Crash every 2 turns in GEM/Late Medieval - Is it normal?

pipeora

Chieftain
Joined
Feb 19, 2011
Messages
23
Location
The Spiral Minaret
PC Spec:
Intel Xeon 1220 (whose performance between Core i5 and i7)
8.00GB RAM
GeForce GTX 560

C2C Version: SVN 7771

I'm about to annex the whole land of United States in our world with Apache while starting exploring the dogfighting Old World. But my ambition was almost strangled with this annoying crush each 2 turns.I have tuned down the graphics setting to the lowest but wouldn't help a bit.
As for the error information sometimes it pops up "memory allocation failure" and sometimes the game just terminate without a message.

I would like to know if there's anything to do to fix this situation? Forgive me if I missed any existing solutions of it in FAQ or other threads. They're too scattered to fit into my dumb search ability.:crazyeye:
 
If you look at the first post of Bugs thread you would see what it is we need to find out if the problem is something we can fix or not. Basically we need a copy of the save and mini dump file. It may also help to mention what size map and other options your game has.
 
Please update to current SVN 7780. 7771 had a problem that has been addressed.

JosEPh
 
Upgraded to 7780 and recalculated the savefile. It does not work still. Now it crashes like each turn. I wonder if this is the absolute hardware threshold for such a scale of game? (GEM 35civs in C2C, with Rev and MinorCiv on)
 
I'm on the case here. MAFs would probably be a result of overwhelming the system yes but usually when they take place you don't get any data saved on the mini so right there it tells me something more interesting is likely afoot. So I'm looking into it now (sucks though that it takes so long to look into these kinds of huge games but at least I'm on top of things elsewhere so it's not bad timing to introduce this task today.)
 
Thanks for the responsible and gracious efforts from the C2C team. So it seems there's no immediate solution to this MAF. I would hibernate this set of game until the day C2C team helps us with improvement on the software side, or me help myself on the hardware side. :)
 
I've had the recalculation on this game running for over an hour or two at this point - on the debug dll this is not abnormal for a large game but it does say this game is enormous!

Still, I'd not be surprised to find some kind of fixable something at the end of the rainbow.

This game is a good example of why I was wanting to take us down to a max of 35 civs.
 
@pipeora,

You stated this was a GEM map. Is it the map from the Map folder for C2C, there is only 1 working GEM map for C2C at the moment. There is a 2nd GEM in the Map folder but it has Not been updated for over a year now.

Also it should be clear now that starting a game with 35 AI And having Rev On is not a good idea for C2C especially on a GEM map that may or may not have been updated.

JosEPh
 
I can SEE 30 civs on the list (and it looks like any more than that would run off the page) - this is really an epic game with huge turn times to the point even I would find difficult to play. Not finding much of anything outside of the standard to be expected asserts at the moment. But I AM beginning to suspect a game of this magnitude would be memory problematic for any system.
 
So I'm thinking the RAM and Graphics is enough but there's just a LOT to process here and I suspect the chip may not be quite up to the job. I've run 3 turns without problem though they each take about 20 minutes apiece on the debug dll. Just a bit too much imo.

And I'm now thoroughly convinced that running only 35 civs max is far more appropriate than 50. 35 with some room to turn up to 5 of those into Non-Player Civs sounds about right. More than 30 civs in the list won't show up on the screen at a normal resolution.

I do wonder too if that's part of the problem. Last round I objected to being introduced to another civ and I wonder if that might be it - one more on the list and it's crash city. I'm running my last turn here before moving on but if I get that offer again I'll accept it to see if that's partly the issue.
EDIT: nah... didn't offer again. Ah well. I doubt it's the problem anyhow but still... MAF's that take place with greater regularity than every 100+ rounds (which seem to be fairly normal lately on all settings) are usually a good indicator of the system not being up to the demand. The mini I was able to look into had a crash in the exe which could well be a MAF from such a situation but doesn't say much more to our limited ability to see what's taking place in the core engine.
 
Thanks very much for the detailed diagnosis from you guys!
I have pushed the game forward for about 5 turns with two saves and game restarts each turn (beginning and end of turn). Great willpower (and some weekend idle mood) took effect here. ;) The game on my PC could not simply sustain more than 2 rounds for AI actions. And "memory relocation failure" frequently appears during attempts to save games.
As @Thunderbird stated that the processor is the crucial shortfall here - will there be any instant suggestion to enhance its performance (e.g. some BIOS tweaks or a software)?
@JosEPh_II It is the GEM bundled in the new SVN with introductory messages like "redesigned for C2C V35" or something.
 
Top Bottom