Genuinely forces you to contemplate... hmmmm
>>55533405
Genuinely forces me to contemplate why they compared it to a 950 instead of a 960 and how the hell they got a 950 to use 140w.
Oh wait. This is the same company that claimed this.
I still don't understand how this company has such a blind, rabid, slavering fanmob to defend their lies.
>>55533405
Love getting my money's worth when I spend a$440 on a graphics card
>>55533405
Genuinely forces me to contemplate if they are using the lowest power usage reading rather than an average or if they specifically chose "med present" so as not use its much higher average of 160+w.
>>55533405
Genuinely forces me to contemplate how anyone with half a brain can be so forgiving of AMD and yet so critical of Nvidia for lying.
>>55533405
Hello sirs this is an spomsored AMD forum. Please take discussion to /v/.
>>55533551
Welcome to 4chan. You might want to talk to your supervisor on how to post here as you are clearly new.
>>55533422
Its Polaris 11
and those figures are total system power consumption
tech illiterate shitposter
>>55533527
Its average, they demoed it live. This is from back in January. Read the Anandtech write up.
>>55533682
>Its Polaris 11
Testing done in 12/2015 and it is not Polaris 10, which they are fixing to launch early in the next year. No. It's Polaris 11 which they haven't even gotten mocked up hardware for yet.
You sir, are the perfect example of an AMDrone making up any excuse to forgive AMD lies.
>>55533724
>has no idea what hes talking about
They showed off the die. It was Polaris 11, a 50w~ design that addresses entry level and mobile.
Tech illiterate sub human children need not post here.
>>55533405
DELETE THIS
>>55533405
install gentoo
>>55533753
>They showed off the die.
When did they show off the die? Months AFTER this interior lab testing?
>has no idea what he's talking about
Oh the irony!
>>55533682
>This is from back in January
>AMD Internal Lab testing as of Dec 2, 2015
You are claiming that they are even lying on when AMD did the testing?
AMD lying about everything
>>55533405
>>55533422
>>55533724
You guys do know they were talking about the what ever chis is used in the RX460?
>>55534042
chip*
>>55534042
>rx460
[citation needed]
I find it hard to believe they are advertising about any chip or gpu other than the one they launch after they present such an advertisement.
The demos were clearly faked. It was funny to see AMD marketing and shills meme about perf/watt too, especially since poolaris is equivalent perf/watt to maxwell despite poolaris being on the 14nm process and nearly half the size of gm204.
The same meme hype train will start up soon for vega if the rumor about the release being moved up to october is true.
>>55534069
going to have to look for the demo bench, but this was polaris 11 they were showing off, either the lowest end gpu they are making or a laptop part.
>>55534069
this ad is from december when amd announced poolaris, they explicitly said that this fake demo was polaris 11 (which is only being released as the 460 in dgpu form and the other 4 variants of poo 11 are being used for igpus)
>>55534101
>going to have to look for the demo bench
Don't let me keep you from providing a link. Or are you expecting someone else to prove your claim for you?
Calm your tits it's for the 460
>>55533814
They showed the die to people who actually matters.
>>55534069
>I find it hard to believe...
No one gives a fuck what you believe. The power draw proves it.
>>55533405
shows off a p11 slide
>>55533422
talks about p10
just another day on 4chan
>>55534150
>shows off a p11 slide
Still waiting for a citation when all I get is repetition.
>>55533682
>and those figures are total system power consumption
Yea I doubt it. 86W is way below a decent gaming PC. That's the kind of power consumption you got at idle with the thing running.
Interesting how one of the AMD's supposed demo cards had DVI despite the final product not having a DVI connector.
Even more interesting is that several non-reference 290x and 390x cards had similar configuration: http://core0.staticworld.net/images/article/2015/07/strix-r9-390x-100596078-orig.png
>>55534250
fury non-x also had the same configuration of ports. wonder why AMD would lie about something like this? what was there to gain by being deceitful?
>>55534299
did they ever said it will come with dvi?
