I had GPus running in 90s for years. my old 9800m GS lasted 5 years at 93 C temps. My previous HD5870m ran around 80s overclocked, and is still alive after 3+ years. My original 680m died within a week, at sub 90sC temps, but the replacement still runs, at around 83C temps when gaming.
GPUs are designed now to throttle when overheating to avoid serious damage. I think 680m and 780m throttle at 93 degrees C.
-
-
-
I too have been suffering from the same problem have this Alienware 17 and i play Watch Dogs no matter for how long i play for 20min or 5 hours at a stretch the laptop always stays at 94C never went about that i did contact Alienware they replaced my laptop and the problem still exist on my new one looks like these cards are meant to get this hot lets hope there is a fix for this the spec will be below if that helps.
- 4th Generation Intel® Core i7-4710MQ processor (6MB Cache, up to 3.5GHz w/ Intel® Turbo Boost)
- 16GB Dual Channel DDR3L at 1600MHz (4x4GB)
- NVIDIA® GeForce® GTX 880M graphics with 8GB total GDDR5
- 17.3 inch (439.42 mm) 120Hz WLED FHD (1920 x 1080) TrueLife 400 Nit Display w/3D Bundle
- 1TB HDD+256GB SSD Boot
- 5G WiFi Broadcom 4352 802.11n/ac and Bluetooth 4.0
- Slot-Loading Dual Layer Blu-ray Reader (BD-ROM, DVD±RW, CD-RW). -
Even when pushing the GPU to 99% on my end for half an hour or so, I don't go above 90. I mean, GPU utilization percentages are still a mess but at least the heat doesn't seem too bad. -
I wanted to do a comparison between R327.23 driver since there were reports of much better performance with this driver but not seeing any difference so far between that and our latest branch. I did a test run with 3DMark using modified 327.23 driver using the .inf file shared by Manu369 on our GeForce forums and R340 driver and essentially got the same score with both drivers. In testing games, Am I missing anything?
R327.23:
R340.52:
-
IanC and steviejones133 like this.
-
Just release the GTX 880MX/980M already and stop the nonsense with the Kepler GK104 that should have been replaced a long time ago.
GTX 880M, another card in Nvidia`s long record of milking the market while taking forever to push out new architectures.
March 2012 was when we saw the first Kepler cards. We are now soon in August 2014 ffs
Go tell Huang to bring out the big guns instead of this nonsense. Many people are waiting for the GM204, among them many GTX 680M owners that had no reason to "upgrade" to GTX 780M (only 1SMX more) or the GTX 880M (780M rebrand).
/rant -
nightdex likes this.
-
DumbDumb likes this.
-
I highly doubt thats gonna happen but if it did happen we would be lucky.
-
-
880M on stock vbios won't pass 87C in my machine with auto fans normally. Max temp with forced full fans doesn't even hit 80C. Stock vbios throttles though. Modded vbios will go clear up to 93C boosting the whole time at 993MHz while the stock vbios drops to 954MHz and in some cases all the way down to 888.
If you're seeing 90C with a stock vbios, I'd change your paste.
Pidge, I would definitely read the thread linked there, we have all documented our issues all over the place and I don't have my machine right now so I can't do anything with digging up all of my specific issues right now.
Sent from my HTC One_M8 using Tapatalk -
Pidge, could you also comment on this nasty flag which basically make the 880Ms run even worse than an iGPU? The fact that it's undocumented and users have to jump through hoops just to remove the flag (and even it's not known for certain what removes it) is completely unacceptable, and highly unethical.
TBoneSan likes this. -
Just a small update from me...
Recieved my Alienware 17 with a gtx 880m fitted....
Installed bf3 and made sure all updates where installed, launched msi afterburner. Initially gpu temp was 36c in idle from just powering it on...
Loaded up bf3 and within 10 min max it was at 93c and the it had gone from 993mhz to 421mhz!!!!
Then it would go back to 993mhz and back down to 421mhz every 15secs... So I turned it off and thought it would be the paste job from the factory, so I removed the heatsink and to be fair they had done a good job, so I re pasted it and reinstalled and exactly the same thing happened...
I don't think it's good for any gpu to run at such high temps, even if they can....
