MAF! Who wants the saved game functions in the SDK?

Maybe we can start a pettition (like the guy from Poland). Just get some signatures, a whole lot in the best case, send the stuff to Firaxis, and, just wait and see what comes out:) .
 
Well, this damn error pretty much sucks. I uninstalled & reinstalled everything & still get the error. At a relatively small filesize too - only 860kb - and early in the game - Rennisance Era. My above average system specs don't seem to be helping me at all in this regard.
 
Well, this damn error pretty much sucks. I uninstalled & reinstalled everything & still get the error. At a relatively small filesize too - only 860kb - and early in the game - Rennisance Era. My above average system specs don't seem to be helping me at all in this regard.

People report crashes anywhere from 750KB to 1300KB.

A file size that large, that early is just crazy....map and number of Civs and graphic settings are way too high. The first post in the ViSa Thread list some settings to help the MAF.
 
Someone I know said that it could bebecause, the game (not any mod) either doesn't work together with the ram, or the memory sims used for the ram don't cooperate well enough with the rest of the motherboard, or both, the last one might be the problem withme,because I should have had 512 mb ram, but my pc only reads 256. I'll see if that may the answer once I fixed that. But if it is the first one, which is the problem then we should contact firaxis, but it could wel be the last one which is the problem, since (for so far I've read) TAfirehawk doesn't seem to haveany MAF problem, or am I wrong about that?
 
Someone I know said that it could bebecause, the game (not any mod) either doesn't work together with the ram, or the memory sims used for the ram don't cooperate well enough with the rest of the motherboard, or both, the last one might be the problem withme,because I should have had 512 mb ram, but my pc only reads 256. I'll see if that may the answer once I fixed that. But if it is the first one, which is the problem then we should contact firaxis, but it could wel be the last one which is the problem, since (for so far I've read) TAfirehawk doesn't seem to haveany MAF problem, or am I wrong about that?

I can pretty much guarantee that it's not a problem with my RAM, nor my RAM 'cooperating' with my mobo. Both are high end components, and do very well at various GPU & CPU intesive stress tests & memory testing.
 
Hi, all

I've just started modding warlords (playing gigantic maps with 36 civs), and I started getting this error quite regularly.

One thing I found that cuts down the number of them I'm getting is to litter the python code with gc.collect() calls (that's the python gc module, not the global context instance), particularly after any bit of python that has made a lot of use of the SDK/python interfaces.

But, there still comes a point when the game fails. I've tracked the failures I'm getting now to CvPlot::write, usually when attempting to write the m_apaiCultureRangeCities array (though not always).

I'm still trying to find a cure (hack) to this problem (currently running a patched exe where I've made changes to the PE header, though, tbh, I'm not very hopeful for this approach)
 
Hi, all

I've just started modding warlords (playing gigantic maps with 36 civs), and I started getting this error quite regularly.

One thing I found that cuts down the number of them I'm getting is to litter the python code with gc.collect() calls (that's the python gc module, not the global context instance), particularly after any bit of python that has made a lot of use of the SDK/python interfaces.

But, there still comes a point when the game fails. I've tracked the failures I'm getting now to CvPlot::write, usually when attempting to write the m_apaiCultureRangeCities array (though not always).

I'm still trying to find a cure (hack) to this problem (currently running a patched exe where I've made changes to the PE header, though, tbh, I'm not very hopeful for this approach)

It sounds like you hit the nail on the head with this one; i had a longer reply but it was on the other PC which has gone to computer heaven. :)
 
patching the exe (I increased the heap and stack values) did appear to modify the behaviour - but only to make the crash occur every 10-12 turns instead of within 2 or 3 of reloading the game.

So I've created my own FDataStreamBase based class which saves the map data to a seperate file - this seems to have worked, though I'm now getting a CTD (caused by a stack overflow) which I'm investigating. Don't know yet if the latest failure is caused by some inadequacy in my 'map saving' method, but at least I appear to be making progress.
 
I am doing my best getting a maf at an unmodded game. I even replaced my VGA with an earlier (read: crapier) version, which makes me get an maf earlier. Once I'll get one, I'll try to upload it for y'all.:goodjob:
 
Play huge, max civs, marathon and use the advisors a lot and then, later in the game they may start happening (it's a while since I played a 'vanilla' game), but I'm pretty sure I got a few even before I got hooked by Sevo's mod.

