Issues with GTX 780 - Freezing and Display Driver crashes

13539 Views 31 Replies 9 Participants Last post by indiyet, May 2, 2016

xioticin

Discussion Starter · Aug 11, 2013

I just recently purchased the 780 and have enjoyed the boost in frames over my 560 superclock, but I have run into a few issues. I do not know if they are specifically the card or what but I am becoming more and more concerned.

Earlier this week, while I was playing Borderlands 2, I had an issue with my display driver crashing. At the time, I was running 320.49. This happened about 4 times within that day alone and was the first time I felt that there might be an issue. Just tonight, I had issues with League of Legends locking up while playing ARAM. It sounded close to what its like whenever your CPU overclock goes unstable and makes that odd electronic buzz sound through the speakers, or at least that what it has always done to me. The game would freeze up for a second, then become responsive, then freeze again for 3 seconds or more, then about 10 seconds after is normal again. This happened about 6 times within 15 minutes.

I am writing this as I am playing games as its the easiest way for me to repeat the issue. This wasn't my first ARAM game either, but my 3rd and none of the other times did this happen. I have reset my GPU overclock from +80 +175 with 100% power target. My CPU is still at 4.5 as that has proven to be stable so far, but I might drop that soon too to truly test this issue.

I really don't want to RMA unless I have to and believe that most of these issues are always settings/user related and not the hardware. Just looking for help wherever I can get it before I start contacting EVGA. The system in my sig is what I am using. CPU is at 4.5 and ram is at 1600. I will be checking in periodically to report what I find and look at replies.

See less See more

1 - 20 of 32 Posts

Jameswalt1

· #2 · Aug 11, 2013

Can you clarify your overclock statement, was the gpu overclocked during the multiple crashes or not?

Also, in general the first thing I would personally do is put the cpu to stock speeds and the gpu to stock speeds and use the system like normal for a few days and see if the problem persists.

I'm assuming you've done a clean install of the latest beta driver...

xioticin

Discussion Starter · #3 · Aug 11, 2013

You assumed correctly. I actually did a full system reformat whenever I upgraded my card.

After I made my post last night, I continued to play ARAM but with my gpu at stock speeds. It boosted fine still and ran totally normal. I did also go in and change the power management mode settings in Nvidia control panel to prefer maximum performance. I was unable to get it to hang again.

I continue to this moment to look up reasons this could have happened since League of Legends is not intensive at all and I played Crysis 2, BF3, and Borderlands 2 as well and did not experience this exact issue. I never had a Driver Crash in Crysis 2. What I did find is my issue might have been my memory overclock being at +175. I have the 3A version of the bios meaning Elpida memory chips. They do not overclock nearly as well as Samsung chips. With that knowledge, there is a possibility that my memory was just going unstable for some odd reason in League.

See less See more

Zero 7hree

· #4 · Aug 11, 2013

Have you tried any of the beta drivers, I am using the 326.41 beta driver for my 780 GTX, I did have prior issues with the 320.49 w/ LoL, I would freeze or the game would just die at least once a night or when I tried to minimize.

xioticin

Discussion Starter · #5 · Aug 11, 2013

I will try uninstalling my nvidia drivers and clean them with driver sweeper then installing the beta after I restart. I will get back to you after a few benchmarks on stability.

xioticin

Discussion Starter · #6 · Aug 11, 2013

So far seems stable with stock clocks on GPU and 4.5 on CPU.
This is my heaven score:

I have to go to work. I still am not done testing this issue and will be back at it in about 8-9 hours. I appreciate the suggestions thus far. I really do not wanna RMA this card if there isn't really anything wrong with it.

I will be able to respond here but not test anything while I am at work.

See less See more

NateST

· #7 · Aug 11, 2013

Unfortunately just because you're stable in a bench does not translate into stable in game X. I can bench Valley @ 1226/1880, but game stable clocks are around 1150/1750. I would game at stock CPU and GPU settings for a significant amount of time, then overclock one, then the other. Narrowing down will easily tell you if it's your card, your OC, or something entirely different.

xioticin

Discussion Starter · #8 · Aug 11, 2013

I will continue to test more tonight but with my CPU at 4.5 and my gpu at stock I did not have the issue anymore. I'm not out of the woods yet but I feel it was the memory overclock. I don't even see a huge increase in anything except benchmarks so I'm ok at stock card settings. I'm starting to see what you were talking about benchmark vs real world performance though.

Zero 7hree

· #9 · Aug 11, 2013

good luck buddy hopefully you get back up there again, I haven't oc my 780 yet but the issues I did have with crashing haven't happen since the beta drivers

xioticin

Discussion Starter · #10 · Aug 11, 2013

I really appreciate your optimism! I'm hoping it was just me being unaware about chip changes. Everything was fine for sure after a clean install of the beta drivers and stock clocks. Could possibly even be my mobo that's causing an issue with my oc's lol.

xioticin

Discussion Starter · #11 · Aug 12, 2013

So the girlfriend played nothing but league of legends while I was at work, and she said she had no issues at all. Nothing but smooth gameplay for quite a few hours. It is really starting to look like it was only my memory overclock that was causing the issue. It was so very selective though. Tomorrow I plan on just overclocking my core as I didn't do that initially and see how high I can take only that. I will report if any major issues happen.

