Crash 1.0.2 (Out of memory maybe?)

  1. 8 years ago
    Edited 8 years ago by HeroesandvillainsOS

    I don't really know how to interpret these things. Here's a rpt file from a CTD I just had after about an hour and a half of play. At the very bottom I'm seeing something about 'out of memory.' I've heard about this kind of crash, but it has never actually happened to me before. I just got a new PC yesterday and so much for it's maiden voyage. :) If that's the case and ALiVE isn't to blame, does anyone know how to fix that? Here's the rpt:

    https://www.dropbox.com/s/jja9pnngiepw3x4/arma3_2016-03-26_11-26-59.rpt?dl=0

    Again I don't want to pretend I know how to read these rpt's. If anyone would be willing to open it up and give me their thoughts I'd appreciate it.

  2. Edited 8 years ago by HeroesandvillainsOS

    Oh goodness. Google is leading me to showing hidden operating system paths like "page file." I should probably wait for an expert opinion!

  3. Yeah i only have a 125gb ssd for W10 and arma+mods and got the run out of memory crash alot. Had to move the paging file storage location to my bigger hd and all was fixed. Just google how to edit paging file location if your prob is same as mine.

  4. IIRC a memory leak was introduced to Arma in update 1.54 and still hasn't been fixed.

  5. Well, I haven't had an issue at all since 1.56 and this was on my crappy PC. I just got a new one which has a 500 GB SD so I wonder if maybe I just didn't set the computer up right? Are there any protocols for setting up a new PC to allow for better memory usage? I'm 100% positive it's not my mods. The only thing that changed is the computer.

  6. Edited 8 years ago by HeroesandvillainsOS

    And I'm not even positive it is a memory leak. I can't pretend to know how to read rpts. It just says out of memory near the bottom. Any help in interpreting it or any steps I can take that I can try would mean a lot. I've spent months on these missions and it would kill me to not be able to actually play them.

  7. Edited 8 years ago by HeroesandvillainsOS

    Oh. If it helps my exit code was "unable to init DXGI" though I don't see how I can have a memory leak and an issue with my GFX driver. I'm reading stuff about adjusting malloc settings but who knows. Man I hate troubleshooting this stuff. It's like taking batting practice in the dark.

  8. 7 years ago

    @GhostNI Yeah i only have a 125gb ssd for W10 and arma+mods and got the run out of memory crash alot. Had to move the paging file storage location to my bigger hd and all was fixed. Just google how to edit paging file location if your prob is same as mine.

    Ugh. This is still happening. I may just try to disable Page Filing but I hear that's a bad idea (I have only the one 500GB SSD so I can't move it). Any thoughts on that?

    Last night I was greeted with "DXGI Device Removed" exit code. I feel like I must have missed something really simple in the setup of the new PC. I never had this problem on the old one. I'm using the newest binaries too so this is happening on both the RC and experimental ones.

    Anyway, for those curious, here's a rpt: https://steamcommunity.com/linkfilter/?url=http://pastebin.com/PhBCvh71

    I'm open to any other suggestions if anyone has any ideas. Perhaps I should roll back my driver?

  9. Edited 7 years ago by SavageCDN

    You should never disable your page file despite what many forums might say about performance... especially in Windows 7 and above. This might have been the case with XP and earlier but nowadays it is useless (even if you have a ton of RAM).

    That exit code is basically saying you've removed your graphics device - I assume you didn't pop open the case and yank out your vid card while it was running? :) Did it recover on it's own or did you have to reboot?

    If you have an NVIDIA card try rolling back to an older driver (6 months or more)

    http://lifehacker.com/5426041/understanding-the-windows-pagefile-and-why-you-shouldnt-disable-it
    edit: added link

  10. Edited 7 years ago by HeroesandvillainsOS

    Well yes, I actually did rip open the case and yank...

    Ha! Yeah that would be something. The computer actually begins acting really strange after the CTD's. Specifically, I can't exit out of certain programs such as Steam (I can't X out on the X symbol or maximize/minimize the page, etc). I have to restart the computer to get it to work properly again. Most of my other programs worked like normal so afterwards so who knows?

    Also, during one run last night before the crash, about 10 minutes into the mission, my frames dropped to about 1 and it was a literal slideshow. That's never ever happened to me before. I could not recover from it at all. Had to reset the computer then too. I don't get it. My temps are perfectly great (60c was the highest for my GPU. CPU cores are in the 20's).

