![1464993327669[1].jpg](https://i.imgur.com/EMWWXvJ.jpg "1464993327669[1].jpg")
Where were you when Nvidia decided Fermi needed a second chance?
![xcobpc[1].png](https://i.imgur.com/gLZc5Fw.png "xcobpc[1].png")
I was at home, laughing
>>54912972
It's time to bring all the old Fermi memes, m8s.
>>54912985
>full size gtx 1080
>in a laptop
Kekeronie
I don't get the whole fermi ordeal.
I had 570 sli and temps were around 75*c...
The only thing screaming was my 700watt psu
![FermiTitle[3].jpg](https://i.imgur.com/XpLbDsw.jpg "FermiTitle[3].jpg")
Indeed it did Nvidia, indeed it did.
Are people seriously too stupid to realize that reference coolers always run hot with the shitty blower fans that are loud as fuck? This is seriously nothing new yet it's being meme'd up everywhere
Is this your first gpu release or something? Wait for 3rd party cards with better coolers if you want lower temps you fuckheads
Thanks to advances brought by Nvidia's latest Pascal architecture, we'll be able to cook faster than ever.
Thanks, Nvidia!
>>54913057
>5XX
They fixed it a bit, the real housefire was the 480
>>54913100
It's a 180W card, it should not be throttling or running hotter than reference maxwell.
>>54913240
>running hotter than reference maxwell
It's not. And since there were very few reference Maxwell cards, this wasn't seen as and issue because the 3rd party cards came to market first.
And the 3rd party cards with decent coolers will allow them to run much cooler without throttling
Just like literally every other release
>>54913240
Power draw has nothing to do with temperatures, density does.
Once we reach 16B+ transistors chips these blower coolers won't be enough anymore.
>>54913284
Typically the lower drive current per transistor is enough to offset the prospective increase in thermal density. Pascal's clocks counteract all process gains.
Tom'shardware doing high resolution oscilloscope measurements is really invaluable to the review process. Seeing 16nm Pascal draw power like, and even worse in some cases, than 28nm Maxwell was surprising to say the least. Nvidia is a pretty scumfuck company, but they have had a decent track record for the most part in managing power with their last couple architectures.
http://www.tomshardware.com/reviews/nvidia-geforce-gtx-1080-pascal,4572-10.html
>>54913284
1. Bigger fins and heatpipes (4 slot coolers?)
2. Watercooling/Phase change cooling
Otherwise enjoy your housefire
>16 posts, 8 posters
You're not even trying with the bait threads anymore
>>54913021
they already did it with the 980
>The 180W boundary is never crossed without overclocking.
Pascal is super power efficient as Toms says, Nvidia has always been more power efficient than GCN HOUSEFIRES
http://www.hardocp.com/article/2016/05/27/from_ati_to_amd_back_journey_in_futility/
>In the simplest terms AMD has created a product that runs hotter and slower than its competition's new architecture by a potentially significant margin.
AYYMDPOORFAGS are lying pieces of shit as usual and mad after Kyle exposed their Poolaris HOUSEFIRES running inferior Samshit 14nm process on inferior GloFo fabs
AYYMD, UNCOMPETITIVE & IRRELEVANT
>>54913515
This
>>54913515
Completely agree.
>>54913515
Dat Dank Shitpost tho
>>54913515
here's your reply
>>54913515
Here's some more (You)'s
You strike me as a guy in desperate need of them.
>>54913007
>>54913186
what is this wood screw meme?
>>54913570
>>54913584
>>54913603
>BUTTMAD AYYMPOORFAGS DETECTED, UNABLE TO COUNTER FACTS WITH THEIR RETARDED LIES, SUPPORTING A DYING, IRRELEVANT COMPANY THAT IS JEALOUS OF OTHER SUCCESSFUL COMPANIES
>>54913636
I'm so sorry I never intended to make you cry.
>>54913655
The only that is crying is you as your shitty company goes bankrupt and you'll have no drivers or hardware support to go to in the future
AYYMDPOORFAGS, unable to think, unable to comprehend simple things
>>54913675
It's all right, let it out, no one will judge you.
