Asus 780 GTX OC Crashes Shuts down Computer

S

Sephrioth

Weaksauce
Joined
Mar 13, 2010
Messages
108
I installed a 780 GTX in my computer (ASUS P6T DH Deluxe, i7 Overclocked at 3800mhz). But when I start UDK or some games, the computer gives a black screen after a few minutes and shuts down (No errors or blue screens). I have re-installed Windows 7 and even tried out beta drivers 326.41 (So it probably isn't a software problem.) Could it be the PSU (Coolmaster 850W?) Or could the card itself broken?
 
up the voltage if you can, it could be the open GL based games causing issues, in my case when crashes like this, it tends to be missing some voltage on the gpu or the cpu was not set properly(timings, speeds,voltages don't forget these tend to need a small bump over time)
 
What voltages should I increase, and by how much?. In the mainboard bios the voltage of PCI express is set to Auto at this moment.
 
pci-express voltage is for the slot itself, not the gpu, to control the gpu voltage or voltage offset(as I believe the 600 and 700 series cards call it) you have to use EVGA Precision, MSI afterburner, or Asus GPU Tweak. Not enough voltage or to high of a clock can cause these type of issues to happen.
 
I have already tried increasing the voltage in MSI afterburner, and even on the stock clock it still crashes. Crysis 3, Metro Last Light and Unigine Valley run fine, but UDK (Unreal Engine 3) crashes after a minute and shuts down the PC. I still need to try out Bioshock Infinite, which is a Unreal Engine 3 game
 
It seems a little extreme to start increasing the voltage just to get the card to play games.
I'd start by temporarily disabling your CPU over-clock and see what happens.
By the way, do you have both the 6 and 8 PCI power cables connected to your gtx 780?
 
http://hardforum.com/showthread.php?p=1040061986

it is a "known" bug apparently. And no, is not directly "extreme" as you put it, the card or any card for that matter is set with a certain voltage from factory, does not 100% mean it will be perfectly stable no matter what at that clock with that voltage with all drivers under all circumstances(should but nothing is perfect) so a small bump will not kill anything, hell its better then having to reduce clocks to see if maybe was set to fast for a bad chip(which MANY have had happen for "premium" cards)

Both my Radeon 7870s overclock and declock massively different, both also require different voltages to be stable, that is just the way the cookies crumbles, and his issue is GPU related I can near 100% know that.
 
I have connected both cables to the card. I have also tried to lower the processor overclock to 3.4 Ghz and bump all the voltages a little bit (Memory, Pci-E) but it still crashes. I will try the system without a processor overclock, but I am not sure what that will prove. Maybe that the card itself isn't faulty. The system has always been stable without the 780 GTX, under heavy processor load.
 
ok pci-e voltage will do nothing to help graphics card, memory voltage also will not do much of anything for the graphics card. The only thing that will is core voltage or voltage offset(depending on which GPU program you are using) as I have mentioned. Maybe as previous post stated, it is a known issue and Nvidia is working on it, there was a semi-fix in the post
 
I have tried without the processor overclock and with lower memory sertings, and it crashed and shut down almost instantly when I started UDK. Even on 2.4 Ghz the system crashes.
 
i have a feeling that there's a nasty issue plaguing the GTX 7xx line of nvidia cards. I recently (3 weeks) ago upgraded to a GTX 770. For the most part, everything runs great but if my computer is left idle for an extended of period of time than it either BSOD's or reboots (at the moment i start to use it). Several of the windows memory dumps that I've analyzed point to the nvidia drivers (nvlddmkm.sys with bugcheck 116). Nvidia has official said that they are aware of the problem and are working on it. (i saw the post on geforce.com)
 
What GPU was in your computer before you replaced it with the GTX 780?
What brand is the 780?
Is it reference or a vendor OC or alternate cooler?
What is the temp on the GPU before everything crashes?
Do you use a fan profile in Afterburner?
Have you tried upping the power slider and temperature slider in Afterburner?
 
darkpark said:
i have a feeling that there's a nasty issue plaguing the GTX 7xx line of nvidia cards. I recently (3 weeks) ago upgraded to a GTX 770. For the most part, everything runs great but if my computer is left idle for an extended of period of time than it either BSOD's or reboots (at the moment i start to use it). Several of the windows memory dumps that I've analyzed point to the nvidia drivers (nvlddmkm.sys with bugcheck 116). Nvidia has official said that they are aware of the problem and are working on it. (i saw the post on geforce.com)
Click to expand...