Just to let you know how it's going with my efforts in 'fixing' this - I've now managed to play a game (gigantic/marathon/36 civs) to 1000AD without a single crash (related to the bad allocation error, at least). The furthest I'd been able to get before was around 300AD.
 
Nice to know Nev. What portion of the existing file(s) do you change with what patch file? I'm not as computer literate as some on this forum. I would really appreciate some simple fixes that could allow me to get to the post 300 AD years.
Thanks..
 
Sorry, patching the exe didn't work.

But, I do have a solution (the FdataStreamBase one I describe above) which I'm working with the ViSa team on getting ready to release. It appears to work fine, but there are severeal issues that I have to resolve before it is suitable for general use. So, other commitments allowing, I should be able to release it to the community soon.
 
Does adding lines into the CivilizationIV.ini file increase stability? I did manage to get another 70 years of play with that in, and changing the texture res to medium vs high. Should I switch down to low and give it a whirl?

virtual char * ReadString() = 0; // allocates memory
virtual wchar * ReadWideString() = 0; // allocates memory

Is the line you are talking about?
 
Hi everyone,

got the same problem as you do, but also within a 'vanilla' saved game on a huge custom created earth map (320x160).
I've got a 4MB dual-core CPU and a 4GB internal memory, so this definitely can't be a hardware problem. The saved game exceeds the max allowed file size of 500 kb (its size is currently 1254 kB, or 930 kB when zipped), so unfortunately I haven't succeeded in submitting the saved game file onto this forum.
Does anyone have an idea on how to post this and how to report this bug in the regular CIV4 game to Firaxis (patch level 1.61) (and will this urge them to react?)??

Thanks,
 
Hi everyone,

got the same problem as you do, but also within a 'vanilla' saved game on a huge custom created earth map (320x160).
I've got a 4MB dual-core CPU and a 4GB internal memory, so this definitely can't be a hardware problem. The saved game exceeds the max allowed file size of 500 kb (its size is currently 1254 kB, or 930 kB when zipped), so unfortunately I haven't succeeded in submitting the saved game file onto this forum.
Does anyone have an idea on how to post this and how to report this bug in the regular CIV4 game to Firaxis (patch level 1.61) (and will this urge them to react?)??

Thanks,


Tried to get it rarred (makes it even smaller then zipped:) ). I know I might be talking nonsense but atleast I'm helping a bit:crazyeye:
 
Yesterday, I had this same memory allocation error during a long multiplayer session with a friend. We were on a direct ip connnection (same room on a LAN) playing with 6 other civs on a large map.

After reading through this thread I went back and checked my save game files just out of curiousity.

The date modified property on the initial BC4000 autosave file is 2/15/07 at 4:14PM. The recovery file generated when I rejoined the game had a date modified time stamp of 2/15/07 1:51AM. Civ had been running, under load, for the duration (approximately 10 hours). The filesize of the initial autosave was about 120kb and the file had grown to 680kb at the time of the crash.

We were both using an "unmodded" retail version of civ4 and warlords with the 2.08 patch. I've had similiar crashes to the desktop before when using huge maps for long periods of time, but not this MAF error.

I have 2 gb ddr2 800 RAM, ATI X1900XTX video card with 512mb of memory running on an AW9D-MAX motherboard (975X chipset/E6300 processor). Even with twice the RAM, a new video card, and a dual core processor I seem to have more trouble running civ now than I before. :)

I tried to included a copy of the save game files in case they might be of help to someone out there but the upload file size on the forums is 500kb. If anyone wants to take a look at just let me know.

-Greg
gregstadermann@gmail.com
 

Attachments

Alexman is probably the best person to notify. He promised to look into this on behalf of Firaxis.

One way of posting would be to use 7 zip to chop it into 500 kB chunks and then add a fake .txt, say auto1.1.txt, auto1.2.txt and so on. To decompress, just remove the .txt extension and 7zip will do the rest.
 
I have been plagued by this problem since the game came out. I defiantly had a MAF in non-modded game.

Is there any more progress in creating a fix for this?
 
I don't know if the issue I'm having with my game counts as an MAF or not. My game completely crashes when I try to "next turn" during year 1545. This is playing Warlords with 2.08 installed. Shall I post the save game file?
 
Back
Top Bottom