Solved System Locking Up/Screen Distortion

[rebelagentm]

My home built desktop system is having two possibly related problems. Sometimes when the system is left on overnight, it will lock up and become completely unresponsive. This can happen as often as every day to sometimes only once a week.

About a week ago, while doing some normal computing, the screen distorted and became unresponsive. I hard rebooted the machine and it worked for about two minutes before the same thing happened. Did a hard reboot again and the machine was fine for hours. That evening, while listening to iTunes, browsing the web, and moving some photos from one folder to another, the screen flickered a couple of times. It also looked as though the Windows interface changed back to Basic. I opened SpeedFan to take a look at the temperatures and saw that my GPU's temp was quickly rising, so I shut the system down for the night.

Here is what I have tried so far in terms of troubleshooting:

• When these problems started last year, I ran FurMark to test the GPU. The test didn't even last a minute before the screen distorted and the system locked up. The temp got up to around 75 C.
• I ran the FurMark test again today. The temp reached 94 C within about 1-1.5 mins. I ran the test for about nine minutes total. The temp never exceeded 95 C, the system never locked up, and the screen showed no artifacts.
• I ran Prime95 for 8.5 hours and had no errors. According to SpeedFan, the temp never exceeded 80 C. Some cores did hit 92. I also had a co-worker run Prime95 with my proc in his machine and it did not report any errors either.
• I ran MemTest86+ for nearly six full passes and no errors were reported. I also ran the Windows memory testing tool and again, no errors.
• I have had both the motherboard and the graphics card RMA'd, and the manufactures said in both cases that they were fine.
• I tried installing a different graphics card in my system (an NVIDIA GeForce 9600 XT), but I could never get the card to work. During the reboot of the driver installation, the system would always lock up at the Windows 7 screen.

Some additional details:

• According to SpeedFan, my GPU normally sits around 48-52 C and my CPU sits around 38 C with individual cores sitting around 38-50 C. This is under normal load.
• The system is not overclocked in any way
• All drivers and software are fully up-to-date
• I am running Zone Alarm 2012 as my security solution. There are no infections on the system.
• Everything is seated properly
• The case is spacious and I don't think cooling inside the case is an issue as I have two friends with identical cases and far more hardware and they have no problems
• The case, fans, vents, etc. are all clean and free of dust
• My system specs are in my profile

The only thing I can think of at this point is that a cable near the GPU fan sometimes gets in the way. But if that was the case, wouldn't I probably hear a noise or something if the GPU fan got jammed?

Sorry for the long post, but I'm at a loss here.

 

[mgorman87]

You say your CPU gets up to 80 degrees and some cores hit 92 on Prime95. According to the Intel site, the max temp for your CPU is 67.9°C. I would get that problem taken care of first as it could be the source of your problems. If you keep overheating your CPU it will eventually fail.

Try running with the case open and a small fan blowing into the case and see if the problems continue

 

[rebelagentm]

Correct. I wasn't sure how accurate SpeedFan is though and I didn't know if that would be normal while running under Prime95.

I am going to try re-seating the proc as well since I think there may be too much thermal paste applied.

Still, would that cause what seems to be all graphics related problems? Especially since the GPU seemed to overheat the other night?

Thanks!

 

[mgorman87]

It could. There is no way to know for sure.

 

[rebelagentm]

I reseated the processor as there did seem to be too much thermal paste applied. This did drop the CPU's average temp about 5 degrees while idling. I have not tested it again under Prime95, but given that the idle temp didn't drop much, I wouldn't expect it to do good enough on Prime95.

I contacted Intel and they are going to send me a new processor. They did however tell me that the memory I am using is not compatible with the i7 930. It runs at 1.65v, but the Intel rep told me that it should not exceed 1.5v give or take 5% and that using anything over that could possibly damage the processor or greatly reduce its life.
Processors — Intel® Core™ i7 desktop processors frequently asked questions

Interestingly though, OCZ says that the memory I have is specifically for the i7 family, even though it runs at 1.65v.
OCZ DDR3 PC3-10666 Reaper CL9 Edition Low-Voltage Triple Channel - OCZ

Even though Intel specifies DDR3-800/1066, I also saw a post on NewEgg from OCZ that said this: "This memory will run 1333 on i7, but you must increase your memory multiplier, or overclock your buss speed. We sell that as enthusiast memory, and it does require some tweaking to get there."
Newegg.com - OCZ Reaper Edition 12GB (3 x 4GB) 240-Pin DDR3 SDRAM DDR3 1333 (PC3 10666) Low Voltage Desktop Memory Model OCZ3RPR1333C9LV12GK (Post is at the bottom)

What does everyone think? Do I need to buy new RAM that is 1066? Do I need to change the memory multiplier as OCZ suggests? Has this been the cause of my problems all along? I certainly don't want to damage my new processor once it arrives.

 

[karlsnooks]

Follow the advice of Intel.

Follow the advice given to you and remove the cover from your computer.

Direct a table/desktop/floor fan directly at your cpu and graphics card.

Download, install, run CoreTemp to keep an eye on your temps.

Link to CoreTemp in my sig.

If after performing ALL of these measures, you still have BSOD's then:
http://www.sevenforums.com/crashes-debugging/96879-blue-screen-death-bsod-posting-instructions.html

 

[rebelagentm]

Thanks, karlsnooks!

I should have mentioned that I already tried the fan. It helped keep it cooler of course, but I do not believe this is an issue of heat not dispersing from the case quickly enough. The case is large and is not crowded. All fans are operating. I have a friend with a nearly identical setup, except he has even more hardware crammed inside his and he does not have heat issues.

I will have a new processor coming soon, so I will of course test with that.

I will look into getting some new RAM for my system as Intel suggested.

I'm just curious though. Why does OCZ advertise this as an appropriate product for my processor? That is a bit dishonest it would seem, though it would not be the first time a company was dishonest of course. What about their suggestion of changing the memory multiplier? I am not suggesting that I do that. Just curious if this a normal/accepted practice. After all, overclocking is not "recommended" either.

And I'm guessing CoreTemp is better than SpeedFan?

Thanks!

 

[karlsnooks]

good.
Until you get the recommended ram, then I'd keep a fan blowing on the unit (case removed, of course).

I do know the whys and wherefores behind the ocz recommendations.

I prefer CoreTemp because I find the installation and use thereof to be much more straightforward. I also install the coretmp gadget, increase the size of the display to 150% (my eyes aren't the youngest) and keep only the temps displayed.

 

[rebelagentm]

Thanks for the suggestions, karlsnooks!

 

[rebelagentm]

After changing the RAM and processor, and re-installing Windows, my system has been stable for about two weeks.