No, Nvidia isn't making an RTX 2080 Ti SUPER

While I agree with the overall statement... LOL, you're using TweakTown as a source?

Yeah, I don't like using them as a source due to their history of unreliability, but I know for a fact that he was at the Super briefing.
 
There is little point in NVidia producing a super version of the 2080 Ti as it would be very expensive for very little performance increase.

They would also have the problem of finding a use for TU102 chips that do not make the grade for cards that use the full fat version.
 
This comes out one day after AMD confirm there won't be a big Navi yet.

I bet they would have otherwise.
 
The conference where this question was answered was a couple of weeks ago, while AMD have not revealed any further information regarding future Navi yet since 5700's launch. The article is in response to recent forum rumours of a new Ti Super I think, NVidia would have known technical limitations of non-EUV 7nm processes means AMD couldn't have a true big die competitor this year very long(years) ago.
 
Last edited:
The conference where this question was answered was a couple of weeks ago, while AMD have not revealed any further information regarding future Navi yet since 5700's launch. The article is in response to recent forum rumours of a new Ti Super I think, NVidia would have known technical limitations of non-EUV 7nm processes means AMD couldn't have a true big die competitor this year very long(years) ago.

I'd imagine Nvidia know more about Navi than AMD tbh.
 
In this area both companies will at least be able to give solid estimations of the costs of parts and the hard limitations of the silicon available to competitors to be fair, given they are both regularly paying customers and close partners of the companies that dictate those limits(TSMC & Samsung).
 
I'd imagine Nvidia know more about Navi than AMD tbh.

In this area both companies will at least be able to give solid estimations of the costs of parts and the hard limitations of the silicon available to competitors to be fair, given they are both regularly paying customers and close partners of the companies that dictate those limits(TSMC & Samsung).

Basically, what tgrech said, Nvidia will know roughly how much the RX 5700 series will cost and what AMD's margins are.

What Nvidia can't guess is how efficiently AMD can use their silicon and what IPC/clock speed improvements they can expect. They could easily imagine what 7nm Vega would look like, which is why Jensen could be so confident when the Radeon VII launched.

Nvidia could guess what Navi would be like to a certain extent, but I doubt Nvidia thought that Navi would be this good. The main point about Navi is that AMD's design changes have addressed their weak points, resulting in huge gains over Vega in lots of Nvidia favouring titles.

As it stands right now, AMD is still responding to Maxwell, the architecture that gave Nvidia the power lead and clock speed lead they needed to take the entire mobile market and fully assert their dominance. AMD is starting to make a comeback with Navi, but they still have a lot of catching up to do.
 
What I meant was you can bet there are spies and back handers going on etc.

It's one thing taking AMD's word for it, but Navi could have been a lot slower and Nvidia could have just left things as they are. I'll admit there's a lot of science and maths involved, and you could work it out roughly but yeah I bet they have someone inside leaking secrets.

I think Navi is a great tech, and god it must be so much cheaper to make also. I just wish they would have done something like IF that can add small cores together like multi CPU. That may not work, I'm no rocket scientist, but it would have been cool.

I've noticed huge drops on 1080Ti and Titan second hand prices now. Not that I care, I never buy things then worry about losing my cash (Ive lost it the second I hand it over tbh) so yeah whatever.

I'm almost upset in a bored kinda way that Navi isn't better. I mean, is it better than a Vega 64? of course. Is it worth changing out to the tune of nearly 400 quid? absolutely not. I'd rather find a way to water cool my Vega (Devil, no official blocks) and get it screaming. I might go the AIO route at some point.

I'm never bitter about stuff, so I am glad Navi came about. Sure, it's not made high end RTX more affordable but it's given the middle of the market a bloody good shake.
 
I think Navi is a great tech, and god it must be so much cheaper to make also. I just wish they would have done something like IF that can add small cores together like multi CPU. That may not work, I'm no rocket scientist, but it would have been cool.
Oh without a doubt everyone is working on MCM GPUs(Here's some of NVidia's published work from 2017 if you want a dive into some of the technological hurdles in that area), and by all indications we will hit the inter-chip bandwidth to be able to at least pass a small GDDR bus via a separate IO controller in the latter half of 2020, the first use case for this will seemingly be the console APUs, but for multi-GPU die chips it will take longer as that won't really be sensible till we have the bandwidth to interconnect(and/or feed depending on topology) large-ish GPU chips so we can go beyond what's currently economically possible, as you do lose some efficiency in terms of both performance and bandwidth going this route.
 
Last edited:
Maxwell? Nah the power efficiency is better than that. Navi is actually pretty competitive tbh on that.

Navi is competitive with Turing in terms of power consumption, but it is worth remembering that AMD is one die shrink ahead of Nvidia. 7nm VS 12nm

Navi's power efficiency is way better than Maxwell, but again, Navi is two die shrinks ahead. What I mean is that AMD needs to catch up to the technological advantage that Nvidia gained starting with Maxwell. Maxwell surprised everyone at the time and AMD needs a Maxwell-like leap to catch up.

Thankfully, Navi is a huge leap for AMD and brings them close to where they need to go in terms of efficiency and performance. With AMD's partnership with Samsung, we know that power is going to be a huge focus for Radeon.

Lots of Maxwell's secret sauce likely stemmed from Nvidia's focus on mobile chipsets. Yes, these efforts are mostly considered a failure outside of Nvidia's Nintendo Switch design win, but what Nvidia learned in developing their mobile chips paid off in both the laptop and desktop markets. Maxwell was a turning point for Nvidia, and AMD is only now starting to turn things back in their favour.

I hope this clarifies my point.
 
I was very surprised when I saw the power consumption numbers. I know AMD have a node advantage, but for the first time in years AMD are actually competitive with Nvidia. Performance per watt is about the same. It goes against the rumours that AdoredTV was quite confident in, that Navi was another Vega and the hardware team couldn't wait to be rid of it and move on to the next architecture.
 
I was very surprised when I saw the power consumption numbers. I know AMD have a node advantage, but for the first time in years AMD are actually competitive with Nvidia. Performance per watt is about the same. It goes against the rumours that AdoredTV was quite confident in, that Navi was another Vega and the hardware team couldn't wait to be rid of it and move on to the next architecture.

AMD has done a great job with Navi, especially after considering that the RX 5700 has the same number of stream processors as an RX 480. AMD's work on Navi has paid off tremendously.

I can't wait to see what AMD has coming next.
 
I hope RDNA Next or whatever they end up calling it just focuses on performance and features. Power consumption is already pretty great, I just want to see massive gains in performance. If they could get within 5% of a 2080ti I would call it a win(provided power consumption doesn't sky rocket again).
 
Back
Top