So I rang dell this morning and they are considering if they will swap out the 880m for a 780m... I thought I would never say that in my life,, DOWNGRADING.... But in this case I don't think I have a choice.
To say I'm disappointed is an understatement, I sold my aw17 with a 780m to get this, and the one I sold had a better CPU and 120hz screen... Massive kick in the balls for me...
I don't overclock or benchmark ever, just play games
Thanks Nvidia for milking your customers with the most inefficient gpu I have ever used..
Iankamlesh likes this. -
I bet you that they are waiting for AMD to put out the upcoming R9 M295X before even considering announcing the GTX 880MX/980M.
Same story with 7970M. Always speaking last. AMD may have been the reason why GTX 680M turned out the way it was.
Story time:
Anyone that remember this leak?
http://forum.notebookreview.com/gaming-software-graphics-cards/649052-hurray-nvidia-600-series-not-just-fermi-kepler-87.html#post8423165
Then 7970M came along with P6000 score,
Suddenly GTX 680M scored P6000+ too....
Wouldnt surprise at all that Nvidia is waiting to see how R9 M295X perform first, before deciding final specs for the GTX 880MX/980M.IanC likes this. -
IanC likes this.
-
I think there is a 125 000mW (125W) limit on the GTX 880M.
The card should have never been made. Nvidia already had the performance lead with GTX 780M. Increasing the heat and power draw from 100W (780M) was just silly.
But thats what you get from getting greedy to try to milk the market some more. Huang was still not satisfied with the 780M so he had to rename it and increase the clock further to try to fool customers with a shiny 800M name.
And it blew up on their face with this thread among many other. Unstable card probably due to high thermal strain on the silicon, fine tuned to the limit to allow these high clocks, but rarely works like it was intended. Maybe they even paid less attention to the binning, to try to sell out existing GK104 stock to prepare for GM204 launch. Who knows...
Sad to see the customers suffer. Its they who have to deal with this crapIanC likes this. -
780M TDP was 122W, 880M is definitely higher.
Well I don't feel as bad about my temperatures seeing 3 people including Pidge with AW17 and 94C
What a disaster!
Sent from my HTC One_M8 using TapatalkIanC likes this. -
I remember Meaker saying 780M was 110W, while 880M would have to run at 125W TDP in order to not power throttle. But then of course keeping temps under control with a 125W card would be difficult.
-
Eh I'm finding a mix... Tech power up has the 780M listed at 122W but Anandtech has it at 100W.
880M has to be over 120 either way.
Sent from my HTC One_M8 using Tapatalk -
That was kinda my point
n=1 is right. GTX 780M is at 100-110W. Its def not 122W.
GTX 880M is at 125W.
Abandon all hope. I wonder what is Nvidia`s moral limit on how hot chips they are prepared to sell in an already constrained environment aka notebooks. For the sake of more green dollars in their wealthy accounts. -
Is the 880M at 125W in the default vBIOS? From what it sounds like the default vBIOS still limits the 880M to 110W, hence we see all these crazy TDP throttling issues.
-
Ethram could maybe verify with his cards.
The theory involving throttling the card with lower power target than needed to keep the temperature at bay (like that worked lol) is also interesting... -
-
When my cards are given freedom to run at 993MHz, they hit those temps although it takes an hour or two.
Sent from my HTC One_M8 using Tapatalk -
-
They are all over the place once the temps hit 90+, prior to that its solid at 993mhz
Thanks
Ian -
Sent from my HTC One_M8 using Tapatalk -
Mines stock bios, I have never overclocked any of my cards...
I will check later, your probably right at first it's probably between 993mhz and 954mhz, once it hits the 90+ it drop is to 421mhz and then flicks between 954mhz and 421mhz. It will boost back up to max 993mhz when the temp drops below 90c
It's not what I expected, but my temps are crazy... -
Guys so i contacted Nvidea regarding the heating issue and this is what they said [ This is Shreyas, assisting you with the issue that you are experiencing.
I understand from your email that you are facing overheating issues with the card. I apologize for the inconvenience.
I am sorry to hear this, however we do not have any users/customer directly report the issue to NVIDIA. In forums I see many users have posted, however no user has posted it to us.
We only provide the basic chipset for the OEM's, the cards are then customized and then installed on laptops. Hence we do not have much control over the manufacturing and assembling the GPU's in the laptops.
