• 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.

780GTX SLI keeps locking

steveoz

steveoz

steveoz

Associate
Joined
2 Jan 2014
Posts
223
Location
Wiltshire, England
Hi Guys,

It seems I have been having this problem for a while with planetside 2 where the game will randomly lock, sometimes it will play for hours, other times it will be a short period, and sometimes not at all.

I have kind of put up with it, but I just got Titanfall, played it three hours last night no problem, played for about 1 hour today and it locks up, then 30 mins it locks up.

Each time my PC totally freezes, and I have to do a hard reset. I thought the PS2 issues was with the poor coding as other people have had it too, but with it now happening on titanfall too, I am now thinking it really is an issue with my machine.

The machine is only about 6 months old, spec is:

MSI Z87 G45 mobo
16GB Ripjaw 2133mhz ram
4770k currently running stock boost clock of 4ghz
2x Zotac GTX780's in SLI running Nvidia surround
CM liquid cooler

The errors I have seen are bellow:

Display driver nvlddmkm stopped responding and has successfully recovered.

This was one of the first lockups today where I managed to ctrl+delete and then it said titanfall.exe has stopped responding.

On the subsequent two lockups:

The description for Event ID 14 from source nvlddmkm cannot be found. Either the component that raises this event is not installed on your local computer or the installation is corrupted. You can install or repair the component on the local computer.

If the event originated on another computer, the display information had to be saved with the event.

The following information was included with the event:

\Device\Video3
CMDre 00000000 00000454 01200000 00000004 00000010


I have done a clean driver install of the latest Nvidia drivers.

My top GPU hits 79 degrees and throttles very slightly as I have it set at 79.

My lower GPU is normally around 60.

If anyone has any ideas on what to do to try and troubleshoot this it would be much appreciated.


Cheers!
 
Hmm what word is blacklisted here?

Did you just add that Sig or did I miss it before?

I cannot find that PSU only XFX with near same name.

No matter what make/model of PSU old or new it can be dodgy.

If you exhaust all other checks you may need to borrow a PSU to try it.

Do both GPU's work on their own?
 
Last edited:
Correct!

I'm now wondering if VRAM is the issue, given they have 3gb each and I am running ultra / insane textures on games with a widescreen res.

I checked and VRAM is utilised 100% in both of these titles.

Perhaps the VRAM is maxed out and then the drop in performance or response makes the OS trigger the old recovery process for the display adapter which fubars everything?

Sigs always been there :)

I've ran them not in SLI before when my bridge died (well I wrecked it when I took it out of the computer by standing on it!) didn't seem to have any issues.

I've ran titles such as ACBF, ARMA III and lots of others with no issues.
 
Are you sure it uses 100% VRam and its not Caching it?

Also it would then fall back to using System Ram then HDD Page File.

Why is this PSU vendor blacklisted (I still have no clue who it is) lol.
 
Last edited:
I have to say for a high spec pc a better brand psu would be a good investment, had a look on there site for more info on it but they dont seem to have the 1000w any longer
 
The few times I've experienced regular lock ups in more than one game, it's always been a power issue. If it's the PSU I think it is (discontinued one listed as competitor name PowerStation Black Edition 1000W Silent ATX2 Modular Power Supply) then it produces 20A on two of the 12V rails and 30A on the other two. It might be worth trying swapping the PCI-E power cable around so you've got a 30A and a 20A on each GPU. Don't know if it would be better with the higher rated cable in the first power slot or the second on each card so may have to try a few configurations.

Personally though, I'd be looking to swap it out for a known solid PSU - Superflower, XFX, Silverstone, Corsair, etc. Your PSU may be up to the task, but an own brand make that only produces 900W on the 12V for a 1000W rated PSU doesn't look a top quality item to my eyes. It may cost a bit to replaced with a good one, but iit's cheaper than replacing a GPU/CPU/motherboard if something goes very wrong.
 
