More system info needed

[jj]

at these days the CPU info measured in Ghz dont give a decent view of the players system speed.
More info is needed, like the CPU brand/type, real speed (because an Athlon 3000+ is not 3000 Mhz, number of cores in the system and maybe some information about the memory: free physical memory and total physical memory. Maybe some simple cpu speed benchmark applet would be better.

here are some examples why:

• - A 1,8 Ghz Core2Duo could be faster then a 3 ghz Celeron.

• - Big maps and lots of units need a lot of processing power, in fact, Spring needs a lot of CPU power.

• - You need enough free mem to play spring without lag.

• - Its so sad that players get kicked ingame while the battle is already going on for an hour, because they lag really bad

To send Betalord a request email, click HERE
and please vote!

[SirClickAlot]

I agree that the CPU frequency numbers are insufficient.

The applet approach seems a lot more useful to me than dry numbers. The benchmark could be made to simulate precisely the kind of strain spring puts on a machine and so give an accurate and useful performance grade. The benchmark could perhaps even involve running the spring engine itself on some scripted scene.

This would mean that it's still easy to filter people with subpar computers from eg an autohost. This would probably not be the case if the grade simply consisted of the raw info: processor type, memory etc.

[iamacup]

benchmark is more usefull, windows vista (oh em gee they had a good idea?) has something similar because they realised that you couldnt compair cpu speeds anymore.

[Ifein]

at these days the CPU info measured in Ghz dont give a decent view of the players system speed. [...]

Ack.

• - A 1,8 Ghz Core2Duo could be faster then a 3 ghz Celeron.

Afaik Spring doesn't support more cores. (Is that still the case?)
So in this case the 3ghz Celeron is still faster (assuming you don't run any other major program besides Spring).

[Cabbage]

duel core helps alot actualy, i can get 4998 units ingame at 70/80% cpu usage - spring crashes if i go over that. It's completely unplayable FPS wise ofc unless i minimise all settings, but thats a gfx card issue, not CPU (the actual simulation only slows due to CPU usage).

im certain an equivilent single core CPU - athlon64 2.2ish ghz would have a fit long before that number was reached ;P

[Cabbage]

And a 3ghz celeron is much worse that a 1.8 single core athlon64 or other processors, let alone duel core becuase of its design.

[Neddie]

My dual core works as such, though it waits to about 80% load on the first one before starting to spike the second.

[TradeMark]

I dont think betalord can make such benchmark program. Hes just lobby coder.

[espylaub]

Something has to be done here, my 1.6ghz Centrino consistently has less CPU usage than 2.8-3 ghz desktops, and yet I sometimes get kicked for having a "slow cpu".

[Tobi]

TBH the CPU thing should just be removed.

[Machiosabre]

yeah only jerks kick on cpu anyway, why would you want more of that.

[chlue]

I think the whole disussion about getting the system stats of the other players is useless, because the main problem are wrong gamesettings. CPU-speed gives you a good idea, if the simulation should be able to handle a certain number of units. The main question is, if the other player has set the gamesettings in a reasonalbe way, so the 3d-engine can handle them, too.

Personally I have all settings with the attribut (slow) off and some others quite low. Till now, I never got problems even with my laptop (athlon xp 1.8Ghz with mobile radeon 9000).

But if people swith dynamic water on, test it in singleplayer with ten units and then join a ten player multiplayer game, ...

So maybee an additionall 'settings'-rating could work. So when you read "1.5Ghz, settings 97%" you should be doubtfully.

[TradeMark]

I think the whole disussion about getting the system stats of the other players is useless, because the main problem are wrong gamesettings.

No, i have all settings minimum, and my game starts lagging at speedmetal maps. and greenfields, or other porc maps.

[tombom]

yeah only jerks kick on cpu anyway, why would you want more of that.

Because jj is one of these "jerks that kick on CPU".

Performance depends on so many things it's pretty much impossible to get an accurate measure. Even if it's accurate upon entering the game, you can change stuff while in game that changes your performance. If it's really that important to you, just kick people ingame with really high CPU usage. I look like I have a decent enough system but often get huge CPU spikes for maybe a minute. There's no obvious reason for them and a benchmark thing wouldn't tell you.

[jj]

WAKE UP!

spring needs a lot of cpu power. thats why i started the thread
big maps (16x16 or more) with >=10 players will cook your cpu, while other things, like the videocard are less important (a geforge 5500 will do the job)
network is not very important to, since hosting a 16 player game only uses 128 kbit bandwith or something.

SO DONT CALL ME A JERK!

we need this info to prevent lag in big games

[TradeMark]

What if someone made benchmark to spring.

Like:
1) Create units till the FPS rate is 30 or less.

I think that would be enough, and that would use the lowest settings of spring.

So, what more units you can create, that better your computer is... spring is pretty much about the unit count. Long games starts lag more because there is like 3000 units in the end.

[jj]

maybe its a better idea to monitor your average cpu usage in a minute when the total unit count reaches 1000.

spring saves it in a log file, then the lobby client can use it

[EDIT] Trademarks idea is better! but CPU usage instead of FPS monitoring[/EDIT]

[TradeMark]

Yeah, that CPU usage thing sounds better.

[LordMatt]

Another solution is to play 1v1s on non-porc maps.

[KDR_11k]

And a 3ghz celeron is much worse that a 1.8 single core athlon64 or other processors, let alone duel core becuase of its design.

1.8 GHz Athlon 64s get listed as 3 GHz.