BIOS Is not provided by NVIDIA it is provided only by Dell/alien ware. Hence you would need to directly contact them for the update.
To assist you better please update me with the following details:]
I thought that Nvidia is the one to be blamed but they are blaming the manufacturer in this case my case thats Alienware/Dell and i wonder why dint anyone reported Nvidia if you guys are suffering for these problem. -
There could be some times where the vbios is the one that is messing up the graphic card, but usually its the driver from Nvidia, or in very rare occasions, the chip itself. -
The OEMs distribute the vBIOS, but it is provided to the OEMs by NVIDIA. They won't give a vBIOS directly to consumers. The 780M and 880M vBIOS are more or less generic and can be used on more than one brand of laptop. Say, for example, if NVIDIA releases an updated 880M vBIOS that works better to Dell/Alienware and Clevo, but only one of them elects to distribute it this won't matter since Clevo owners will be able to use the updated vBIOS distributed by Alienware or vice versa.
-
You may be right that Nvidia create the vbios for the customers (OEMs) but surely ultimately its the OEMs that decide what Nvidia writes on the vbios since you can find the same GPU but with different clocks across different brands. GTX 680M was an example where Clevo used 720MHz while MSI used 771MHz.
That was MSI`s choice after some testing with the GT70, that the notebook could handle the extra heat from the extra clock. The same applies for GT cards that goes inside ultrabooks and more beefier designs. The vbios have to be tailored for the specific notebook.
780M and 880M are indeed very similar -
Oh my word. I've upgraded to Windows 8.1 and, while my FPS in DOTA 2 has shot up, the massive stuttering and awful GPU utilization issues (card jumps from 99% to 74% constantly for no reason) is back.
This is maddening. I fixed it once before but apparently the 8.1 upgrade knocked those settings back to being a problem.
EDIT: Upgraded to the latest beta and restarted PC. Problem solved. Still annoying that this is the second time it's happened. -
I feel sorry for all you guys having trouble with the 880M. Is the vBIOS the confirmed culprit? People who flash svl7's vBIOS mentioned an increase in performance, and greater stability. You guys should give that a try.
-
Unfortunately the modded vbios doesn't like to return to 2D clocks which means they are running at 1v and 993MHz even sitting idle at the desktop otherwise I'd run the mod 24/7
Sent from my HTC One_M8 using Tapatalk -
So, correct me if I'm wrong, but it's basically a large round-robin going on here? nVidia blames OEMs and OEMs wash their hands of it blaming nVidia? Nobody's at fault but the cards are broken somehow, and a large number of users (albeit small in comparison to the total number of nvidia GPU users on the market) are stuck with essentially fundamentally broken hardware, with broken performance at stock but overheated and permanently-boosted performance under modification?
Ethrem likes this. -
Unless Pidge manages to get something moving on it.
Sent from my HTC One_M8 using Tapatalk -
I think when 880MX on maxwell comes out, if it's a solid improvement over kepler's flagship (and ESPECIALLY if it uses 100W and not 120W) I'ma sell my 780Ms to those who are in need and try to get 880MX cards. Clearly, that's the best thing to do. Now, just gotta work the corner with some makeup for the money...
-
It'll take a class-action lawsuit for nVidia to get off their asses, and even then nVidia would most likely rather waste millions of dollars trying to win the legal battle instead of admitting any wrongdoing and setting things straight.
-
Yup that pretty much sounds about right. I seriously, SERIOUSLY, wish AMD would come and cut their asses in the GPU race. Let them actually do some bloody work for once. Mantle shook them into high gear with their drivers; letting Thief 2014 beat out mantle using DirectX... but since then AMD kind of shut up and stroked their R9 295X2 (which is really two R9 290x cards, for more money than buying them separate... exactly like the Titan Z) which isn't really doing anything that buying two R9 290X cards doesn't do... except cost the user $300-400 USD extra.
