System Randomly Freezes - Forced to Hard Restart

[JoshDBoyle]

Hello community

First of all, I am posting this here because I am not entirely sure what the issue actually is but it's plausible the mobo is the culprit here based on what I'm going to share in a moment.

A few weeks ago I made a post about an issue where, when playing graphically-intensive games, my new rig would completely freeze. Sound would continue to play (such as music from the game) but graphically everything would be frozen and I would be forced to hard reset the system. At the time, this was happening with great frequency - say within five minutes of playing any game (and I did try with several different games with similar results across the board). An astute forum-goer here pointed out a few issues with my RAM:

- I was using RAM that was "advertised" as 1600 but really could only run at 1333
- I was using three sticks of RAM at 4GB a stick yet my motherboard was dual channel (A1, A2, B1 and B2 DIMM slots)
- I had the three sticks of RAM in slots A2, B1 and B2

I went into the bios and turned the frequency on the RAM down to 1333, removed the B2 stick of RAM altogether and moved the B1 stick of RAM into the A1 DIMM slot. After doing all of these, my system became quite a bit more stable and the system freezes were drastically reduced in frequency (say once or twice a day during days where I would game for several hours). However, this does imply that the problem is still there, albeit with less frequency. As a result, I'm reaching out to the community here to see if anyone has some insight on why this would occur. I provide all of the above as a history on the issue in case it helps.

Here are my spec:

- Asus p8p67 LE
- i5 2500K 3.3 GHz/core CPU
- Diamond HD Radeon 6970 2GB GDDR5
- Intel x25 40GB SSD (for boot and system software)
- Western Digital Black Caviar 1TB 6GB/s HDD (for everything else)
- 8 GB Kingston DDR3 1600 MHz RAM
- Thermaltake Black Widow 850W PSU
- Windows 7 Professional 64bit

Here's a bit more information:

- When I was installing the motherboard, I thought I had it seated perfectly but when I turned the system on for the first time it came on for a second, there was a strong burning smell and the system immediately turned off. I disassembled everything and found that, fortunately, only a few of the contacts around the motherboard holes towards its corner were burnt off (I don't claim to be a tech guru so let me say it a different way: there was no damage visible anywhere on the board except around two of the holes where screws go through wherein a few of the little "bumps" around the perimeter of these circles were burnt off). I reseated the board, hooked everything back up, turned everything on and it all worked great (aside from the system freezes).

- The system ONLY freezes when gaming. I have left the system running in Windows overnight and for days at a time with no freezes.

- I strongly doubt it is due to any kind of overheating. I have checked temps on everything and nothing is running abnormally hot.

- The time at which the system freezes can sometimes happen within the first few minutes after entering a game or I could play for hours upon hours and it will finally happen.

- The Windows event that gets thrown when the system freezes is of ID 41 and of type "Kernel-Power". I've done some googling on this and it seems that many different things can cause this and that this is just a generic event when the system fails altogether.


If there is anything else I can think of that might help, I'll come back and post it promptly. For now, that's all I've got. Anyone have any ideas? Thank you so much for your time in advance!

 

[JackNaylorPE]

I had the three sticks of RAM in slots A2, B1 and B2 .... I went into the bios and turned the frequency on the RAM down to 1333, removed the B2 stick of RAM altogether and moved the B1 stick of RAM into the A1 DIMM slot.

Read your manual ...... Memory should be in the 2nd and 4th sots from the CPU .... A2 and B2

From where you started, remove B1 and your gold.

If it was advertised as 1600, it should be 1600 ....what's it say on the sticker ? One thing I notice with some RAM / MoBo combos is that you can not manually set it at 1600 ..... the only way to get it 1600 is to set the XMP profile which bumps the BCLK.

Freezing during gaming is usually a sign of temperature or instability under load ... as in the other thread, we'd need to know ur CPU and GPU temps ....we'd need to know the Event Viewer messages.

Put the Modules in A2 and B2 and run memtest86+ from http://www.memtest.org/

Then move on, if it passes, to OCCT GPU and CPU testing, look at the graphs when done....watch the temps / voltages for 1st 10 minutes

 

[JoshDBoyle]

Thanks for the reply, Jack.

I moved around my RAM to have a 4GB stick in slots A2 and B2 like you've said. I also managed to run memtest86 and everything passed just fine but I did notice something that doesn't seem right (although it may be a lack of understanding on how memtest86 works on my part). On the screen during the test, as it lists out some system data such as the speed of your CPU and the size of your L1, L2 and L3 cache, I noticed that the line that begins with the word "Settings" read as follows:

Settings: RAM : 596 MHz (DDR3-1193) / CAS : 9-9-9-24 / Dual Channel

596 MHz? Is that just the frequency memtest tests at or is my memory set to run at that frequency?


Anyway, I'm now moving on to GPU testing to see if I can find any issues there.

 

[JoshDBoyle]

Both GPU and CPU testing with OCCT are finished and there were no errors. I ran both for the full hour.

Any other ideas?

