Revisions working/not working in MP

[45°38'N-13°47'E]

Which gamespeed? I usually don't expect OOS before Medieval Era

 

[Aquila SPQR]

We love to feel the slow flow of time so we always play on the slowest - Eternity. On previous versions (1041 and 1037 IIRC) we had OOS from time to time, on both ancient and classical eras (we did not reach medieval yet on this mod) and on our last game on 1041 we had a strange and annoying problem when we had an OOS every time the host player cut down the forest. It was absolutely predictable - "I'll cut the forest on next turn, prepare for OOS". In current session (we started a new game) we did not cut any forest yet, so I hope we won't get these OOS errors again.

We've just ended on about 170th turn, but on etertnity we did not reach far, we've just founded our second cities. Tomorrow we'll continue.

 

[45°38'N-13°47'E]

New game with rev1049, blitz gamespeed large map, revoultions and mostly suggested options as described in the other thread, 2 players; we're currently at turn 205/600 (halfway Middle Age) and we've gone OOS at turns 132, 134 (this one was probably due to a wrong setup), 150, 159, 174 (then no more for 31 turns). More will follow.

 

[45°38'N-13°47'E]

Turn 284/600, halfway through Renaissance; we've experienced some more OOS at turns 211, 240, 242, 252, 267 (then no more until we saved at turn 284). More frequent but there have been many events (war, revolts, etc), which might have caused these OOS. Still playable anyway. Some OOS have been solved with recalc (Ctrl+Shift+T), most haven't. No wars involving human players anyway, except against barbarians or against rebels.
More will follow.

 

[45°38'N-13°47'E]

Turn 303/600, almost finished Renaissance. 36 turns without OOS, last one was at turn 303, the previous one was turn 267. Not too bad until now.

 

[45°38'N-13°47'E]

Played up to turn 346/600, some more oos at turns 317, 336, 343, 346. But the last three ones might have been caused by some changes I've made by mistake to the mod on one PC only. Next session I'll restore the original version of the mod on both computers and I'll see if it gets any better.

 

[45°38'N-13°47'E]

So, here we are, up to turn 385. More OOS at turns 356, 360, 371, 375, 379, 385. Definitely more frequent. But, I think the cause might be the upgrade code. From the logs, most OOS happened when I decided to upgrade all my units at once. So next thing I'll try is to upgrade each unit one by one. Painful, I know, but I want to make sure this is the source of current troubles. If it is, maybe I can find something in the code.

 

[45°38'N-13°47'E]

Well, it looks like upgrading isn't what's causing OOS. But I've noticed short disconnections when a turn is beginning. I'm not sure if they depend on my LAN or on the mod or on different computers speed, but I've noticed than my pc is faster than my wife's and even when I start a turn, she's still locked because her laptop is still calculating AI moves. But if I wait some seconds we reconnect and we can play. We had some OOS for a few turns in a tow, but then we played 12 more turns without OOS with this method.

 

[45°38'N-13°47'E]

Turn 438/600, no more OOS since 35 turns. Really surprising. Year is 1942, start of Modern Era. It looks like OOS are greatly reduced when you let the game cool down at beginning of every turn. I mean, we don't touch anything when our turn starts for one minute or so, it gives time to both computers to perform some calculations that are probably made when the turn starts. When we are sure calculations are over, we start playing (diplomacy, production selection, and so on). Until now, it looks like this is the best thing we could do to prevent OOS.

 

[45°38'N-13°47'E]

Up to turn 446/600 still no OOS. But we had to stop there because a never ending turn, waiting for other civilizations to play. I was trying to debug what was causing this loop but I couldn't find a fix. Then I just raised the UPT limit (it was 40, now set to 50) and it worked! Probably some unit couldn't move becasue of the UPT limit. More will follow, hopefully we can play some more and I'll let you know how it goes.

 

[45°38'N-13°47'E]

Turn 458/600, OOS. First time in almost 60 turns. Looks great halfway through modern era.

 

[45°38'N-13°47'E]

Well, I finished the game somehow; I've got one more OOS at turn 490/600. Then we played quickly to the last turn almost without doing anything else that selecting techs and things to build and moving some random units. No wars except in the last 10 turns when a mad AI declared war to both human players launching some nukes. I've built and launched a Spaceship, no OOS. I've built Eden Project, no OOS. In the end, we've reached turn 600/600 and I won a Mastery Victory and the game went OOS when displaying final score. No OOS for 110 turns, although we didn't play at the full mod capacity. At least it means that AI part works. There have been revolutions, UN votes, wars, lots of things going on but still no OOS. We've also played 25 turns more after the end of the game so that we can test what happens when someone builds the Ascension Gate. It worked, and still no OOS.
It's the first time from what I know that someone finishes the game in multiplayer. There have been definitely been some OOS, but I think I can say the mod is stable. I'm already playing another game and running some test with another player. Next step will be testing a longer game or a game with more than 2 players. I'll let you know.

