Alienware 17 + 780m Issues

Hey everyone, I got my new 17 about a week ago and LOVE it. I have an issue with the 780m though.

It seems to be throttling when playing ANY game. I have looked into it a bit and it seems to be setting or driver based. I am running Windows 8.1 (clean install form a Win7 factory OS), and have upgraded to the latest Nvida beta drivers (337.50).

At first, as soon as I launch any game, I can watch temps and power states fluctuate wildly. In Nvidia Inspector it immediately jump to 89-91c in any game, and the power states go from P8 to P5 to P0 back and forth. The temps don't really ever go down. And the clock speeds fluctuate wildly as well.

I then found a power setting in Nvidia Control Panel and set it to Maximum Performance, instead of Adaptive. This helped, temps actually went down, to the mid 70s to mid 80s, and performance was smoother, but still a little janky. Not up to par really. Is there any way to turn off all the power saving features and let the 780m live up to it's potential? It still feels gimped, and I have to set things to Low to keep frame rates within something that doesn't look like a slide show. I tried multiple games too, Old and new, and same thing has been happening. Thanks!

 

i have my 780m set to max performance and high performance power setting and I don't see mine reach into the 80s...usually sticks within 60-70 deg C.

it's either the GPU or the thermal paste on the heatsink isn't working.

 

To really get the most out of that 780M, you will need to re-paste it and flash the modded vBIOS provided on TechInferno for a small fee of $5. I think you can access it for free, but you must meet some forum/post requirements.

 

probably should just test it after replacing the thermal paste first to make sure the card is functional and the temps are down to the average before flashing any VBIOS.

 

Yeah, I agree. But if it's only reaching mid-80's, it shouldn't be throttling due to the temperature, which is why I suggested flashing to a modded vBIOS. The default temperature limit is 93C or 95C.

 

fair enough but if you compare it to my 780m's operating temps they're in the 60-70s and mid-80s is high for everything on stock so i still think in the best interest of eliminating probable causes that the thermal paste on the heatsink should be replaced and then retested for temps.

run something like precision X or afterburner and have the temps displayed in an OSD so you can see it as you play.

 

Is it really the optimal choice to repaste a laptop that is only a week old to me? Should I just call Alienware tech support or something?

To me, watching the temp spikes, when it's switching power states, during game play (which in my opinion it really shouldn't be doing), it seems to raise in temp to the highest it can go, then it drops the power setting instead of cranking the fans. The fans are blowing, but don't seem to go full speed. It seems to just drop the power state and makes it feel like it's downshifting and whatever game I have going at the time slows to a crawl, then ramps back up. This typically happens over and over every couple seconds.

 

And now it just started doing this, after launching a web browser. I closed it after I noticed the fans cranking up, and now it's just sitting there at desktop and not dropping..

 

Have you tried restoring the system to factory state using AlienRespawn? This may be related to an update or driver.

 

Yea it could be a driver issue

Sent from my SGH-T999 using Tapatalk 2

 

I tried to install AlienRespawn and it wouldn't install, said I didn't have enough disk space. I couldn't choose where I wanted it to save to before it ran it's initial stuff, so it just errored out and uninstalled itself.

I've tried the newest WHQL drivers from Nvidia and the newest beta drivers and get the same results.

I'd rather not had to reimage it. It came with Win7 Home Premium and I wiped it and put Windows 8 on there that I had a copy of. That's a whole day process and there's about 200GB worth of downloaded Steam games on there as well already. Sigh.

 

what games you playing that are showing these high temps?

 

It may also be related to Windows 8. There are so many things that could be causing abnormal operation.

You should attempt to reinstall the latest NVIDIA driver, but this time, go into the installation menu and click "Custom" installation as opposed to "Express" installation and do a "clean install." This will completely remove the driver first, and may prompt a restart before re-installation.

 

For problems with nvidia drivers, I recommend to use ddu uninnstaller. Do not use nvidia. Restart and go to safe mode, let ddu do the job. ddu restart to Windows again . Install nvidia with custom adjustments (clean install)


Display Driver Uninstaller Download version 12.6.2

 

yeah, only a few months ago i've been using the custom option due to the fact i don't have a 3d monitor/display and hence don't need the driver for it.

 

From what I see the NV card may be suffering from a combination of power throttle and heat. From what I understand the power throttle will not occur on temp alone but is some sort of algorithm using power draw in the equation?

