NVIDIA Readies GeForce RTX 4070 SUPER, RTX 4070 Ti SUPER, and RTX 4080 SUPER

[lexluthermiester]

For anything faster than RTX 4050 (better known as RTX 4060), using 1080p chart is ill-advised. 2160p chart would be feasible for 4070 Ti and higher; 1440p chart, for everything below them.

Spoiler: As you can see, the difference at 4K is way bigger

View attachment 318028

View attachment 318023View attachment 318024View attachment 318025View attachment 318026
I'd not expect anything better than plus 8 percent bang per buck on average compared to non-Super semi-predecessor. Don't forget that Leather Jacket Guy got a wee greedier than he used to be. Most likely discounted "old" Ada GPUs will be much more feasible, especially outside the States.


"Just buy a 4090."



P.S. If for some reason the miracle happens and they release some 4070 Ti Refresh/Super with 16 GB VRAM without that 12HPWR nonsense and the price won't be too sky high I'll go ahead and grab it. Never had any RTX series GPU in my personal disposal, and I'm still with my wish to make my RX 6700 XT a backup GPU.

Believe what you will. Your graphs do not match current gen cards. We're talking about the RTX 4000 series.

 

[Beginner Macro Device]

Your graphs do not match current gen cards.

Implying they could.

The point is the only generation with "Super" SKUs had no "Super" SKU faster than their non-Super predecessor by more than 15 percent. We're currently into 2023 and even 5 percent FPS/$ gain is enough to make "Super" Ada SKUs more feasible than their RDNA2/3 competition all around. 8 percent would be the ultimate hard counter. There is no point for Huang to do anything better than that.

 

[PapaTaipei]

So it's all speculation and guesswork..

I'll be interested when specs and pricing are confirmed, I can speculate all I like right now about what configurations and pricing could be, likely fairly accurately too, but it wouldn't be news.

It's still interesting for a lot of buyers. Including me. There are good reasons to belive a refresh is on the way, and with such good news we can wait a couple of months to make a choice on a 1K euros ballpark GPU. Btw ppl whining about too many SKUs dont make any sens, you want to sell as many chips as possible, even the ones with defective SMs that's totally normal.

 

[Dr. Dro]

This rumour is no surprise given this brand does have a history of launching too many SKUs every generation and screwing over the customers with more you buy more you save philosophy.

No one's being screwed over, at least not any more than AMD is screwing customers with that 6750 GRE of theirs, that come in both RX 6700-equivalent and RX 6700 XT equivalents at roughly half their MSRP... Now that's some good screwing they're doing to their value minded customers... if only these weren't 3 year old cards already.

Ada's been around for a year, if a refresh isn't released within the next 3 to 6 months, it may be a generation that won't receive one altogether. I fully expect Blackwell 50 series to ship within 1 year to 18 months from now.

 

[R0H1T]

Who didn't see this coming, unless you're probably blind as bat

Also not sure if anyone brought this up, but did JHH have this in mind when of thought of RTX ~ Resiniferatoxin

 

[PapaTaipei]

Who didn't see this coming, unless you're probably blind as bat

Also not sure if anyone brought this up, but did JHH have this in mind when of thought of RTX ~ Resiniferatoxin

Wow, I just read that wiki page. Very interesting.

 

[cvaldes]

It's important to stress that SUPER is a marketing term not an engineering one. In no way does SUPER equal a certain percentage boost. However Nvidia marketing used the SUPER descriptor on a previous generation doesn't require them to use it the same way again. As most of us are well aware, Nvidia doesn't maintain any sort of consistency between generations.

Of course, GPUs are measured and quantified by a bunch of numerical metrics (cores, frequencies, etc.) so Nvidia really can't just slap the SUPER moniker on the exact same part. About the only given is that the SUPER SKU will slightly outperform the normal model and that it will be more expensive.

That's pretty much all anyone sane can expect. LOL, maybe the only honest thing the SUPER adjective is good for is the price.

 

[Prima.Vera]

So 4070ti Super vs 4080 ?
Is this a joke?

 

[lexluthermiester]

The point is the only generation with "Super" SKUs had no "Super" SKU faster than their non-Super predecessor by more than 15 percent.

Taking your suggestion, if we look at the 2000 standard to the 2000 "super" models then to the "ti" models, the jump in performance is about 15% each jump. But that was the 2000 series. We're two generations ahead now. The "ti" models of the 4000 series are a consistent 20% faster than the non-ti models. Logically, we should reasonably expect that these new "super" models should be a similar jump of 20%, even though they did the "super" and "ti" naming convention backwards from the 2000 series.

So 4070ti Super vs 4080 ?

No. It'll be the 4070Super not the 4070ti Super.

Is this a joke?

NVidia is just selling off binned dies that don't meet the required specs for a certain model range, but can fit somewhere inbetween. So for example:

They have dies that didn't qualify as a full 4090, so they're going to bin them as a 4080Super.
They have dies that didn't qualify as a full 4080, but are better than a 4070ti, so they're making a 4070Super.
They have dies that didn't qualify as a full 4070, but they're better than a 4060ti, so they're making a 4060Super.
There is even more fine-grain examples than just these as well.

This is a binning practice we've all seen for decades. It's not new and if we understand what is really going on, not at all surprising. It's just standard business. Nature of the machine, as it were.

