Interesting Problem...

I'm not quite sure what the issue is but maybe you guys can help.

So, I've been playing some rather intense (graphically-speaking) PC games over the past few months, and I've begun to have to one major problem...hard locks. It happens very infrequently, without notice, and I have to hard reset my computer in order for it function again.

My PC Specs are:
OS: Windows 7 Ultimate
CPU: i7 260 2.6ghz
Graphics: Nvidia GTX 580
Memory: 4GB

Now, daily functions (word processing, browsing, watching tv shows, etc.) doesn't seem to produce any sort of failure. However, I get into a game (FEAR 3, Deus Ex: HR, Dead Island, and now Rage) after a period of time (like an hour-three hours) the game will hard lock. I try CTRL+ALT+DEL, alt-tabbing, but to no avail. I have to hard reset every time this happens.

I have checked everything, literally. I ran a benchmark on my graphics card, natta. I ran a test on my CPU, nothing. I have no idea what the problem is, and it only seems to occur within a game. I wasn't able to test it thoroughly, but it could have something to do with temperature. The max is 90 degrees on a GTX 580. I ran a temp. monitor and found it goes to 93 while gaming. I wonder if it's overheating and shutting down?

Any help would be appreciated. Thanks!

Check that the driver you have is not the shit version one nVidia released that made GPUs fucking melt. Down- or upgrade accordingly.

GPUs are generally hardy as fuck but 93 is pretty fucking steep. Check your GPU's thermal specs.

I should've stated I have the latest beta drivers.

As for thermal specs, the max nvidia set is 97. But, I read on forums that max on gaming should be 87. Nothing should be in the 90s. Haha. Big problem.

I could try upping the fan speed. Before then, I'm gonna check the airflow in my case (I only checked during idle and it was cool). If it's just blowing around hot air that's not gonna be too well.

Yeah, downgrade your driver. Google for a consensus on what version to go for, it's a pretty common problem.

Windows 7 has event viewer just like XP, the folder directory is different I think, in xp it is found in administrative tools, event viewer. In windows 7 it is called logs. You might see something logged when the failure occurs. Your system bios may also have a temp shut down feature for the cpu. I hate problems like this they are so hard to track down. I replaced memory two times on this machine I am now using and it worked great for almost two months and started the blue screen crap again. It appears I have a bad memory controller on the motherboard, problem is nothing is flagged when it dies, I can use one ram slot and everything is normal, makes no difference which slot, that alone says it has to be the controller not the slot.

No it doesn't. How old is your motherboard? How old is your RAM? Have you ever flashed your BIOS? Are you saying 1 stick works fine and 2 don't?

5 years, 1 year replaced same ram after using for 2 months . Yes it fails using two sticks makes no difference which slot or which stick you use, often you also need to change the slot using one stick to get a good boot again, then it runs flawlessly with that one stick. Last week I tried it again with two and things seemed to be normal with one exception, all large downloads tripped memory errors and caused drive checking when booting. That happened before and drove me crazy while trying to download UDK sdk, usually 1.5gigs , the down load would finish but was always corrupted. It does not have dirty slots or connection problems. There is one chip on the motherboard that once had an aluminum heat sink bonded and it slowly lost the adhesion and is no longer sinked, I'm thinking this may well be the controller. The thing that pisses me off is when it runs good and shows no ram problems that drive corruption occurs and you don't realize it happened until you boot the next time. Also the bios has the last update which is quite old. I need something newer but don't have the money for now and in fact this machine runs everything I want so I will probably just go cheap again and find a motherboard that will use the memory I purchased, I hate to spend money for something that I can't use.

Yeah 5 year old motherboard is tragically outdated for the RAM, and is likely to be picky as fuck about the channels you slot the ram into, on top of the BIOS probably being too old, you might have to up the voltage, change the RAM timings or operating frequency to get stable performance. I am confident it is a hardware issue, but less confident about it being the problem you think it is. I'm not ruling it out, but it seems less likely.

You could probe your North Bridge temperatures using some diagnostic software like Lavalys Everest to make sure it's not running at obscene temperatures (over 50C), just to make sure though.

But yeah, you can get a cheap motherboard, it is usually the least costly component of a computer anyway. But yeah that old a motherboard is a legitimately huge problem for tinkering with many RAM settings too. But if you can, try that out first.

Only setting for RAM is timing ratio and it is 1.1 or 3.4. I have never changed it, the thing ran for two months without a problem before it blue screened. The RAM is a matched pair of 1gig sticks certified for this motherboard. I just reinstalled the Intel chip set again and thinking about shutting down and installing the ram again. It never quit last time, only the storage errors were taking so long to do a disc check, the HDD is a one terabyte so you have to go have dinner or lunch while waiting on it to clean up.

I conducted another stress test, of sorts. I started up Rage, and while Rage was running I checked the airflow inside my case and it was perfectly cool. The GPU temps monitored were at 90s+. I decided to fix the fan speed at 85 instead of the range of 65-75 during intense gaming. After that was done, I started Rage back up, and I stepped out for 2 hours.

When I came back, I discovered my computer was at the login screen. I logged in and found that "Windows encountered a problem and had to shutdown." A BSoD error. The minidump indicated it was a "Driver_Power_State_Failure." I've also found the file that seems to have been the problem: ntkrnlpa.exe, inside the minidump. However, it sems to be related to regular system functions. I have no idea why it would cause such problems.

Monitoring the last few minutes of my computer's up-time the GPU reached 90-91 degrees.

Can anybody make something of this?

Did you downgrade your driver?

I have had some driver issues with new nVidia but when a problem crops up they are usually pretty quick to fix it. Going backwards always solved it for me.

Yup. I'll give it another go. Hopefully, I didn't burn out my card.

I purchased a new card several years ago from one of the cheaper vendors, well it never was a great card but it did give me the shaders I needed for some of my software. So I'm sitting here playing a game with earphones and I hear a loud pop, I looked around the room thinking one of my power convertors blew up but I saw nothing and no smoke so I just let it ride. This happened two more times in the next week and the last time I noticed my mouse cursor was getting all crazy at times, my first thought was the mouse was dirty or bad so I swapped that out and it still looked bad on screen. By that time I figured it must be the graphics card because I have had a few go south while using them. I pulled it out and replaced it with another one I had laying around and sure enough it had an entire row of capacitors with the tops blown open . Tried warranty but got nothing but a run around so I made sure I never purchased another card from these people. It really wasn't anything more than inferior parts used to build the card.

Problem solved.

It wasn't due to my card overheating. It had to do with a leftover system file from an Nvidia driver update. The system file was supposed to be cleaned but was not. It's a minority of people who experienced this issue but at last it's over.

The only reason I actually discovered this rare phenomenon was when I went to check the BSoD error in the action center. There was an error listing and inside there were several errors all of the same type across the span of a few months (the few months I've been having problems with crashing). I looked it up on google and fixed!

Played Rage for 5+ hours. No crash.

Well today I got a small windfall from the VA, I will be able to purchase another cheap motherboard very soon, I am tired of waiting on this machine to do anything, suppose I shouldn't complain too much since it has continued to function even if it is crippled. I have every service just about disabled for start up and this morning I turned it on and it croaked and had all kinds of errors, I think it was due to shutting off some services that shouldn't have been turned off. So far it has ran flawless all day even after several reboots .