• Competitor rules

    Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.

NVIDIA ‘Ampere’ 8nm Graphics Cards

Honestly, when I read comments comparing the expected performance of the (so far fictional) AMD 6900XT with the (so far also fictional) 3080 Ti, based on (estimated) mad gains in nvidia perfomance and AMD having solved all their (presumed) scaling issues and implemented competitive hardware RT (which they haven't revealed much of anything about yet), and making guesses about the prices based on how nvidia has to release by September with a price drop a couple of months later to beat AMD to market (on ranges of cards neither one of them has even announced yet), it feels a lot like building castles in the sky...

I'm not here to stop you, hell, it's not my forum! It all just seems a bit of a stretch to make predictions about.
I would rather think that which is in line with how it has been historically for the most part then assume the price will go up because you know, it’s nvidia and that is what they do.

Next 3080Ti ONLY £1999! Yolo, lest do it, don’t worry about the biggest economical disaster in our life times or anything...
 
We’ll see, but I think you will be wrong on this occasion. We will get much better RT performance at the very least. Also at the very least we will get a 40% bump from 2080Ti to 3080Ti imo.

Even if RT performance is a 2X increase it still won't be nearly enough. Worse still any RT increases will mostly get eaten up by game devs trying to use it even more.

As to the 40% bump from 2080 Ti to 3080 Ti, this is possible but if NVidia go for it this will result in very expensive cards.

I think Ray Tracing will only really get there on the generation after Ampere.
 
A lot of the hype for 3000 series is from us 1000 series owners refusing to be gouged.

If they gouge again for 3000 series... so be it, at least the jump in performance is more in our favour.
 
I can't see it being cheaper from this gen to the next, I just can't see a reason why it would be. Anything above the performance of the next consoles will be in a price bracket above, with each step you take up the ladder there will be diminishing returns in performance. The performance gains will be marginal, while the marketing will push the benefits of the new technology which can give 'up to' a percentage improvement in games, whereas only new AAA titles will take advantage. I still think it will be a more worthwhile proposition than Turing and the super series, but anybody just waiting for a big reason to upgrade will be a bit disappointed unless it caters specifically to them/their games.
Personally I bought the 1080 on release, I wouldn't have stumped up the extra for the 1080ti anyway although every top grade consumer card released by Nvidia has proven to be a good long term investment. If the XX80 comes out months before the XX80TI, I might just wait for the TI, stump up the money and retire from chasing performance for a few years because of how robbed I feel, even though past experiences show it's a good decision!
 
Who cares what the die name is? It's like if Ampere comes out and GA106 is named the 3080Ti then saying it's low end.

If the performance is there the die name is irrelevant.

The die name relates to the perfomance of the chip amongst all the available dies for a given GPU architect of course it is relevant:confused:
In your example consumers would be getting a rubbish deal because they are not getting the fasest graphics card for a given architect since GA102,103 and 104 would all be faster than your hypotetical 3080ti based on GA106.
 
Who cares what the die name is? It's like if Ampere comes out and GA106 is named the 3080Ti then saying it's low end.

If the performance is there the die name is irrelevant.

First off, I care and I'm sure many other cares. I would call the 3080ti a low end card if the die said GA106 as that would tell me that there is a GA104 and GA102 and even a GA100 above it, all in the same family. It doesn't mean that the card is then automatically bad but high end it is certainly not. Call me pedantic if you want to, I'm fine with that but die names dictate the tier of the chip and a GA106 die by Nvidia's current standards dictates a lower-end chip.
 
The die name relates to the perfomance of the chip amongst all the available dies for a given GPU architect of course it is relevant:confused:
In your example consumers would be getting a rubbish deal because they are not getting the fasest graphics card for a given architect since GA102,103 and 104 would all be faster than your hypotetical 3080ti based on GA106.

If the GA 106 performed the best then it's a high end chip regardless of name. I have no idea how it could be low end just because of a name even if it would be the fastest available gpu.
 
Even if RT performance is a 2X increase it still won't be nearly enough. Worse still any RT increases will mostly get eaten up by game devs trying to use it even more.

As to the 40% bump from 2080 Ti to 3080 Ti, this is possible but if NVidia go for it this will result in very expensive cards.

I think Ray Tracing will only really get there on the generation after Ampere.
We’ll soon see I suppose. I do agree RT is many generations away from where it needs to be, but I do think Ampere with DLSS 2.0 won’t tank fps nearly as badly as Turning did.
 
The die name tells a different story IMHO.

The Die size tells the real story and it's a high end die size. And aren't you forgetting it's an RTX, not a GTX card?

First off, I care and I'm sure many other cares. I would call the 3080ti a low end card if the die said GA106 as that would tell me that there is a GA104 and GA102 and even a GA100 above it, all in the same family. It doesn't mean that the card is then automatically bad but high end it is certainly not. Call me pedantic if you want to, I'm fine with that but die names dictate the tier of the chip and a GA106 die by Nvidia's current standards dictates a lower-end chip.

Internal die names do not always indicate the tier of the chip. Since Fermi there have been changes every generation. And if you do go by Nvidia's current standards, then the GA106 will not be low end cards. The chip will used in the mid to high end cards in the $350 to $550 price range. The GA116 would be the mainstream cards covering the $249 to $350 range.

So it beats every 2080 in every single game?

The 2080 and 2070 are discontinued.
 
Back
Top Bottom