New High end computer keeps getting blue screen of death!

Landoflemon

Hello! Built a new computer in the last fortnight, it's my first attempt at building one so there are a couple of problems that I've been trying to iron out and succeeding, but there is one problem which is persistent and very annoying: Blue Screen!
It happens once or twice every single day.

Here are my Specs:

Core i7 860 (not overclocked)
Radeon Sapphire 5850 toxic
Asus p7p55d-e motherboard
2 x 2gigs of Corsair Dominator DDR3 1600Mhz Ram
850W Corsair modular PSU
2 x 500GB WD Caviar Hard Drives
Coolit Eco Watercooling
Windows 7 Ultimate

My temperatures are not the problem, the CPU is stable at around 33 degrees all the time. I've a feeling the problem is related to the RAM, because after it restarts from a blue screen windows generally will not successfully load unless I take out the 2 sticks of ram and swap them over?!

I've now taken out 1 of the sticks of Ram to see how it works, and if there are changes to my blue screen situation.

I read something about p55 motherboards and 2 sticks of ram and 4 dimm slots, saying that it will work with one stick but I'll need to increase the voltage if using 2 sticks. Is this true??

Thanks for any replies.

Landoflemon

Deano12345

Right, first off, run Memtest86+ to check to see if the RAM is faulty. If its not that I have a feeling it could be a problem with the DIMM slot on the Motherboard itself. Asus' P55 boards have been known to have problems ( Cue Solitaires entrance :P )

Landoflemon

Thanks for the speedy reply, I'll run memtest and post back here if anything shows up, Since I got the computer parts very recently maybe I can just ship the motherboard back to Dabs? My other reason being that it supports crossfire but only with both PCI express lanes in X4 configuration, and I didn't realise before the substantial hit that would take on performance in crossfire.

Deano12345

Landoflemon wrote: »

Thanks for the speedy reply, I'll run memtest and post back here if anything shows up, Since I got the computer parts very recently maybe I can just ship the motherboard back to Dabs? My other reason being that it supports crossfire but only with both PCI express lanes in X4 configuration, and I didn't realise before the substantial hit that would take on performance in crossfire.

If its opened and not faulty, I doubt they will replace it out of the kindness of their hearts unfortunatly. If you try and put it through as a repair, you'll still have to send it to Dabs who will send it to Asus, Asus will look at it and find nothing wrong and send it back.

Now, if the MB is faulty, then you might have a chance.

Crossfire at 4x isn't a huge drop either, a 5850 probably wouldn't saturate a 4x lane by itself

Landoflemon

Hi guys, just ran the memtest utility that was suggested. Immediately the screen filled up with red:

It should be noted that this image is not actually from my computer but is for illustrative purposes. My results looked exactly like this.

So my next step was to find which of the 2 ram modules was the fualty one, so I did 1 pass with one stick, and there were no errors. Then I did 1 pass with the other stick and again no errors. However when I ran the test once more with both sticks in again the results had many red lines and errors. I am thoroughly confused. Anybody have any idea what to do next??

Thanks for your help and suggestions so far.

Solitaire

Suggests an issue with either the mobo being screwed (memory slot or channel damaged) or (less likely) memory controller not getting enough voltage to remain stable. Try retesting the modules one-at-a-time only using the other slot (the one you didn't use for the first two tests).

duffman85

Landoflemon wrote: »

Hi guys, just ran the memtest utility that was suggested. Immediately the screen filled up with red:

It should be noted that this image is not actually from my computer but is for illustrative purposes. My results looked exactly like this.

So my next step was to find which of the 2 ram modules was the fualty one, so I did 1 pass with one stick, and there were no errors. Then I did 1 pass with the other stick and again no errors. However when I ran the test once more with both sticks in again the results had many red lines and errors. I am thoroughly confused. Anybody have any idea what to do next??

Thanks for your help and suggestions so far.

I had a similar problem with ram which memtest said was faulty(i.e. lots of errors) but which turned to be quirk of the mobo/ram. It was stopping my new build from starting.

