NVidia Ada Speculation, Rumours and Discussion

[DegustatoR]

From the following link, it seems like taking advantage of the OMM and DMM engines will require using special Nvidia SDKs for Nvidia Micro Mesh's. So the BVH improvements and the alpha-testing improvements may not be realized by most games. Not sure how widespread the adoption of these technologies will be, but I'm guessing adoption won't be good.

I was wondering this myself. It may be the case where these will eventually be absorbed into DXR like it has happened with mesh shaders and VRS. Or it may remain Nv only tech in which case its usage will be limited to games which will partner with Nv for their RT implementations.

It kind of is? Your sentences are poorly constructed and filled with grammatical errors. It makes it impossible to understand what you're attempting to tell.

I was pretty clear in what I was saying over the last several replies to you - there are no 3090 performance equivalent in 30 series at 1/2 of price. This isn't how it was with Titan lineup and thus it is different in what new cards are hitting 3090 performance levels. English isn't my native language and I'm writing too many things in many places right now. Do you understand what I'm saying?

Nah, they need to save the 4080Ti name to fill that enormous gulf between the 4080 and the 4090. Based on the gap between these GPU's, the 4080 12GB is clearly a 4070.

With this possible option coming later they could've always make a 4080 Super or 4080Ti 20GB or whatever.
It is less of an issue in such lineup adjustments than making the original lineup weird from the start.

The comparison points I was referring to were GTX 970 = 780Ti, GTX 1070 = 980Ti, RTX 2070 = 1080Ti

Yeah and all of these had their Titans in place of a x90 part with the same performance.
I'm sure that 4070 will hit 3080Ti performance as well. The difference now is that 3090(Ti) is actually faster.

 

[hurleybird]

Feels like Nvidia's Fury X moment (hitting an architectural brick wall with scaling), minus the memory capacity limitations and plus some new features.

 

[Below2D]

I was pretty clear in what I was saying over the last several replies to you - there are no 3090 performance equivalent in 30 series at 1/2 of price. This isn't how it was with Titan lineup and thus it is different in what new cards are hitting 3090 performance levels. English isn't my native language and I'm writing too many things in many places right now. Do you understand what I'm saying?

I get it now I think, but this makes the 90's cards even worse value propositions than the Titans.

 

[trinibwoy]

I tried watching the live stream earlier and ended up watching 45 mins of AI nonsense. Where’s the GeForce reveal? I looked on YouTube and just found a 1 min video with a few 4090 specs.

 

[Below2D]

I tried watching the live stream earlier and ended up watching 45 mins of AI nonsense. Where’s the GeForce reveal? I looked on YouTube and just found a 1 min video with a few 4090 specs.

He starts the live stream talking about Geforce.

 

[pjbliverpool]

I tried watching the live stream earlier and ended up watching 45 mins of AI nonsense. Where’s the GeForce reveal? I looked on YouTube and just found a 1 min video with a few 4090 specs.

It was right at the start for about 10 mins.

 

[davis.anthony]

I've asked a friend of mine who has a YouTube channel that analyses frame rates in video if he can run the Digital Foundry teaser through his software to see if he can obtain any frame rate data.

Specifically the 3090ti vs 4090 footage.

 

[Flappy Pannus]

I tried watching the live stream earlier and ended up watching 45 mins of AI nonsense. Where’s the GeForce reveal? I looked on YouTube and just found a 1 min video with a few 4090 specs.

You can still use the slider on live youtubes, you know.

 

[Bondrewd]

RDNA3 could have a very significant lead in that area if their >1.5x perf/w claims for RDNA3 are good

Well, yea, that's the point really.
Focus on raw PPA above all else.

 

[troyan]

Okay so no-where near 4x RT performance then. Cyberpunk maxed out (one of the heaviest RT tests available presently) is generally less than 2.5x faster when using DLSS3 on the 4090 vs DLSS2 on the 3090Ti. So probably a little under 70% faster with no DLSS (based on the DLSS2 vs DLSS3 test showing a roughly 50% uplift for 3). So fast, but below expectations and certainly slower than it was portrayed in the video. The 4080 12GB (4070) is probably going to struggle to keep up with the 3090Ti in many current games even with RT. Not bad if it were priced like a 4070 but not even close to worth the 80%(!) price increase for a performance level that we would generally expect from a new generation in this category.

DLSS performance is 1080p. I bet in 4K with raytracing it will be more around twice the performance.
The new RT mode in Cyberpunk shows the performance uplift:

Should be 2.5x faster with DLSS performance.

 

[DegustatoR]

I tried watching the live stream earlier and ended up watching 45 mins of AI nonsense. Where’s the GeForce reveal? I looked on YouTube and just found a 1 min video with a few 4090 specs.

You can just read these here instead: https://www.nvidia.com/en-us/geforce/news/
All the fancy graphs are there anyway.

 

[pjbliverpool]

I'm sure that 4070 will hit 3080Ti performance as well. The difference now is that 3090(Ti) is actually faster.

If you mean the GPU that Nvidia will end up calling the 4070 then I doubt it outside of DLSS3 cases. Based on this chart the 4080 12GB is only trading blows with the 3090Ti and according to TPU the 3090Ti is only 9% faster than the 3080Ti. So unless the 4070 is much closer to the 4080 12GB in performance than a x070 usually is to a x080 then I'd expect it to be slower than the 3080Ti.

https://twitter.com/x/status/1572257012263657475

 

[DegustatoR]

If you mean the GPU that Nvidia will end up calling the 4070 then I doubt it outside of DLSS3 cases. Based on this chart the 4080 12GB is only trading blows with the 3090Ti and according to TPU the 3090Ti is only 9% faster than the 3080Ti. So unless the 4070 is much closer to the 4080 12GB in performance than a x070 usually is to a x080 then I'd expect it to be slower than the 3080Ti.

https://twitter.com/x/status/1572257012263657475

In PS4 era games without RT or anything modern like ACV and TD2? Maybe. We'll see.

 

[Below2D]

The fact that NVIDIA makes the numbers so nebulous is extremely annoying. SHOW us the performance WITHOUT DLSS. Then show us DLSS 2.0 vs 2.0, then 2.0 vs 3.0.

 

[Bondrewd]

The fact that NVIDIA makes the numbers so nebulous is extremely annoying

NV sits with gobs of unsold Ampere inventory so without funny slides Ada would be a very hard sell.

 

[neckthrough]

The fact that NVIDIA makes the numbers so nebulous is extremely annoying. SHOW us the performance WITHOUT DLSS. Then show us DLSS 2.0 vs 2.0, then 2.0 vs 3.0.

I think the charts already show the latter two. The strict non-DLSS comparisons are academic and not very useful for games that support DLSS.

As usual independent reviews will dissect things with far higher fidelity and sprinkle in their own flavor of editorialization.

 

[gamervivek]

https://twitter.com/x/status/1572268575385329664

 

[Jawed]

The fact that NVIDIA makes the numbers so nebulous is extremely annoying. SHOW us the performance WITHOUT DLSS. Then show us DLSS 2.0 vs 2.0, then 2.0 vs 3.0.

Strategy by NVidia to conceal the real performance of Ada from AMD? All the marketing is in place, gamers are primed, AMD doesn't know what it's aiming at...

 

[Clukos]

Probably the least exciting reveal from Nvidia the past 5 years, definitely worse than Turing. Maybe prices will normalise once Ampere stock dries out.

 

[davis.anthony]

Not only is this card not going to fit in a lot of cases, but there's also going to be a lot of cases that can physically fit the card but won't be able to thermally deal with the heat dump inside the case from it.