>>55533422
>>55533405
>>55533459
holy delusional
There is a live demo of this using a Kill-A-Watt wall draw device telling you total system draw you fucking moron. The P10 variant with nearly twice the SPs naturally consumes more (~240W total system draw). The Polaris 10 cards weren't aimed towards this "medium settings 1080p60" area whatsoever.
https://www.youtube.com/watch?v=H9ZIcztgcNY
It's rather clear that AMD had the hardware way before the launch date, they likely had the silicon Q2 last year.
>>55534250
The card has a DVI header that board partners can attach DVI out to if they choose to drop the same PCB.
>>55534316
So the 950 was the total system usage not just the video card or the "architecture" as the ad misrepresents?
>>55534250
Hilarious how a budget card like the 480 doesn't even come with DVI, which is what most older monitors will use.
>>55534250
Interesting how one of the Nvidia's supposed demo cards had wood screws despite the final product not having any wood screws.
I was very upset about this in shop class.
>>55534316
>The card has a DVI header that board partners can attach DVI out to if they choose to drop the same PCB.
the point wasn't to complain about there being no dvi connector, but to show that the demo card was clearly not a genuine polaris 10 reference card. by the time they showed that polaris 10 demo (which was in april or may iirc) the PCB would have already been finalized and in production.
>>55534344
kek
>>55534344
>nvidia being deceitful and lying to customers excuses amd being deceitful and lying to customers
that's some great whataboutism there pajeet
>>55534316
>The Polaris 10 cards weren't aimed towards this "medium settings 1080p60" area whatsoever.
The RX 480 is a P10 card, yes? It is not aimed at the 1080p60 are whatsoever? Really? What is it aimed then?
>>55534378
Both company lies, but AMD never had fucking wood screws.
>>55534398
They actually did. The fury x2 that was originally shown when AMD announced the fury x and nano was a non-functional mockup. The final product was completely different.
>>55534398
AMD just has cards that kill motherboards, no biggie
>>55534420
Thats normal to have changes. Thats why its not final. Also what about when Nvida saw cut a PCB o a demo card? You can't find stuff like this with AMD.
>>55534444
Find me a video like this from AMD
https://www.youtube.com/watch?v=sRo-1VFMcbc
>>55534444
Find me a video where they cook an egg on a AMD
https://www.youtube.com/watch?v=ASu3Xw6JM1w
>>55534448
>Thats normal to have changes. Thats why its not final.
lol, not on the level the fury x2 went through. they changed the layout of the board completely and replaced many of the power delivery components with different ones, most likely because they were either glued on fakes or just weren't sufficient for the design in the first place. http://images.anandtech.com/doci/10279/RadeonProDuoBareD_StraightOn_4c_5inch.jpg
> Also what about when Nvida saw cut a PCB o a demo card? You can't find stuff like this with AMD.
more whataboutism, as i said, nvidia being deceitful doesn't excuse amd doing the same thing. you're also using one example from 6 years ago, amd was doing this shit as recently as 2 months ago with fake furyx2 pcb and the fake polaris 10 demo card.
>>55534459
>>55534471
samefag AMDummy getting so mad :^)
Sorry you buy inferior AMD products.
>>55534483
>most likely
Oh so you just making shit up?
>6 years ago
Same mother fuck runs the company now. They just got ass blasted so much they try not to do it again but guess what? 3.5 was not 6 years ago
Later mother fuckers!!!
>>55534483
yes not the fact that lisa stated that cooler master couldnt find a way to cool the vrms on the place they had them...
obviously you dont know nothing...
when ferrari asked for the evolution of 430 the 458 the initial draws was almost the same as the 430 because of the air ducts and cooling system... they had to literally redesing the whole car in order to fit the final design into the chasis
>>55534459
>>55534471
>unironically memeing about thermi in 2016
you realize that the 290/290x/390/390x all drew more power than the 480 and 580, right?
https://www.techpowerup.com/reviews/NVIDIA/GeForce_GTX_480_Fermi/30.html
https://www.techpowerup.com/reviews/AMD/R9_290X/25.html
kek, 290x draws more across the board
>>55534503
and remember the "N" stands for quality!
