Sajin

erikdfier
 
Indeed, I thought so too.  But, its not just Benchmarking is unstable, Playing both Crysis and Metro 2033 crash to desktop w/driver failure.  

Downclocking the card clears all this up correct?

erikdfier

Brimy

Daz1967
I got my EVGA GTX 980 SC from Scan UK yesterday as an upgrade to my 16 month old EVGA GTX 780.

I'm very pleased with the overall performance increase of around 25-30% in the games and benchmarks I've ran so far but I've encountered two driver crashes while benchmarking unfortunately ("Driver has stopped responding and recovered" errors) which I never had on my GTX 780 unless I overclocked too far (I used a core offset of +100 MHz and a memory offset of +50 MHz). This SC model has a default core clock of 1,240 MHz according to GPU-Z v0.7.9 but does boost to 1,342-1,367 MHz during benchmarking. I have only been testing on the out of the box settings and have not overclocked anything (even my CPU is running at stock 3.5 GHz). I always do this to ensure everything is stable.

I was using the v344.11 WHQL drivers but I upgraded to the newer v344.16 beta ones yesterday evening. Also, because my PSU is modular I am using two separate 6+2 pin PCI-e leads to my GTX 980, which is how I had it set up for my GTX 780. Each lead does have two 6+2 pin connectors though so should I be using one and both connectors rather than two individual leads? I figured using two would spread the power load. Is that incorrect?

I normally play games with v-sync enabled so I am unlikely to be stressing my card to the extent I have during the benchmarking but I'm a bit concerned at the unstability issues at the moment. My previous GTX 780 has been 100% stable for the last year once I stopped overclocking it too much.

Advice?

If you have to downclock the GPU for it to work then the package has false advertising which is something i doubt a company like EVGA would want to do/risk.
 
You should have used two separate 6+2 cords and see if you get the same problem with stock clocks. You can use a twist tie to hole down/made to look presentable the extra two 6+2 pin leads on either cable. 
 
If you still get the error either contact EVGS's support via CHAT to see what they think or just RMA it pronto.

Brimy, this applies to me as well.  I am currently NOT using Two power Cords.  Im using a single PCI-E 6+2 pin + 6 pin to power this card.  

erikdfier

Brimy

erikdfier

Brimy

Daz1967
I got my EVGA GTX 980 SC from Scan UK yesterday as an upgrade to my 16 month old EVGA GTX 780.

I'm very pleased with the overall performance increase of around 25-30% in the games and benchmarks I've ran so far but I've encountered two driver crashes while benchmarking unfortunately ("Driver has stopped responding and recovered" errors) which I never had on my GTX 780 unless I overclocked too far (I used a core offset of +100 MHz and a memory offset of +50 MHz). This SC model has a default core clock of 1,240 MHz according to GPU-Z v0.7.9 but does boost to 1,342-1,367 MHz during benchmarking. I have only been testing on the out of the box settings and have not overclocked anything (even my CPU is running at stock 3.5 GHz). I always do this to ensure everything is stable.

I was using the v344.11 WHQL drivers but I upgraded to the newer v344.16 beta ones yesterday evening. Also, because my PSU is modular I am using two separate 6+2 pin PCI-e leads to my GTX 980, which is how I had it set up for my GTX 780. Each lead does have two 6+2 pin connectors though so should I be using one and both connectors rather than two individual leads? I figured using two would spread the power load. Is that incorrect?

I normally play games with v-sync enabled so I am unlikely to be stressing my card to the extent I have during the benchmarking but I'm a bit concerned at the unstability issues at the moment. My previous GTX 780 has been 100% stable for the last year once I stopped overclocking it too much.

Advice?

If you have to downclock the GPU for it to work then the package has false advertising which is something i doubt a company like EVGA would want to do/risk.
 
You should have used two separate 6+2 cords and see if you get the same problem with stock clocks. You can use a twist tie to hole down/made to look presentable the extra two 6+2 pin leads on either cable. 
 
If you still get the error either contact EVGS's support via CHAT to see what they think or just RMA it pronto.

Brimy, this applies to me as well.  I am currently NOT using Two power Cords.  Im using a single PCI-E 6+2 pin + 6 pin to power this card.  

Yes i understand that but i have noticed every "rig" is different and functions the same. What may work for you rig may not work on mine, especially with drivers. Ruling out all possible causes should be looked at no matter if it does not apply to someone else, is all i am saying.
 
For example i have switched motherboards last year and i never re-installed my OS, i just deleted the previous drivers and installed the newer ones and all worked well, i tried that for a build for a customer and only a fresh OS install worked for him. Another, example, i have the XSPC rs360 kit from late 2012 to now and the X20 750 v4 has worked since then to now, someone i know bought the ex240 kit and his pump died after 4 months.
 
So in my experience it is best to go through the process of elimination to rectify or narrow a cause to a problem.

Hey Brimy, I should have clarified, that my statement was more of a question than a refute to your point.    Now, I would like to know what is the Cord system that is best to deliver to the GTX 980 regardless.  So, do you think...  Dual 6 pin+2pin Connections, instead of the Single 6pin+2pin +6 pin?  I'm open to suggestion.

erikdfier
Brimy, using dual 6+2 pin power cords had no effect on stability.

