help...computer crashing - one my cards dying?

i was playing witcher and my computer froze up black screened i restarted drivers didn't load got a message saying "directx 9 failed to load", restarted again drivers loaded but its freezing and crashing now, reinstalled older drivers same problem, cards did NOT overheat...were working great 75C steady.....should i just disable one of the cards in device manager and see if problems go away and that card is defective? something is obviously wrong =/ very sad... witcher 3 killed my pc lol

there is stuttering on windows load and freezing

i have been playing witcher 3 and dying light all day with zero problems....

 

confirmed: disabled 2nd card no more stuttering or freezing

is there any way to diagnose this maybe bad SLI cable or something? although cable is brand new...would a bad sli cable cause it to freeze on the desktop?

 

Try removing one of the cards and see which of them works fine without SLI. If both are OK, than it is strange issue. Think cable is not a culprit, most likely some kind of a GPU failure or a software issue.
You can also try clearing CMOS in BIOS, and try boot with another HDD\SSD to see if it behaves the same way on different OS version - say if you are running Win 8.1 try Windows 7 system disk.

 

how could it be software issue if they ran good all day, the previous 780m ive had for over a month already has been running with no issues, this 2nd one i have had for literally only 1 week; i bought it used from china

i will try disabling the other card in device manager and see if it works just one the 2nd one

EDIT: disabling card1 and using card2 only results in a "failed to load directx9" upon startup

EDIT2: you can actually see where it hangs during bios post for 3 extra seconds where it didn't before; the card still loads up windows and works but something is apparently wrong with it that it will not initialize the drivers correctly, and nvidia control panel won't even open on card2

 

Try going back to a good driver, like 345.20, instead of using the new piece of garbage from NVIDIA. Clean with DDU first. Just because the new driver is supposed to be good for the new Witcher game doesn't mean it is a good driver, or that you must have it to play the game. I often find that the "Game Ready" label is nothing more than meaningless marketing hype.

 

did DDU in safemode, installed older driver, same problem card2

 

hmm... doesn't sound good. I've never seen a bad SLI bridge cause that kind of error.

 

i contacted the chinese seller regarding a return/refund

so much for my first SLI system....will get another card ...happens

its possible the card was just abused in it's past life

sure is nice to have 2 video cards....at least i have a working computer still with a 780m

 

Yes, once you have multi-GPU it is really super hard to settle for less. I know I struggle with accepting single GPU.

Let us know how the exchange goes. Probably should try both cards in the primary slot solo just to confirm it is the one you just bought.

 

I would try this.. I think it's probably a bad SLI bridge...

 

780M are known to be stable cards, it can be a SLI bridge.

 

it could...but it's brand new and the card stumbles and freezes even with SLI disabled

 

Best way to know for sure (without buying an SLI bridge and waiting for it to arrive) is to remove both GPUs and use them one at a time in the primary PCIe slot and see if one of them has problems.

I had one of two new 780M cards die within one hour of installation in an M18xR2 that I upgraded for a fellow back in November. It worked fine for about an hour, then all hell broke loose. The eBay seller exchanged it promptly and everything has been fine since then.

 

to do that i would have to go through taking it apart and exchanging the heatsinks on the cards...way more effort than im willing to put in this lol...way i see it is card2 failed pretty definitively...not going to go through the effort of ripping the laptop again unless i already have the replacement 780m

 

I know your pain well, the problems you describe are exactly what I went through when one of my 980ms died. Thanks to the help I got after I posted my problem (thanks Mr. Fox) I did what they are suggesting you do, and id the right dying card (I had it wrong). I would say try what the others have suggested, if nothing else you will learn to break down your pc quicker.

 

i've already ripped this laptop apart more times than i ever wanted to..install ram...install cd hd adapter...install new cpu...install new gpu...install 2nd gpu....im pretty good at it and i hate doing it

as an aside, i've had no negative experiences with the new 352.86 driver ..i modded it as per mr fox's instructions and it installed and runs great

 

ok now my 2nd GPU is having the exact same problem...computer crashed now failing to load drivers on startup .... you have got to be kidding me

what are the chances of 2 GPU's failing within a week of each other

 

ok confirmed ...i have had 2 video cards fail in an identical way within 2 weeks of each other...the 2nd one actually starts up in windows as "disabled" unlike the first failure which starts up normally but doesn't load drivers just like the other one

im honestly worried about putting another set of cards in this computer because i have no explanation for this unless its just bad bad luck

id like to add that for 6 months prior to that the 560m was running just fine with no issues and im actually about to put the 560m back in...don't really have any choice at this point

780m #1 was in the computer for 1 week before failure
780m #2 for less than 2 months before failure

neither card was running hot, both on liquid metal paste

 

I hope you've applied the LiquidUltra properly and it only covers the core.

Have you tried the 560m? I've gone through quite the list of parts before I narrowed it to being the MoBo (my brother's, not mine). Might be your MoBo failing and burning those GPUs (not that Chinese aren't re-sending defective parts, just to be sure, I've dealt with a lot of those). With that said I wounder if it's safe to try the 560ms as well.