 

[Leaps-from-Shadows]

I would suggest running the Furmark GPU benchmark/tester program. 1.9.0 is the newest version. It's pretty much the hardest load you can put on the GPU. Run the burn-in test overnight with the Fullscreen, Dynamic background, Burn-in, and Post FX options checked. Note that the hair dryer fan noise might get a little loud.

It's also entirely possible that something else was damaged when your mainboard shorted, and only manifests itself when the system is fully stressed. As soon as I saw that something actually melted off the mainboard my first thought was "RMA the board."

 

[JoshDBoyle]

Will do, Leaps-from-Shadows. I'll run it overnight tonight. As far as RMAing the board, I'd have to look and see what the time limit is after purchase for being able to do so. I do know that eventually I had plans to upgrade the board anyway as this particular board doesn't offer full CrossFire and SLI. I can't remember exactly why this was but it had something to do with the fact that one PCI slot was full x16 speed while the other was only x4 speed and thus running a second card in Xfire would be bottlenecked by the second PCI slot.

I'll let you know the results of the test.

 

[lun4tic]

I had the exact same problem, and it did turn out to be a RAM/MB problem. I was using a X58A-UD3R motherboard with A-DATA RAM. I would crash during games and during Prime95's blend test.

As it turns out, the motherboard didn't like four sticks of RAM, period. I attempted to use them in either a dual-channel or triple configuration and both resulted in crashing. According to the manual, I should have been able to use them in the first through fourth slots, for dual-channel, or 1st, 2nd, 3rd, and 5th for triple channel. Neither worked, so I pulled a stick out and tested three sticks in 1st, 3rd, and 5th slots, and that worked like a dream.

I don't know if it's an issue with a motherboard or the RAM, but Memtest came up clean for every stick.

BEWARE OF FURMARK. It will completely blow your PSU if you're not careful and could damage other parts in your system. I hope you read this before running your tests.

As far as SLI/Crossfire go, most motherboards offer a second x16 lane running in a reduced mode due to bandwidth concerns. I'm not experienced with Crossfire, but I wouldn't bother with SLI. nVidia's drivers rarely recognize it without a lot of uninstalling/reinstalling, and you're forced to create a new SLI profile for almost every game you play, otherwise you'll be playing with only a single card. It's much easier to run with a single, more powerful card.

 

[Leaps-from-Shadows]

Note that both ATI and Nvidia have instituted auto-throttling when your system runs FurMark. The risk of damage is minimal.

 

[JoshDBoyle]

So I ran FurMark a handful of times and the GPU is running like a champ. The GPU starts from idle at around 60C and ramps up gradually to around a maximum of 82-84C where it then sits for the duration until FurMark has run its course. No issues and no hiccups.

After this testing, I think we've got a strong indicator towards a MB/RAM issue so I'll list out those items here again in hopes somebody has some in depth knowledge of them:

Motherboard: Asus P8P67 LE B3 Revision
RAM: 3 x 4GB Kingston HyperX DDR3 1600mhz

As mentioned, I've only got two of the three sticks in the MB right now and they are in slots A1 and B1 with slots A2 and B2 being empty. I've got the third stick sitting here on the desk starring at me.

 

[JoshDBoyle]

After doing some more research, I'm now strongly thinking that the RAM is the culprit here. In looking through my mobo's manual, I came across a RAM compatibility chart and the RAM I have is not listed. The part no for my RAM is as follows:

KHX1600C9D3K3/12GX

The closest match in the table for my mobo for 1600 RAM is the following:

KHX1600C9D3K3/6GX

Another thing I noticed is that, for every single Kingston RAM that is compatible with my motherboard, NONE of them say 4GB or more. ALL of them are exactly 2GB a stick.

Could the above cause system freezes after some unspecified amount of time?

 

[i7ftw]

I had a horror build that i spent hours trouble shooting in which i tought it was a driver or a hardware issue. Everything on the build was brand new so it was hard to pin point the issue. I then tryed booting and running using only one stick of ram and there i found the prob, one stick ran great the other blue screen. Both together it ran and booted but would freeze or blue screen eventually. I had the same issue mb saw the ram at 1333 and not 1600.

run a stress test or windows scan the memory
what drivers are you using with the 6970?


I72600k
Sabertooth P67
Intel 320 ssd
HIS HD6950
Corsair 1600 2x4 9-9-9-24
Corsair 750tx v2
Windows 7 64

 

[chavez1]

Windows 7 Ultimate 64 bit Randomly Freezes
No BSOD just random freezes with screen on and no keys or cursor repsond, all I can do is use a hard reset to restart. Then I get the standard windows was not shut down properly messege at start-up. I have been having this issue/s and have tried all the same standard help people have had to offer, in all the dfferent post I've read online.
To actually fix this issue I ended up having to go to the device manager and check for driver updates on all my devices, one at a time. I think the ones that had updates that fixed this issue were the ACPIx64-based PC under the computer drop down and the IDE ATA/ATAPI controllers especially the ICH ATA and SATA AHCI ontrollers (I think these made the difference). I checked them all, but these were the only ones that had updates. Once those were updated and the computer was restarted, it has stayed on. It's been on for a day now and no problems so far.