Review: GeForce RTX 2060 and 2070 SUPER Founders edition

What's wrong with the founders card - it's not a blower? The cooler is good. The pcb and power will be good. Realistically any after market card will give you a negligible o/c advantage unless it's under water. I can't see why anyone would spend more then $400 on a 2060S, particularly as the much faster 2070S is only $100 more?

 

But if they still have chips for the 20 series and now they have better binned chips, why wouldn't they just try to push them out? And they are most likely holding onto them just to see what AMD does. Nvidia doesn't give a crap how close the next release is, if anything they will push it back in order to get more sales on this 20 series. However I feel as though if they do release a 30 series early next year, it's going to be Non-RTX first.

 

Not really. Since usage of RT and tensors in gaming was a "new approach" with no way to test (as DXR was not yet released) , it is more likely later RT development/games revealed more performance was required at all price points.
The obvious solution was to introduce the "Super" versions at the same price points, and no better time to do this around the Navi introduction.

 

Better binned chips are available at Titan RTX. It's all speculation anyway at this point. We'll see soon enough !

 

Yes, it's a replacement, 6-9 months later. It's new, as in was not available before.

It's not like these cards are even just the same thing with higher clocked frequencies, they have different amount of cores and everything.

Basically it just seems like you're upset that they are replacing the 2070 and 2080 rather then lowering the prices of them and having both the super and non-super cards, which would be very close together GPUs, price wise. it wouldn't make much sense to do that as you'd have one card, whichever is better value, selling more then the other card, you'd have too much stock, and either have to lower the other card that isn't selling, or stop production of one of them.

I gave many examples of same-architecture-generation replacement cards with same prices, the only difference in those cards is that they are labeled differently. For instance, someone who bought a GTX 1080, 1 month before the cheaper, but very similar in performance card, the 1070 ti came out, would be the same situation and just as "upset" as it is right now.

It really, and truly, seems you have an issue with the naming, not the cards, since you don't have an issue apparently from what i can gather about previous generation cards.

So, for instance, if the 2080 ti currently available was actually a Titan

And they released a 2060, 2070 and 2080 all the same as they currently are

But 6-9 months later, aka now, released a 2060 ti, 2070 ti and 2080 ti, all with better specs, more cores, and better bang for your buck, as replacements, just as a 1070 ti was a replacement for the 1080 at a lower price, you'd have no problem.

A 2060 Super is a 2070 replacement at a lower price, a 2070 super is a 2080 replacement at a lower price. Sure, they aren't exactly the same performance, just as the 1070 ti was not exactly the same performance of the 1080. And a 2080 super, when released, looks to be the same thing as what would have been a 2080 ti (if it didn't already exist and the titan was what is currently a 2080 ti) to get closer performance to the titan.

And the whole "beta" thing makes no sense. No one who used these cards were in "beta", they used these cards because they wanted the performance of the cards that they knew they'd get and the features of the cards they knew they'd be able to use. The fact that 6-9 months later, newer, better performing cards are released doesn't change this fact. If these were "beta" cards, then all non-ti cards of previous generations are "beta" cards, and at that point, my original statement is true still: Nothing new has happened here that hasn't been happening.

So ultimately it seems you have an issue with the naming of the cards rather then the history of everything that has happened the last many years.....

Pretty sure the 2070 super is nvlink capable, given the picture provided in this article anyways.



That bulge looks like a cover for an nvlink connector, and isn't on the 2060

@Hilbert Hagedoorn do you still have the card and are you able to verify this?

 

Your logic is flawed.If that was the case, then why would anybody ever bother with any aftermarket cards in the first place?Anyone could EASILY settle down with the stock one Nvidia releases.Why things don't work like that?Because ppl want more and more.Better designs, better clock speeds, better OC capabilities, better thermals, noise levels, so expect ppl to pay more.I wish things were as simple as that but they are not.

 

It's better than before, thanks to competition.
I won't need GPU update for a while (I hope), even so it's not my price point.

People should know companies's practices and choose based on that, but people also have the right to complain and disagree of companies's practices, then maybe change something.

 

@Aura89 : I have to disagree about "New" in terms of GPU. They are just different cut of existing GPUs. nVidia had enough of chips that did not match existing products, therefore this release.

