• Welcome to TechPowerUp Forums, Guest! Please check out our forum guidelines for info related to our community.

NVIDIA GeForce RTX 2080 Ti Captured in Its Sleek, Green, Metal Glory

Behold the sleekness

60d32f2d4c93e405ab3dd21b17ffc8224bae5ede60d0508bb94328548ec93620.jpg
 
TDP is 250W, same as Pascal.
265 to 280 ish, tdp means shit, it will use slightly more power than pascal when OCd, titan v had a 250 watt tdp but you could make a titan v draw heaps more than 250 watts while ocd
 
Could very well be the case, base clocks are extremely conservative. 754nm^2 is simply massive and binning is very difficult, clocks generally go down with die space even if everything else remains unchanged.

The fact that the 2080ti has the same TDP as the 1080ti can only signify that on average it runs at a much lower clock speed. Or perhaps, this time around Nvidia will take some "liberty" with their TDP ratings, they did massively improve cooling after all.

12nm vs 16nm should mean something, and some optimizations, they know
 
and some optimizations, they know

Hardly any, it's the same node with a higher reticle limit allowing larger dies. And I suspect that doesn't come for free, power and clocks are likely to be affected to a small degree.
 
12nm vs 16nm should mean something, and some optimizations, they know

Maxwell was faster and more efficient than Kepler on the same 28nm node, but it should come as know surprise who is spinning to the negative here.
 
Maxwell was faster and more efficient than Kepler on the same 28nm node

Maxwell wasn't 750mm^2 with double the transistor density, a small detail that got lost along the way, I presume.
 
Last edited:
"Sleek" is synonymous with "glossy" and "graceful lines." The card is really no more glossy than your average graphics card and there's absolutely nothing graceful about the design that's fundamentally a box with some bolted on bits to obfuscate that fact.

Yes, dual fan is a step up from the previous blower design but the HSF is still wholly unremarkable compared to AIB alternatives.
From someone who's worked in the design industry for many years I would disagree with your analysis, the card has a sleek and stylish appeal.
 
Maxwell wasn't 750mm^2, a small detail that got lost along the way, I presume.

Not to mention dumping of its FP64 units.

I believe it was mentioned by one of the tech sites that Turing is doing the same. Dumping FP64 units to make room for RT cores.

Turing is Volta (exchanged FP64 units for RT cores)
 
Maxwell wasn't 750mm^2, a small detail that got lost along the way, I presume.

Of course not, my Maxwell is still a pretty large 601mm2, especially when on 28nm. Doesn't change the fact that Maxwell was better on the same node.

Not to mention dumping of its FP64 units.

I believe it was mentioned by one of the tech sites that Turing is doing the same. Dumping FP64 units to make room for RT cores.

Turing is Volta (exchanged FP64 units for RT cores)

Volta is packed full of DP goodness and is an even bigger chip, yet still using less power than 1080 Ti:

https://www.hardocp.com/article/2018/03/20/nvidia_titan_v_video_card_gaming_review/18

"Following similar trends from Pascal, NVIDIA’s TITAN V based on the Volta architecture seems to be very power efficient. It has the same TDP of 250W as the GeForce GTX 1080 Ti FE video card. In our system full peak gaming Wattage hit 406W in Kingdom Come: Deliverance. This is less than the peak of 414W for the GeForce GTX 1080 Ti FE video card. Less power, yet 30-40% more performance! "
 
Last edited:
From someone who's worked in the design industry for many years I would disagree with your analysis, the card has a sleek and stylish appeal.
Let me guess, you worked on the Fiat Multipla?
 
this marketing crap is getting really boring. "sleek" is in the eye of the beholder, and personally i don't give a rats ass about the looks of a card as long as it does its job of rendering images fast enough. this topic looks like the fast and the furious franchise a bit more each day. i'll give it to nvidia tho, they sure know marketing and the sheep they have as followers.
 
Let me guess, you worked on the Fiat Multipla?
Even that is sleeker than this HSF. :laugh: Ugly AF but definitely sleek:
Fiat_Multipla_silver_front.JPG


Older Silverado/Sierra is more like this HSF:
1992-gmc-sierra-64k-57-l-350-chevy-v8-like-chevy-silverado-1.JPG

A box...on wheels. No graceful lines whatsoever. And where the paint fell off that old pickup, that's the plastic bits on the new HSF. Considering you're paying $200 premium for that HSF, one would think they would get a lot more value back for it. You know, like water cooled or something. But no, that's the "Founders Edition" tax. Feed NGREEDIA lest it starve. What a shame that would be.
 
Last edited:
265 to 280 ish, tdp means shit, it will use slightly more power than pascal when OCd, titan v had a 250 watt tdp but you could make a titan v draw heaps more than 250 watts while ocd
Keyword is "while OCd". You do know manufacturers spec is only to spec, right?
Nvidia cards in the last 4 generations or so have been very well regulated when it comes to TDP. There is a pretty hard and well-working limit on the power consumption exactly where the TDP is.
 
