Running into major problems on my first attempt at SLI

[Kelston]

So I built a new computer finally and decided i'd finally jump on SLI and with my luck apparently crashed head first into a brick wall.

I narrowed it down to SLI being the problem, however, I don't know how to further narrow it down to determine if it is motherboard, SLI bridge, or something else.

Symptoms:

When I first built the system, got all the drivers up and had everything installed. It seemed just fine. Installed a game (World of Warcraft, go ahead and laugh), played for about 2 hours without a single problem. Performance was wonderful (but then again it's WoW). No hitches. I go to bed, and that's when the problems started. Next day, I turn on the computer, try to log into Windows, doesn't work, get the dreaded nvlddmkm BSOD. Reboot into safemode, uninstall drivers, figure since i'm on the new beta 270s i'd go back to the most recent 266 WHQLs. I try to install them, computer says no and BSOD me midway through the driver install (0000003b system service exception on the ntoskernel).

I rage. And then I panic. And then I settle down, go through standard diagnostics. Figure maybe bad ram. Run 2 hours of passes of memtest. No problems. Run Western Digital's HD diags, no problems. Ran sfc /scannow, no problems.

So I go back in, successfully install the 266 drivers still uncertain as to why that 0000003b happened (though most google results come up as generally driver problems so I figure hey, video card drivers are bad so i'm tryin to fix it right here).

266 drivers are installed. Everything seems okay. Get Dragon Age 2 installed play that for a few hours, no problems. Play WoW again (I raid, sue me). No problems. Shut down my computer to get some food. Turn it back on a couple hours later. PROBLEMS, DEAR GOD WHY.

My mouse becomes unresponsive. Attempts to move it are completely laggy and jerky, it's basically unusable. So I shut down again. Figure since major issues are stemming from video cards, I take out the second one. Boot up, viola. System starts up without issue, no mouse problems. Everything seems perfect.

Do some googling. Some suggestion say maybe i'm an idiot and put the SLI bridge the wrong way (though most people seem to suggest it doesn't matter which way and there doesn't seem to be any indication in the manual that there's directionality) so on two separate boot ups I try the SLI bridge both ways. Same issue, my mouse is unusable. So more searching says that I could try running SLI without the SLI bridge.

Apparently not, trying to SLI without the bridge results in my system starting up in basic VGA mode with Windows telling me that there was a problem with both video cards so it disabled them both.

Individually, I can verify that both cards are functional. I don't know if it matters, but I popped in a PCI-X sound card into the PCI-X slot the 2nd video card occupied for the time being. Figured it would at least do some very minor testing (?) maybe. The sound card seems to be functioning.

What can I do to further determine exactly the root of the problem? Is it a bad SLI bridge? Is it a bad mobo? Is it something else?

System specs:

i5-2500k
Asus P8P67 WS Revolution
8GB DDR3-1600 ram
2x 64GB SSDs (1 boot, 1 apps)
1x 1TB HDD
Theoretically 2x GTX 460 1GB (currently on 1)
X-Fi Titanium HD (occupying the slot)
Windows 7 Pro 64 (non SP1 since tech sites are recommending not installing)
Corsair TX 750W

 

[Raswan]

I had a problem kind of like this, and it turned out to be a bad RAID card. Go figure. You don't list one with your specs, though, so I'm assuming you don't have one?

 

[Kelston]

Nope. The only raid device I would have is one built in to the mobo itself.

What i'm finding most frustrating about the issue is that on the initial install of the drivers and enabling of SLI, it will run through standard benchmark programs and run games for hours without issue. But when I turn off the PC and turn it back on it decides I angered some patron saint or techno deity and fails on me giving me my issues.

 

[Rifter]

Reinstall windows. I know it will suck but the last time i had SLI issues with a buddies 470's thats what it took to fix it.

I tried everything else including new ram, RMA'ing the mobo, new hard drive, new PSU and nothing worked. Then re installed windows and that system has been flawless for over a year now.

 

[Kelston]

Before I go through that (I wouldn't mind trying it if it had a chance of fixing it since i've been diagnosing so often I haven't really settled in yet), i'm actually curious if it matters what slots I use for SLI.

That's one theory I have running.