They've made a total of one improvement card over their 7970 flagship, while nVidia has made... like five. 770, 780, titan, titan black, 780Ti; when the original 680 could easily trade blows with a 7970. nVidia's getting sloppy; intel's gotten complacent and all but forgotten what computational improvements per architecture mean and AMD's just sitting there, waiting, trying to squeeze more power out of what they've already got, which is almost straight at its power limits (AMD FX-9570 for $900 anyone? How fast that dropped to $300...). I seriously hope they shape up, ship out some stuff that make intel/nvidia put away their e-peens and go "=O... we need to counter this!" then have a great battle of the titans over skylake and maxwell versus whatever-amd-is-bringing-I-do-not-know-their-codenames -
Well, so long as schmucks continue to buy nvidia enough that it continues to maintain like 4-5 times the market over AMD, even when AMD has the clearly superior product in cost/performance terms, if not straight performance AND straight cost, what do you expect?
Although I do admit maxwell is a sweet one, and I'm surprised nvidia released a pretty well optimized and uncut GM107. -
-
I see on this forum posts about dying 6990M, 7970M, 8970M all the time but it's extremely rare you see a dying nVidia card. And then you read those posts and these people are already on their second card since they got the machine!
My 260M cards were still quite functional but in need of a repaste and that M17xR1 got the crap beaten out of it since I got it in 2009. The reason I say were is because I broke one with a vbios flash a few months ago. AMD cards just don't have the reputation for life expectancy that nVidia cards do and there is really no excuse for it whatsoever.
Yeah 880M is a dud. NVidia released one dud in the last three flagship generations while AMD released 3 (probably 4 because M290X is a straight up 8970M rebrand).
Sent from my HTC One_M8 using Tapatalk -
Well I take that back about nVidia being reliable. Just my luck that the same day I post that my 880M slave card froze the system to the point it rebooted and now nVidia control panel doesn't show 2x anymore and Hwinfo sensors only show GPU memory allocated and a 2.5Gbps link speed for that card, nothing else. Wow...
Sent from my HTC One_M8 using Tapatalk -
LOL xD got warranty?
-
I just got this machine in like May... yeah its got a warranty... but I'm going to be without a machine for god knows how long now.
The card doesn't report its temp which is a red flag and GPU-Z shows revision FF which is a classic sign of a dead card. -
MickyD1234 Notebook Prophet
.
NV and AMD have historically aimed their products at different market segments. AMD was the choice of the enthusiast, someone that would work with the card and drivers to get the best - and way cheaper. NV on the other hand have been aiming at the 'noob' market. Less educated in the intricacies of a PC (and bang-for-buck on GPU's) and want plug-n-play. Hence GFExperience. Clearly aimed at the user that does not want to mess with driver settings.
Hopefully NV will get their eye back on-the-ball soon... -
Sure, they seem to be creating issues now but that's usually not the case. AMD burned me too many times and, until I start reading reviews that say Nvidia are behind in terms of performance/stability again, I won't switch. The 880m is an outlier but it still performs fine for me for the most part.
Although, that being said, my 880m is underclocking from 80-86 degrees (max temps so far). It drops from 993Mhz to 980 for prolonged periods, even when not at 99% usage. It's weird. I've also seen it drop to 960 at times. Again, not when being challenged either. This is Bioshock Infinite at 60FPS on mostly ultra which is hardly the most punishing test.transphasic likes this. -
980MHz is not underclocking... That's still boosting. My slave card drops below stock clocks... down to 888MHz at times... That is underclocking. Boost isn't guaranteed and you're still boosting at 980MHz, that's extremely impressive.
So... my card came back to life. I've never seen anything like it. Filed the RMA with Xotic, was browsing the web and suddenly the pop up came up that my system was SLI capable, I enabled it, rebooted, fired up Bioshock Infinite and left it sitting for 15 minutes while I went to go do other things, came back and the master was 79C @ 954MHz, the slave was at 82C @ 901.5MHz but it was running. Positively baffling.
I posted about it here
http://forum.notebookreview.com/sag...ager-np9377-owners-lounge-33.html#post9732395
GPU-Z went from showing FF and I couldn't even get the core temp of that card with HWiNFO to now it shows everything its supposed to and works again. Took apart the machine and looked for shorts, couldn't find one. I'm totally at a loss as to what happened but I'm just going to say at this point its safe to assume its on its way to failing and just go through with the RMA.
Pidge from Nvidia has asked that user experiencing problems with the 880m list them here..
Discussion in 'Gaming (Software and Graphics Cards)' started by DumbDumb, Jul 16, 2014.