Good/Bad, I can't judge just now. They are pretty close to current ones, therefore it may look like replacement. But that remains to be seen too. Same way as price mark.
I do expect that 2060 S will land in similar performance as 2070 and be cheaper enough to move performance per $ to desired direction.

 

No clarification on SLI yet for the 2070S. I can't find any info if it can be paired or not.

 

Steve at GN got confirmation of NVLInk works on 2070S strait from Nvidia.

 

There's not a single, non-super RTX 20 series GPU with:

2176 cores
2560 cores
or
3072 cores

Yes, they may different variations of the TU106 and TU104, but that doesn't mean they aren't new, as there were no configurations like these before.

It'd be like going to AMD with their CPUs and saying "you only have one product"(in april 2017), because the Ryzen R3 1500x(not sure about below, less cache and etc.) to Ryzen R7 1800x all have the same die on it, but with different frequencies and parts of some disabled. Ryzen 5 1500X came out over a month after the 1800x, is it not a new product, simply because the die is the same?

You guys seem to be getting stuck on naming of products and naming of the architecture behind them. I never once stated its a new architecture, i never once stated there's brand new tech never seen before in the super series of GPUs.

But are the products new? Yes. There were no GPUs 6-9 months ago with these configurations, regardless if it's the same, or slightly altered die, that you could have gotten. Again, they are not new architectures, and they are not just a rename with same stats and same amount of cores etc. Are they a reconfiguration that has never been released in their current configuration? Yes, because they are new.

If they are not new, then i'd like to know where i could have gotten any of these configurations before now? Because that's the only way the PRODUCT is not new.

I really don't understand your guys' fixation on trying to apply one meaning of a word to every possibly meaning of the word. Again, no, these GPUs are not new architectures, i never stated they were. But there are no facts to back up that they are not new products, as though they are rebranded products you could have had before.

An 8770 is not a new card, because it's a rebranded 7790. There's nothing new about it, no change in cores, no change in architecture, not even a change in frequencies. The 260x as well can be considered not a new card, with only a slight boost in frequency.

An RTX 2070 Super is not a new card, because it's a rebranded...........?

 

The 2060 super seems like a nice upgrade from my 980ti although the current gpu prices are still crazy, I paid £500 for my 980ti now I'm looking at "midrange" cards for that price

 

Same silicon. Different cut. It is result of binning. It is not any newer or older. It is question of yields and quantities.
Later launch only means that there was not enough chips with given properties.

Ryzen 3000 will be best example. 3950X will not use any newer chips than 3700X / 3800X. Only differently binned.
- - - -
To add, this launch may be just clearing of those differently binned chips. And nVidia may already be readying 7nm EUV.

 

So then what's not new about it? Where could i have found these GPUs before?

That is my question, my one and only question. If they are not new, as in a new product, not a new architecture, then that means i should be able to get it before hand. So since it's not new, where? Please tell me where i could get a Turing series GPU with 2560 cores before July of 2019. If that can't be answered, that's fine, because that's the entire point, only new 20 series cards released July 2019 and after have that as an option.

So that's my question. If it can be answered, great! But it's the only question i want answered, until then, i will no longer be replying to any statement made toward me regarding these not being "new", as you'll have to prove where these products were with their configuration before you do.

 

Its all

Ti is going out the window - Your next name for the top card will now be called Super Duper. I find the whole naming jargon pathetic

 

2070/2080/Quadro 5000... It's no different than having HD 7970 and then releasing HD 7950. Existing GPU from which laser cuts different portion away.

People were getting them till now. I am not arguing that it is new product. But that does not make silicon new. Quite contrary, those chips were sitting somewhere for very long time... waiting for critical mass to create that new product.

Same as intel releasing KF chip where they had defective iGPU that they disabled. Or KS which could clock higher. Same chips, different binning.

 

You bought RTX, could barely use it for RT and now you are burned

 

I dont think a single person who bought an RTX card had any illusions about RT. The highlight of Nvidias pre-release RTX presentation was a miserably performing RT demo that was savaged by critics and the tech media alike. Everyone from day 1 knew what they were buying and it certainly was not for RT. Unless they had some perverse inclination of using a 2080ti on a 1080p display.

 

The problem is you're having to assume that these new cards are using "old" chips. When there is no way to prove it either way.

 

Yes they do. Most gamers are kids that know no better.