Nvidia might give 7nm GPU teaser on GTC

How much for 3080ti 2k,3k, or more?

 

Over 9000!

 

All depends on how well Navi does and what it's price is. Fingers crossed hey.

 

Awesome, this is going to be interesting. So are high end or any form of Navi also going to be at GDC!?

Man 2020 is going to be insane or near the end of this year! Let's GO \m/

 

There is a reason 2080Ti launched alongside 2080 instead of the usual 6-9 month delay and that's because 3080Ti is just around the corner.

 

Nah, the reason is that in their internal tests Nvidia had noticed their RTX raytracing implementation is more or less unusable with a card less powerful than the 2080 Ti. Without it in the market, the market would have had no card that could run Nvidia raytracing reasonably (at 1080p 60 fps anyway).

 

@Hilbert Hagedoorn

"GTC (Graphics Technology Conference)"

It's "GPU Technology Conference"

 

I wonder if the newly updated CryEngine's implementation of their own version of real-time Ray-Tracing that's hardware-agnostic was developed in tune with Nvidia to utilize RT cores alongside general compute ...

 

Lol my gtx is over 9000!!!

 

if they give a 7nm teaser, its going to be a refresh of TU with a 102b, 104b, etc for release in september.

Still waiting on that fully enabled 2080*

 

If it's true, they basically admit that they failed and kill the Turing series after a few months.
at least 7nm should help them push clocks much higher. but I expect the RTX theme to still remain strong....

 

GTC is a conference focused on AI and ML. Pretty sure nothing will be said regarding to gaming cards!

 

In past years it was AMD to adjust their prices after Nvidia, not otherwise. Price performance for Turing and Vega/Polaris is equal. Vega and Polaris are even slightly more expensive.

 

Unfortunately for Nvidia they've pretty much already set their pricing : performance with the RTX. To release another series offering greater performance but at significantly lower selling point would not sit well with a significant many, and ask some awkward questions.

AMD, on the other hand, have a very real opportunity. The ball is most certainly in their court providing they can deliver on price : performance. The buying mass is generally a sensible beast and tends to buy wisely once they're informed.
Here's hoping for both Navi (in all it's forms...high end too) and Matisse.

 

What I was thinking. Anyhow some 7nm news would be good would give some timelines to speculate on.

 

How is it any different with AMD? Vega VII is priced similar perf/$ as 2080 except you aren't getting all the value-add features. If they were to come out with something significantly cheaper perf/$ why wouldn't it raise the same questions? In fact I'd argue Nvidia has a better shot because they can do a die shrink and just say "7nm allowed us to drop the price" where as AMD already pulled that trigger.

Personally if either one dropped their price I'd just be like "cool" and probably still not buy it. The 2080Ti would have to be $650-700 to justify it's performance imo and Vega VII is a sidegrade for me. Would be nice for people on older cards though.

 

I'd suggest that Vega II should be looked, primarily, at as a demo of a 7nm die shrink process - which matches Nvidia's 2nd most powerful card - and as a pointer of what additional gains might (hopefully) be achieved if more of the architecture was also at 7nm i.e. Navi. I agree that Vega II's pricing is steep though a significant portion of that is due to there being 16GB HBM2 and uses additional to gaming were also highlighted by AMD.

I'd think we're all wishing for a return to the 650-700 range for a high end card, and, that CryEngine's recent RT demo can show we can get acceptable results in real time without the need for an additional proprietary hardware accelerant e.g. Nvidia RTX.

 

CryEngine's RT demo is interesting but it's still a faux implementation. SVOGI has shown promising results in the past, yet integration into actual games has been non-existent due to various issues that multiple studios/companies have tried to overcome but have not done so.

I also don't like the idea of branding DXR/RTX as "proprietary". "RTX" as in RT cores/Tensor are just a method of accelerating DXR. If Nvidia branded it's tessellation hardware as "trianglegenerator9000" no one would say tessellation was proprietary.

I agree with your point about HBM though. I think AMD overall has more flexibility for pricing but I think Nvidia also has some leg room with the 7nm shrink.

 

imagine if the 3080 etc have no rtx what so ever ...