Blue screen of death despite prime95 and memtest stable?

[Diasper]

A while back I had my AXP at 11x200 but then decided to tweak it some more especially with a small fsb increase. I got a small boost upto 10.5x214 which was both prime95 and memtest stable - 23hrs and 8hrs respectively and proceeded to back it down to 10.5x213 just to give me a little more peace regards stability. RAM I think was memtest stable up until about 218mhz at 2.5-2-2-11 - not bad for CAS3 value RAM while my cpu will do 2.3ghz albeit needing 1.85V.

However, it randomly Blue Screens of Death complaining of a pagefile corruption despite it seeming stable otherwise. I've backed it down to 210x10.5 but I still get BSODs although perhaps not as frequently. What gives? How come it can run prime95 and a variety of other stability tests no problem yet still be unstable otherwise?


Just further note, it seems to happen alot when playing quintessential - although I have it open most of the time so that might be mute point. Anyway could this be a softare problem or a hardware problem not related to overclocking. I've had SATA corruption issues with my NF7-S recently out of the blue although have since updated the bios and drivers to the latest.

 

[Hyperlite]

if your memory and processing system is that stable, then your bsod's must be caused by something else
btw, thats pretty crazy that your getting 2.5-2-2-11 out of that ram...try 2.5-3-3-10 and see if that helps at all.

 

[RussianSensation]

Sometimes even a virus can cause BSOD or a driver incompatibility. So you can update to different videocard drivers, clean spam, and run some type of virus scan.

 

[CheesePoofs]

Are your sata ports locked? If they arent' , the increased FSB could easily cause the corruption and BSOD's you're talking about.

 

[MDme]

I agree. it could be the SATA being overclocked. usually SATA 1 and 2 are not locked while SATA 3 and 4 are. could also be the 1/2 multipliers since I've heard boards that "emulate" 1/2 multis. but this is unlikely since you are prime and memtest stable.

 

[AkumaX]

but it's a bizzzarton

i say run memtest86+ and put it on test 5 w/ cache on only

 

[Diasper]

Just to make this more bizarre. The SATA corruption happened when i was just overclocked at 11x200 out of the blue when I had been running it happily for over a year without problems. I had not installed any new hardware - the only changes would have been slightly updated drivers but that would seem unlikely.

Originally, I thought my hardrive had gone bad or had been hacked and ended up with anew hardrive but it turns out the NF7-S has problems with SATA corruption - linky. but why it'd suddenly happen out of the blue after working fine for a year I have no idea. Also, the reported problems tend to occur at higher fsb/overclocks than what I'm doing or else by trying to RAID drives which I'm not doing.

Moreover, now I doubt it would be any malware as I run a fair few scanners, namely: AVG7, Windows Antispyware, Ad-aware, Spybot, Spyware Doctor, spyware blaster, X-Cleaner, ewido and A2. They always come up clean.

I just plain don't understand. Perhaps I'm still seeing SATA corruptions as to keep having errors in the pagefile might suggest that little hardrive corruptions are still happening ever so often. I really hope that's not the case as last time I lost many GB of data - I couldn't rescue anything off the computer including photos as they had all been corrupted. I couldn't even download drivers and programs to burn for a fresh install because they might get corrupted on the way. Well, since updating the bios and drivers I haven't been finding any obvious corruptions like I did before so I thought I had got this problem sorted but maybe it's still happening although imperceptibly.

Other than that I think SATA should be locked and because it's a nForce 2 board it only has 2 SATA ports - I tried googling for the info but couldn't come up with anything. I think perhaps SATA isn't quite native to the NF7-S as it was unfortunately it was immature tech when they introduced it and so used some workarounds which didn't include a proper implementation.


I'm flumoxed. SATA corruption or driver problems?



I'm going to try and run Prime95 on blend again and see what happens as the BSOD didn't seem to happen straight away after overclocking but rather seemed to start some days later. Just some extra info: cooling-wise I idle about 47C and Prime about 59C while I also have a 120mm blowing over my hardrive so heat shouldn't be a problem.

 

[TrueWisdom]

Yeah, if the SATA port isn't locked then that's definitely what's causing the page file corruption--after all, prime95 and memtest show that everything else is stable. Since the pagefile depends on RAM and your HDD, and we know the RAM's not an issue, we have to assume it's the hard drive.

A good way to test this is to drop down your FSB back to stock speeds and run an HDD-intensive test. (The manufacturers normally provide one.) See if it passes that--just to make sure that the overclocking hasn't physically damaged your drive. If it does, bump your speeds back up to 213x10.5 and see what happens when you run the test again. If the port is locked, then it should chug away at that test without receiving any errors.

If it's not locked, then you'll probably need to run at 133/166/200/233 etc. FSB to ensure you remain at a workable ratio, but even then, overclocking poses a risk, as those ratios don't always hold tightly unless they're locked. Also, you should definitely back up your data! Use that DVD burner of yours and make sure nothing valuable's on the system!

 

[Diasper]

Just to update this i turned it down to 200 x 11 (which I originally had it running on fine for 1.5 years and then corrupted) and it was fine - no more BSOD. Then when I went back upto 213x10.5 it was fine - no more BSOD.

However, I then tried 220x10 to see if it could do it - it didn't post and complained of a key windows config file not being found - I though ok and turned it back down to 213 x 10.5 but then the random BSOD came back every so often.

So I then put it back down to 200 x 11, got into windows, restarted and put it bacxk to 213 x 10.5 and it's been fine for 3 days running without any BSOD.

Is it just me or is that really strange behaviour?




ps I didn't get round to carrying out the Hitachi Fitness Drive - it'll be the next thing I do

 

[Duvie]

Sometimes running just those two test are not enough alone....prime may isolate the cpu and memtest the memory but once you start running the cpu, with the ram being accessed and using the gpu you can still have errors...It is more of a system test that is really needed after you have isolated te other components with the test you have run....

That is why I run FH for a week straight while doing all my normal activites...a bit of encoding, watching DVDs, AVI,s, CAD work, DVD shrinking, downloading files...and many times a bunch of those at a time....

 

[Fern]

Originally posted by: Duvie
Sometimes running just those two test are not enough alone....prime may isolate the cpu and memtest the memory but once you start running the cpu, with the ram being accessed and using the gpu you can still have errors...It is more of a system test that is really needed after you have isolated te other components with the test you have run....

That is why I run FH for a week straight while doing all my normal activites...a bit of encoding, watching DVDs, AVI,s, CAD work, DVD shrinking, downloading files...and many times a bunch of those at a time....

True that!

I had my Corsair XMS @ Cas2.0 and it passed memtest for 24hrs. Got these errors you mention in HL2. Set the Cas back to (spec) 2.5 and never saw the problem again.

Mem timings don't mean squat in gaming anyway as far as I can tell.

Fern