>>55534335
The "ad" or the Radeon update is somewhat misleading in that it is extremely opague; but it becomes elucidated by the live demo.
>>55534363
The PCB also had DVI pinout. AMD may have decided to change the board cooling solution and connectors last minute. Regardless, from looking at that picture, no one can know if that was a genuine P10 card. IIRC it was leaked by a journalist during a demo without context (I believe it was running Hitman).
>>55534387
It's aimed at entry level VR, as stated on slides. The 470 (also P10) is intended to max out games at 1080, and the 460 (P11) is designed for "esports" and 1080p gaming.
>>55534534
we call it thermi because it holds the record for being a fucking fusion reactor passing 100c at ease
>>55533405
I couldn't give less of an iota of a single shit how many Watts my _desktop_ GPU is drawing.
![polaris-5[1].jpg](https://i.imgur.com/evdMq4G.jpg "polaris-5[1].jpg")
What did he mean by this?
>>55534603
>it becomes elucidated by the live demo.
You mean the youtube link where they people making the video had no information other than what AMD gave them, including not knowing what version of polaris was being demo'd?
>>55534603
>It's aimed at entry level VR, as stated on slides.
Perhaps I am remembering this wrong but I thought the 480, at launch, was advertised as a low cost 1080@60fps card or "designed for premium gaming beyond HD?"
>>55534101
>Laptop part
>86W
Yea no. That's a whole laptop worth of power on its own.
>>55534618
perf/watt is a big deal when you need to pack 8 gpus each into a few dozen machines for gpu mining or other gpgpu compute stuff. it's part of the reason why nvidia is so dominant in that market, amd has been an entire generation behind them in perf/watt for a few years now.
>>55534626
>up to
>all this damage control in light of the vulkan Doom update
>>55534069
i honestly find it hard to believe anyone could believe that was the 480
>destroys the 970
why would they be comparing it to the 950??
>>55534626
he was talking about 470 and 460 (470 being 2.8x over the previous gen of amd)
>>55534669
And that is the crux of how AMD blinds it fanmob. AMD is going to win because Vulkan has "potential" to replace DirectX. Devs have the "potential" to use async for multiple GPUs. The 480 has the "potential" to better than its Nvidia competition. So they'll just believe the speculation based upon AMD's potential and disregard any inconvenient evidence.
>Wait for Vega even though Polaris is turning out to be turd
>DX12 will favor AMD even though half of the current games on DX12 perform better on Nvidia cards.
And then there is the perennial
>They are using beta drivers. AMD cards will be better with new drivers by which time new cards are launched.
>>55534709
>why would they be comparing it to the 950??
That's one of the questions. Another would be why would they be advertising for a gpu that is not the one coming out subsequent to the advertisement?
>>55534712
>AMD is going to win because Vulkan has "potential" to replace DirectX
DX12 or Vulkan doesn't matter, AMD are the only ones with GPUS that support modern APIs
>>55534712
>And that is the crux of how AMD blinds it fanmob. AMD is going to win because Vulkan has "potential" to replace DirectX. Devs have the "potential" to use async for multiple GPUs. The 480 has the "potential" to better than its Nvidia competition. So they'll just believe the speculation based upon AMD's potential and disregard any inconvenient evidence.
it's funny how AMD and it's fanboys still can't learn this. devs will not do extra work to extract a 10% performance gain from hardware that 1% of it's customers use. the same thing happened with the terascale architectures, bulldozer and GCN. it won't change with the poolaris GCN rehash. AMD will have to do the heavy lifting themselves if they want their cards to sell.
>>55534637
A lack of information does not equate to them lying to you. The demo was the only official piece of information at the time, which was released in January during the RTG Summit. AMD was simply exhibiting that they had a technology that focused on power efficiency.
