Ladle 79

[sinewav]

It would be nice if everyone could upgrade to .4x clients, but I get, like, 15 fps with that version. I'm stuck in .2x for the next few years I'm sure.

[Monkey]

There have been massive changes to the netcode in 0.4.

Finally a reply from someone in the know, thankyou Z-Man. I guess I should have searched more for code changes but hey, I'm a little lazy

0.2.8 clients on 0.4 servers run on an emulation layer that, no doubt, causes extra CPU load, and that is on top of the already higher CPU load of the new code

I'd rather it get better usage of server resources than bandwidth

Is v0.4 really so much more of a CPU hog? I remember helping out some people a few years ago with their servers (Yellow Submarine and Wildcat IIRC) and Arma dedicated didn't use that much CPU, even on systems that were nothing special. Certainly people were of the opinion that bandwidth was a much scarcer resource. Admittedly times have changed a little but while bandwidths have increased, so has CPU power.

...0.4x clients, but I get, like, 15 fps with that version.

I currently run v0.2.8 client and it uses virtually no CPU at all. My system is nothing special by todays standards. By far the limiting factor for me is my graphics chipset and especially the open source driver that it uses. I haven't been able to successfully compile v0.4 so I haven't been able to try it but can it really use so much more CPU to the point that it seriously affects your FPS? *OR* does v0.4 require more from your GFX chipset?

The difference here is not between old and new. It's between "old with all the bugfixes that apply" and "new with all the bugfixes for the old code, but maybe some of them incorrectly merged or causing side effects with the new code, and maybe not all of the bugfixes for the things that were changed or added".

So if I understand everything correctly, the best chance of having least lag is by using either v0.4 clients with v0.4 server (on a system with enough CPU power) OR v0.2.8 clients with v0.2.8 server (on a system with enough bandwidth). Mixing clients and server could (but may not) cause extra lag due to possible side effects of code and extra CPU requirement.

My head is now officially mashed.

[takburger]

It would be nice if everyone could upgrade to .4x clients, but I get, like, 15 fps with that version. I'm stuck in .2x for the next few years I'm sure.

never gonna happen even if I have the computer for it. Sorry for 0.4 devs but I really don't like the version.

[Light]

I'd rather it get better usage of server resources than bandwidth

Is v0.4 really so much more of a CPU hog? I remember helping out some people a few years ago with their servers (Yellow Submarine and Wildcat IIRC) and Arma dedicated didn't use that much CPU, even on systems that were nothing special. Certainly people were of the opinion that bandwidth was a much scarcer resource. Admittedly times have changed a little but while bandwidths have increased, so has CPU power.

I actually don't know. I don't use 0.4 servers. My point was that if you're going to try to make one better, and you had to choose, that I think it was the wrong choice to go with bandwidth.

[ogo]

armagetron.kicks-ass.net seems to be down Z-man, any word on this? Didn't get the chance to download recordings yet

[Z-Man]

Could be that the server didn't like it when it (along with the router) was briefly disconnected from power while I fixed a socket the other day. I'll check on it in a couple of hours.

[Lowkey]

thx z-man! works again!!!

[woof]

thx z-man! works again!!!

Thanks a bunch! Your hard work is appreciated

[aP|Nelg]

It would be nice if everyone could upgrade to .4x clients, but I get, like, 15 fps with that version. I'm stuck in .2x for the next few years I'm sure.

Whats the graphics card? On this desktop: I can get anywhere between 30-80 fps (usally around a constant 50/60 though) on 0.4

There have been massive changes to the netcode in 0.4. The native message format has been completely overhauled. 0.2.8 clients on 0.4 servers run on an emulation layer that, no doubt, causes extra CPU load, and that is on top of the already higher CPU load of the new code (development focus was on conserving bandwidth and improving on extensibility).
The difference here is not between old and new. It's between "old with all the bugfixes that apply" and "new with all the bugfixes for the old code, but maybe some of them incorrectly merged or causing side effects with the new code, and maybe not all of the bugfixes for the things that were changed or added".

What about 0.4 clients running on 0.2.8/0.2.9/sty+ct/sty+ct+ap servers? Or even 0.4 clients running on 0.3 servers?
Also, why cant you run an 0.3 client on 0.4 servers?