The mobo/bios wasn't detecting the voltage automatically from the ram modules. In the end setting the ram voltage manually in the bios instead of leaving it on auto solved the problem.

Landoflemon

Hi Duffman, Solitaire.

About an hour ago I did spot that the memory was at about 1.54Volts and on the memory itself it was printed 1.65volts, so I changed from auto to 1.65 in the bios and memtest then passed with both memory sticks in the system!

I'm now just using the computer as normal to see if it will blue screen on me, but thank you very much for your suggestions. I'm very new to this scene and am still making my beginners mistakes!

Thanks for everything

Deano12345

Landoflemon wrote: »

Hi Duffman, Solitaire.

About an hour ago I did spot that the memory was at about 1.54Volts and on the memory itself it was printed 1.65volts, so I changed from auto to 1.65 in the bios and memtest then passed with both memory sticks in the system!

I'm now just using the computer as normal to see if it will blue screen on me, but thank you very much for your suggestions. I'm very new to this scene and am still making my beginners mistakes!

Thanks for everything

Glad you got it sorted

Dublin_Gunner

Landoflemon wrote: »

Hi Duffman, Solitaire.

About an hour ago I did spot that the memory was at about 1.54Volts and on the memory itself it was printed 1.65volts, so I changed from auto to 1.65 in the bios and memtest then passed with both memory sticks in the system!

I'm now just using the computer as normal to see if it will blue screen on me, but thank you very much for your suggestions. I'm very new to this scene and am still making my beginners mistakes!

Thanks for everything

Good stuff. Silly BIOS or SPD setting the RAM voltage incorrectly.

Run a few tests over a few hours to to make sure its stable still - Prime95, 3DMark, or even a few games to stress it out.

Hopefully you sorted out the problem

Landoflemon

Nope, just blue screened on me there again when running Pro Engineer :mad:

zappb

Something very strange going on with that memory / motherboard.

It could be a faulty motherboard - If its not the CPU Seating (Re seat just in case) - Try the stock Heat Sink Fan for awhile to take out the water cooling variable, or if its not the memory (try and borrow another 2 gig stick to be sure) - and it couldn't be power related with that beast of a PSU.



Anything weird happen with the Windows 7 Install? Any driver issues or are you running a RAID setup? What about the Toxic card? (Fabulous card btw...) - Can you try and older graphics card for awhile?

If your not monitoring the 5850's temps - keep and eye on them just in case just in case its overheating?

when it blue screens are you doing anything in particular?

Landoflemon

Hey Zappb, the 5850 is idling at around 32 degrees and I just played a few minutes of bad company 2 and it was up at 62 degrees but that is probably standard enough yeah?

Ok, next time it blue screens, probably at some stage over the weekend I'll take out the processor and reseat it to make sure, install the stock cooler, and put my old radeon x300 that came with my old dell dimension 5100 in instead of the 5850. That's going to be depressing.
I don't think it's to do with the water cooling as the i7 is idling at around 32 degrees which seems fine to me, but I'll change the cooler anyway just so as I can rule that out. I only have 1GB of other ram from that dell that I can install instead of the corsair dominators, but this won't run windows 7 so it sort of looks like I'll have to stick with the corsair ram. Unless I install XP during this testing period.

I could be doing anything when it bluescreens, replying on Boards.ie, checking Email, just sitting at the desktop. Interestingly it's never Blue screens when playing any games, like Crysis Warhead or Bad company 2.

Perhaps it's worth noting that performance is always very smooth and no problems there, which is what makes it more of a surprise when it does crash.

Actually, one thing I forgot to report, a few days ago all 4 cores went to 100 percent when using pro engineer, not doing much just opening a file, and everything went funny shades of yellow and pink and very slow to respond. I can't remember if it did blue screen that time or if I just powered it down. That is the only time I've had that happen.

Landoflemon

Also, I'm not running a RAID setup, and there are no driver issues as far as I can tell. Installation of Windows 7 went without a problem.

Dublin_Gunner

It could, even though unlikely, be a faulty PSU.

Do you have a lower end vid card to test with?