>>55534637
Raja mentioned the 480 as a 1440p capable card (benchmarks show that the 480 is at it's limits at that resolution, memory bus simply isn't capable), but the 470 is mentioned as the 1080p60 card (further that it would be able to max out games). The 480 is repeatedly hailed as an entry point to VR.
>>55534643
The reason nVidia is winning in GPGPU is not because of perf/watt; that factor is rather insignificant because compute kernels run faster per watt on GCN from design since GCN 1.0, it's just an extremely ALU dense design. GPGPU is extremely popular on nVidia because they pioneered high level GPGPU APIs through CUDA, and is, at least at the moment, far more mature than OpenCL, bearing association to the majority of GPU compute libraries.
>>55534725
since their target audience isn't 12 year olds incapable of comprehending something that is more complex than a 12 times table it doesn't matter if retarded children like you don't understand.
>>55534741
>A lack of information does not equate to them lying to you.
Misrepresentation is not just as deceitful as lying?
>Raja mentioned the 480 as a 1440p capable card
Capable=designed as its primary use? Thank you for proving that misrepresentation is just as deceitful as lying.
>>55534709
because it was the polaris 11 show not the polaris 10...
https://www.youtube.com/watch?v=hKEcLxqofIQ
>>55533405
>86 Watts
What's that power usage from the PCI-e lane alone?
>>55534764
no shit, i was pointing out the nvidiot's stupidity
>>55534745
Thank you for the classic example of an ad hominem argument.
Or perhaps you can explain what they meant in way that even a 12 year old can understand though I expect you'll stay to such offensive and vague defending of obvious bullshit from AMD.
>>55534785
get back to hanging off nvidia's 3.5mm long dick
>>55534779
Allegedly that's what the system draws from the wall outlet.
>>55534780
sorry anon i have lost quite a lot of iq these days trying to argue with them
>>55534794
>offensive and vague defending of obvious bullshit from AMD
Yeah, that's what I thought.
>>55534795
https://www.youtube.com/watch?v=oNA6fll2DDQ
"allegedly"
no, that's exactly what it was.
>>55534741
>The reason nVidia is winning in GPGPU is not because of perf/watt; that factor is rather insignificant because compute kernels run faster per watt on GCN from design since GCN 1.0,
you're confusing perf/watt with raw performance. GCN has better performance clock for clock with an equivalent NVIDIA card in these applications, but it draws 2-3x the power to achieve that.
>>55534816
Did AMD provide the meters or were they verified to be functioning as they should be by independent reviewers?
>>55534814
>obvious bullshit
[citation needed]
>>55534824
that would be such an nvidia thing to do
>>55534832
Because AMD is above lying or misrepresentation? See pic in >>55533422 for an example.
>>55534534
Its not about how much power it draws, its about how HOT it got.
>>55533438
>being Australian
>>55533438
You never get your money's worth on AMD.
>>55534837
maybe throwing in 8gbps ram as opposed to something more standard was a last minute thing?
all these conclusions you're jumping to and all this shitposting you're doing, i think you deserve a raise from nvidia. they probably don't care about you enough for that though.
>>55534839
who cares about how hot a card gets, especially a reference card?
if you're not setting up a custom fan profile for any gpu you buy then you're doing it wrong. more often than not they are misconfigured out of the box and don't even ramp up past 20% until the card is in failure territory.
>>55534860
>who cares about how hot a card gets
nvidiots, everyone. justifying their housefires as usual.
>>55534860
I don't know, why don't you go ask all the rabid fanboys who think that 80c is hot for the RX480? or 90c for the 290?
I think you're forgetting that Fermi would easily reach 105 - 110c
>>55534859
>maybe throwing in 8gbps ram as opposed to something more standard was a last minute thing?
Even so, it was deceitful to claim it was "unprecedented" even if it was accidental. That's the same thing we put the 970's 3.5gbs to the coals for.
