Mystery game crashes, 2+ hours.

Every game crashes after 2 hours. Sometimes they make it as far as 5 hours, but never less than 2.

Vista 64
OCZ 1066 Reapers 4GB, underclocked.
E8400, usually OCed to 3.6 (400 FSB) but running it stock makes no difference.
Xigmatek S1283
Sapphire 4870 512mb Toxic
Corsair 750TX
G15 keyboard
G5 mouse
A couple WD SATA drives
An HP DVD drive
An X-Fi Extreme Gamer Fatality Pro whatever sound card.

Max core temps with overclock never break 60
Max GPU temp never breaks 60. (Yeah Toxic edition cooling is that good.)

Vista crash reports usually just list the game client, and sometimes both the game client and the ATI driver.

It's always a CTD, and usually I have to open task manager before I get the mouse cursor back.

So it's not heat.
memtest finds nothing. I would not expect a memory issue anyway as it NEVER CRASHES under 2 hours.
Closing out the few items I can in the system tray does no good. (I keep a clean system)

Performance remains excellent, right up to the no-warning crash. I have kept Hardware Monitor running in the background, so I can assure you IT's NOT HEAT :)

I have a few vague theories...
I'm thinking some sort of memory leak caused by some driver conflict, probably mouse, KB, or sound?

I'll keep working on it, not really expecting anything here. If you have an opinion let me know, or any other troubleshooting suggestion.
24 answers Last reply
More about mystery game crashes hours
  1. Have you runned Prime 95 64 bits if it fails then you problably have an hardware issue with your rig. try upping the MCH (North bridge chipset) and RAM voltage a bit with your cpu at stock and run the test again.

    if Prime95 works for a couple of hours straight min 6 then try a GPU benchmark like 3d Mark or my new favorite Fur Mark

    Good luck
  2. Thanks, I was considering Prime95 overnite tonight anyway. I've never had a problem running it briefly for heat tests, but I've never just let it go.

    I'll take the voltages off auto and give those a whirl as well.
  3. Having a similar crashing (bluescreening) problem with a lab computer over the past couple of months. Reloaded Windows XP and all software, drivers etc. and swapped out all hardware except the MB, video card and PSU. Realized each time I bring the system home and test it overnight, no blue screen of any kind. Tonight I have swapped out the video card and testing stability with a different card. As usual, system is running in my house for many hours with no BS/system crash. Our lab has a lot of machinery running on that particular 110v circuit breaker and that is in adition to x-ray machinery running on the 220v in the room. The circuit is definately 'taxed' with all the machinery on that lab circuit. I believe I have found the source of the crashing problem with my lab computer. We use variable speed plaster grinders, air pressured machinery, film developers, and a host of other lab equiptment in that room and all drawing at various times from the circuit/breaker.

    Anyway, try moving the system to another room, or at least to another circuit breaker in the house. If I am wrong about my system, I'm am left with the MB and PSU and posibly the CPU, but I think I had another CPU in there when this all started a couple of months ago.
  4. I built a computer for a friend once that seemed to work fine under normal conditions. A few hours of load testing under 3DMark06 would cause it to eventually reboot. It turned out the motherboard wasn't supplying enough voltage to the RAM. I looked up the proper voltage for the memory, upped the voltage a tiny bit in the BIOS and everything worked smoothly after that.
    If I had to guess I would say it is probably not a memory leak since those usually cause degraded performance over time as the system runs out of memory. Since yours is running fine up until the issue, I'm betting against that being the culprit.
  5. Thanks guys. It's the whole timing issue on this that is really nuts. I've had this issue for some time. It's not as inconvenient as it sounds because I often don't game so long in a single sitting.

    2-4 hours is quite a bit for an old arthritic guy like me, and so these crashes are usually just welcome breaks :)

    I should add that the system remains stable after each CTD and the game can be restarted NP, and would run for at least two hours again.

    Prime95 64-bit ran overnite with no errors. The max core temps were 50 and 51C. Board temp (NB I guess) was 38C.

