Rampage IV Formula PCI-E lane switch light randomly shuts off, desperate need of help

[clownshoes]

I finished a new build about two weeks ago and am having a recurring issue that makes absolutely no sense to me.

I'm running a quad R9 290x setup using an ASUS Rampage IV Formula motherboard. My problem is this. The GPU in the first PCI-E slot or the first PCI-E slot itself (not sure which one it is) goes dead for no apparent reason.

I originally thought it was related to crossfire because the first time the card/PCI-E slot went out, it was at the exact second I enabled crossfire between 2 of the 4 cards on the board. After that first incident though, the first card/PCI-E slot has been dropping out randomly.

When the card/PCI-E slot drops out, the computer responds as if I had removed the card from the computer. It doesn't output any video. The card doesn't show up in device manager, catalyst control center or GPU-Z. Also, the Rampage IV motherboards have PCI-E lane lights that light up when a card is present / detected by the motherboard. When the issue crops up, the lane light corresponding to the first PCI-E slot goes out.

The ONLY way I've been able to remedy the situation is by completely powering down and unplugging the computer (power down --> turn off both PSU switches on back (running 2 PSUs in the system) --> unplug computer from wall --> hold power button to drain any residual). I've tried just about everything else I could think of (clear CMOS, update bios, reinstall video driver, reinstall OS) and this power cycling is the only thing that has worked.

Power cycling the computer doesn't always work on the first try either. Sometimes I have to do it 3 or 4 times and leave the computer unplugged for several hours.

When I get it working again, it usually works for a day or two before the problem randomly pops up again.

Does anybody know what might be causing this? Does the power cycle as I was doing it reset something (any fuses in the power supply that reset)?

I greatly appreciate any input. I'm completely lost.

Thanks

 

[spacin9]

There's a 4-pin molex connector right below the IO and above the first PCI-e slot. I believe it serves as extra power to the PCI-e slots on SLI/Crossfire motherboards. Believe it or not, esp. with the Rampage Extreme that I have (it's a 6-pin PCI-e on the Extreme), that connector does make a difference with multiple cards for extra power to the PCI-e slots. It's an awkward place to have it... I have a backplate on my Classy and there's like zero room to plug it in... but if you haven't I suggest you do.

Other than that...check your power supply, check to see if all those PCI-e slots are on via that little red box (next to the 24-pin connector) Those are obvious.. but it might be RMA time IMO.

To be honest.. with 4 of those monsters... I would switch out to one of those Rampage Black Extremes or a regular Rampage Extreme. That board just doesn't seem beefy enough to handle 4 four cards. We'd all love to see those benchies when you get it all worked out. Also, I know the color might put you off... but the second gen P9X79 WS-E runs with four 16X slots. I have to the non-E WS and I love it... it's a beast.

 

[clownshoes]

There's a 4-pin molex connector right below the IO and above the first PCI-e slot. I believe it serves as extra power to the PCI-e slots on SLI/Crossfire motherboards. Believe it or not, esp. with the Rampage Extreme that I have (it's a 6-pin PCI-e on the Extreme), that connector does make a difference with multiple cards for extra power to the PCI-e slots. It's an awkward place to have it... I have a backplate on my Classy and there's like zero room to plug it in... but if you haven't I suggest you do.

Other than that...check your power supply, check to see if all those PCI-e slots are on via that little red box (next to the 24-pin connector) Those are obvious.. but it might be RMA time IMO.

To be honest.. with 4 of those monsters... I would switch out to one of those Rampage Black Extremes or a regular Rampage Extreme. That board just doesn't seem beefy enough to handle 4 four cards. We'd all love to see those benchies when you get it all worked out. Also, I know the color might put you off... but the second gen P9X79 WS-E runs with four 16X slots. I have to the non-E WS and I love it... it's a beast.

Thanks, greatly appreciate the suggestions.

I have the 4 pin molex plugged in but I was sort of thinking the same thing. Even with the 4 pin molex, the 290x cards are so power hungry that maybe the board isn't designed for them.

I just checked out that P9X79 WS-E and the specs on it are awesome. I'm really much more concerned about functionality than I am looks so if the motherboard turns out to be the problem, I'll definitely consider that board as a replacement.

I hate to tear down the computer (all under water so it's a pita to do so) but I'm starting to think that's the only way I'm going to figure this out.

