Zombiewarpig

The temps are still good. I am at 45 - 48c idle and 57c - 70c gaming with no crashing still. I ran a mem test on the gpu and it goes into the black screen before the end of the test so I am sure its a ram chip as the core is solid. I watched some videos on resoldering those by hand and it looked like a lot of work. I am tempted to sell it on Ebay as is with full disclosure of the situation to try and recoop what I can for a new one which is around $300 so I am only out a couple hundred bucks. Thanks for the info and replies:)

So I downloaded the 1075v 900/1300 bios and have not had a crash yet. So I am worried that this two year old card is tired and now it needs more voltage to run proper. I would imagine its on borrowed time. Any clue why this would fix it and should I be looking for a new card? Thanks

Thanks for the response and I will try the external monitor test tonight. Its a 17 so I only have one gpu and set to peg in the bios. Ran furmark lasnight and gpu never went over 74c and stayed around 71c with no issues.

No one has any suggestions? Maybe post is to long? 1. Dell 7970m Upgrade in M17x R3 2. Random problem with screen turning one color ( Red, White, Blue ) 3. Tests all come up as passed 4. Happens on internet or playing games at random 5. Goes weeks without issue 6. Can benchmark and its never happened during Update: So I just ran Memtest and Furmark for an hour and no issues but earlier it crashed to a black screen this time on facebook.

So I upgraded my laptop a year ago from a 6970 to a 7970m. It is a dell version and one day my laptop while playing BF4 crashed. Weird thing is its not a BSOD but a random color. I reistalled the whole OS and rethermal pasted the cpu/gpu with diamond. Now it ran fine for weeks and then out of no where on the internet it went to a white screen. I rebooted and ran every test I could and even benchmarked the crap out of it with no issues. Then I had some windows updates and it restarted and went to a light blue screen. Lastnight I started up Rust and with in a minute it went to a pink screen. I ran it all night and today testing it and no crashes. WTF is going on? I am not OC it at all and it idles in the 40's and reaches like 70c max. I keep it on a dual 140mm cooler at all times and blow it out few times a month so its spotless. Any Ideas on what I can do? I am pretty sure its the video card. Should I use a modded bios to undervolt or overvolt it?

I would also like to know what laptop is running sli 880m's on full blast at 70c. I have an AW 18 and she is sitting on a CM SF-19 with dual 140mm turbine fans and I used ic diamond. I am getting 75c to 85c depending on the game. I seem to hover around 80c with the fans on my cooler at 30% as it gets loud at anything above. I also am in my basement office and it's always 68F or lower. Either my laptop has issues or I am missing something, I also am running the mod-vbios and I was throttling with lower temps in most games before that.

Well my 780m went up in smoke and that's how I ended up here in the first place. Never even touched that card and lasted all but 4 months and I have read a lot of issues with the AW 17. Now I have a beast that got his nuts chopped off.

I noticed somethings with the mod as well. I am not getting shut downs on post but I am still getting throttling in a way, I will explain. This is just about gaming as this mod seems geared for benchmarks and for that it works 100%. It seems that on stock the clocks and gpu usage would throttle to what was needed to maintain my 60fps and it did that. I would really only see clock throttling in game when my gpu usage was not at max and that makes sense, why have the clocks at full blast if the gpu usage was at 65% as it only adds heat. I was running at around 38 - 42c idle and 75c in game on stock. Now I am seeing 42 - 54c idle and up to 90c+ on the mod. The idle is all over as you have to disable shadowplay in the startup or else it will not down clock and if you want to use it then you have to enable and restart every time. So in conclusion it seems to me that on stock the 880m seems to use the clocks and gpu usage in a sort of balancing act and only using what is needed with vsync enabled. While running at max clocks on the mod the cpu usage does not drop accordingly and remains higher then needed in some cases driving the temps through the roof. Now I know that the driver has a huge impact on performance and the latest beta shows this as it cripples the card with the mod vbios. So did Nvidia push the chip further then it should and the 780m was the sweet spot? Did adding 8Gb of vram somehow have an effect even at the driver level? I ask that as most games only report 4Gb when looking at the card. Does the benchmarks reveal the limits Nvidia set to prevent overheating or are the benchmarks really the problem? I mean its where you would overclock and run the card at max in a different way then your typical game. I myself think I will be going back to stock until things get sorted out as If I don't have my Laptop on my cooler in my cold basement then I wouldn't be able to game under 85c and that I do not feel comfortable with.

Same thing is happening to me. I can oc the core and its a hit or miss if its going to work and even when it decides to only one of my cards has while the other stays stock.

Thanks

Yeah never had secure boot on and don't use it. I did have legacy on as well but turned it off because I don't like to see the credits screen on boot up that came with the modded vbios. I really do appreciate the modded vbios but wish there was a donation option for leaving out an extra screen on bootup with these credits, hint hint.