 

[N/A]

There isn't going to be a 4080 based on 102. Potentially a 4070 Super with 8704 cuda and 4080 super full AD103.
the binned 102 are in RTX 5000 Ada and those are not necessarily a bad batch of dies just laser cut.

 

[Cheeseball]

Pretty sure if ever there was a RTX 4070 Super, it would definitely be in-between the 4070 and 4070 Ti as the Ti moniker is usually the top version (e.g. 2080 vs. 2080 Super vs. 2080 Ti). So almost guaranteed to be 192-bit again and possibly $650?

Perhaps this is for the mobile line up as there is a huge performance gap (like 40% to 50% faster) between the RTX 4070 and RTX 4080. Maybe this is where a 4070 Super would fit in.

 

[Camm]

Nvidia would be better served dropping prices on 4070/Ti and creating a SUPER tier above or between the Ti.

Trying to insert a Ti SUPER is likely just to do more brand damage going into the future.

 

[PapaTaipei]

There isn't going to be a 4080 based on 102. Potentially a 4070 Super with 8704 cuda and 4080 super full AD103.
the binned 102 are in RTX 5000 Ada and those are not necessarily a bad batch of dies just laser cut.

RTX 5000? Your comment make 0 sens.

Nvidia would be better served dropping prices on 4070/Ti and creating a SUPER tier above or between the Ti.

Trying to insert a Ti SUPER is likely just to do more brand damage going into the future.

I don't think they care about brand damage since they have a monopoly. Besides if people are ready to pay 1K for a GPU why would they lower their prices?

 

[gffermari]

People buy nVidia gpus no matter what. But the interest has declined since the prices have been stable for nearly a year.

The supers will bring the interest back.
It doesn’t matter if they are good or vfm.

 

[N/A]

RTX 5000? Your comment make 0 sens.

Quadro Ada generation GPU Variant AD102 / 12800 Shading Units

 

[lexluthermiester]

Quadro Ada generation GPU Variant AD102 / 12800 Shading Units

Those are not RTX 5000.

 

[ratirt]

How was it with previous gens. Is super better than Ti? 2000 series super stack somewhere between Ti and regular(2080 Ti and 2080) right?
All in all it all comes down to price.

 

[Assimilator]

RTX 5000? Your comment make 0 sens.

Do research before being rude.

NVIDIA RTX 5000 Ada Generation Specs

NVIDIA AD102, 2550 MHz, 12800 Cores, 400 TMUs, 176 ROPs, 32768 MB GDDR6, 2250 MHz, 256 bit

www.techpowerup.com

 

[PapaTaipei]

Quadro Ada generation GPU Variant AD102 / 12800 Shading Units

My bad, sorry.

 

[ratirt]

They have dies that didn't qualify as a full 4080, but are better than a 4070ti, so they're making a 4070Super.

I thought Super has lower performance than a Ti version.
The Ti Super is a new creation though.

 

[lexluthermiester]

How was it with previous gens. Is super better than Ti? 2000 series super stack somewhere between Ti and regular(2080 Ti and 2080) right?
All in all it all comes down to price.

With the Geforce 2000 series it went 20X0, 20X0Super, 20X0ti. It would seem they're doing things in a different order this time.

Do research before being rude.

NVIDIA RTX 5000 Ada Generation Specs

NVIDIA AD102, 2550 MHz, 12800 Cores, 400 TMUs, 176 ROPs, 32768 MB GDDR6, 2250 MHz, 256 bit

www.techpowerup.com

Well crap. I thought they were naming the workstation card something else(along the line of A5XXX). This begs the question: What are the next gen Geforce RTX card going to be named?

I thought Super has lower performance than a Ti version.
The Ti Super is a new creation though.

They did that with the RTX 2000 series, but as they have already done the "ti" variants, it's only logical for the "super" card to be up next. Of course, with NVidia's janky-asre naming scheme, who knows at all...

 

[Chrispy_]

The vanilla 4060 is the model that needs the most love from a "Super" refresh, but it's already using the full AD107 die so there's no room to grow. 24 SMs on a puny 128-bit bus is why last-gen mobile parts are doing better than the desktop 4060; Laptop 3060's get 30 SMs on a 192-bit bus, and they get 25W more power. The 4060 is a step backwards in all ways except energy efficiency, and Nvidia aren't even responsible for that bit.

This is the problem with Nvidia selling their lowest-tier part with a mid-tier name, they've f*cked the naming scheme for refreshes. Unless they use AD106, there's no way to stop the vanilla 4060 Super from being a waste of effort, as it'll just be a 5-10% clock speed bump at the cost of efficiency.

 

[John974]

Waiting for the ultra to come back, never understood the "super".

Back in the day you had the vanilla, ti and ultra.
Also liked the vanilla, gt and gtx.

4080 super ti ultra extreme might be new and great for marketing?

 

[Double-Click]

Super is just the new lingo they use when it's a refresh - ironically, not so super.

 

[stimpy88]

Super is just the new lingo they use when it's a refresh - ironically, not so super.

It's to con certain customers who think they will be ahead of the curve with all these UE5 and high RT games. UE5=RTX5, the quiet part nVidia is certainly never going to say out loud.