• Competitor rules

    Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.

System Crash (new GTX1080)

Will set to gen 3 (currently on auto) and give it a whirl. Not to sure what you mean by the PCIe frequency though, is that a bios setting?

Cheers guys.
 
Not to sure what you mean by the PCIe frequency though, is that a bios setting?

Just a wild stab in the dark :p

I haven't specifically looked for it on an Intel board recently, and from googling it's perhaps tied to bclk on Intel??? but yes. I had to lower mine to 98MHz ages ago, on an AMD-based PC with a Radeon HD5870, or I'd get a grey/green screen of death.
 
Yep, sounds like it. There should be a workaround though...

There shouldn't need to be a workaround. I thought there would be one, but there wasn't one I could find. There was a fault that cost me many hours of my time. I tried everything I could, but I still had a graphics card that wouldn't work in my system.

If only graphics cards were like washing machines or TVs. If it doesn't work properly, the retailer or manufacturer fixes/replaces it.
 
There shouldn't need to be a workaround. I thought there would be one, but there wasn't one I could find. There was a fault that cost me many hours of my time. I tried everything I could, but I still had a graphics card that wouldn't work in my system.

If only graphics cards were like washing machines or TVs. If it doesn't work properly, the retailer or manufacturer fixes/replaces it.

Try the rainforest next time
 
GTX1080 under heavy load via GPU Boost 3.0 uses a tremendous amount of extra PSU juice! That's just the GPU throw in the CPU+Ram etc etc I think this PSU is marginal for this type of GPU hence me saying back it all back to no overclocks see how stable it is then.

All PSU's degrade over time with natural use any ocing just degrades them faster if you used a probe on the 12V rails like some of the more thorough reviewers do the results on older units would be eye opening!! they lose up to 30% in normal use so ocing introduces even more degradation. PSU's hate extra heat & load its been said by experts they lose around 10% per year if driven harder so a 3 year old PSU = 30% degrade!
http://www.johansondielectrics.com/ceramic-capacitor-aging-made-simple#.Um_e0fmkr9k
http://www.tomshardware.com/reviews/power-supplies-101,4193.html

Don't listen to any expert that tell you a PSU will degrade 10% per year. The 10 year warranty some PSUs have would be a bit of a problem if that were true!

In general a reasonable (pretty much most of OCUKs inventory, even the cheaper ones) PSUs performance many degrade 0-1% per year even with 24/7 usage. It would be almost impossible for a PSU to work if it degraded 30% as you'd hit a complete component failure well before it got to that level.

Overclocking specificity has nothing to do with degrading.
As long as you are using a PSU that is suitable for your usage it shouldn't be a factor. If you were overloading it then that would be a problem and that would cause deterioration in the PSU.
The PSU doesn't know how you are using the power, it just supplies it. 400 W on a stock system is the same as 400 W on an overclocked system.
Temperature is a main factor for degrading, so a PSU in a hot case is likely to have a shorter life span.

Overclocking is better with a good PSU that can maintain a stable power supply. But that has nothing directly to do with PSUs degrading.

PC PSUs mostly use electrolytic capacitors. And they are the ones under the most stress. They are in general reliable components if they are used with in spec and not faulty due to manufacturing defects, which will tend to show up very early in usage life. (not counting cheap ones)

Ceramic caps used in PC PSU aren't under the same sort of stress as the electrolytic capacitors so ceramic cap ageing is very unlikely to be an issue with these.

As for power usage with a GTX 1080, I wouldn't say it's using a tremendous amount of power, even my current rig overclocked won't go over 650 W from the wall and that has two 1080's. Even with the more power hungry CPU his PC isn't going to get close to that with 1 1080.
 
Last edited:
All stabs in the dark welcome. I've had some dark weeks with mysterious computer behavior before but this is quite unreal.

Have you tried running the GPU fan at 100% and run the benchmark.

This will keep the card as cool as possible and at least remove the possibly of a temperature related issue. If that fixes the problem it would suggest a GPU issue.

Also I guess you could try powering the GPU using a different power cable.
Even try powering it from the Molex cables using an adaptor(s). It's possible that a particular connection is faulty and the 1080 doesn't like it.
 
In general a reasonable (pretty much most of OCUKs inventory, even the cheaper ones) PSUs performance many degrade 0-1% per year even with 24/7 usage. It would be almost impossible for a PSU to work if it degraded 30% as you'd hit a complete component failure well before it got to that level.

