Random BSODs and Program Crashes

[r1zeek]

I've recently built a new machine, had problems at first with dodgy memory and RMAd them, had the replacements running for about 2 months all fine.

Today my machine has started to randomly crash background apps "Superfetch has stopped working" "Windows Destop Manager has stopped working". This would be followed by a BSOD where it would show it was dumping "physical memory to disk" before restarting. There's nothing else specific on the BSOD, it happens v quickly so can't see anything.

Suspecting this was a virus I formatted and installed Windows 7. Good clean install and still the same problem, applications randomly crashing followed by a BSOD. I haven't worked out any reason, they generally happen when browsing the web or virus scanning. Did a full test on 3DMark06 to see if it was stress related and nothing. I moved the ram about, took one stick out etc and doesn't seem to be anything helpful.

I built my machine in September out of the following:
Intel Core2 Q9650 (Quad Core)
Asus Maximus II Formula mobo
Corsair 2x 2GB DDR2
XFX 8800GT
Corsair Modular 720w PSU
Western Digital 7200rpm 500GB HDD

All drivers are up to date, nothing is overheating (NB is running aroung 39 at idle) so I really can't think what it is. Possibly a motherboard issue?

[peterb]

Download and run memtest 86+ just to make sure it isn't a RAM problem.

http://www.memtest86.com/

[mac124]

BSOD's tend to be memory related (but not always) so memtest is always a good idea. The first thing i would do though is double check all the bios settings for memory are correct. If you haven't set them manually then it could well be running the voltage too low and / or wrong timings.

[r1zeek]

Cheers lads, thanks for the advise.

Been testing all the memory, I've tested both sticks individually and they both pass in memtest. However when they're both in together memtest won't get past 1 percent (crashes almost straight away). It doesn't matter if they're in dual channel mode or not they always crash.

They're Corsair Dominator PC8500 by the way. I've tried manually upping the voltage to 2.1v but it is exactly the same effect. I've tested 1 stick of Dominator and a stick of generic ram and they both run fine so I feel i've eliminated the possibility of any problems with the slots.

Need to go and make sure the memory timing is set exactly as it should. Anyone got any other ideas?

[r1zeek]

With the ram set to 5-5-5-15, voltage at 2.1v (tried at 2.2 and 2.3) with the FSB set to 1066Mhz it's still unstable. Is this dodgy RAM? (eligle for an RMA?)

[r1zeek]

Even with a single piece of the Corsair Dominator running, it still caused the BSOD. I've now got a generic stick of DDR666 in there, running memtest to see if it causes a crash.

I'm really quite stumped as to what is going on whether it's the memory or the motherboard?

[peterb]

Tricky - I would still tend to suspect the memory. The problem is that most memory sellers are memory assemblers - they buy the actual memory components from various sources and assemble them, and while the memory chips from different mfrs may have the same characteristics on paper, they may be slightly different in practice, so sticks from the same mfr may have very slightly different characteristics unless they are matched or built from chips from the same source and batch. I have come across this problem. (funnily enough, with Corsair memory)

However it could be the motherboard - and sadly the only way to check it really is substitution. Memory is the cheapest bit, so you may end up having to buy some new memory (different mfr -0 and get them as matched pairs).

The risk then is that it is the mobo.

Have you tried running something like a linux live CD? That is pretty demanding on hardware, and wold help discount the possibility of an iffy driver in Windows (although you would normally expect a different type of windows error)

You could (if your mobo supports it) try underclocking the mobo to see if you can get a stable system.

[r1zeek]

The memory is definately matched pair, everything had been running fine until the other day when it just started doing it. Nothing had changed hardware wise and I was just using it the way I normally use it.

The DDR666 seems to be running ok, 9 passes on Memtest86 with no errors but we'll have to see how it goes when running in windows.

[r1zeek]

Searched the Corsair forums, found people had similar BSOD issues with this ram, advise was to disable legacy USB support, manually set FSB speed, latency and voltage aswell as possibly increasing the NBv by 0.2 or 0.4

Installed the two modules and tried all manner of combinations of the above and it still won't get past Test #1 on Memtest86 (v4)

When a single module from the pair is installed there are no errors on Memtest, but it still creates instability on the system.

When the DDR666 is installed, system passes memtest and keeps the system stable.

This leads me to conclude the memory is faulty, but I can't see why they pass memtest on their own?

[peterb]

I think I would come to the same conclusion - and the only explanation I can think of is perhaps they aren't quite as matched as you might have been led to believe. Depending on the age and the retailer, you may be able to return them as unsuitable. (IIRC, Corsair memory has a lifetime warranty)

[mightymouse]

have you considered testing out the mobo slots? not sure if its the same issue but i had a problem once and it was because of the slot that was not working in dual mode. just keep trying different combination of slots to see if its really the RAMs if it is then RMA them, corsair is quite good about it.

[CupofCare]

Can you make a photo of the bsod and post it here (i don't know if anyone asked about it yet)

[r1zeek]

Returned the memory to Scan, they tested it on their rigs and it came up with no problems, Michael W from Scan suspects a potential fault with the memory controller on the mobo. System has been stable ever since I installed the single piece of DDR667

[r1zeek]

Sorry to drag this thread back up from the dead but wanted anyones opinion on this matter:

The problem started happening again and some of these blues screens said "Memory Management error" as they were dumping to a file.

I decided to do some more investigating and substituted my Asus Maximus ii Formula for a brand new one - my theory being that if it were a duff mobo this would show up.

However, when testing with 1 stick of anything, it MemTested fine. When there were two sticks (2x2Gb), it would crash within 1 second of starting the first test.

As mentioned before, I had returned the RAM to scan and they hadn't found any problem. I borrowed my mates RAM which was 4x1Gb sticks. When I loaded two into the rig (totalling 2Gb), there was no crash and memtest was fine. Investigating further I loaded up all 4 sticks and lo' and behold, it crashed!

I've tried reducing the FSB of the processor (it's an Intel Core 2 Q9650), upping the voltage for the memory, all the other things I've seen suggested for this motherboard.

I'm thinking it must be the processor? As every other component has been changed and this is the only constant.

Can anyone help by offering their opinion on this matter?