Not to mention dumping of its FP64 units.

I believe it was mentioned by one of the tech sites that Turing is doing the same. Dumping FP64 units to make room for RT cores.

Turing is Volta (exchanged FP64 units for RT cores)


FP64 excuse :rolleyes:
 
lol at people inventing more creative excuses to complain about the design, the design is fine,it's sleek,and I know what that word means. save your energy and go complain about the price of the fe version,not how it looks. If that was the design of the new radeon you'd say it's best ever :laugh:
 
Ha, comparing quite possibly the worst looking car design ever to the new Founders Edition card is laughable. Nvidias FE cards have adorned a classic design for the last three generations, something I can appreciate that isn't going to appeal to all, especially the ones who are anti Nvidia.
 
It's a new design language for sure... And honestly, I can see the appeal. It's all bulky, like a piece of industrial machinery, or perhaps a tank. Makes it look muscular.

But sleek? No way. The previous founders edition cards were sleek. Ford's right. It's a box with some details. It looks like the ass end of a rocker ship, sure. But that doesn't make it sleek.
 
eypw5ts1tih11.png

well I think this is self explanatory and matches what we have heard before. I do not care about total price increase (like 700? 1000$ who cares its +/-300$ for few year spending on something that I probably will use every day),
but I have never been so disappointed with next gen gpus, no doubt RTX will be x6 times faster than GTX in Raytracing ONLY - and that is and will be a non-factor for next few years, and if nvidia could fix all this with drivers optimisation - they would have done it before presentation and show that off as much as they could - instead we got "it just works" and "JiggaRayz" for like x30 times.
the question now is - to buy a used Pascal now or wait a bit? because mining is unprofitable for only 2nd week in a row now (am I right?) - I mean all those miners will have to dump those Pascals (and flush the price regardless the RTX) now or later?
Turing has a completely different SM structure from Pascal.
Estimating performance across generations based on FPU performance is dangerous, just look at the past:
maxwell_gflops.png
 
It's a new design language for sure... And honestly, I can see the appeal. It's all bulky, like a piece of industrial machinery, or perhaps a tank. Makes it look muscular.

But sleek? No way. The previous founders edition cards were sleek. Ford's right. It's a box with some details. It looks like the ass end of a rocker ship, sure. But that doesn't make it sleek.
You see a box, I see style. It's subjective. Thankfully for you there'll be gaudy aftermarket RGB goodness that awaits.
 
Turing has a completely different SM structure from Pascal.
Estimating performance across generations based on FPU performance is dangerous, just look at the past:
View attachment 105808
Titan and 780Ti spent a lot more time on their base clocks than 970/980 ;)
On Max Boost clocks they do 4.1 and 5.1 TFLOPS.
Also, Maxwells had more ROPs which had become a problem for larger Keplers.
 
You see a box, I see style. It's subjective. Thankfully for you there'll be gaudy aftermarket RGB goodness that awaits.

I literally said that I could see the appeal... I never once said that I do not see style. In fact, I complemented the style as muscular. I said it specifically is not a "sleek" style.

But hey, continue to read what you want and ignore the point. :toast:
 
I literally said that I could see the appeal... I never once said that I do not see style. In fact, I complemented the style as muscular. I said it specifically is not a "sleek" style.

But hey, continue to read what you want and ignore the point. :toast:
I guess it depends on your perception of sleek. To me sleek is smooth and contoured and catches the light, the card has what looks like an aluminium outer shell which is molded round the edge of card. It also has a smooth texture which catches light, to me that seems a perfect word to describe the card.
 

i think it might be better to wait until the game is upgraded to allow for RTR to function ... which it isn't as yet. While we are at it, do ya think it might matter if the cards OC 30+% like 9xx or single digits like AMD ? No interest at all what reference cards to not overclocked.

Behold the sleekness

As opposed to the TPU tested 349 watt Vega 64 or 432 watt Fury X ?
 
Last edited:
i think it might be better to wait until the game is upgraded to allow for RTR to function ... which it isn't as yet. While we are at it, do ya think it might matter if the cards OC 30+% like 9xx or single digits like AMD ? No interest at all what reference cards to not overclocked.

Not just Shadows of The Tomb Raider

RPS: Metro Exodus targeting 60fps 1080p for new Nvidia RTX features

After speaking with Metro Exodus developers 4A Games, however, we might finally have some idea of what to expect when Nvidia’s pretty RTX settings are all switched on.

“We’re always going to be pushing 60[fps],” rendering programmer Ben Archard told me when I asked if 4A have any kind of performance targets when it came to implementing Nvidia’s uber lighting ray tracing tech. “But we’ll see what we get. Obviously, there are three cards there and we’ll see what profiles we can get for each.”
He also confirmed that 4A’s 60fps target was with a 1920×1080 resolution as well. “It’s 1080p, yes,” he said. “That’s the goal, but we’ll see how it goes.”
 
Back
Top