The issue is confirmed, the only known workaround at the moment is to revert back to the 314 drivers and use a modified inf file to get them to install. I can confirm that does work.
 
magoo said:
What GPU was in your computer before you replaced it with the GTX 780?
What brand is the 780?
Is it reference or a vendor OC or alternate cooler?
What is the temp on the GPU before everything crashes?
Do you use a fan profile in Afterburner?
Have you tried upping the power slider and temperature slider in Afterburner?
Click to expand...

It is an Asus OC model. I had a 9800GTX in this system. I have tried multle settings in Afterburner, including setting the voltage higher without overclocking.
 
Sephrioth said:
It is an Asus OC model. I had a 9800GTX in this system. I have tried multle settings in Afterburner, including setting the voltage higher without overclocking.
Click to expand...

I think your card may be FUBAR.
Only other suggestion is that your PSU might be the culprit.

Lately it's good luck getting an RMA from ASUS.:eek: If you can send it back to Amazon or newegg or whom ever you bought it from, I'd do that. (ie your still in the 30 day window)
 
What exact model is the PSU? I would suspect that first.
 
Coolermaster has made probably 15 different 850W, thats why I was asking the exact model. For example, if it is a multirail configuration you might be tripping the OCP and you need to mix rails to make it work. Check on the side of the PSU label and get the exact model.
 
I have replaced the Power supply with a single rail 1000w and nnd it solved the shut down crashes. thanks everyone for the help.
 
darkpark said:
i have a feeling that there's a nasty issue plaguing the GTX 7xx line of nvidia cards. I recently (3 weeks) ago upgraded to a GTX 770. For the most part, everything runs great but if my computer is left idle for an extended of period of time than it either BSOD's or reboots (at the moment i start to use it). Several of the windows memory dumps that I've analyzed point to the nvidia drivers (nvlddmkm.sys with bugcheck 116). Nvidia has official said that they are aware of the problem and are working on it. (i saw the post on geforce.com)
Click to expand...

My setup was solid until I installed my GTX770.. now I get lockups and random crashes. The screens turn off like they are supposed to and when I try to wake them up I get nothing.
 
Sephrioth said:
I have replaced the Power supply with a single rail 1000w and nnd it solved the shut down crashes. thanks everyone for the help.
Click to expand...

Awesome. Good to hear it was a simple fix, and that our thought about the psu was correct.
 
Sephrioth said:
I have replaced the Power supply with a single rail 1000w and nnd it solved the shut down crashes. thanks everyone for the help.
Click to expand...

So was your old PSU bad? I thought that 850w psu should do the trick? I'm concerned because I am awaiting a 780 from the step up and I have a thermaltake SP-750m 750watts
 
kanecvl said:
So was your old PSU bad? I thought that 850w psu should do the trick? I'm concerned because I am awaiting a 780 from the step up and I have a thermaltake SP-750m 750watts
Click to expand...


That PSU is decent, it wasn't the wattage that was an issue but more likely the configuration of the +12V rail(s) or an outright failure.
 
850 is plenty for sure. I currently run 4770K, 780 SLI, 3HDDs, and 1SSD with an 850W.

I had a similar problem initially though. My 4yo 850W PSU must have slowly went bad because I was getting reboots under stress as well. I plan on doing a warranty replacement eventually but I have since replaced it with a new 850W of a different brand to just get up and running faster. So far no problems with the new one.
 