Jameswalt1

· #12 · Aug 12, 2013

Yeah memory overclocks can be really tricky. Different programs act totally different from each other with memory overclocks. For example I can run Valley benches all day long with a +600 memory offset, however I cant make it 20 seconds in 3dmark with anything over +200.

Core clocks are way more forgiving and it's easier to find a stable 24/7 core overclock. I usually don't even overclock memory for 24/7 use, only core.

Cyblade64

· #13 · Aug 12, 2013

Reading about the overclock issues just wanna ask -
Im currently thinking of getting 2 x GTX 780 Gainward Phantom GLH editions - are these overclock issues ok if they are done by the manufacturer? I dont wanna go for the higher overclocked ones if they are gonna prove to be unstable and have to downclock them, what are your thoughts on this?
FYI the GLH edition has clocks of - 980 core 1033 boost, and 6008 on the memory, compared to the standard phantom being 902 core 954 boost and the memory lists the same of 6008.

xioticin

Discussion Starter · #14 · Aug 12, 2013

From my personal experience, I see no issue with the factory overclock of my card. I would say you would be perfectly safe with getting a card that is factory overclocked as its like getting a guaranteed overclock. I so far haven't had to downclock at all, just go back to my cards stock settings.

Those Gainwards look massive! I do like that fan design. If you are purchasing them for the oc' go ahead, but if you are purchasing for the temperature control the ACX is a beast as well.

Speaking of temps, ever since I have changed to the latest beta drivers, my cards has gone from 30 idle to 39 idle. I still cannot find what would be causing this! I remember reading that sometimes drivers can cause temps to just idle higher. I have looked for some form of bitcoin miner in my processes and see nothing out of the ordinary.

See less See more

xioticin

Discussion Starter · #15 · Aug 12, 2013

I have kind of struck on something as of last night. It seems I am having a bit of a DPC Latency issue. While using DPC latency checker, I am having red bars every so often and after I trace it, it shows that it is coming from nvlddmkm.sys. I was watching last night while I ran Heaven benchmark and was getting them, but I also seem to get them at idle. This process is related to nvidia drivers from the looks of it. Trying to sort this out further.

Cyblade64

· #16 · Aug 12, 2013

Yeah the Gainwards are a hefty size, 2.5 slot width so effectively take 3 slots each! Im looking at the GLH ones as they arent much more expensive than the standard phantoms and actually are slightly faster than titans, so just wanted to check about the overclocks, wouldnt be planning on overclocking them beyond this. Im looking at the Gainwards as one of my current 580's is a Gainward Phantom and its great so wanted to stick with those

See less See more

speedytech7

· #17 · Aug 12, 2013

I am having the same issues as your are having, I can play HL2 with my GTX 780 for like 45-60 mins. then it will freeze, go black, stutter sound, and then come back at a reduced framerate. I think it is driver related though, so I'll try the beta.

xioticin

Discussion Starter · #18 · Aug 13, 2013

I too believe it is driver related. I have been playing LoL all day long and not been able to repeat the error. If you are idling in the 40's on your temps, it might be google chrome and its gpu rendering/compositing. Just disable it in about:flags and restart pc. I know no one has mentioned that, but I have gone back to my normal low temps by making that change.

The Beta driver so far has helped me regain my confidence in my hardware working fine. Now I am looking into possibly getting a newer mobo as my old one might not be the best for overclocking my i5 anymore haha.

Reactions: 1

speedytech7

· #19 · Aug 13, 2013

Well, I played a good 2 hrs without issue with the beta drivers and adaptive perofrmance turned off, I too am now confident my hardware is solid. I was overly concerned because I just had a motherboard failure on the same computer (bought an identical replacement). That motherboard failure was a broken PCI-E connector and I was worried it took the card out with it. Glad to know it is just nVidia being buggy (lol I just switched from AMD for the same reason, at least nVidia fixes thing in their new betas).

xioticin

Discussion Starter · #20 · Aug 13, 2013

I think both are great companies at this point. My first card was a Radeon

I just swapped to Nvidia forever ago after one hardware failure and stuck with it since. Sorry to hear about your motherboard failure. It sucks to have happen but at least its a clear sign that its the culprit to problems within the system; I wish all issues could be that simple.

My performance has been fine tonight, though I wish my cpu and gpu usage were a little higher(though I probably shouldn't as that just shows my system easy easily handling it) . I do find it hilarious that my card can run LoL only using about 19% and its temp doesn't go above 38 pulling 70-90 frames; I know its not demanding just find it funny.

See less See more

1 - 20 of 32 Posts

This is an older thread, you may not receive a response, and could be reviving an old thread. Please consider creating a new thread.