980ti and Sabertooth z77 Incompatible?

[HallYourBase]

A bit of a long read but in order to supply as much detail as possible I will try to be thorough. TLDR new 980ti causes reboot under load (gaming) not temperature related

 

Here are the current specs before I get into what has been tried so far.

 

Asus z77 Sabertooth

Intel i7 3770k (stock settings no OC) cooled with H100

Corsair Vengeance 4x4 GB RAM

EVGA SC+ AC2.0 980ti

OCZ 1000W Fata1ty PSU

Samsung 850 EVO 1TB

WD Black 1TB

Some Old samsung 640GB HDD from old desktop

Case: Corsair 650D

 

Basically the computer hard reboots when I play pretty much any game. It can be almost instantly after loading a game or after 30 minutes. I watch the temperatures and they never get about 80 degrees C on the CPU or GPU.

 

I have RMAed my PSU, the GPU, and my motherboard, and the problem persists. I have run Memtest and the ram is fine. I recently did a fresh install of windows and fully updated when I got the card as I had a Crucial M4 256GB boot drive before the upgrade. I have tried updating the bios. I have tried various versions of the NVidia drivers. I had the same problem before upgrading as well when I ran crossfire 7950s, but using one card ran fine. So I thought it was a power supply issue, but the new unit after RMA had the same behavior. The 980ti is not listed as compatible with the sabertooth z77 but since the issue occurred with the other GPUs I am at a loss as to what the issue may be. The windows even viewer states event ID 41 task 63, which is usually power related. I have tried using another outlet, another power strip, and no power strip. All produce the same results. I am at a loss as to what the issue may be and have been unable to play games, one of my favorite stress relievers, for 2 months since purchasing the 980ti. Side note ASUS support for motherboards is terrible.

 

If anyone can offer any advice I would much appreciate it, I am about to just buy a new motherboard and processor and maybe power supply to just complete the system rebuild so I can use my machine as I like.

[Guest Kloaked]

Do a BIOS reset and see what happens.

[HallYourBase]

Do you mean clear CMOS with the jumper or load the original BIOS revision? Sorry I am not super familiar messing around in BIOS.

[DotBowder]

Do you mean clear CMOS with the jumper or load the original BIOS revision? Sorry I am not super familiar messing around in BIOS.

I suspect Kloaked just means resetting the settings on the bios by either using a jumper per the motherboard instructions, or disconnecting the power cable and removing the 3v button-battery on your board for 10 or 15 seconds.

 

If you don't see any change after resetting your BIOS, I would definitely try to test your RAM further. You said you had run memtest. How long did you run memtest for? Did you run it in windows or on a boot disk?

 

Have you tried removing all but one ram stick in the DIMM_A2 socket? (DIMM_A2 is the Closest light-brown slot to your CPU and is the slot recommended by the mobo manual for single-stick configs.) This will force your comptuer to use the memory in single channel mode and it can single out a bad stick. (If you still have issues with a single stick of ram in, try swapping that single stick, as, if even 1 of the sticks is bad, you have at least 25% chance of picking it right off the bat.)

 

Also, what speed is the ram at? 1066Mhz, 1333Mhz, 1600Mhz, 1866Mhz, 2133Mhz, etc... Is it running at it's rated speed? Is it under/over-clocked?

 

The issue is definitely weird. 

[HallYourBase]

I suspect Kloaked just means resetting the settings on the bios by either using a jumper per the motherboard instructions, or disconnecting the power cable and removing the 3v button-battery on your board for 10 or 15 seconds.

 

If you don't see any change after resetting your BIOS, I would definitely try to test your RAM further. You said you had run memtest. How long did you run memtest for? Did you run it in windows or on a boot disk?

 

Have you tried removing all but one ram stick in the DIMM_A2 socket? (DIMM_A2 is the Closest light-brown slot to your CPU and is the slot recommended by the mobo manual for single-stick configs.) This will force your comptuer to use the memory in single channel mode and it can single out a bad stick. (If you still have issues with a single stick of ram in, try swapping that single stick, as, if even 1 of the sticks is bad, you have at least 25% chance of picking it right off the bat.)

 

Also, what speed is the ram at? 1066Mhz, 1333Mhz, 1600Mhz, 1866Mhz, 2133Mhz, etc... Is it running at it's rated speed? Is it under/over-clocked?

 

The issue is definitely weird. 

I reset the BIOS, same issue, tried one stick of RAM at a time, same issue. The ram is rated for 1600 and XMP is enabled and shows 1600. I also tried the default setting given by the motherboard which is showing as 1333.

[Stefan1024]

It looks like your system get's messed up by high power draw on the PCI-e slots.

Does the MoBo feature an additional molex plug to feed more power? If yes plug it in.

Maxwell cards in general have a very aggressive power saveing feature that triggers high spikes on the supply rails.

[HallYourBase]

It looks like your system get's messed up by high power draw on the PCI-e slots.

Does the MoBo feature an additional molex plug to feed more power? If yes plug it in.

Maxwell cards in general have a very aggressive power saveing feature that triggers high spikes on the supply rails.

Beyond the normal 24 pin and the 8 pin CPU which are both plugged in securely (trust me I triple checked) I don't see anything else, especially not molex.

[Stefan1024]

Beyond the normal 24 pin and the 8 pin CPU which are both plugged in securely (trust me I triple checked) I don't see anything else, especially not molex.

Hmmm ok, was just an idea becauses by mATX has an extra molex connector and it refuses to boot when all slots are populated and the moley cable is not plugged in.

However I just looked at the pictures and ther is indeed no extra power connector.

[HallYourBase]

Hmmm ok, was just an idea becauses by mATX has an extra molex connector and it refuses to boot when all slots are populated and the moley cable is not plugged in.

However I just looked at the pictures and ther is indeed no extra power connector.

Sorry if that came across as rude or unappreciative. I do thank you for any ideas, because I am at wits end with it. I don't know if I should just buy new motherboard and CPU, in which case I would probably need a new PSU as well as this OCZ one doesn't support the lower power modes of Haswell and beyond to my knowledge ( could be wrong on that haven't really looked). At this point if the 980ti is unusable (for gaming / high intensity anything) it was a waste of money, if I have to upgrade to use it so be it, better than it being a fancy paperweight or just running internet browsers and text editors with it.

[Stefan1024]

Sorry if that came across as rude or unappreciative. I do thank you for any ideas, because I am at wits end with it. I don't know if I should just buy new motherboard and CPU, in which case I would probably need a new PSU as well as this OCZ one doesn't support the lower power modes of Haswell and beyond to my knowledge ( could be wrong on that haven't really looked). At this point if the 980ti is unusable (for gaming / high intensity anything) it was a waste of money, if I have to upgrade to use it so be it, better than it being a fancy paperweight or just running internet browsers and text editors with it.

Don't worry I'm not offended, nor I felt your answer was rude.

 

I think it's a power draw realted issue, but since you allready RMAed GPU, PSU and MoBo they should be fine.

The last thind I can suggest is to reduse the TDP of the GPU to the lowest possible value (60% as far as I know) and test if the problem still persists. If not, it's power related for sure. Than you can go up to 100% set by step and see how long it is stable.

 

Ah yes and set the PCI-e link speed to 2.0. You hardly loose any performance but it is more stable.