Here's the specs on the machine first: Case- NZXT Nemesis Elite Motherboard- DFI Lanparty SLI-D Processor- AMD 4000+ RAM- Corsair 1024XMS 1GB Video Card- XFX 6800 GT 128GB Hard Drive- Western Digital SATA 120GB Optical Drive- Mad Dog DVD-RW (Unknown specific model) 2nd Optical Drive- CD-RW (Unknown model/brand. Taken from older machine). OS- Windows XP Pro While playing games such as Far Cry or Need for Speed: Most Wanted Black Edition the game either randomly kills, kicking you to the desktop. Or, it just freezes. Or, you get a blue screen with stop errors. Originally I thought it was overheating but after chilling the room it still crashes, but less often. This has been an issue since the machine was built, since then there have been plenty of reformats and drivers reinstalled/updated. Anyone know why this constantly happens?
First off; nice rig. If cooling more means less crashing, then you don't need to look further than it being a cooling issue. I've had a similar problem with a previous CPU. It crashed often when I tried to use it for games, until I tried to underclock it (have it run slower than it is spec'ed for). The thing about under and over clocking is that speed is not making the heat, but the increased energy demand in the processor does. By reducing the power sent to the CPU in the CMOS setup, you underclock it. If you try this and notice a significant drop in crashing frequency, then I suggest either you have your processor replaced (it's on warranty, right? it might be a faulty piece), or you get used to having it underclocked (I suppose that since it's a gaming rig, that is not an option for you). If it doesn't change much, your problem might be an overheating video card. Most video cards do not overheat unless you overclock them. Are you using dodgy third party programs that claim to "boost" your video card? If not, then check into a complete case cooling solution: add case fans. If the problem is not, after all, due to cooling, then the culprit might be anything: - Faulty video card (somewhat likely, if it only happens when you play games) - Faulty CPU (intense loads such as video games can possibly increase the chance of problems occuring) - Faulty memory (not very likely, as you would probably notice problems outside of games, although they have a tendency to make them more likely. Sometimes, a few memory cells are broken, and when a program tries to write/read them, it crashes). - Faulty mobo (not very likely either, you would probably have much bigger problems. It might be a problem with the video connector, which I suppose is PCI-E, right?) Of course, I can't do much over the net, especially if the problem is hardware. It might also be software, as Windows is a pile of crap.
Yeah, most of what you said I was thinking of already. Although, it didn't occur to me to underclock it. That's something I'll have to try, thanks. The mobo is apparently known to have compatibility issues with RAM. (As in, it's REALLY picky). Haven't been able to find any technical manuals to tell me what's specifically compatible. Because of this I've been tempted to replace it to see if it solves the problem. Hate to make that leap of faith though and have it not pay off. I keep the room pretty chilled though, usually around 60-65 degrees Fahrenheit (15.5-18.3 Celsius). The tower itself is literally right next to the air conditioner. (Low window in room). Because of this I gave up on thinking it was purely an overheating issue. Not sure on the warranty for the CPU. Have had the rig for awhile. I'll check up on that. Not using any software to "boost" performance. I'll look into getting some additional case fans though. (Although not sure where to put them, all fan room is taken up). And yeah the video card is PCI-E. Although, have switched the card between different PCI-E slots and the problem still occurs.
Ive had similar problems in the past, it only crashed during games, turned out the ram wasnt compatable with the Mobo. Download memtest 86 and run it for a couple of hours, this will show errors if your ram is incompatable or damaged. You could try running the ram at 333mhz too, this fixed the prob on my rig. You could also try prime 95(cpu torture program) to see how your cpu holds up under strain, if your ram is at fault it will crash during this test too.
15 celcius can in fact be too cold for your computer chips to function effectively. It may again be that your motherboard ram isn't functioning. I tried memtest once but in a few cases it doesn't do quite what its supposed to (Id somehow denied a block of memory existed). Howver, its fairly good. Maybe some jumpers on things aren't set correctly, you may want to check if you can/
Well, the room is relatively that temperature. I doubt that it actually cools the RAM/Tower in general that much. But I'll run some of those tests that you guys suggested and tell you the results. Thanks for the advice.
So, I ran Memtest. 36ish hours later, it got to 100+ percent with 4 errors. The errors were: "Error found with word pair #77106444 or 77106443 (Either could be the source)." All the other errors said the same thing but with the following numbers: 77106448, 77106442, 77106451, 77106452, 77106456, 77106455. So I believe I found my problem.
If you can, try a different ram module in your mobo and run the tests again just to make sure as it could still be a mobo fault. There are 2 tests(5 and 7 i think, check with google) that commonly show errors with amd compatible boards. You can also get some decent compatability advice from the ram guy on corsairs website. It could be that you just need to tweak the ram settings in your bios.
