SB 2500k setup was stable, now blue screens. Any ideas?

[Spike]

I built my 2500k setup basically right when they came out and after running OCCT, IBT, Prime, linpak, and several others I can't even remember I found a stable bench at 4.4 with 1.28V.

Since then everything has been chugging along until last Friday night when I got a blue screen while playing BF:BC2. I know that game stresses components heavily so I ignored it, even though I had played MANY hours of the game since declaring the build stable and it had no problems apart from slightly high temps (mid to upper 60's under OCCT, mid to upper 50's under BC2). I randomly increased my volts to 1.285 on Saturday just because, hoping that may have covered the problem. Today I just had another blue screen while just sitting here surfing the web so now I'm starting to get worried.

What do you guys recommend doing? Starting over with the OC to find something stable or should I look elsewhere for the solution? I did upgrade my 6970 drivers to 11.2 but I did that yesterday, in between the 2 blue screens so I'm not sure if that's the issue or not.

Thanks for any help.

 

[Markfw]

I found a problem with my motherboard, and once I stopped using the sata2 ports (went to an IDE controller card and drive) the problem disappeared.

The sata bug may be more serious than they suspect. This is just one thing to try...

 

[Axon]

Mark9hunnid, if it wasn't for bad luck with hardware, you wouldn't have any luck at all.

 

[Accord99]

Does the BSOD have any error message. There is a Microsoft hot-fix that seems to have worked for some Sandy Bridge owners:

http://support.microsoft.com/kb/979444

Also, I noticed in my overclocking experience with SB and a P8P67 Deluxe that when I used an aggressive LLC mode with offset settings for the vcore, while the vcore was stable when 4 cores are loaded, the vcore actually drops as fewer cores are loaded. This resulted in the vcore being too low for when only one core turbos up to full speed and a BSOD. This was observable by Prime95 being stable with 4 or more threads, but quickly blue-screening with only 1 thread.

I went back to the automatic LLC mode, raised the vcore offset a bit higher; this results in the same voltage when 4 cores are loaded but with much less of a drop with a lower load and slightly higher idle voltage. I haven't seen the BSOD since.

 

[Spike]

Does the BSOD have any error message. There is a Microsoft hot-fix that seems to have worked for some Sandy Bridge owners:

http://support.microsoft.com/kb/979444

Also, I noticed in my overclocking experience with SB and a P8P67 Deluxe that when I used an aggressive LLC mode with offset settings for the vcore, while the vcore was stable when 4 cores are loaded, the vcore actually drops as fewer cores are loaded. This resulted in the vcore being too low for when only one core turbos up to full speed and a BSOD. This was observable by Prime95 being stable with 4 or more threads, but quickly blue-screening with only 1 thread.

I went back to the automatic LLC mode, raised the vcore offset a bit higher; this results in the same voltage when 4 cores are loaded but with much less of a drop with a lower load and slightly higher idle voltage. I haven't seen the BSOD since.

Thanks for the link, I'll check that hotfix out. I don't know what the error code was as I did not look that closely. Can I go into the admin tools and check the log, will that have details on a BSOD?

My LLC settings are on auto, I only messed with the vcore and long load power draw max to get this OC "stable"

Mark, I am using some of the SATA 2 ports but I kinda have to as I have 5 SATA devices and only 4 SATA 3 ports. I guess I could just move my bluray burner to the SATA 2 and leave the rest on the SATA 3 ports for now. I'll give that a whirl.

 

[BTRY B 529th FA BN]

RAM usually slowly degrades

 

[Sephire]

Return.

 

[eternalone]

It has begun, the chipset debacle only delayed it. These cpu's may be VERY sensitive to voltage increase. Please return your voltage back to stock wich should be around 0.988v and run the tests again. These chips may degrade very quickly and you dont want your money to go down the drain. Being that its only been 3 weeks or so Id say be very conservative, and wait another month or so before pushing these babies really hard.

 

[Markfw]

It has begun, the chipset debacle only delayed it. These cpu's may be VERY sensitive to voltage increase.

Did you even read my post ? Its wasn't cpu in my case, but motherboard.

 

[Accord99]

Thanks for the link, I'll check that hotfix out. I don't know what the error code was as I did not look that closely. Can I go into the admin tools and check the log, will that have details on a BSOD?

I don't think so, I've looked at the error events that I have but they don't provide the error message from the BSOD.

 

[Spike]

RAM usually slowly degrades

After 3 weeks? Seems odd, I've had dozens of sticks over many pc's and only had to send one back after a few years as it stopped passing memtest. Otherwise they are all chugging along, some a good 7 years old.

Return.

Oh I plan to... at least the mobo that is. And only when they have the fixed chipsets in stock

It has begun, the chipset debacle only delayed it. These cpu's may be VERY sensitive to voltage increase. Please return your voltage back to stock wich should be around 0.988v and run the tests again. These chips may degrade very quickly and you dont want your money to go down the drain. Being that its only been 3 weeks or so Id say be very conservative, and wait another month or so before pushing these babies really hard.

