• 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.

Skyrim black screen in SLI

I can't see why, I don't understand how it is a SLI problem. In fact in this instance SLI is exposing a problem that would otherwise probably go unnoticed.

Aside from this strange (and admittedly very annoying) quirk I've got no complaints about the drivers or performance relative to my previous 6990+6970 setup.
 
Yeah I guess so, got no idea how to use GPU-Z to simulate a similar level of load that Skyrim was applying across all 4 cores though. I guess I could play around with the resolution and detail until it's close?
 
Good luck with tonights fault finding. Post back with your findings please. I'll be monitoring post from the pub, tonight is curry club and the match is on but I'll be back home 10PM ish. Theoretically, the lad above is right. You could use GPU-Z to record gpu voltage, and then when it crashes, reboot and look to see if there was some kind of change the moment it happens. I believe it can record results to a graph, or something similar. Never really used it, I use Afterburner.

Got to get it sorted ready for Dawnguard!
 
Quick update:

Removed the bottom card from my PC, loaded Skyrim and just left it idle (after a short time the camera starts panning 360 degrees around the avatar, I assume this will be sufficient for testing purposes). Had to force a 30 FPS Frame Rate Target to try and approximate 50% load since one 690 can run run Skyrim at 60 FPS without issue.

Left this for about 4 hours and it hadn't crashed by the time I switched my PC off. Not sure how conclusive this is because I could play Skyrim for hours sometimes without a crash. :(

Now to do the same thing on the other card...

Should I switch out the cards completely or use the PCI-E slot disable feature of the motherboard?
 
I imagine you won't need to disable the slot if there is nothing in it. For testing purposes though, you need to do the other card, in the same slot, and then repeat the process for the other slot (like how you do it when your testing ram in memtest)
 
Yeah EVGA suggested that when I filled in a ticket. The SLI bridge is brand new, came with the motherboard. Could conceivably be faulty I guess but seems unlikely. Will try a different one anyway.

Did I mention I hate removing stuff and readding stuff from my PC constantly? :(
 
Update:

Well, think I'm reaching a conclusion...

Removed the top graphics card and replaced it with the bottom one which I had removed previously. Rest of system & configuration unchanged. I fired up Skyrim with the same setup as before (Frame Rate Target of 30 FPS), straight away noticed that this card needs 1.062v and 1.075v for the two cores, compared to 1.050v for both cores for the other card. Manufacturing difference there I guess.

More importantly I played about 20 minutes of Skyrim and in that time I got a CTD with "Nvidia driver has stopped responding and recovered" (or whatever the text says), the Skyrim process just disappeared completely.

So whilst I didn't get a BSOD I think I can say that this card isn't 100% stable, and when it's running in a SLI configuration having one card crash with the other still up probably would result in what I've been seeing - a black screen and unrecoverable graphics situation.

I've left it running again as I did the other night with Skyrim "idle" where it does a 360 degree pan-around, but I'm confident in thinking I need to send this card back.
 
Just crashed again in Skyrim, this time it's a proper lockup. I can see the desktop but it's frozen.

End of the road methinks! Thanks a lot for all of the help provided by everyone.
 
Yeah, sounds silly but a wave of tranquility washed over me when the 2nd card crashed lol

Was starting to think it would be such a narrow set of variables that would cause it that I'd have a nightmare trying to RMA it. The 2nd card would actually play Skyrim for a while, and would seem to run it fine during that time, until something would cause it to crash. I'm guessing that when it was in SLI with the other card having one card crash brought the whole system down.

edit: As a side note - EVGA are awesome. Not only have they not questioned me RMA'ing this card, but I noticed today that the 30 day extended warranty extension period had lapsed (I started doing it when I bought the cards, realised I could do it at any time during the 30 days, and then subsequently forgot - schoolboy error!) and they allowed me to extend it anyway.
 
Last edited:
Received a replacement GTX 690 from EVGA today on advance RMA (love this scheme), popped it in and played some Skyrim for a few hours and didn't crash once :D

Glad we finally got to the bottom of it :)
 
Back
Top Bottom