3XS Edge i7 930 D0 4.0Ghz Bundle BSODs/Errors in Prime95

[JWRPayne]

Build No.: No Record on Invoice/Receipts
Invoice No.: OOE1418288

Hey all,
I have been having a few problems with my 3XS bundle, as of late. I thought it may be worth my while trying to get this sorted on here before I take the time to contact you guys over the phone/via e-mail. The specs are as follows:

Intel Core i7 930 2.8Ghz (@ 4.0Ghz, by 3XS) S1366
Gigabyte GA-X58A UD3R (Rev 1.0) w/ BIOS v.F4 (as installed by 3XS)
6GB Corsair XMS3 DDR3 1600Mhz
Prolimatech Megahalems CPU Cooler
120mm Akasa Apache CPU Fan
MSI GTX 460 OC Cyclone 1GB GPU w/ ForceWare 258.96
Asus Xonar D1 Sound Card
Western Digital Black Caviar WD500 500GB SATA HDD (Primary, Boot, OS)
Samsung HD502 F1 500GB HDD (Secondary, Backup)
OCZ StealthXStream 600W PSU
Windows 7 Enterprise 64-Bit

I’ve had the system since mid-March this year. Apart from the odd BSOD once or twice, it has run flawlessly. I haven’t touched or changed the BIOS settings input by the 3XS team, apart from turning HyperThreading off/on (discussed below). However, the system has recently started BSODing for no apparent reason. This often happens during idle stand-by, with little to nil load on the CPU. Very confusingly, however, extensive gaming sessions with intensive games (e.g. Company of Heroes, Battlefield: Bad Company 2, Stalker: Call of Pripyat etc.) go without fault, error or BSOD.

To begin discussing my diagnostic deductions thus far, I will run through any changes I have made and how my thought as to the source of the problem have changed over the last few days.

I have recently switched off HT in BIOS, as this is known to crash and cause BSODs whilst playing Company of Heroes, since the latest patch (a quick Google will give more info, as I can't post URLs). I assumed, albeit counter-intuitively, that this may have caused such system instability, as everything seemed to function perfectly before around the time of disabling. However, turning HT back on has not resolved the issue (but the crashing, BSODing Company of Heroes returns…).
To investigate further, I decided to blend stress test the CPU under Prime95 with both HT on and off. Each time, a rounding error was reported for one of the cores in under 15 minutes, with all of the cores slowly following with the same error (although, it may be worth noting that speed of the errors was slowed by having HT on). Reports from Prime95 can be provided if this will aid diagnosis.
Temperatures during Prime95 stress testing seem reasonable, with peak core temps hitting 65°C (idle ~26°C) with HT off, and 74°C (idle ~30°C) with HT on. Northbridge temps never go above 57°C (idle ~51°C, which is quite good for X58 systems, I have heard). This, for me, means cooling is unlikely to be the issue (although, I will stand corrected if anyone can show me otherwise).

Thus, I turned to my PSU. I have wondered if maybe the PSU isn’t giving enough power to all of my components. However, running my components (opting for more power consuming variations, if my exact component isn’t available) past the CPC Advanced PSU Calculator gives an estimate of 450W, 25A on the +12V rail, 10.2A on the +5V rail and 4.8A on the +3.3V rail. Not an exact science, I admit, but my PSU more than satisfies these requirements, even if we assume that the CPC calculator has generously underestimated my requirements (see the OCZ products pages, as I can't post URLs).

The only other explanation I have for the new problem is either RAM or my recent GPU upgrade. The problem has seemingly occurred since I replaced my GTX 260 XT my new GTX 460 OC, but I fail how this could be causing the discussed problems (for example, the new card consumes less power than my old GTX 260, as discussed above and in various power consumption tests), unless it could be something like video drivers (the new generation nVidia drivers for the 400 series are bit flakey and awaiting a proper update). But, this would not explain the Prime95 errors also discussed above.

‘How about RAM?’, I hear you ask. I’m yet to test the RAM, so this may indeed be the problem. I am, as we speak, downloading and burning MemTest86+ to run some tests. I will update the thread with anything new I learn.

In the mean time does anyone, 3XS or otherwise inclined, have any idea what is going on or have a way of explaining what is going on here? Although keenly interesting in building and overclocking computers, I must plead a little ignorance to some things and settings, so I apologise if I have missed something that is blatantly obvious. Likewise, if any error reports could be of use (e.g. the BSOD dumps recorded in Windows Event Viewer), these can be provided.

So, does anyone have any further suggestions or ideas? I appreciate any help given.

Best,
Jack

[JWRPayne]

Quickly, for the record, I read through the above 'Power Plan' thread, and my system was already set up in such a manner.

[JWRPayne]

‘How about RAM?’, I hear you ask. I’m yet to test the RAM, so this may indeed be the problem. I am, as we speak, downloading and burning MemTest86+ to run some tests. I will update the thread with anything new I learn.