What the heck are you even saying? Your implying that the new chips don't like volts that are no where near exceeding the intel max VDC?

I don't think so, I've looked at the error events that I have but they don't provide the error message from the BSOD.

Yeah, I looked before going to work and no such luck. If/When it happens again I'll have to pay attention so I can see if the code on that hotfix is what I'm seeing.

 

[MrTransistorm]

Yeah, I looked before going to work and no such luck. If/When it happens again I'll have to pay attention so I can see if the code on that hotfix is what I'm seeing.

One quick way to find it (assuming you're using Windows 7) is to open Reliability Monitor in Control Panel. Click on the day that the BSOD occurred. Look at the list of errors below. It'll probably be either "Windows stopped working" or "Shut down unexpectedly." Select the error, right-click, and choose "View technical details." Depending on how it displays the details, the error code will either follow "The bugcheck was:" or "BCCode".

You can also find the errors in Event Viewer, but I find it a bit more cumbersome for finding specific errors quickly.

 

[BD231]

Have you tried loosening your memory timings?

 

[tweakboy]

Yes, might be the SATA bug your experiencing with the Sandy Chipset.

 

[BoomerD]

Reset your CMOS to stock. See if that cures the problem. If so, proceed slowly with your overclock.

 

[deimos3428]

Does your ambient temperature in your computer room change seasonally? (Mine does; it's in a renovated attic.) If that's the case you may not be able to get the same overclock all year round.

 

[Spike]

One quick way to find it (assuming you're using Windows 7) is to open Reliability Monitor in Control Panel. Click on the day that the BSOD occurred. Look at the list of errors below. It'll probably be either "Windows stopped working" or "Shut down unexpectedly." Select the error, right-click, and choose "View technical details." Depending on how it displays the details, the error code will either follow "The bugcheck was:" or "BCCode".

You can also find the errors in Event Viewer, but I find it a bit more
cumbersome for finding specific errors quickly.

I'll give that a glance today and see what I can find out.

Have you tried loosening your memory timings?

The ram is running all stock settings and voltage. I may drop them if nothing else is the problem but that would probably be one of my last steps.

Does your ambient temperature in your computer room change seasonally? (Mine does; it's in a renovated attic.) If that's the case you may not be able to get the same overclock all year round.

The temp does change but it's an SB setup... those chips have only been out for 3 or so weeks. Kinda hard to have that large of a swing in 3 weeks

However, assuming I get this whole setup stable again your point is actually one of my concerns. I am shopping for a new case to replace my old P180 so I can get better airflow during the warm months, especially to my 6970 that likes to hit 90c even now when playing BC 2.

 

[deimos3428]

The temp does change but it's an SB setup... those chips have only been out for 3 or so weeks. Kinda hard to have that large of a swing in 3 weeks

Heh, just a passing thought. Around here, over the past three weeks we've transitioned from the middle of winter to an early spring, with about a 30F difference in outdoor temperatures and a whole lot more humidity.

Obviously it depends on where you live, and also how close your OC is to the edge of stability in the first place. I'd mirror BoomerD's comment and recommend resetting the CMOS and troubleshooting further from there.

 

[Spike]

Heh, just a passing thought. Around here, over the past three weeks we've transitioned from the middle of winter to an early spring, with about a 30F difference in outdoor temperatures and a whole lot more humidity.

Obviously it depends on where you live, and also how close your OC is to the edge of stability in the first place. I'd mirror BoomerD's comment and recommend resetting the CMOS and troubleshooting further from there.

Probably not a bad course. I may actually update my BIOS as well since I'm still running the shipping BIOS and they have had at least 4 new versions out since.

As for the temps we went from low 30's to 50's back down to the 30's now... I just wish the weather would make up it's mind.

 

[Bryf50]

This is wierd. I have the same cpu and motherboard and am encountering the exact same issue. My overclock is stable though 24 hour prime and linpack runs, but still bsods on rare and random occasions. It could be during a game or browsing the web, but it only happens every few days or so. the code it throws up ends in ...00124 and from what I've googled could mean almost anything. For now I've put all my bios settings to bone stock to see if they overclock is definitely the cause of it. Because its such a random and not really repeatable issue its gonna be a pain to troubleshoot.

 

[Juji]

Im running a 2600k which runs 4.5 stable at 1.3v, however i tried to increase my voltage to 1.35 and 4.6GHz, that gave me a bsod
How much voltage would you dare running trough your SB if you were shooting for a 24/7 clock?

I got a new bios update for my p67pro today but i havent tried a new oc with that bios yet.

Oh and Spike, i you want a new case i'd recommend the haf x
Tho if you are throw in another 200mm fan at the top