 

well if the 560m fails then its time for new laptop

i really have no way to test if my mobo is burning the cards but the 560m ran in that mobo for months and months without issue gaming full time

i applied the liquid ultra to cpu and both GPU's and CPU has been running for months without issues..with a hefty overclock too

neither card was ever overclocked by me

 

If you ask me, take your chances and test with the 560m. No need for more money and time wasting. If it burns, you know it's MoBo, and to be honest I'm not sure you'll find one in good working order. I'm on the third one, and either I'm psyched already, or there are indications that it's going to be yet another faulty one. So you might as well start looking around for a new machine.

 

ok so goodnews/badnews same problem with the 560m drivers won't even install....bad motherboard? on the bright side does that mean my other 2 gpu's might still be good?

 

ok question....the default alienware drivers install but none of the others will including mr fox's 345.20 modded...so confused

do i need to mod newer drivers for a 560m on an m18xR1?? i didn't think i did

306.97 drivers from alienware installed and are running just fine...

im just getting an m18xR2 not going to take the chance to put more expensive cards in this just leaving the 560m to see if it burns out too

i take it the fact that the 560m is running fine means the other 2 cards are probably toast

 

Nah, don't even do that. Motherboards are getting hard to find and NVIDIA is doing retarded stuff that causes SLI throttling with the R2. I have an M18xR2 that belongs to a friend and I'm having a hard time finding one now. I got one from Parts-People.com that was DOA and they are having a hard time sourcing one to send for the RMA exchange.

Just get a Panther (P570WM) or completely forget about laptops and build yourself a desktop. Alienware, Intel and NVIDIA are basically ruining everything in the high performance laptop scene with crippled drivers and BGA feces. It's just not worth it any more. It's all turning to crap for the most part. I'm not going to spend any more money on laptops after all of the throttling crap I have had to put up with. I refuse to settle for a high performance computer (laptop or desktop) with one GPU and SLI laptop options are extremely limited.

 

i have to get another 18.4" for my gf anyway and i have 2 780m's that may or may not be good sitting around so ill most likely go with a cheap used m18x and shove all that in there ...desktop is deff not an option for me

EDIT: hmm ok all of a sudden i was able to install latest drivers without issue..weird

 

can anyone comment do the newer drivers make throttling worse? im getting throttling at 90 on the 560m but problem is that the clocks won't bounce back to full when the card cools off

EDIT: ok it really is starting to feel like my motherboard is bad...with the 560m drivers will fail to install randomly even after a full clean then install fine at a later time...and i dont think the downclocking issues im getting are normal

can anyone comment before i end up buying another one...basically with the 560m now its going to 90 and downclocking to 4fps in-game..ive never had that issue before i had the 780m's in it...at least not that i can remember

 

put the 1st 780m back in minus 2nd card minus SLI cable and its working fine....wtf so confused

will continue to test

 

okkk so 2 days now almost its running fine....should i try putting in 2nd card without SLI cable see if it runs ok to try to isolate it?

 

Here is a crazy idea. It may be the contacts MXM card and the slot are dirty. Take a q-tip, soak it in 91% isopropyl alcohol (or greater), and gently rub the contacts on the MXM card and in the slot. Of course, make sure none of the cotton gets left behind. A dirty/finicky contact may be to blame. The smallest changes can make the biggest problems.

 

ok did some more testing last night

p780m by itself works fine
putting in s780m with sli cable - problems
removing sli cable - still problems
putting 560m into secondary slot - problems with p780m ??
putting s780m into primary slot - black screen on boot DEAD
putting p780m back as i started would not work until drivers reinstalled even though same drivers were there to begin with, now working as before

soo s780m definitely dead, but i don't know if the motherboard killed it or the other way around so i will not be trying SLI on this system again

can anyone comment on why putting 560m into secondary would make the 780m in primary appear as malfunctioning? no sli cable obviously was used

EDIT: also i had my SLI cable on backwards apparently but it ran like that for a week so i dont know if that was somehow responsible for what happened?

 

Conflict in PCI IRQs and commands trying to take each other over would likely cause conflict. Is your dead 780M under warranty?

 

Yeah, some kind of conflict and/or short on the second slot.

 

well several days of usage only problem is computer freezes at least once a day usually 2x requiring a restart

games work without issue its freezing during regular desktop usage

 

so im still having hard freezes with the computer when gaming but im not quite sure how to diagnose it, there are never artifacts it just hard freezes and works fine on restart

i adjusted cpu and ram just to see if maybe one of those was responsible but im not sure how to diagnose at this point other than just throwing another GPU in which im reluctant to get if i possibly have a bad motherboard

EDIT: raised FLEX VID a couple points, downgraded video drivers to 345.20 and disabled HWINFO and no more crashes today.....

 

well i feel slightly stupid but one of those 3 things definitely eliminated the crashing i guess now i have to go back and figure out which one

 

Probably flex VID set too low.

 

well i have it set on 15 now and i lowered the OC to 4.0 whereas before i had it on like 10 with 4.5 so that doesn't really make sense but i guess GPU wouldn't randomly freeze my comp like that?