I have the AW 18 with a 4910 @ 4.1 and dual 880m's so I don't have a single Mhz to spare. I just wanted to confirm that what that gentleman was saying was indeed true. I think there is something that is just off with the stock bios and the modded one as well. I have noticed things since flashing that have concerned me that I didn't have an issue with with the stock though. Sometimes my temps run away while gaming and after a restart I can play for hours on the same game without the same issue. With stock vbios I was gaming around 75c and never had a huge throttling issue after the A07 Bios only in benchmarks. Now I am hitting 90+c in certain demanding games and a re-paste with diamond did nothing to change it. I am starting to think that while higher benchmarks are nice to look at, I don't know that this extra heat is worth a few extra fps in a real game. If I am playing a game that needs 98% cpu use then I need a dual 140mm cooling pad to stay at 80c. If I play without a cooling pad then it sounds like my laptop is going to take off and the temps do just that. I miss my oc 680m that ran 65c max. The 780 in my AW 17 was a hot little prick to but not this hot. I keep my laptop in my basement office that is around 66F on this cooler so that tells you something. Second thing I noticed is I have to disable shadowplay in my startup or the gpu's stay at max speed. While not a big deal it has been know to get stuck and the internet also brings the gpu's to max Mhz as well which I am not sure is needed. I don't think I will go back to stock but wait for a revised 880m bios and see what the future drivers have in store. The new beta though is just terrible and halves my cards on all games although benchmarks are only 5 - 10% slower. I would be curious if the new beta driver cripples a 880m on stock bios as well.

Yeah it was only when I went back that OC became broke. It allows me to change the clocks but nothing happens when I do now. The new beta driver was causing my gpu usage to throttle around 60% even though my core was at full. I was getting just over half the fps I was before. Now going back things seem to be running like they should and I don't oc anyway so not a big deal, I tested what the other guy was saying and it seems to be true. I can oc everything to the max hit apply and play a game like nothing happen.

I did the same thing and also noticed I can not oc the mem or core after reverting back. I have not went back to stock vbios though. Are you having throttling issues still with stock and can you oc now?

No problem glad I could help:)

Hey guys, I am enjoying the new vbios on my cards but had a question. On some games now my cards reach as high as 80c and then the fan goes into high gear and it drops a few degrees. Is the 880m safe running at that temp while gaming or should I think about re pasting? Thanks

I found the direct culprit that was preventing my 880m from down clocking when under 0% load. Disable Nvidia Capture Server Proxy in your startup. This is Nvidia's shadow play and for whatever reason it pushes the clock to 100% under 0% load. It still takes 20 seconds after boot up to downclock but works everytime. If this worked for you then I would be happy if you let me know.

Better performance running benchmarks or gaming? I went from a 680m to a 780m and now dual 880m's. I can say that for Gaming the dual 680m's are going to be the best even with a lower clock I would think as they are around 30% faster then a 780m. But if I had to choose between those two then it would come down to what you want to do with them. Gaming wise the 880m stock vs a 780 stock are going to be almost neck to neck depending on the game. With a the vbios mod the 880m for sure why not free oc. Or you could use the 780m vbios mod oc it as well to match the 880m but from what I saw it still may not perform better game wise. If you are going for benchmarks then test all of them and see.

Hey, Well for me I had to remove it from startup along with shutting it down when used in order to down clock. I would just remove everything you can from startup besides security and what not and open inspector. As soon as its running now and I close it my cards drop, but not before I removed it from startup as well as taskbar.

Well indeed I have this problem of the cards staying at full power at idle. What can I do to fix it? I fixed it thanks for the reply and the mod svl7. Turns out that Nvidia Experience was in my startup and preventing my downclocking. Dropping -8c at idle makes me happy.

Does that have to be? Is there a way to let it downclock when not used?

Shouldn't they throttle down when not under load though or are they meant to just run full blast 100%.

Well I made the jump and flashed both my 880's. I can tell you that I did get around 1000 jump on 3dmark11 but I have no head room to oc my cpu/gpu. As soon as I try and make the slightest bump my system freezes on the first 10 sec of benchmarking. My twins are running about 5c hotter right off the bat as well. So I have pushed the computer to the power max with my 4910mq @ 4.1Ghz and 880m's with mod bios. But now I know that its as good as it gets and thanks for the vbios, I will donate as soon as I can:) NVIDIA GeForce GTX 880M video card benchmark result - Intel Core i3/i5/i7 4xxx,Alienware 01W2J2 P13763 36th place Any reason why 3dmark does not recognize my cpu? And there is no adapter mod for me is there? Thanks

I re ran it and score even higher with just a small +50 bump to the core and +250 on the mem. Scored much higher this time. Weird thing is after all the game benchmarks I ran both gpus stayed at 1044Mhz the whole time and only 3dmark11 is causing these throttling issues. NVIDIA GeForce GTX 880M video card benchmark result - Intel Core i3/i5/i7 4xxx,Alienware 01W2J2 [h=1]P13234[/h]

I have yet to flash but was curious about my score. Am I lower then to be expected for dual 880's. Also once I flash will this have a big impact on gaming performance. NVIDIA GeForce GTX 880M video card benchmark result - Intel Core i3/i5/i7 4xxx,Alienware 01W2J2