Thanks for the replies.

I'll take a look at the PSU set up first.

I just played for another couple of hours fine, then a lockup, I quickly and frantically hit the windows Key and CTRL+delete and got to task manager, the application had stopped responding, but EVGA shows something very odd......

These peculiar figures corresponded with the driver not responding message in event viewer.

weird.jpg
 
sounds like the problems I had a while with a gtx 480, would randomly lock up mainly in BF3

tried various power supplies, etc, even rma'd it but they sent it back as ok

I ended up fixing it by underclocking the ram on the gpu
 
Last edited:
I have ran it on just one but not one and then the other to test.

Jfan said:
sounds like the problems I had a while with a gtx 480, would randomly lock up mainly in BF3

tried various power supplies, etc, even rma'd it but they sent it back as ok

I ended up fixing it by underclocking the ram on the gpu
Click to expand...

Interesting, it's the Zotac 780 AMP which at the time was runing the highest mem clocks etc (that's why I bought them), but I know they had issues with the bios and clock speed on the first round of them so revised it.

It is a pretty high OC from standard so maybe a small underclock will do it then?

I have just tried switching the power round, I already had one 30A 12v on the 8 pin and one 20A 12v on the 6.

What I tried instead was to put all of V3 and V4 which is just 30A but dedicated one to each card for both eight and six pin. I rebooted, tried and driver stopped responding in like 30 mins.

So I ramped up the fan speeds to 100%, but the first card is still hitting throttle, I increased to the stock throttle temp of 80 from 79, and replaced the nvidia driver SYS file.

So far five matches and no lock up, we will see.....
 
So the clock is back to stock.

I changed the PCI link power state, and power options in both Nvidia control panel and windows to be set for maximum performance.

No difference, still random lock ups.

I have also tried changing the SLI bridge (last one had a small dent in it) just in case, and I also switched cards around, put the bottom and the top instead. Interestingly I get better temps this way, so one card must have better past or just be a better card. Still temps aren't the issue anyway, as I tested with the side of and both cards never got anywhere near close to throttling and it still locked after an hour or so combined with the above changes.

I now uninstalled and re-installed the Nvidia drivers, updated the mobo bios and all other mainboard drivers to ensure they are fully up to date and compatible with windows 8.1.

I have also now switched the SLI bridge from the front connector to the one closest to the rear.

I don't think it is power, as I have been running an hours furmark test with both GPU's maxed out and no issues or dips in power at all. Also when I switched them to run from one single 12v rail with only 30a it wasn't less stable, just the same.

I'm actually wondering if it is even the GFX cards now? I've noticed the Nvidia error I talked of on every start up, seems a common thing.

The display adapter stopped running has only happened maybe a few times in the past couple of months, as the other times it has been a complete system lockup.

I am wondering now if it could actually be the RAM causing this but who knows could be one of a million things.

I'm going to run memtest now and take it from there, also do a drive scan and make sure everything is good there.

Cheers!
 
Sorry actually I have just thought of something else.

My GPU's are both slightly different, which is understandable as no two are really the same, but one runs at a core clock of 1150 while the other runs at 1163, I noticed this when I originally installed them but didn't think much of a 13mhz difference.

Could the difference in the cards rendering albeit very small cause an imbalance or a stability issue? Should I try just clocking the higher card to the same as the other card?

**EDIT** According the the GPU bios they should be boosting to 1059, so I guess that the Nvidia driver / boost is boosting to what it thinks each GPU can handle, could this be a factor? Still rock solid on a burn in test with both GPU's at 99% usage, power consumption on each is at 88%
 
Last edited:
Test the cards individually, but I'm 99% sure it's the PSU. Store-branded junk = no good for your high-end rig.
 
More than likely is the power supply, but just try a big under clock , think I ended up with something like -200mhz under clock to get it to work 100%, and make sure there the same speed if poss
 
You must log in or register to reply here.
Back
Top Bottom