There isn't some sort of program I can run in the background that records everything that's happening is there (so the next time the problem occurs maybe it can point me in a direction)?

Also, regarding those PCI-E lane switch lights (red box you mentioned), all the lights are on when the computer is working but when the issue crops up, the light corresponding to slot 1 goes out. Does this indicate a problem with the card more so than the motherboard? I was also thinking it might be a problem with the power supply (if the power supply isn't supplying power to the card in slot 1, the lane light would go off wouldn't it?).

Anyway, thanks again for the suggestions.

 

[spacin9]

Thanks, greatly appreciate the suggestions.




Also, regarding those PCI-E lane switch lights (red box you mentioned), all the lights are on when the computer is working but when the issue crops up, the light corresponding to slot 1 goes out. Does this indicate a problem with the card more so than the motherboard? I was also thinking it might be a problem with the power supply (if the power supply isn't supplying power to the card in slot 1, the lane light would go off wouldn't it?).

Anyway, thanks again for the suggestions.

You know the only way is to swap out another card. Pain in the butt... I think there's a way to plug-in to like a laptop for BIOS monitoring and overclocking. I don't know if that would even help you.

That's 2000-2400 dollars in video cards alone. I wouldn't even mess with it anymore. Get the Black Extreme..or the Red Extreme; problem solved. Get the WS-E and paint it lol.

 

[clownshoes]

You know the only way is to swap out another card. Pain in the butt... I think there's a way to plug-in to like a laptop for BIOS monitoring and overclocking. I don't know if that would even help you.

That's 2000-2400 dollars in video cards alone. I wouldn't even mess with it anymore. Get the Black Extreme..or the Red Extreme; problem solved. Get the WS-E and paint it lol.

ha ha, yeah. If I knew it was the motherboard this is exactly what I'd do (really digging the WS-E and it's cheaper than the other boards to boot). I'm just not sure if the problem is the motherboard, video card or maybe even the PSU.

Because the problem is intermittent and random, even if I do tear down the computer I'm really not sure how to go about testing the parts.

Only thing that would work is to remove the GPUs from the water loop and reinstall the stock coolers on them. Then run the computer until the problem pops up again and swap in a working GPU...

but the thought of removing those waterblocks, reinstalling the stock coolers, then removing the stock coolers and reinstalling the waterblocks again makes me want to beat a puppy with a baby seal.

I'm gonna see if I can find one of those monitoring programs first. I've looked through the windows event logs each time the problem occurs and can't find anything.

Anyway, thanks for the suggestions.

 

[spacin9]

I had the Rampage Formula. I had to RMA it because the first PCI-e slot refused to run at anything other than 8X. Maybe there is some sort of defect with those boards and first PCI-e slot.

I would kick that puppy dead square in his chops, take a deep breath and rebuild.

 

[Raja@ASUS]

To debug this:

1) Update to the latest UEFI (if not done already).

2) Load optimized defaults - don't overclock the system in any way (keep memory below DDR3-1600). If using multiple DRAM kits, then use a single kit only and see if that helps.

If neither of the above helps, type out a full parts list (include all connected devices). Might need to RMA the board, but worth checking everything out first.

-Raja

 

[clownshoes]

I had the Rampage Formula. I had to RMA it because the first PCI-e slot refused to run at anything other than 8X. Maybe there is some sort of defect with those boards and first PCI-e slot.

I would kick that puppy dead square in his chops, take a deep breath and rebuild.

This is interesting spacin. Were you running a multiple card setup while having issues with the first PCI-E slot? I only ask because I was running a single card setup on the board for a few months before I added the quad GPUs. I didn't have a single issue while running just one card on the board.

To debug this:

1) Update to the latest UEFI (if not done already).

2) Load optimized defaults - don't overclock the system in any way (keep memory below DDR3-1600). If using multiple DRAM kits, then use a single kit only and see if that helps.

If neither of the above helps, type out a full parts list (include all connected devices). Might need to RMA the board, but worth checking everything out first.

-Raja

Thanks Raja, I'll give this a shot. If it does solve the problem, what exactly does it tell you?

 

[spacin9]

This is interesting spacin. Were you running a multiple card setup while having issues with the first PCI-E slot? I only ask because I was running a single card setup on the board for a few months before I added the quad GPUs.

No I was running SLI. But it didn't matter. Single card... two cards. First slot was 8X only. Gen2 or Gen3. I figured I could live with it... until it gnawed at me enough to RMA it.