I agree with this, and everything else you stated.

The power capability of a PSU isn't likely to reduce gradually over it's lifespan. The semiconductors in the PSU will either work or they won't, they generally don't "weaken". Electrolytic capacitors do tend to degrade slowly, but until certain capacitors reach a critical point in their life, a PSU will continue to provide pretty much the same level of power output as it did when it was brand new.

Again, it's not entirely impossible for the OP's PSU to be causing the problem, but it sounds more like an incompatibility issue between the graphics card and the motherboard. It may well be something that can't be fixed with a driver or a BIOS setting. I`ve experienced this kind of problem, and it's very frustrating and difficult to diagnose.
 
I am at a loss now, I have tried the following since my last post:

Fans at 100% throughout - Crash after 6 mins Witcher 3

Completely different power cable from a different socket on PSU - Crash after 3-4 minutes Witcher 3, crash after 25 minutes Prepar3d.

After a full case cleanout and reassembly yesterday as well as a brand new OS on a clean HDD I think I am spent. All my parts are from OCUK but obviously I can't just RMA things based on hunches so I feel my only option is a refund on the card and go back to the 780ti for a couple of years until I can do a brand new system. Annoyingly this has knocked my confidence in building my own rigs, something I've been doing for 10 years. Usually with these kinds of things I eventually find a solution but I am running out of ideas and I am sure you guys are too.

Unless I just happen to have two DoA cards in a row? I've heard the EVGA FTW 1080 cards had a major problem with crashing and lots of RMA's. Is another RMA worth trying or maybe a swap to a different manufacturer? (MSI for instance).
 
Last edited:
Unless there's something special about GTX1080s I'm not aware of: If you were to switch to another manufacturer, the only difference as far as your motherboard is concerned would be the GPU's BIOS.

Under that assumption, you could just flash to an Asus BIOS (to match the board?)

Edit: It appears there are a two versions of the G1 BIOS:

86.04.17.00.66
and
86.04.17.00.FD (newer?)

Nvm.. just read that you already tried that :(
 
Last edited:
I am at a loss now, I have tried the following since my last post:

Fans at 100% throughout - Crash after 6 mins Witcher 3

Completely different power cable from a different socket on PSU - Crash after 3-4 minutes Witcher 3, crash after 25 minutes Prepar3d.

After a full case cleanout and reassembly yesterday as well as a brand new OS on a clean HDD I think I am spent. All my parts are from OCUK but obviously I can't just RMA things based on hunches so I feel my only option is a refund on the card and go back to the 780ti for a couple of years until I can do a brand new system. Annoyingly this has knocked my confidence in building my own rigs, something I've been doing for 10 years. Usually with these kinds of things I eventually find a solution but I am running out of ideas and I am sure you guys are too.

Unless I just happen to have two DoA cards in a row? I've heard the EVGA FTW 1080 cards had a major problem with crashing and lots of RMA's. Is another RMA worth trying or maybe a swap to a different manufacturer? (MSI for instance).

I have had a PSU to be faulty where it was able to supply enough power for a lower spec system without any issue and then I added a newer GPU that used more power and it would have similar issues to what you are having.

It was an Antec unit rated at 1200w with loads of headroom (done on purpose to allow such upgrade) and because I had it for a few years it was out of warranty when I found the issue. So a new PSU did resolve my issue but not sure how common that really is.
 
Hardware troubleshooting isn't difficult:
1. Test the card in a different system to rule it out. Based on the number of B-grade 1080GTX's for sale the card is probably fine.
2. If the card is good then either the system (CPU/RAM/mobo) is bad or the PSU is bad. Swap out the PSU to confirm.
3. RMA / replace faulty parts as needed. Done.

It's not a compatibility issue or software issue. If it was the card wouldn't work for 5 minutes then crash. It wouldn't work at all.
 