Well, unexpected shutdowns are results of problem with computer memroy, overheating and other serious issues. Do check your RAM for possible issues, solve overheating issue. Identify and repair errors first. If that doesn't work, then try subsequent steps given below:

Run WIndows Memory Diagnostic:

1. Click Start, type "Windows Memory Diagnostic" in the Search Box.
2. Press ENTER.
3. Click "Restart now and check for problems" button.

Solve Overheating

Are you using a laptop or desktop computer? If you're using a laptop, use a laptop cooling pad. Do service maintenance regularly. Clean the fans, all components inside your laptop or desktop.
 
JohnSmith420 said:
Well, unexpected shutdowns are results of problem with computer memroy, overheating and other serious issues. Do check your RAM for possible issues, solve overheating issue. Identify and repair errors first. If that doesn't work, then try subsequent steps given below:

Run WIndows Memory Diagnostic:

1. Click Start, type "Windows Memory Diagnostic" in the Search Box.
2. Press ENTER.
3. Click "Restart now and check for problems" button.

Solve Overheating

Are you using a laptop or desktop computer? If you're using a laptop, use a laptop cooling pad. Do service maintenance regularly. Clean the fans, all components inside your laptop or desktop.
Click to expand...

What is this? Are you a bot or something? Did you even read his post?


Anyways could somebody please explain why the multi-rail PSU caused the problem and why a single rail fixed it?
 
Could be gpu was drawing too much power from a single rail that couldnt handle it. More than likely was just a small component was bad and outright failed had this happen with a 1000watt once and was only running a 480 and i7 750 rig. Got a 750tx to replace it and still running 5 years later on my buddies rig he uses everyday.
 
Sephrioth said:
I have tried without the processor overclock and with lower memory sertings, and it crashed and shut down almost instantly when I started UDK. Even on 2.4 Ghz the system crashes.
Click to expand...

You have a bad video card/powersupply (not enough AMP to the 12+ line)/bad pci express 1 X16 slot. Did you try moving it to PCI Express 2 X16 slot.

You did a clean install and ran the system at default speeds and get the same results. You might want to focus on the power supply and pci express slots next.
 
smaudioz said:
What is this? Are you a bot or something? Did you even read his post?


Anyways could somebody please explain why the multi-rail PSU caused the problem and why a single rail fixed it?
Click to expand...

Lets say GPU requires 30A on +12V rail.

Power supply A is multi-rail with 18A on each 12V PCI-E rail. Since it technically has two connectors on each cable, you could hook up the GPU with just one cable. However, in order to supply the GPU at full load during gaming, you would have had to use one six or eight pin from two separate cables from the PSU to combine 18A+18A to exceed the 30A requirement.

Power supply B is a single rail, it supplies 42A to all +12V cables. This means you can plug just one cable with two PCI-E plugs and be fine, but it also means the over-current protection is much less sensitive.

Either PSU will work, but PSU A requires a little attention to detail in order for it to do what you want. This is the scenario assuming his PSU wasn't bad outright.
 
LigTasm said:
Lets say GPU requires 30A on +12V rail.

Power supply A is multi-rail with 18A on each 12V PCI-E rail. Since it technically has two connectors on each cable, you could hook up the GPU with just one cable. However, in order to supply the GPU at full load during gaming, you would have had to use one six or eight pin from two separate cables from the PSU to combine 18A+18A to exceed the 30A requirement.

Power supply B is a single rail, it supplies 42A to all +12V cables. This means you can plug just one cable with two PCI-E plugs and be fine, but it also means the over-current protection is much less sensitive.

Either PSU will work, but PSU A requires a little attention to detail in order for it to do what you want. This is the scenario assuming his PSU wasn't bad outright.
Click to expand...

Thanks for the info, but I have a 680W PSU with 4 12v Rails (18A on each one) and I use one cable that has 2x6pin (one of them is a 6+2) to connect to the two 6 pin sockets on my GTX 570 and it is fine. I am however looking at getting a graphics card that will require either a 6 and an 8 pin connector or two 8 pin connectors. So in that case what should I do? Should I use 2 PCI-E power cables instead of the 1 I am currently using?
 
You must log in or register to reply here.
Back
Top