>>54913515
>written before the event
>by a site butthurt for not being invited
>pure speculation written like it's true
And then they wonder why they stop getting free samples and invitations...
>>54913685
Well, it's pretty obvious you have no facts to counter real facts, so keep on being a loser in life with no income and no money to afford anything
>>54913707
Take as long as you need, we're all here for you.
>>54913735
Buttmad? Good, let the butthurt flow through you
If it's not crashing its not running too hot.
>>54913751
You sure have a lot pent up, this might take a while.
>>54913773
>The 180W boundary is never crossed without overclocking.
Awww, there there, you were proven wrong and still too much pride to admit you are wrong and have a small dick & impotent, unable to satisfy any pussy
>>54913796
Whew, sudden topic change.
What are you talking about now?
>>54913515
>hardcp being relevant
>>54913814
Hey, it's a small dick & impotent person
>>54913829
Just checked, everything seems to be in place for the time being.
But seriously, what were you talking about?
>>54913850
Small dick & impotent person, what are you talking about?
>>54913515
>I get my reviews on hardware from cyclists who are so irrelevant they aren't invited.
>>54913021
gaming laptops are the ultimate meme
>>54913876
I was planning on asking you what were you talking about concerning the 180W thing, but you seem pretty interested in my genitals for whatever reason.
Wanna talk on /gif/?
>>54913904
Small dick & impotent person, there's no need to talk about that anymore, it's on the Toms page
You were proven WRONG, but yeah, pride goes before a fall
>>54913614
Literally the first result on google. http://semiaccurate.com/2009/10/01/nvidia-fakes-fermi-boards-gtc/
>>54913926
You were so adamant on talking about it, why the sudden change of heart? But since you seem to like my dick that much, who am I to halt your interests.
I love the samefaf replying consistently to his own posts after 90 seconds really highlights how bad this thread is. it started off so good too :(
>>54913614
>being this new
>>54913987
The thread was always garbage, senpai.
>>54913987
It's just one Nvidia fan destroying a AMDcuck with cold facts and logic.
Pretty funny to see
It's just not Kyle who thinks AMD has produced a shitty inefficient hot chip, many professionals and experts thinks so too as well.
>Nvidiots
>Buying the GTX1070 for 699 USD
Not even surprised.
Fuck off AMDrone shill poorfags
The 1080 is the most powerful card ever made, and it would be a mistake to purposely not buy it
>inb4 muh price
>Inb4 powerdraw
It doesn't fucking matter
>>54914260
Lol beat by two 200 dollar cards.
>>54914265
Um, it's not?
>>54914260
>1080 is the most powerful card ever made
Yet the 980ti still hold all the records
>>54914271
>implying
>>54913100
Hello, Reddit!
>>54914280
Well gee, probably because the card only came out a few days ago and custom boards aren't even out yet.
>>54913176
I remember when Hardocp did their 480 review and had a video showing noise and temps on 480 SLI, two wedged together without space in between
Fucking shit got so loud and hot he shut the bench down prematurely
>>54913404
Fury X showed that closed loop gpu cooling should definitely be a thing for flagships
It has no downside except finding a place for the radiator
Temps and noise are good as fuck
>>54914281
>2x Radeon Rx 480
>not $200
>in shitty CF
>barely edging out a single card in a game AMD has historically performed better in
>AMD fanboys desperately clinging to a shitty game no one would have cared about if not for DX12
That's all fine and dandy and you go ahead and play the only game where AMD wins, and I'll play the other 99.99% of games where the 1080 performs better at with less issues than CF has
>>54913007
I never thought this day would come. I really miss the constant bickering between Fermi causing house fires and when ATi drivers really were shit back then.
>>54914371
Why are you so upset that you need to justify your purchase and meaningless brand loyalty in this thread?
