Multiplayer

For now we've got 3 player. Only one of us has the possibility to run a 24/7 server, but he's got only 256k upload bandwidth. In the prev game it was barely enough to two of us.
Hope new players capable of running 24/7 server will join when "officially" made the MP game.

BTW 30$ is not so much if we count with a couple of people. ;)

I might be able to set up a server during Christmas Break, that you guys could use, 24/7.
 
Sounds great! :goodjob:
Probably we start the MP before that, but migration should not be a problem.
BTW what's going to be the config of the server?

Something cheap, an older computer, but it can still play RoM. That isn't important though, as all the other computers make the calculations, and send it to the server.
 
If my DSL wasn't so crappy (100kb/s) I would be glad to make my desktop a RoM server and play/mod on my laptop.
 
Something cheap, an older computer, but it can still play RoM. That isn't important though, as all the other computers make the calculations, and send it to the server.

I think that is not right. Clients should handle the graphics and communicate with the server, and every count shold be made at server side because of synchronization. In human turns it requires not so much counting power because the gap between human and computer action-reaction is huge, BUT in AI turns the counting requirement is huge in RoM.
Actually RoM's bottleneck is the "slow" AI (Better BtS AI).

(Please correct me is I'm mistaken... It's only a theory, but this is the stanard logic for client-server architecture)
 
If my DSL wasn't so crappy (100kb/s) I would be glad to make my desktop a RoM server and play/mod on my laptop.

It doesn't need a fast connection to be a server, or to play Civ4 MP. Civ4 was made so dial-up games played fairly smooth.
 
I think that is not right. Clients should handle the graphics and communicate with the server, and every count shold be made at server side because of synchronization. In human turns it requires not so much counting power because the gap between human and computer action-reaction is huge, BUT in AI turns the counting requirement is huge in RoM.
Actually RoM's bottleneck is the "slow" AI (Better BtS AI).

(Please correct me is I'm mistaken... It's only a theory, but this is the stanard logic for client-server architecture)

That is normal client server relationship, yes. However, Civ4's implementation is far from normal. What actually happens is that every computer makes it's own set of calculations, and sends them to every other computer. If the calculations deviate from what the other computers calculate, you get an OOS error. It's a decentralized system. I'm not even sure you need a server.
 
I'm not even sure you need a server.

Yes, I thought so too.

[irony]Can't someone remod the damned multiplayer system?[/irony]
 
That is normal client server relationship, yes. However, Civ4's implementation is far from normal. What actually happens is that every computer makes it's own set of calculations, and sends them to every other computer. If the calculations deviate from what the other computers calculate, you get an OOS error. It's a decentralized system. I'm not even sure you need a server.

Damn. It sound bad. Every client communicate with each other, not only one server? That's mean quite a big bandwidth requirement. And of course it requires more firewall configuration. Is this apply to pitboss games, where there is actually a running server?
How can it possible work, how can two "client" calculate the same result, when randomness is part of the game engine?
 
Damn. It sound bad. Every client communicate with each other, not only one server? That's mean quite a big bandwidth requirement. And of course it requires more firewall configuration. Is this apply to pitboss games, where there is actually a running server?
How can it possible work, how can two "client" calculate the same result, when randomness is part of the game engine?

I have no idea how pitboss games work, but as for standard MP, I wrote about it a few months ago.
 
E_Pluribus_Unum and me have finished in september a game lasted almost 8 months (playing once per week) and are planning a new game; possibly with pitboss+meeting once a week (about 20 GMT). Settings are still to be discussed and as he said, we're waiting for the MP forum to have a better organization and have some time to test RoM 2.8. Stay tuned if you want to join in, as we'll surely start a game in the next weeks! :D
 
Damn. It sound bad. Every client communicate with each other, not only one server? That's mean quite a big bandwidth requirement. And of course it requires more firewall configuration. Is this apply to pitboss games, where there is actually a running server?
How can it possible work, how can two "client" calculate the same result, when randomness is part of the game engine?

I've tried running a pitboss server on a VERY OLD celeron 800 Mhz and had no problem for the beginning of the game. I admit I haven't played very long so I don't know how it behaves when there are lot of units, bigger savegame files and more players. But it looks like Afforess is right, it looks like calculations are done on the client. If I have time, I'll do some other test on the subject playing a little longer to the modern era.
 
Previous posts in other therads sais that 2.8 is not good for MP game.
I'll contact 45°38'N-13°47'E and hopefully we can make some tests to confirm that.
It the statemenst stands than we've to move back to 2.5.2 for MP games.
 
Previous posts in other therads sais that 2.8 is not good for MP game.
I'll contact 45°38'N-13°47'E and hopefully we can make some tests to confirm that.
It the statemenst stands than we've to move back to 2.5.2 for MP games.

Actually it's not RoM 2.8 that's the problem, it's RevDCM 2.6. RecDCM 2.6 isn't MP compatible unless you disable revolutions in the custom game options screen, then it will be playable.
 
Actually it's not RoM 2.8 that's the problem, it's RevDCM 2.6. RecDCM 2.6 isn't MP compatible unless you disable revolutions in the custom game options screen, then it will be playable.

Thanks!
What about Barbarian civ option? If I'm not mistaken that is also the part of RevDCM. Should we isable that also?
 
Actually it's not RoM 2.8 that's the problem, it's RevDCM 2.6. RecDCM 2.6 isn't MP compatible unless you disable revolutions in the custom game options screen, then it will be playable.

I suppose I could live without Revolution, although I'd prefer playing with it. If Rev isn't fixed and everyone else agree, we could surely play with Rev off in our MP game.

Edit: I suppose I should stop starting every phrase with "I suppose" :lol:
 
I'm curious about HOW broken RevDCM 2.6 is for multiplayer - we heard that for the other versions and haven't had all the much trouble with it, even playing with revolutions on. It may depend on your tolerance, a few OOS a night is worth it to play, as long as we can keep the game up. If it's freaking out with every attack, then no, that won't work. I'm curious, please do keep me informed how it works, what the problems are and if they can be worked around. We may try it later, right now we are on our usual almost turn 300, where the game flies apart and one of us is clearly going to have a miserable game and we start a new one... hehe
 
Back
Top Bottom