline.

It would be great if we could get a major LAN/Internet cafe to do some baseline testing BIS or 505 might be able to sponsor this or it might be better to use the new Demo at a university as a student project to gather some data. It is certainly the kind of thing a university may be interested in.

A university is probably the best idea as BIS could also use said university in the future for similar work. And Published work on capabilities would help with VBS2 hint hint.

LAN V LAN

Two to Four LAN/Internet cafes, this is doable I can think of a group of LAN cafes that might participate. Would make a good launch for ArmA 1.05 (hint to Marek)

Easy National Test

Only people from the same country. perfectly doable

Normal Regional Test

Only people from the same region. What we are sort of doing anyway.

Hardcore inter regional destructive testing

Some from the US some from Oz some from Siberia, some from Asia and some from Europe with servers in each region is the hardest test for any game. Speed of light means at least 1 second of lag between those regions. Watch any game die the desynch death.

Any way my thoughts on how to do it.

Kind Regards Walker

[Blake 0]

Detachment Kyllikki will do some serious testing this week as ArmA hits the shelf here. I predict 40+ people connecting to our server mostly from same country so I think it can give some reference on performance. Our OFP server could host 30+ players without any major problems and I expect ArmA to top that.

[Metal Heart 0]

Quote[/b] ]Speed of light means at least 1 second of lag between those regions.

If there was no other limiting factors it would only take roughly 66ms to the other side of the globe on a straight cable. Triple for curves on the route and double for the trip there and back, it would still be "only" about 400ms.

[walker 0]

Quote[/b] ]Speed of light means at least 1 second of lag between those regions.

If there was no other limiting factors it would only take roughly 66ms to the other side of the globe on a straight cable. Triple for curves on the route and double for the trip there and back, it would still be "only" about 400ms.

Hi Metal Heart

I do not dispute your theoretical figures.

But as I am sure you yourself recognise you are assuming perfect connections. The reality is that even the BBC with its Uber fast dedicated connections for TV looks at 1 second plus delays across the pond UK to US and back.

Various factors contributing to latency in any game prevent us from reaching the ideal. Not least is: the drunkards walk that a packet takes, often far exceeds the crow flies distance our theoretical packet would take just going round the curvature of the earth. Then there is time spent in switches and routing and traveling miles in electronic chips. Finally there are are bottlenecks like duff hardware and sudden increases in traffic load beyond what is expected that max out pipes and shift the overload to new routes and all the sundry things that cause packet loss, acks never received, rerouting etc. etc.

The only real way to find out what the real figures are is to do the real world tests. Preferably over some time and in various conditions.

That is why I put forward that framework.

What we need now are:

BIS or 505 to voice the need for this and maybe publish a request on the forums for a bit of free research.

Then we need a university computer student or four in need of a class project.

There is a dearth of this data generaly particularly in a proper academic form (do not believe me? try and find some on the web), so maybe someone could even get an academic paper published.

Kind Regards Walker

[vipermaul 246]

Ok what this community needs is someone to host a 100+ stress test night.

Give a minimum of 5-7 day notice. So everyone can plan to be there.

Create a mission CTF or CTI mission that allows 50-60 slots each side.

We may need two different days if you can manage it.

I am not satisfied with people saying awesome 60 man servers when the announced goal was 100+ at least.

Thoughts? Any daring admins want to promote their server?

This will be a great service to the community.

[CouchMonkey 0]

Given that ArmAInteractive.com is now free, sign up and we can do it right. Our battledays are 17:00 to 21:00 GMT. Custom CTI mission baseline of 48FPS and an average game time of 2.5 hours.

Remember, the first hurdle is actually getting people on the server, the second is making it playable.

[nm-rosenrot 0]

when you run #monitor as admin in the server.. youll see the fps, memory, bandwidth etc..

be intresting to see some results and what hardware/server is hosting the game..

for me if a server is quite full and the fps drop below 20 using #monitor, then the max number of players should be reduced..

[Blake 0]

Quote[/b] ]for me if a server is quite full and the fps drop below 20 using #monitor, then the max number of players should be reduced..

FPS below 20 on server is nothing bad. On OFP game was smooth on less 10 FPS.

[nm-rosenrot 0]

yes, 20fps was just a number that came to mind, as we tested our server with 50 players, and quite a large coop map(alot of AI etc) the fps was up and down from about 7fps - 15fps depending on what was happening i guess.. smaller maps with no AI such as ctf did run alot better tho.. cpu load on our server was about 50%-60% running arma dedi and some other apps..

but to me, a well run server whatever the hardware spec.. should keep the game running smooth which means a balance between the max number of players and a good fps.. giving the player better gameplay in your server..

we have set our 100Mbit p4 server to a max of 32 players..

[Bad Pilot 0]

I played on a server where i was the 74th person on there. It was fine.

[vipermaul 246]

FYI -

I played on a server that had 85 players on it. It was using the seemingly popular Bezerk Missions and although I had some D'sync I can say I had no lag. We played for about 1.5 hours before I had to leave. It was great.