As it stands, I am unable to use the first PCI slot on my motherboard because the heatsink blocks it. So what I tried was putting video cards in 2+4.

 

[Keysplayr]

This only happens when you shut down for a while (systems cools) and then startup again?
Cold bug comes to mind.
Next time, after a gaming session and you need to get some food, don't shut the PC off. Just let it idle for a few hours. Go back and fire up a game and see what happens.
Also, didn't the new intel 1155 boards have SATA bug? Data corruption via SATA controller and were recalled?
Is this one of those boards? Or did you get the new revision?

 

[Kelston]

I doubt it's the cold bug since it happens even when I had shut it down and pretty much started it back up immediately (less than 2 minutes).

It seemed to have no problems when I had accidentally fallen asleep and went into sleep mode (guess I forgot to turn it off this function, since I usually do as I had some weird issue with it on my last system) and woke it back up again, but that was one of the initial reinstalls and I had yet to shut it down at that point.

Though, if it's the cold bug, what's the solution? I had always figured the cold bug only tended to pop up for guys running the super extreme cooling solutions while i'm running some half-assed probably poorly applied thermal paste thinger.

I have one of the new B3 revisions, it's what resulted in being so delayed in building the system.

 

[notty22]

I would re-position your cpu cooler, so you can use your first pci-e slot.
Your bios may not be able to cope with 2 video cards, outside the first slot, which is usually used in every configuration.

 

[Kelston]

I would re-position your cpu cooler, so you can use your first pci-e slot.

I'll have to use the stock cooler to test this theory either later today (doctor's appointment!) or tomorrow.

Unfortunately, I can't reposition the cooler i'm using to access the first PCI-X slot. With most mobos (maybe all?) designs, manufacturers tend to throw in an x1 PCI-X slot between the CPU socket and the first true x8/x16 PCI-X slot which acts as a spacer (and I guess some people use, maybe!?). However, for the WS Revolution, this first x1 slot is removed in favor of making room for 4 PCI-X x16 slots.

I guess the ambitious design clashed with the general monstrosity that is the Noctua NH-C14. Any valid mounting direction of the heatsink does not grant access to the first slot.

So while I don't know if slots matter (I guess until i'm able to test it), let me be the first to say: Noctua's NH-C14 blocks the first PCI-X slot of the Asus P8P67 WS Revolution.

It doesn't seem to be mentioned anywhere and I picked it up based on general good reviews of the cooler's performance especially given the low sound signature.

 

[Elganja]

do you have another Sli bridge you can try? they are fairly cheap on ebay, it may be worth a shot

 

[Kelston]

Yeah, I tried one and unfortunately, I did not win.

 

[FalseChristian]

I'd reformat and reinstall everything. This always solved problems for me.

 

[notty22]

I understand what you mean about the slots positioning.

I would try updating to the latest m/b bios, and also try running each card on its own. In 2d surfing the web, booting and 3d gaming, to see if you get any of the odd behavior. bsod's , driver crashing. To eliminate a possible bad card.

 

[digitaldurandal]

I understand what you mean about the slots positioning.

I would try updating to the latest m/b bios, and also try running each card on its own. In 2d surfing the web, booting and 3d gaming, to see if you get any of the odd behavior. bsod's , driver crashing. To eliminate a possible bad card.

Solid advice here.

I would add reinstall windows. Especially if it hasn't been that long anyways. Sometimes something can go wrong during intial driver installs and windows can be borked.

Check each card seperately to make sure both cards are good. If one card does not function properly in 3d mode you will have many issues. I had this issue when I went SLI 285s. Was having issues and eventually found out that one of the cards was bad and would crash when used alone after a few minutes, replaced it and had no more issues.

If each card works on its own then I would say the issue is one of these things:

Motherboard BIOS
PowerSupply
Temperature

Good luck

Oh also if your board allows it use a card only in the second slot to see if the issue is a bad pcix slot.

 

[stahlhart]

Are you running SLI with or without the NF200? If your board is like mine, you should have either an 8X/8X option or a 16X/16X one. I ultimately chose 8X/8X as it was more advantageous for cooling with the addition of a discrete audio card.

edit: Wait -- is that blocked first slot limiting your options? Sorry I missed that...