>>55534879
>Fermis
>tfw
>>55534879
if you let your GPU reach 80c, 90c or 105c in the first place then you're a mouthbreathing retard (which is why there are so many people complaining about high temps on AMD cards: stupid people naturally buy defective rebranded products)
any non-braindead person knows they should be setting up a fan profile with afterburner or speedfan to maintain temps under 70 or 60c.
>>55533405
>86w
Off the PCIE?
>>55534927
that's the entire system, the gpu is 50w tdp
>>55534758
>Misrepresentation is not just as deceitful as lying?
A lack of information does not equate misrepresentation.
>Capable=designed as its primary use? Thank you for proving that misrepresentation is just as deceitful as lying.
??? Yes? It was designed for "premium VR" according to slides. Of course this is very opague, and a lot of people overanalyze it to extrapolate performance because their lives revolve around computer hardware for whatever reason, but the common folk will have no such gripe.
>>55534818
I think that's rather tenuous. cgminer MH/s runs much, much faster per watt (450MH/s peak on my 7950 stock whereas CUDA mining on a 680 rakes in only 100MH/s). Of course, that depends on the task at hand, but any task that can be run concurrently at load with many threads will massively favor the design of the SIMD-V pipelines in GCN.
>>55534933
>A lack of information does not equate misrepresentation.
Presenting information in a manner that would be easily misunderstood by the common audience is misrepresentation.
Didn't we have a recent thread about the 1060's comparison with the 480 starting at .8 that was loudly proclaimed by AMD fanboys as similarly misrepresentative even if there was not a lack of information?
>>55534956
>Didn't we have a recent thread about the 1060's comparison with the 480 *using a bar graph starting at .8 that was loudly proclaimed by AMD fanboys as similarly misrepresentative even if there was not a lack of information?
Fixed.
>>55534339
Dual-Link DVI is what you need for 120 or 144Hz, Pajeet.
>>55534956
Okay, but neither was misrepresentation. In one case was a distorted bar graph that had (supposedly) correct information on it. The AMD demo was that simply implies that AMD had a card that decreased system power usage. Please don't associate me with an "AMD fanboy".
>>55535037
>distorted - to change (something) so that it is no longer true or accurate
>not misrepresentation - to describe ([ . . .] something) in a false way especially in order to deceive someone : to give someone a false idea about (something [. . .])
Yeah, I'm done wasting words on someone determined to re-write the English language whenever it the common usage of a word is inconvenient.
>>55534669
Curry nigger from "windowz tech support."
>>55533812
this
>>55534927
kek
>>55534733
GCN is used on all the consoles, Nvidia has a minority of the whole gaming market
Also, Vulkan is supported on pretty much every platform, specially on the mobile market where optimization matters
>>55534872
You realize that AMD cards run way hotter than Nvidia cards
It's pretty sad when the 480 is both slower than the 1070 and runs hotter and uses more power.
>>55534688
So much this, shilling was already in control again and now we got a shitton of threads again
It was never this awful
>>55535388
>mobile market
>console market
Pick one.
>support
>use
Pick one.
>>55534924
Fermi couldn't get down to 90 at full load even with the fans at 100%
>>55535388
consoles have their own custom OS, graphics API and other special configuration. most games still aren't parallelized despite the xbone and ps4 having two 4-core CPUs clocked at 1.8ghz either.
>>55535414
>moving the goalposts
Nice try Rasheed, but you're not getting paid for this thread
>>55535432
>Fermi couldn't get down to 90 at full load even with the fans at 100%
[citation needed]
thermi was a housefire but it wasn't 90c at 100% fan speed, even a shitty blower cooler can dissipate enough heat to prevent that (otherwise every reference 290x would have died within 5 minutes of use, since they drew more power, and therefore had to dissipate more heat as a result)
>>55535438
The PS4 runs pretty much the same hardware as the XBOne, optimization for both consoles is pretty similar
Also, the XBOne runs DX12
>>55535481
The GTX480 drew more power at stock than the 290X
>>55535484
>Also, the XBOne runs DX12
it doesn't, dx12 was not a thing when design for the xbone started. if there's dx12 support at all it'd be from a shitty wrapper built over the low level graphics api that it's os exposes.