>>54914260
> buying a $700 or $800 card that will have all the value preservation staying power of the GTX 980, thanks to 1080 Ti and Vega coming out in 6-8 months
shiggy
>>54914416
I don't have brand loyalty, I just want the best
If Vega comes out and beats the 1080 and 1080ti, I'll gladly switch to it
The other anon said that a $200 card beats the 1080, when it doesn't, and two barely do it in an AMD optimized game
>>54913176
I have the 480 in my system right now. With water cooling, that thing sits at 65°C under load. But the GPU can easily handle 100°C, so whatever
>>54914518
>I have the 480 in my system right now
Prove it and time stamp it
>>54914420
hey anon, can you explain the difference between the 1080 and the upcoming 1080ti? isn't it just an incremental improvement?
>>54914550
1080 Ti will actually be based on GP104 instead of the cut down GP100.
>paying $600+ for a mid-range chip
>>54914578
what's the difference? isn't the 1080 already ridiculously powerful, hot as it gets? maybe it will be mid-range in a year but right now?
I ask all of this naively. I'm trying to build my first gaming PC after 8 years of being out of the PC market and, well, it's daunting to say the least
>>54914229
They don't call it the Fanboi Edition for nothing.
>>54913377
>up to 311w power peak power consumption in game
>in game
>not furmark
How far behind the 980 ti is it?
>>54914870
>peak
Means literally nothing. Average is what counts.
>>54914759
1080 is only 180-200W and uses 256b GDDR5X for 320 GB/s bandwidth.
Vega (and GP102??) is expected to get around 1000 GB/s from a 4096b 2GHz HBM2 memory subsystem and will be maybe 500+ mm^2 dies instead of the ~300 mm^2 for the GTX 1080.
This will probably require water cooling but at least be a small card like Fury X again FWIW.
>>54914870
Notice how tight the average and peak power are during Furmark? Nvidia's firmware is causing entirely different behavior during this bench than in any other workload. The degree of variability while gaming is equal to the 28nm Maxwell cards despite 16nm Pascall having lower working voltages.
Nvidia has pulled a Samsung/Qualcomm and is writing firmware to recognize benches and alter their behavior accordingly.
>>54914895
Only the average matters for accumulative power consumption, peak draw, and how frequently the GPU hits peak draw matters for analyzing the architecture and power delivery on the board.
>>54914895
high peak draw actually means 2 less obvious things:
> slightly reduced lifespan of VRMs
> voltage sag and potential subsequent unreliability of other components sharing system power
It's better to throw bigger caps on a GPU and be more rigorous about transient peak draws.
>>54914998
Sounds like somethin that custom boards will fix ehen they're released. Refrence design calls for only like 5 phases in power delivery.
>>54914998
Peaks in line power like that aren't always caused by a lack luster power delivery on the board. If you actually match up power over time with the bench as its running you can see a picture of how the GPU is operating internally. Firmware with anything that has a modern implementation of a turbo core is going to try and keep a chip around a certain power target, but periodic peaks where the limit is breached still happen. This is true of all GPUs and CPUs. Intel and AMD both give themselves gratuitous margins of overhead on their mobile chips for this very reason.
Nvidia knew that Pascal would break their power target in a terrible way during a heavy bench like Furmark so they had to have written a protocol specifically for it.
>>54914949
ok, that sounds sick.
but what I'm hearing is that, since I'm fortunate enough to be able to do so, I could buy the 1080 and evaluate the 1080ti in a year and possibly just sell the then-old 1080 for a moderate loss.
so I don't think I will wait, but that's only because I know I can afford to not wait. nonetheless, thank you for informing me! :)
>>54912972
Uh, all nvidia cards do ~80C. That's literally what the thermal limit is set at. The fan will maintain whatever speed required and clock will be at whatever level required to keep that specific temperature.
The only way to get lower temp is to either manually set the thermal limit or to use a third party card with a different thermal limit.
>tfw my ati 5850 is idling at 84 degrees celsius
>>54916764
Dust it and change the thermal goo
>>54916764
Clean your fucking fans then.
>>54916830
>>54916838
It was already idling at 75 degrees when it was brand new, though.
>>54916853
So does it get up to like 110C while under load or something?
>>54916875
Yeah, about there.
>>54916908
So upgrade, you've gotten plenty of life out of it
>>54916853
>>54916908
Damn son, that's retardedly hot, is it a reference card?
>>54916924
One day, man, one day.
>>54916991
Fuck knows.