    Perhaps it's some sleep state timer firing off, like maybe the secondary hard drive. I don't know why that would cause a game crash. I'll try deactivating all sleep mode stuff tonight.
  6. If you have not already, I would do as others have suggested and be sure to set the RAM to specs in BIOS including voltage and timings. You might try as suggested bumping the RAM and NB voltage. I have tried everything like I mentioned. My lab system is running here at home all night, since 10pm last night. So, when I take it back to the office on saturday, I am going to use a new PSU cord and plug it into a completely different circuit breaker. You can test your hard drives with this software:
  7. The HDs seem to test fine.

    I can't seem to get any manual voltages or timings to run with the RAM now. Auto works fine, but these sticks never would run at their rated 1066. I'm going to replace the RAM and see if that resolves it.
  8. RAM is often the cause of system crashes. Especially when running software applications (games).

    Anyway, for what it is worth, my lab system ran all night and day (24 hours) without crashing. Runs beautiful and flawless at home. So, I can attribute the random BS crashing to the electrical current situation at our lab. I am thinking one of the assistants plugs in a 'heat sealer' right into a six socket wall extension next where I plug in the computer. There is an autoclave on that very socket too as well as a heat sterilization unit. The monitor on the system in the other room on the other side of the wall shakes like mad when she uses an articulator to vibrate plaster into a mold from time to time during the day. I did mention two x ray system run off the 220v in the lab where this system has a 'stability'problem. I am going to run a good quality extension cord to the a different plug away from where all the the machinery is plugged in. Except for an x-ray processor the computer will have to share with. The processor has moving parts too and auto switch off and on. Sorry if I have hijacked this thread. Didn't mean too. I have my problem solved by myself 8)
  9. For the past several months, I go out to lunch and come back or my wife will say the lab computer has "writing" on the screen (Blue screened). Random crashing during the day when left running all day. I traced down a lot of stop errors and even at one point did a total manual repair of the system registry. Try that. I swapped everything by now except the MB, HD and PSU (and possibly CPU, can not remember exactly when the problem started). Also reloaded Windows XP and all apps, drivers etc. Changed video cards this time I have it at home (nothing wrong with ATI 800GTO that was in there). I was scratching my head over this one. I get it home and ream, rap and rotate it and it runs perfect. Take it to my lab and it crashes. I almost replaced the MB, but I wanted to keep my NF4 socket 754 and not go down to standard NF3 socket 754 I have as a spare. The system is already slow enough handling large image files we deal with. I'm going to tear it down if it don't stop!
  10. Sounds like a UPS might help with that. Or maybe if you encased the comp in lead :p

    Don't you have some pretty big magnets in that room? Sounds like you might.

    I'll get my G.Skill PI Black Monday or Tuesday.
    I tried disabling sleep, tried several arcane BIOS settings, and installed SP2.
    This is a fairly fresh installation, but it probably does have a few things to clean up. I'll try that.
  11. Well, I plugged into a different socket, dfferent circuit breaker at the lab this morning. System ran 9:30am-3pm without crashing. I believe I have solved the random system crashing problem. I look at it on the positive. This guy will require a lot more time than I spent to fix his problem. :lol:

    Big problem :)
  12. LOL
  13. okay people running p95 is not the same as running the game. one loads the cpu, the other loads the system. whats the difference? the power intake. try the psu, that might be your faulty link.
  14. I think I have isolated the issue a bit more. The crashes seem to happen at save points. Not the ones that happen at zone changes, but the ones in-between. Since there is usually no on screen notification of these, it took a while for me to figure it out.

    So, it's an I/O issue.

    I had dismissed the hard drive because one of the crashing games, (Fallout 3) was installed on the secondary drive. Now I realize that Fallout 3 saves happen on the primary drive.