>>55535501
>The GTX480 drew more power at stock than the 290X
wrong.
https://www.techpowerup.com/reviews/NVIDIA/GeForce_GTX_480_Fermi/30.html
https://www.techpowerup.com/reviews/AMD/R9_290X/25.html
>>55535481
my experience with gtx 470 with that stock blower cooler is different, it runs over 90c with 100% fan speed when under heavy load
>>55535447
>>moving the goalposts
Moving the goalposts? Do you even understand what that means? First you made a point about consoles but instead of elaborating you went onto a different point about mobiles, ie. completely irrelevant or "a different goalpost."
Then you argued about "support" as if it was synonymous with actual use. Very few engines actually use Vulkan (only one as far as I am aware and that is only in one game) even though console support it. Everything else is using DirectX.
You are the perfect example of the type of "blind AMD fanmob" I was describing earlier, ie. the goalpost.
>>55535620
both consoles and adreno is amd tech....
quallcomm still pays royalties to amd for this so no its not moving goalposts amd has literally their tech on EVERYTHING
>>55535662
>adreno is amd tech....
not anymore, adreno was still a VLIW architecture when qualcomm bought it. they pretty much redesigned the entire thing into a modern unified simd architecture, completely different from the original 2005-era architecture they purchased.
>quallcomm still pays royalties to amd
no they don't, qualcomm purchased amd's entire mobile gpu division. the IP is wholly theirs.
>>55535390
>AMD cards run way hotter than Nvidia cards
Except that's not true in majority of cases. The housefire meme started right in Nvidia's backyard. Those cucks are trying to turn it back at AMD but that's not working.
>and runs hotter
Ok you might want to show some actual bench comparisons.
>uses more power.
Again where are those benches.
> is both slower than t
Costs 2 times less though.
>>55535685
>Except that's not true in majority of cases.
more power drawn = more heat to dissipate
>>55535662
>both consoles and adreno is amd tech
Straw man? No claim was raised that they weren't.
>>55535722
oh wow what a nice conclusion you made there. Sadly it doesn't pertain to what I'm asking.
>>55535740
are you retarded? the card is going to be hotter if it's drawing more power. this is basic physics.
>>55535760
Not the same anon but how does the 1070 (150w) and the 480 (150w) compare in producing heat?
>>55535775
they're roughly the same in power draw so they produce the same heat, the 1070 just performs much better.
Do I buy my cousin's used 970 (still in fine condition) for $200 or buy a new Radeon instead?
>>55534669
Is he a poo?
>>55535790
>they're roughly the same in power draw so they produce the same heat
Really now?
1070
>Idle 30
>Load 76
480
>Idle 35
>Load 83
http://www.guru3d.com/articles_pages/nvidia_geforce_gtx_1070_review,9.html
http://www.guru3d.com/articles_pages/amd_radeon_r9_rx_480_8gb_review,6.html
Perf/watt are not equal regardless of architecture. As we learned with Fermi, leakage affects heat as well.
>>55535824
Funny how the 1070 is way faster than the 480 but uses way less power.
The 480 tech is just garbage compared to Nvidia's.
>>55535824
are you really trying to compare temperatures when the cooler and fan profile is different?
>Perf/watt are not equal regardless of architecture. As we learned with Fermi, leakage affects heat as well.
https://en.wikipedia.org/wiki/Conservation_of_energy
any energy you use comes back out as heat, you can't magically create extra energy from thin air like you're trying to imply. the only difference to make when all things are equal is how well the cooler dissipates heat.
>>55535685
It's true though. AMD's cards have been hotter and less power efficient than Nvidia's for like 3 generations now.
>>55535884
>are you really trying to compare temperatures when the cooler and fan profile is different?
Yes. It proves the point that heat is not equal based simply on being used at the same draw.