See how stable it would be with a basic system, i.e CPU, 1 stick ram, 1 HDD, low end vid card.

That would consume less power, and may provide stability.

Do you have other ram to test? (as it may be the ram). It may be worth while RMA-ing both the ram and the motherboard, and have them replaced.

Also, have you tried moving the DIMMs to the other 2 ram slots?

Solitaire

Yeah, did you ever try to run Memtest86+ while the memory modules were in the other memory slot? That sounded like a likely contender earlier!

BigEejit

I had a gigabyte mobo that was VERY picky about the ram I put in it. I changed just about everything that could be changed and just could not get it running reliably .... one boot everything was fine, next boot it blue screened. i was convinced it was the ram and RMA'd it and they sent it back a month later saying it had worked with no issues.

So, did you check the mobo manufacturers website, is your ram down as compatible? Can you check the ram in another box somewhere?

Finally, after a lot of bsod's my windows 7 installation was buggered and failed with bsod's all over the place ...even after I sorted the ram problem. I had to do a fresh install to get it all settled again.

Landoflemon

Hmmm, ok just had a look through the Asus compatibility list for the mobo and my Ram is not on it How particular are motherboards in general about the ram that is running on them? I think I may have been just a little too lazy when picking out the motherboard and I didn't know there were different levels of compatibility that the ram had to comply with.

Just ran another 3 passes in memtest 86 for the last hour and no errors showed up with both sticks installed.

Solitaire, which slot orientation are you referring to? Currently the ram is in slots 1 and 3. I put the ram in slots 2 and 4 at one stage the other day when testing and it didn't boot into memtest or anything at all. When I tested the sticks on their own it was in slot 1, do you mean test them individually in slot 3, or which slot?

Also, after me setting the voltage to 1.65v (which is the recommended voltage), and it still bluescreened, I then went and gave it 1.685 volts to test it like that. I must admit I did this without much thought, and nearly just out of frustration and curiosity. It is safe enough to increase the voltage just like that yes? Or do you think I should set it back to 1.65v and just reformat everything and install win 7 again?

Dublin_Gunner

Landoflemon wrote: »

Hmmm, ok just had a look through the Asus compatibility list for the mobo and my Ram is not on it How particular are motherboards in general about the ram that is running on them? I think I may have been just a little too lazy when picking out the motherboard and I didn't know there were different levels of compatibility that the ram had to comply with.

Just ran another 3 passes in memtest 86 for the last hour and no errors showed up with both sticks installed.

Solitaire, which slot orientation are you referring to? Currently the ram is in slots 1 and 3. I put the ram in slots 2 and 4 at one stage the other day when testing and it didn't boot into memtest or anything at all. When I tested the sticks on their own it was in slot 1, do you mean test them individually in slot 3, or which slot?

Also, after me setting the voltage to 1.65v (which is the recommended voltage), and it still bluescreened, I then went and gave it 1.685 volts to test it like that. I must admit I did this without much thought, and nearly just out of frustration and curiosity. It is safe enough to increase the voltage just like that yes? Or do you think I should set it back to 1.65v and just reformat everything and install win 7 again?

1.685 should be perfectly fine, probably up to 1.7/8 would be fine for overclocking (not saying you should do that though!!)

Motherboards usually aren't that picky these days, and just because your modules aren't on the list, doesn't mean they wont work - just that Asus haven't tested them with that motherboard.

ON your mobo; channel 1 = slot 1+3, channel 2 = slot 2+4

Landoflemon

Thanks for the reply DublinGunner, just doing more testing of the ram on another machine at the moment. It's reasuring to know that just because the ram isn't on the list doesn't mean it wont be compatible.

Landoflemon

I ended up swapping the motherboard with somebody else who was interested in swapping. I now have an EVGA p5 ftw.

Before exchanging motherboards I found the solution to my problems by pressing the memOK! button on the motherboard until it established it's the memory timings that were stable, and then I had to save these changes in the bios.

The new motherboard is working fine now too (except for a sound problem, but I'll work around that). Thanks to everybody who had an input while trying to help me, just thought I'd write to put an end to the thread.