Re-installed old GTX 650 OC 2gb version and ran the metro 2033 benchmark again, set it to loop 20x and the system ran flawlessly.  
 
As for the GTX 980, This looks like a bad card or the 344.16 drivers are causing problems with the new architecture, or by chance its overloading a rather robust PSU... Regardless, Its back in the box to be returned.

erikdfier

enigma442
Just looking at your system specks do you mean your storage drive is 1 terabyte and not 1 gigabyte?
Also you may very well have an unstable card.

Good eye man, fixed.  1gb storage, that would be pretty lame.

slickrick289
I want you to be aware of something erikdfier. Going through this thread ive acertained that this is the exact problem I'm having albeit with a 780 Ti instead. All my games run fine except for two of them which are Crysis 3 and either one of the Metro Redux games. Both those games are severely flawed at the moment. If you head over to their respective forums, one in Steam and the other at the new EA HQ site, many people are getting constant crashes with stable hardware. After the recent patches, the Metro games have become very unstable. I spoke to a developer through steam and they are blaming Nvidia for the problems. All I know is that after the last few patches the game crashes randomly through playthrough. Crysis 3 is a lost cause as the developers have abandoned that game already.

In summary, I don't think it's your hardware, I think its those games. You said so yourself that those are the only games giving you issue. Go to those threads and you'll see what I'm talking about. I unfortunately cannot post them here.

Brimy

erikdfier
Brimy, using dual 6+2 pin power cords had no effect on stability.

Re-installed old GTX 650 OC 2gb version and ran the metro 2033 benchmark again, set it to loop 20x and the system ran flawlessly.  
 
As for the GTX 980, This looks like a bad card or the 344.16 drivers are causing problems with the new architecture, or by chance its overloading a rather robust PSU... Regardless, Its back in the box to be returned.

Alright well now i can understand you starting an RMA.....those things suck.

erikdfier

enigma442
Just looking at your system specks do you mean your storage drive is 1 terabyte and not 1 gigabyte?
Also you may very well have an unstable card.

Good eye man, fixed.  1gb storage, that would be pretty lame.

Back in 1983 10gb HDDs were a seriously expensive thing, crap i just showed my age. Never make fun of where you came from man

Brimy

slickrick289
I want you to be aware of something erikdfier. Going through this thread ive acertained that this is the exact problem I'm having albeit with a 780 Ti instead. All my games run fine except for two of them which are Crysis 3 and either one of the Metro Redux games. Both those games are severely flawed at the moment. If you head over to their respective forums, one in Steam and the other at the new EA HQ site, many people are getting constant crashes with stable hardware. After the recent patches, the Metro games have become very unstable. I spoke to a developer through steam and they are blaming Nvidia for the problems. All I know is that after the last few patches the game crashes randomly through playthrough. Crysis 3 is a lost cause as the developers have abandoned that game already.

In summary, I don't think it's your hardware, I think its those games. You said so yourself that those are the only games giving you issue. Go to those threads and you'll see what I'm talking about. I unfortunately cannot post them here.

Do you and erikdfier have a CPU OC on your system? This can also produce these crashes in certain games. As i recall, going from the 335.23WHQL driver to the 337.50 beta and up i had to add voltage to my 4.5Ghz and 4.7Ghz OC on my 3570k just to play CoD Ghosts for just to load. So it could be the game and if yo have any OC and not really the GPU.

erikdfier
Testing a 3rd Game.  "The Vanishing of Ethan Carter." Literally just came out within the last hour on Steam.  Its installed and running, will likely put some hours into the game, unless the game crashes.
 

Misterbond
Same problem here
 
Got my EVGA GTX 980 SC today, made a clean install of the drivers (344.16) and played Assassins Creeds 3 for about 1 hour and worked flawless. Then I tried do make some benchmarks with Heaven Benchmark 4.0 (using stock clocks).
First run was perfect, but when I tried the second run the software started to freeze and then crashed.

Now when try to play Assassins Creeds 3 or TF2 or any other game that is 3D the game crashes (sometime even at the menu)

Tried with 344.11 but the crashing continues. Also, noticed that after crashing Nvidia's driver lock my Core Clock at 999 MAX, that way I can play without problems, but as expected, with fps impact... so... its really a OC issue

Before the 980 I had 2 x 660ti in sli, so I dont think that its the PSU, my PUS is a CoolerMaster SilentProM 850w
Already tried putting the power limit up to 125% using MSI Afterburner and EVGA Precision but same problem.


RMA looks like the only path now :(
 
Here are some pictures to have and idea of the problem
At the first one you can see some blue blocks... wth o0
The second one you can see that the CORE CLOCK goes higher than the BOOST CLOCK, is that possible? Also is it normal to my Power Consumption be at 61% while the core clock is at 1392?
 

 
 
 
 

erikdfier
 
That sucks. I have no advice about the situation you have as it seems worse than my own here since I have benchmarked Heaven for HOURs; however we have similar issues for example, no problems until benchmarking crashed the driver, then everything DX 11 became unstable.  I wouldn't be surprised if this whole issue was just bad drivers for our hardware setups and will be easily worked out in a couple months.