>any energy you use comes back out as heat
At the same rate at the same draw? Does a 16nm node produce the same heat as a 28nm node at the same draw?
You have way, way, WAY oversimplified things and seem to be determined to present yourself like a complete idiot on the subject to avoid admitting to a simple error.
>>55533405
>even amd uses intel cpus for internal testing
>w-wait for zen
>>55535915
>Yes. It proves the point that heat is not equal based simply on being used at the same draw.
which is the result of the 480 having a weaker cooler, not some magical property of amd's architecture or process.
the real question to ask here is how AMD could have possibly made such a terrible architecture that it's nearly 1/2 the performance of an equivalent nvidia chip despite equivalent power consumption and process.
>>55535958
>which is the result of the 480 having a weaker cooler, not some magical property of amd's architecture or process.
No claim was made that there was any magical property only that it was DIFFERENT which is contrary to the idiot claim:
>they're roughly the same in power draw so they produce the same heat
Different cards, obviously, will produce different levels of heat even on the same power draw.
>>55535973
>Different cards, obviously, will produce different levels of heat even on the same power draw.
so now you're saying AMD's design violates basic rules of physics and that you can get more energy out than you put in. kek.
>>55536002
>so now you're saying AMD's design violates basic rules of physics
Let's get this straight. We have two cards using the same amount of energy. One is a 28nm node card and the other is a 16nm node card. But if one produces more heat at the same 150w it somehow "violates basic rules of physics?"
>>55536025
yes. heat is energy. if one card emits more heat than another despite equivalent power draw and no other factors accounted for the difference (like the different cooler) then you would have what is essentially a perpetual motion machine. you would be able to generate more electricity from the heat it put out than the electricity you used to power the device.
>>55536025
Are you truly this retarded?
>>55536048
>yes.
Because you don't get less resistance meaning less energy is wasted in heat with a smaller node?
>The benefits of shrinking
As a high-level abstraction, you can pretend that each layer has zero thickness, including the insulative layers in between the layers that you use to draw your circuit. However, reality is not so kind. Layers have non-zero thickness, and their thickness doesn't scale to perfectly match when you change your value of lambda.
This means that if you move to a smaller value of lambda (a smaller process node), interesting things can happen. For instance, suppose you have a capacitor made up of two stacked squares of metal. The capacitance (analogous to resistance) of that capacitor is directly proportional to the area of the plates, and inversely proportional to the distance between them (and also directly proportional to a material proprty that we will treat as constant for simplicity).
Nor are there concerns with leakage wasting energy as heat regardless of node size?
>The tradeoffs of shrinking
>All is not sunshine and roses, though. Smaller transistors have higher leakage currents, and they can create new headaches.
>Leakage current is often compared to a leaking faucet: no matter how hard you try to close the valve, it is never quite completely closed. So it leaks a little bit. All else being equal, smaller transistors have to contend with higher leakage currents. There are all sorts of tweaks that can be done to the materials to combat this, but those can only do so much.
>Leakage current translates into static power consumption, the power that a chip consumes even when it is sitting and doing nothing.
>>55536100
See pic
>>55536158
>Because you don't get less resistance meaning less energy is wasted in heat with a smaller node?
you still don't seem to understand this. how much energy you 'waste' does not matter. ALL of the energy has to come back out as heat. energy does not cease to exist just because a GPU used it to calculate 2+2 for each pixel on your screen.
>>55536193
>how much energy you 'waste' does not matter.
Because that waste does not take the form of heat? What form does that wasted energy take?
>>55536216
all of the energy the card uses comes back out as heat regardless of how what % of that you waste with bad design. are you a burgerfat or something? you can't really be stupid enough to think the 'not wasted' energy just disappears into the aether right?
>>55536245
What is heat? A measurable difference in temperature?
>The transfer of energy from one body to another as a result of a difference in temperature or a change in phase.
If we have energy converting from one body to another with a different resulting temperature are they both the same amount of heat? Are you trying to define heat without regards to temperature?
