Tuesday, October 17th 2023
NVIDIA Readies GeForce RTX 4070 SUPER, RTX 4070 Ti SUPER, and RTX 4080 SUPER
NVIDIA is rumored to be working on a refresh of the higher end of its GeForce RTX 40-series "Ada" series, according to hongxing2020, a reliable source with NVIDIA leaks. The company could be bringing back the SUPER brand extension that it introduced with the RTX 20-series. As many as three SKUs are on the radar—GeForce RTX 4080 SUPER, GeForce RTX 4070 Ti SUPER, and the GeForce RTX 4070 SUPER.
There is no word on when the company plans to release these, or what their specifications are, but we can certainly speculate. The current RTX 4080, while based on the AD103 silicon, doesn't max it out—it uses 76 out of 80 SM (streaming multiprocessors) available on the silicon, but we doubt if those extra 4 SM could drive up enough performance to make a whole new SKU, especially given that the 256-bit memory bus of the AD103 is maxed out. We predict that the RTX 4080 SUPER could be based on the larger AD102 silicon that physically has 144 SM that the current RTX 4090 uses 128 out of. NVIDIA has the opportunity to pick an SM count such as, say, 96. AD102 also has a wider 384-bit memory bus, giving NVIDIA the option of either giving the RTX 4080 SUPER the same 24 GB memory configuration as the RTX 4090, or even 20 GB, across a 320-bit memory bus.
As for the RTX 4070 Ti SUPER (an oddly named SKU that combines both the Ti and SUPER brand extensions), considering that the current RTX 4070 Ti maxes out the AD104 silicon (60 SM), it stands to reason that this new SKU could be based on AD103, with a lower SM count than the 76 of the RTX 4080. It remains to be seen if NVIDIA utilizes the full 256-bit memory bus. As for the RTX 4070 SUPER, there exists a rather wide gap between the current RTX 4070 Ti and the current RTX 4070. While the RTX 4070 Ti uses all 60 SM on the AD104, the RTX 4070 only uses 46. This gives NVIDIA the opportunity to get closer to the RTX 4070 Ti in SM counts, and bolster against AMD's Radeon RX 7800 XT.
Sources:
hongxing2020 (Twitter), VideoCardz
There is no word on when the company plans to release these, or what their specifications are, but we can certainly speculate. The current RTX 4080, while based on the AD103 silicon, doesn't max it out—it uses 76 out of 80 SM (streaming multiprocessors) available on the silicon, but we doubt if those extra 4 SM could drive up enough performance to make a whole new SKU, especially given that the 256-bit memory bus of the AD103 is maxed out. We predict that the RTX 4080 SUPER could be based on the larger AD102 silicon that physically has 144 SM that the current RTX 4090 uses 128 out of. NVIDIA has the opportunity to pick an SM count such as, say, 96. AD102 also has a wider 384-bit memory bus, giving NVIDIA the option of either giving the RTX 4080 SUPER the same 24 GB memory configuration as the RTX 4090, or even 20 GB, across a 320-bit memory bus.
76 Comments on NVIDIA Readies GeForce RTX 4070 SUPER, RTX 4070 Ti SUPER, and RTX 4080 SUPER
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.
They can reserve a small number of GPUs for the new SKUs to keep consumer mindshare fresh.
As for pricing, these SKUs will slot in somewhere between existing products. Nvidia's product labeling has been wonky at times but one might guess: Normal < SUPER < Ti (price lowest to highest). But a 4060 Ti shouldn't be more expensive than a vanilla 4070. If they release a 4060 SUPER, it might be between the 4060 and the 4060 Ti.
The Ada Lovelace generation is a mess. Nvidia has sort of painted themselves into a corner concerning VRAM size, VRAM speed, power consumption, bus width, memory bandwidth, and price.
It would be naïve to expect major performance improvements based on the current product stack and established pricing. One thing you can count on: there will be no bargains. The most realistic expectation would be a few more cores enabled with a marked up price.
Or maybe yes it will use a AD102 but heavily cut down to 86 SM and 256bit.
The existence of the 4080 Super could serve a purpose of clearing faulty AD102 dies and,
create a chance for officially lowering the 4080 MSRP so to clear all the dead stocks of 4080 which no one wants to buy (very poor value).
On the other side,
The current 'Upsell the 4090' practice works extremely well and The leather jacket won't let anything happen to it.
So the 4080 Super maybe just 10% better than the 4080 at best, so it will never threatens the 4090.
All the reports that gaming cards are flying off the shelves because people are repurposing them for AI servers proved unfounded, and there's no stampede for expensive cards to run Stable Diffusion and similar home AI applications. But I still wouldn't put it past Nvidia to (artificially) create a shortage of gaming cards - focus the manufacturing on AI accelerators, and then raising the prices of gaming GPUs unreasonalbly to make up for the loss of sales numbers. And explain proudly that the reason is AI explosion, new focus of Nvidia they are wholly embracing (screw the gamers).
You can see the space between 4060, 4060 Ti, 4070, 4070 Ti, 4080 varies a little bit but not much.
A +15% boost seems far fetched. A more realistic figure would be between +7% and +10%. From a performance standpoint, unless your eyes are glued on the FPS counter on the monitor, it's barely enough for the average person to tell. +7% is going from 59 fps to 63 fps. And it's not like your eyes are going to say "hey, we hit 60 fps" if the FPS counter is turned off.
But I'm sure these AIB partners will be happy if someone upgrades from a 4070 to a 4070 SUPER just to gain 4-5 fps.
Nvidia's datacenter business blew gaming out of the water in their last reported quarter. There are only a finite number of wafers available.
Let's say you're a baker with 1 kg of bread dough. You can make two 500 g loaves that you can sell at $6 apiece or four pizzas that you can sell at $9 apiece. What are going to do?
Remember that Nvidia is a publicly traded company. Their primary responsibility is to increase shareholder value. They aren't going to do that by continuing to sell cheap bread loaves when the pizza demand has skyrocketed.
This isn't specific to the electronics industry. When there is a finite amount of production capacity, some care has to be put into the product mix decisions.
Nvidia certainly has the option of changing prices. However this goes back to the shareholder value point. Shareholders don't like to see eroding gross margin. A more strategic approach might be to clear out inventory of older models (with some sort of promotion or regional discounting) before introducing the new SKUs so customers can't do a side-by-side price comparison at the store shelf because the old model is no longer for sale.
I don't see Nvidia dropping the MSRP of the 4080 SKU. This is their bread-and-butter premium card. And remember that there are other forces in play. A strong dollar is leading to higher local pricing in Europe.
But like I said earlier, there will be no bargains this time around.
All I'm saying is that for Nvidia this time they don't have to feel embarrassed about it, like they did with crypto earnings which they tried to hide into other sectors, and were then caught when crypto crashed and clearly showed what was fuelled purely by it.
Nvidia can now fully embrace the hype, and advertising cutting production of gaming cards, raising prices will send clear message AI is the focus now.
And that every bussiness in the world should hurry up and order their AI server cards or be left behind in a pre-AI world!Is it though?
I remember people rushing to buy GTX 1080, RTX 3080 before the cryptomadness price hikes, but I have seen only contempt for RTX 4080 and the fact that it went backwards with price / performance compared to RTX 3080 - 50% performance increase for 70 % price increase! Similar as with RTX 2080.
I know more people that bought "lobster-au-thermidor" RTX 4090 than the "bread-and-butter" RTX 4080 one.
And I looked Steam Survey for September 2023, which confirms my anecdotal observations:
NVIDIA GeForce RTX 4090 0.71%
NVIDIA GeForce RTX 4080 0.51%