Anyhoo, I'd say the same as others, if you are OK with doing it go for a repaste. Also make sure that the original pads are all there. I have seen two machines with pads missing from the factory! I know the 780m runs hotter than mine for a lot of people but I would not expect much higher than 70c on stock settings. Right now I run a 10% OC and it never gets hotter than 65c

If you want dell to fix this then they will want to replace the heatsink completely (assuming they think it may be heat) as it comes with new pads/paste already applied. You will need to be running the original OS and the latest dell NV driver when you call, so the re-install is going to show if you haver a software issue anyway.

The P0 in Windows happens for me if I have NV power management set to 'maximum preferred'. Setting it back to adaptive does not work until after a reboot.

Good luck

 

I am going to try a fresh driver uninstall and clean install first.

Would a repaste void warranty? I wouldn't mind doing it, I am an "IT Professional" (lol, god i used that term) so I know what I am doing, it's not a biggie, I just don't want to void any warranty since this unit is only a week old to me (manufactured March 11th 2014 though)

I'll post results after redoing the driver.

 

Nah, a re-paste will not void anything. Just take your time and do so properly.

 

The AW guys are quite clear on dismantling and replacing parts - not an issue as long as you don't break it in the process, and any original parts are replaced before calling for support. First time I have seen a user manual that show you how to tear it down!

Yeah, I call myself the same, that cringworthy 'catch-all'!

I've been purchasing hundreds of dell laptops over the last 10 years and the on-site is golden. Dealing with phone techs was easy. I didn't expect the same level of service on a consumer product and have been pleasantly surprised. Dell did let me down big style in the server room though, nearly lost my job on that one and I replaced the Dell kit with HP...

Good luck with it.

 

Ok, just did a driver clean reinstall. Went with the WHQL 335.23 drivers. Same temps. Tried it with multiple games, including Diablo 3, Titanfall, Civ5, WildStar, Minecraft. All of them would crank the GPU to temps of between 85C and 93C, even with one spike at 95C.

It's also idling at desktop in the mid-50's.

 

Looks like the heatsink is your next target. That idle is the same as mine but I have the 3D model so no iGPU for low power draw. Sure sounds like an overheating issue to me.

After that I would just call it in with such a new machine ....

 

You could just need a re-paste. I wouldn't go as far as replacing heat sinks just yet.

 

oops, that is what I meant to say

 

If the OP is not comfortable doing the repaste then maybe they should take it to a shop that does custom PC's to do it for you as they have more experience.

Sent from my One using Tapatalk

 

LOL, I am fine with doing it myself. I've actually done this sort of thing thousands of times before, I've been in IT for a long time, so it's a no-brainer doing it. I just didn't want to void my warranty or anything. It also kinda ticks me off that a brand new laptop needs a repaste, but that's a different discussion.

 

you should be fine then. if you see your max temps drop into the 60s-70s during gameplay (don't bother with furmark) then I think you're fine, you can breathe a sigh of relief. if the temps are still the same then at least it can be eliminated as a main cause (though it might have contributed initially).

 

when you guys are repasting the gpu, what are you doing with the vrms? There covered with the gray thick thermal tape, are you guys taking it off and repasting these too or just leaving it be? Need to make sure the heatsink is making contact , the tolerances are pretty big.

 

Yeah, mine had some sort of thermal putty strip all along one edge - well the replacement I got a few weeks ago did. It was missing from the original! I only noticed the VRM's getting into the high 90s when HWInfo started to report the VRM's as well as core temps

It was around 3mm~4mm thick and you can double up pads. Luckily they are at the edge and you can see the contact , but it should be reusable anyway...

 

The first AW 17 I got, I didn't repaste and temps were pretty high. The 2nd one (in sig) I did a repaste on CPU and GPU and my temps are crazy low. Def worth it.

 

I ordered some Arctic Silver 5 and Cleaner. My supply apparently is empty. LOL, should be here Tuesday.

 

So yeah, it needed a re-pasting. So much so, that the small thermal compound from the factory MISSED THE GPU!!! LOL! I'm surprised this thing worked at all...

You can see the thermal compound square UNDER the thermal pad on memory above. It wasn't even ON the die.


The underside of the heatsink:


I was dumbfounded when I pulled the heatsink off. Someone needs to calibrate a machine at the factory.. lol

But, IT WORKS now, put a fresh coat of Arctic Silver 5 on it and it's purring at about 55C under full load now. I did the CPU while I was at it and that dropped a good 5-8C as well. Runs like a champ now!

 

glad to hear it all worked out....now you can chuck in the unlocked vbios LOL

maybe I should get someone to do the same cos 55 deg C under load sounds pretty good.

 

I figured as much. Glad it's resolved.

 

yeah, we were right on the money this time