![perfwatt_3840_2160[1].png](https://i.imgur.com/eG0i25J.png "perfwatt_3840_2160[1].png")
>>55533405
lol?
>>55535775
>>55535760
if the heat isn't being dissipiated it's going to stay hot.
also, please stop using the stock cooler 480 as an example, it's a piece of shit
>>55536158
Power = heat
There's no other place the heat comes from
>>55536296
>if the heat isn't being dissipiated it's going to stay hot.
No claim was made to the otherwise. The question was:
>If we have energy converting from one body to another with a different resulting temperature are they both the same amount of heat?
Is that supposed to be interpreted as a no?
>>55536301
>Power = heat
Straw man. No claim was made that it wasn't. Only that the RATES of heat are different. Now would you like to answer the question or should I assume from your failure to do so that you are conceding the point?
>>55536316
>Only that the RATES of heat *production are different.
Just in case you get confused or forgot that the idiotic claim was:
>roughly the same power draw produce the same heat
>>55536316
spare me the sententious talk. it's picayune and unctuous
>>55536316
>Only that the RATES of heat are different.
Rates of heat? What does that mean?
Heat is directly proportional to power, there is nowhere else the heat is coming from.
The only thing that changes is the rate of heat dissipation, i.e. how fast a cooler moves the heat away from the chip.
The amount of total heat generated is always equal to the power usage, no matter what cooler you use.
>>55536390
>sententious
It took you that long to look up a thesaurus? Pathetic.
>>55536391
See >>55536381
I knew, fucking KNEW, you were going to be a pedant.
>>55536398
hey buddy, go rope yourself :^)
>>55536398
It's ok, you can admit you have no idea what you're talking about.
>>55536417
>>55536414
One of you losers asked "Rates of heat? What does that mean?" But it seems neither of you want to discuss rate of heat production as if you know it is a losing argument.
>>55536398
>needing a thesaurus
>this nigger actually believes that "sententious" is such an obscure word
what a faggot you are
>>55536429
The rate of heat production is equal to the power usage, end of story
>>55536451
>The rate of heat production is equal to the power usage
[citation needed]
I'm pretty certain two circuits with the same power draw but different resistances will produce heat at different rates. Unless you can provide evidence to the otherwise? No? I thought not.
>>55536264
>>55536316
What? I assume it's clear that thermodynamically clear that what comes out of wall has to be converted into heat if we assume that the work done is negligible (fans). The heat transfer argument is largely irrelevant because the size, material, and environments of the cards are very similar, even in 28nm designs, so the heat transfer compared to each other is also very similar. Some one screencap that idiot for retard hall of fame.
>>55535799
Honestly, you either wait for 1060 or buy the 480 imo
>>55536481
>I assume it's clear that thermodynamically clear that what comes out of wall has to be converted into heat if we assume that the work done is negligible (fans)
It is. See the second part of >>55536316
>Power = Heat
Stop bringing up irrelevant and already noted as irrelevant arguments.
> The heat transfer argument is largely irrelevant because the size, material, and environments of the cards are very similar, even in 28nm designs, so the heat transfer compared to each other is also very similar.
Except one, we are not comparing two 28nm designs. The claim was far more general. Are you trying to move the goalposts now?
And even if I do humor you and allow you to move the goalposts that still doesn't change simple laws of physics. You too may see >>55536472 if you want to argue two circuits with different resistances and the same "power draw" would produce the same rate of heat.
>>55535287
Why so mad bro? Did you get a witches kiss this morning?
>>55533724
>>55533941
>>55534208
http://www.anandtech.com/show/9886/amd-reveals-polaris-gpu-architecture
Tech illiterate retard BTFO
>>55534483
>more whataboutism
Literal /pol/ anti-russia shills in /g/
>>55536472
The heat production is always the same. You are talking about where the heat goes and the rate of dissipation.
Looks like the 480 power usage is looking even worse now with the 1060 specs released.
>>55534843
How post the latest patch
Nvidiots buttmad