To update what has been said so far, after nearly 6 hours of MemTest86+ testing, my RAM seems to be rock solid. Previous experiences with dodgy RAM has usually kicked up a fuss within a few minutes using this test. Admittedly, it's not a 24/7 test but, to be honest, doing this well so far when there is evidently a problem elsewhere can only be a good sign. It also isolates the issue towards the CPU, in my eyes (again, please correct me if you think I'm wrong).

Does anyone have any thoughts on the matter?

[JWRPayne]

And to drive home my point about the Prime95 failures, the errors are appearing within one minute of blend testing with HT off, and two minutes with HT on. It's always two cores that fail at exactly the same time, generally always core #1 along with core #2 or #3. Is this an RMA case?

If you've just picked up on the thread, please read down from the opening post. Any help is massively, massively appreciated.

[JWRPayne]

Right, I have now removed all of the USB devices from my back panel, bar my mouse and keyboard (for obvious reasons) and, so far, Prime95 has run for 12 minutes without error.

Is this looking like a power issue?

[JWRPayne]

...and a reboot later, its BSODing again. Nothing has changed. Really think this is an RMA case.

[kalniel]

Go back to the original build - put the 260 back in and see if you can replicate the error - if you can then you've ruled out the new card. If you can't then you can focus your tests on the 460's 2d mode perhaps.

[Nzastudios]

The GTX 400 series I reckon are happier with 7series PSUs and up.

[JWRPayne]

Go back to the original build - put the 260 back in and see if you can replicate the error - if you can then you've ruled out the new card. If you can't then you can focus your tests on the 460's 2d mode perhaps.

Alas, the GTX 260 has been sold on. However, there is a GTS 250 that I can get hold of, that I know works, so I'll try running some tests with that.

Hopefully, like you said, that will rule out the new GTX 460.

[PeteG@Scan]

You may wish to cross post this to the main scan care section as we don't have experiance of working with these bundles in system support! We only handle the systems and not the bundles

As such the isn't much I can suggest as far as specifics go with those overclocks. One thing that does seem odd through is that it only happens at idle. Have you enabled a screen saver or is the screen saver slideshow option running. We've seen this cause the sort of problem your having with 4ghz+ overclocks.

[JWRPayne]

You may wish to cross post this to the main scan care section as we don't have experiance of working with these bundles in system support! We only handle the systems and not the bundles

Ah Pete, my mistake. I thought this sub-section would account for all of the 3XS products. Ideally, I'd like all of the posts to move across. Do you have sufficient moderation power to move the thread to the section you advised? I'd really appreciate it, if you could.

Regarding your suggestion, I have indeed turned off all of the Win7 oddities, as suggested in your power plan thread stickied at the top of this forum. The problem continues to occur, not only when idling, but also gives errors through Prime95 (as discussed in my above series of posts). This suggests to me that the problem lies beyond a simple windows 'option' and is hardware or driver related.

To update a bit, I have formatted to see if this resolves the issue. Again, errors are chucked out my Prime95 within one minute of blend testing. Likewise, I have removed all unecessary items from the PSU (i.e. storage HDDs, Blu-Ray drive etc) and it still BSODs. So maybe this isn't an issue of insufficient power?

I am yet to swap out the GTX 460, so I'll report back when I know more about that.

[PeteG@Scan]

I'd be suprised if it was but then again the discription matches :-/

OCZ psu's are pretty middle of the road performance wise but it certainly shouldn't be pulling 600w's worth so it should handle it. Saying that a power sag could explain the cores dropping, maybe check raise the voltage on the PLL & Vcore slightly.

Anyhow I'll move it over to the scan care section now for you.

[Jay]

does it have speedstep?

[JWRPayne]

It does indeed, but it is disabled by default.

Formatting seems to have resolved the idle BSOD issue (maybe it was a really bad, repeating indexing/checking error that Win7 performs whilst at idle?), but the Prime95 failures still come very quickly (within two minutes). The system seems to be stable enough. I'll try notching up the voltages a bit to see if they stabilise things out a bit.

Indeed, I didn't think this system would kill my OCZ PSU. I went for this particular one as, according to the reviews, it is just a rebranded top-end Zalman unit. Saying that, the fan controller has broken, so it's running flat out constantly now... joy of joys. OCZ RMA here we come.

Does anyone know what stability test Scan 3XS use for their systems/bundles? I'm willing to accept that Prime95 can be a little pedantic at times, but would like to verify my system stability in the same way Scan 3XS does.

[EzyRyder]

There is nothing pedantic about Prime95. It is the ultimate test and used by people who overclock their computers and want maximum stability. My suggestion would be to save the overclock settings in the BIOS and try the processor at stock speeds and voltages to see if it still has the same problem.

[JWRPayne]

Well, I ran it once and it chucked up errors within one minute. I closed Prime95 down, started it up again, and it was fine for 6 hours (at which point I stopped the test). This pattern can replicated consistently. Excuse my scepticism, but surely that isn't entirely correct?

Regardless, I did try as you suggested and the same errors appear at the same time and in the same fashion (i.e. within one minute, two of the four cores, typically core #1 and either #2 or #3). Any ideas?

Thanks for all the replies, by the way. Much appreciated!