    I'll roll back the GFX driver.

    Question: I noticed Steam decided to install my games in a unique path. I never use Program files (x86) (should I?), but I used a folder on my C drive called Games.

    Inside that path would normally be games/steam/steamapps/common/arma3 but for some reason my game is installing in games/steamapps/common/arma3. Do you think that's an issue?

  11. Edited 7 years ago by SavageCDN

    Shouldn't really matter as long as Steam and the game starts up properly. By default it will use \Program Files (x86)\Steam\Steamapps\Common.

    You can also move your Steam library after the fact to a different folder/drive.

  12. Edited 7 years ago by HeroesandvillainsOS

    So I reinstalled Arma completely fresh, did a clean install of my graphics drivers and ran -malloc=system and I still just had another Out of Memory Crash. Same exit code. "DXGI Error Device Removed" and again it happened at almost exactly an hour on the nose of server time.

    This specific error spammed my rpt for almost 15 minutes consecutively before the crash. Literally thousands of lines that said this (with the only change being the map location):

    13:15:05 WARNING: SuppressCache::Trace called with huge input. Function exited to prevent program stall inside. tBeg[165.000000,0.964884,0.000000]; tEnd[-0.262677,17799.156250,166.386459]; ammoHit(0.000000); distance(166.386459) 13:15:05 In last 500 miliseconds was lost another 2 these messages.

    Does that mean anything to anyone? Is it possible night missions are broken and causing memory/DXGI crashes? My mission is turning into night time at the exact time of all these crashes and I'm guessing TRACE means "tracer" rounds. Or, perhaps is it referring to the "TRACE" squares you can enable in ALiVE? If so that would be crazy because TRACE is not enabled in this mission.

    Anyway, I can't believe how long this error went on for before the OOM crash. Does anyone think it's related?

  13. Tupolov

    30 Mar 2016 Administrator
    Edited 7 years ago by Tupolov

    https://forums.bistudio.com/topic/186794-warning-suppresscache-crash/

    Protip:

    Run only CBA and ALiVE and test a vanilla ALiVE mission.

  14. Edited 7 years ago by HeroesandvillainsOS

    @Tupolov https://forums.bistudio.com/topic/186794-warning-suppresscache-crash/

    Protip:

    Run only CBA and ALiVE and test a vanilla ALiVE mission.

    Thanks Tupolov. I'm sorry but the "solution" posted in the thread doesn't make sense to me.

    We solved it (more or less) In the Init is @ace twice called this let run the ACE advanced balistics against it self

    I don't use ACE and what he typed is gibberish (to me).

  15. Edited 7 years ago by HeroesandvillainsOS

    Looking at that, the only mods we have in common are CBA and AiA. I'll try vanilla but I have my doubts this is a mod issue. That error cannot come from ACE.

  16. Edited 7 years ago by SavageCDN

    In that example the guy had @ACE mod loading twice which caused a bunch of issues leading to the OOM error. "Trace" refers to something else.. not ALiVE related.

    It's also possible that your OOM error and the DXGI error are not related - usually DXGI problems are fixed by re-installing the GPU drivers or rolling back to an older version.

  17. Edited 7 years ago by HeroesandvillainsOS

    Is it possible I'm running out of memory because of error spam? You should see how many times that error repeated. I'd estimate a thousand times before I ran out of memory and crashed. I have a suspicion lunching -nologs would "fix" the issue but I really want to sort this out.

    Exit code is still DXGI Error Device Removed and inside the rpt, I also saw unable_to_init_dxgi.

  18. Edited 7 years ago by HeroesandvillainsOS

    Any idea where I can find the source of the "trace" errors with feedback tracker down?

    EDIT: Plenty of people reporting this error in EXILE. I'll do some googling for a concrete solution.

    EDIT2: Only about half a dozen reports of this error I can find and no solution or possible cause has been shared.

  19. @HeroesandvillainsOS  Is it possible I'm running out of memory because of error spam? You should see how many times that error repeated. I'd estimate a thousand times before I ran out of memory and crashed. I have a suspicion lunching -nologs would "fix" the issue but I really want to sort this out.

    Those errors get written to file so I would doubt they would be stored as well in memory but that's an easy thing to test :)

  20. How do I test it?

  21. Newer ›
 

or Sign Up to reply!