    Also, WoW has never crashed on me, and it is installed on the secondary drive but of course does not save :)

    So, perhaps the WD 320GB AAKS has some issue, although HDTune was happy with it.
  15. i HATE it when HDD fail. my system, when i first built it, had a bad HDD. there would always be some illusive issue in XP preventing me from doing this and that. i replaced after months of trying to fix it and all my problems went away.
  16. Installed the HD fan between the cages of my case, and installed my nice new G.Skill PI Black RAM. I think that OCZ RAM is crippled. Guess I'll RMA... It's subjective and all but I'll swear I gained a lot more speed than I should have going from CAS 5 to CAS 4.

    I'll know sometime tomorrow if it's all better.
  17. OCZ high performance can be a biatch. Their highest rated stuff is notorious for requiring voltage increases to the NB as well as the RAM to stabalize and act nice. If you ask for an RMA their support may ask if you tried increasing voltage to the NB. Lower CAS latencies create faster response times. That is true with all the RAM timimgs settings in BIOS. Higher transfer rate plus lower timings along with stability is a good thing 8). The system I mentioned (Lab) is now perfectly stable.
  18. FO3 is one of the offending games? Somehow, I'm not surprised...

    What games exactly are known offenders? Maybe theres a common theme we can figure out (high demand, same developer, etc)
  19. Proximon said:
    Installed the HD fan between the cages of my case, and installed my nice new G.Skill PI Black RAM. I think that OCZ RAM is crippled. Guess I'll RMA... It's subjective and all but I'll swear I gained a lot more speed than I should have going from CAS 5 to CAS 4.

    I'll know sometime tomorrow if it's all better.

    your not very specific about the ram you have. there are many black Pi series.
    do you have this
    if so, i have the exact same stuff and i OCed it to ddr2-880 (fsb 445 with a 1:1 ratio) without touching the voltages. its good stuff
  20. Yep, It's at 1.9V, 4-4-4-12. Same problem.

    I'll probably try a fresh install next.
  21. Update:

    Complete driver removal in safe mode followed by the installation of ATI drivers 9.2 fully resolved this. This had been done earlier (9.1 I think) but it worked this time.
  22. My 3870's would not even load up CAT 8.12 Vista 64. Fail. Cat 9.1 fixed the issues I as having and 9.2 worked fne also. I loaded up 9.3 yesterday and had to reinstall and remiove the included sound driver to get my system sound back. There were a boatload of issues with Vista 64 last year. Hotfixes and copmplaints on AMD's forum boards. Here's the response I got when I sent in a trouble ticket related to CAT 8.12, the driver directly before CAT 9.1. I'm geting tired of all the trouble with ATI's Vista 64 driver although the last two issues have worked out pretty well for my needs. Glad you managed to get your system running stable. here's ATI's response to my inquiry about CAT 8.12 which never managed to even install with 3870's xfired/Vista 64

    From review, the issue appears to be related to the latest Catalyst 8.12 as reported. Some related issues similar to the symptoms that you have described are already addressed by AMD Software Engineering. In the interim, revert or roll-back to Catalyst 8.11 or previous from or until the issue is resolved with future Catalyst 9.x releases.

    We have noted the issue and have requested high priority. However, we encourage you to report your issue to AMD Software Engineering through the Catalyst Crew Feedback Form. Please encourage others who have experienced the issue to do so as well.

    Please report your issue to AMD Software Engineering through the Catalyst Crew Feedback Form for further consideration.

    Catalyst Crew Feedback

    ATI strives to continually improve our drivers and software, and we invite you to tell us how.

    If you are experiencing a problem with our latest driver release, or if you have suggestions on how to make our drivers better, please take the time to submit your feedback.

    This program is intended to gather feedback for specific driver releases and not to solicit regular support inquiries. All support inquires will not be responded to.

    NOTE: If you have an issue that requires a technical response, please contact Customer Care.

    Continue to the Catalyst Crew Feedback submission pages.

    Thank you for choosing AMD!

    AMD Global Customer Care for ATI products
  23. Hmmm thanks. I did post over on the ATI forums, but perhaps a bit more input from me might help them out.
  24. Yeah, if you want an informed answer, send your question to ATI Customer Care, not the usual suspect imbesiles over there on their boards.
Ask a new question

Read More

Motherboards Games Product