 

[45°38'N-13°47'E]

Another thing I've noticed when playing MP is that problems might arise when the speed of the players rigs are very different. The more advanced the game, or better, the more calculations required, the harder this difference affects the game. I've seen my pc ready for a turn while my wife's laptop still needed 1 minute or so to end all its calculations for the same turn. This might lead to disconnections. We were playing with rev 1049-1050, things might have been improved after that revision because I've scaled up training times for units so less units are in the game, greatly reducing calculation times.

 

[45°38'N-13°47'E]

Another game with RyoHazuki, rev1054, we've reached turn 485/600, no OOS in the last 25 turns (and counting); during the game we've suffered from time to time of multiple OOS, mostly around every 10 turns but some times they were probably caused by some different settings between us. As a workaround, before starting a game we've saved a full copy of the mod somewhere else so that if needed we can both copy-paste it again just to make sure we're still using the very same set of files. This mostly solved the problem when we had repeating OOS. All in all, we had around 30 OOS in 485 turns until now and some of those might have been avoided but since we were testing the game we didn't bother.

 

[45°38'N-13°47'E]

Well, we have finished this game (this time RyoHazuki and me), turn 600/600. 5 or 6 more OOS between turn 485 and 600 but all in all very playable, especially when we needed to resync the game but used the shared save folder which speeds up a lot the reloading of the game after an OOS. Second game I finish at blitz speed; next attempt will be with 3 players and then a longer game too (normal/epic gamespeed).

 

[45°38'N-13°47'E]

Started a new game with 3 players, huge map, blitz speed. Turn 137 and 2 OOS until now. There have been wars and lots of fighting with barbarians (30+ units stacks trying to conquer a city - and failing, the city was defended by 8-10 units). So far I don't see any significant difference from a 2 players game. Let's see how the game goes on and if more OOS will pop up. Revision is 1065.

 

[rtt4a]

Hi, dont know if this is exactly correct forum to report but we have problem with rev 1065 and hot seat MP.
Clean instal and no modifications except mega civ back and auto saves wont load. Game crash without any messages Tried to fix installation couple times but no change..

Weird thing is that I can load auto save in same game session but after closing civ and starting new session auto saves crash

Edit.. ok, getting more weird, I noticed that game wont crash if I load save game first and after that auto save. So if I have any saved game then after loading that I can load latest auto save without crash

 

[45°38'N-13°47'E]

Hi, dont know if this is exactly correct forum to report but we have problem with rev 1065 and hot seat MP.
Clean instal and no modifications except mega civ back and auto saves wont load. Game crash without any messages Tried to fix installation couple times but no change..

Weird thing is that I can load auto save in same game session but after closing civ and starting new session auto saves crash

Edit.. ok, getting more weird, I noticed that game wont crash if I load save game first and after that auto save. So if I have any saved game then after loading that I can load latest auto save without crash

I'm not sure because I think I only tested hotseat a couple of times; but it might be something similar to what happens in PBEM: IIRC if some saves are being overwritten, they don't work. I suppose that both players autosave when it's their turn, so probably a savegame by player1 gets overwritten by the same turn savegame by player2. I'm afraid I don't have any solution for this; at least for PBEM was a BTS limit, nothing to do with RAND.

 

[rtt4a]

hmm.. yes you are right about owerwriting. It seems that in hot seat auto saves work +1 turn principal.. loading 0 turn auto save means really beginning of turn 1 and so on.
Have to test more..I did some testing and loading auto saves in hotseat seems to cause weird things even if game wont crash. Couple of times game loaded but after moving units it freezes.

Did two autoplay test runs, 50 turns both, without any problems. Auto saves crash but loading saved game first fix thatDont understand..

 

[45°38'N-13°47'E]

hmm.. yes you are right about owerwriting. It seems that in hot seat auto saves work +1 turn principal.. loading 0 turn auto save means really beginning of turn 1 and so on.

That's how autosave works, not just hotseat.

Have to test more..I did some testing and loading auto saves in hotseat seems to cause weird things even if game wont crash. Couple of times game loaded but after moving units it freezes.

Did two autoplay test runs, 50 turns both, without any problems. Auto saves crash but loading saved game first fix thatDont understand..

Sometimes it's autoplay that's causing weird things, but I really don't know why autosave should crash but works fine after loading another save.