Am running Asus X99-S board - 2 x 290 gpu (not clocked) and 5820@ 4.2 (i.e. simply changed multiplier in bios to 42 so it runs slower until required :)) PSU is 1000 w Superflower platinum.
First board was faulty but second (which I am running now) has been stable. (problem with first was the well documented problems that arose some times with ref. to usage of rear 3.1 slots ; which I now try to avoid using to avoid any such problems arising!)
Your issues are not something I have any experience of so doubtfull I can assist :( but I cannot see that it will be the gpu (can you try it a n other machine?) and rather that it is problem with the board and your overclock.
If I read your previous posts correctly was it not the case that all was stable when you reverted to standard clocks for cpu ?
Only comments I found ref . overclocking which might give some clue was the comments mafe in Kit Guru's review which I have linked to below and which might (most unlikely I think- but you never know) be of help.

http://www.kitguru.net/components/motherboard/luke-hill/msi-x99a-sli-plus-motherboard-review/9/

Good luck!!

B.......... :confused: have just read carefully (your) post 35 where you stated if failed at stock - my apologies ; the above will more or less be useless. - One can but try!!
 
Last edited:
Hardware troubleshooting isn't difficult:
1. Test the card in a different system to rule it out. Based on the number of B-grade 1080GTX's for sale the card is probably fine.
2. If the card is good then either the system (CPU/RAM/mobo) is bad or the PSU is bad. Swap out the PSU to confirm.
3. RMA / replace faulty parts as needed. Done.

It's not a compatibility issue or software issue. If it was the card wouldn't work for 5 minutes then crash. It wouldn't work at all.

My GTX460 worked for about 30 seconds under load, then the system would freeze.

Tested it in 2 other PCs, and it worked perfectly. Purchased a cheap motherboard as a temporary replacement while my motherboard was away for repair. GTX460 worked perfectly in the temporary motherboard. Motherboard was returned after thorough testing, no fault found. Tested it again with the GTX460, crashed within a minute. Sent both motherboard and graphics card for testing/repair (fortunately they were both Gigabyte). GPU was repaired. When it was returned with the motherboard, I tested the GPU with the temporary motherboard and got a blank screen. Put my original motherboard back in, and the graphics card worked perfectly. About a year later I upgraded my PC with a new MB/CPU. It was no surprise that the graphics card had to be sent away again when it didn't work with the new motherboard.

Your advice (points 1 & 2) seem logical, but that logic did not apply in the case of my GTX460, and troubleshooting that fault was very difficult (as it defied logic).
 
Holding out for some magical beta bios!
Try this:
1: Only use 1 stick of Ram & make sure you have the pagefile enabled on the OS as system. If it still crashes try another stick of ram & so on its possible the ram could be faulty and cause this. I had that last year on a GTX980Ti. Faulty ram caused the GPU to lockup until ram was replaced.
2: Connect only the lowest number of items which use any power with the case side off disconnect all fans except PSU+CPU & only connect the mouse & kb no other USB devices.
3: Clean the motherboard PCI-E slots with compress air by blasting them a few times at close range to eliminate any dust or fluff.
4: Connect the GTX1080 with a dedicated 8 pin molex from the PSU. Try another one do not use the same one.

Obviously do all of this at stock clockspeeds & reset the bios back to std settings for everything before powering on after cleaning the motherboard.

If after all of this you still get the same issue all you can do next is find someone in your area willing to allow you to test the GTX1080 in their motherboard or buy a really cheap used model from an auction site.

If your GTX1080 is still faulty then go back to Gigabyte & ask for an RMA. Not sure how good they are nowadays but a few years ago it took them almost 2 months to deal with mine so I vowed to never buy anything from them again as it failed after a few months & as outside 28 days they made me wait for months.

At this stage all you can get is a like for like replacement from Gigabyte which means either sell secondhand if it does not work in your system, buy another motherboard or buy a GTX1070 from another vendor more reliable like MSI with your GTX1080 used funds.
 
Just an update in case anyone encounters similar problems.

2 RMA'd cards and a swap to a different brand (MSI) and the problem continued. Problem was solved by uninstalling the Realtek audio drivers. Either I had old ones or something else was wrong with them (installed via MSI Live Update every time I did a driver sweep). Not sure if I am to feel incredibly stupid or relieved! Few days now with 5+ hours on Witcher 3 and 3+ on Metro 2033 Redux (both prime culprits for the crashing) and stable. My overclock is back up to 4.4ghz and I can run the card without underclocking.

Thanks for all your help and patience guys!

(Also the 1080 is a beast and well worth purchasing for anyone considering).
 
Last edited:
Glad to hear you've got it sorted. I don't think that's something to feel stupid about, I'd not have thought of audio drivers in this scenario. Relieved is ok though :)
 
Back
Top Bottom