I wouldn't be so fast to claim that, since it seems like Ampere flops are a very different beast from Turing flops, assuming these are FP32 flops at all. I'm thinking something like Terascale to GCN, but in reverse.

According to the benchmarks above, the 8,704 Ampere Cuda cores of the RTX 3080 perform closer to ~ 5,000 - 5,500 Turing Cuda cores, clock per clock. So they are just about 60% as efficient compared to Turing. I'm not sure how RDNA 2 would compare.

Edit - I do remember, years ago, that you were once mislead by Nvidia's PR speech and believed the Tegra portable was 1 TFLOP of FP32 too, so yeah.

Something feels off about the tflop numbers. 3070 seems like it will maybe be a 10-15% improvement over 2080ti, but they are saying that it has 7 more tflops than 2080ti. It's strange to see for sure, usually Nvidia flop numbers are on the conservative side, punching above their weight class and beating AMD cards which have higher flop ratings in gaming performance. But now all of a sudden they are advertising these massive flop improvements. At the same time, AMD seems to be trying to improve performance per flop as much as possible. It feels like a total reversal of the flops marketing from Nvidia and AMD that we are used to. 

That being said, 3070 will definitely beat PS5 and Xbox Series X. Rumor is Series X performance falls in between 2080 Super and 2080ti, and 3070 is above 2080ti. And PS5's GPU is weaker than Series X's. 

However, you also have to remember than 3070 is $500 for just a GPU, an entire PC built around a 3070 will probably cost you around $1200. Series X and PS5 will be $500 tops, they will still be price/performance kings later this year. 

The's more to it than TFlops as for a start AMD TFlops aren't the same as Nvidia's which is why the 2080Ti is so significantly above the 5700xt another thing is features like dlss and such which reduce required power for better output as seen in Death Stranding's PC version you can play and have 4K visuals at 60fps with out actually having to run the game at 4K.

Ampere is using a new kind of shader that is able to execute two instructions per clock. This is similar to hyperthreading in CPUs. However it's of course less efficient in real world applications than having 2 full shaders. Nvidia now proceeds to treat those new shaders as double the shaders from which they also derive the FLOPS. This brings us into a bit of a predicament because now the shader count and FLOPS are not comparable to Nvidia's own cards anymore.

I mean when the price of the GPU is likely $300 more than the XSX/PS5 in their entirety, the power difference should come as no surprise. What I'm slowy realising though is how unprepared it feels for sony/MS to go into this generation with no built in image reconstruction features. It could be that neither wants to speak about it yet as it sounds like a power concession but that's unlikely.

I think that has less to do with the willingness of the console manufacturers and more with the ability of AMD. I'm sure they would've loved a feature similar to DLSS, but I doubt AMD has anything to offer on that part. Which kinda is a shame because DLSS would be absolutely perfect for consoles who already struggle with performance and image clarity. Of course they could've gone with Nvidia, but that would've been a real mess technologically, economically and politically. Less so for Xbox, but still.

Depends on the application. VRAM is a far bigger deal in the professional/render world vs gaming, for example.

GPU's for years have always executed two instructions per clock... Which is why we always used the:
Clock Rate * 2 instructions per clock * Shader Pipeline algorithm to determine theoretical floating point performance...
I.E. Xbox Series X.
Clock Rate = 1825Mhz * 2 instructions per clock * 3328 shader pipelines = 12,147,200Gflop or 12.14 Teraflops.

What nVidia has done here is made each pipeline branch out into two.

Shader counts and Flops have never been comparable, it was always a coincidence that they lined up in comparisons.
It's always been a hypothetical maximum rather than a real-world capability.

Case in point you could pick up a Geforce 1030 DDR4 variant and compare it to the GDDR5 variant, identical Gflops... But the DDR4 variant will be half the performance... Which just reinforces the idea that there is more to rendering a game than how many flops you have.

For example, AMD has always had more "flops" than nVidia... But in gaming that didn't amount to much, why was that? Because games do need things like Vertex and Geometry operations to also be performed, games may use integer texturing as well.
In short it meant that nVidia had better gaming performance... However.
If you were to throw a pure Asynchronous compute task at an AMD GPU, it will switch into another gear and it will beat nVidia hands down, which is why AMD GPU's were such lucrative buys during the crypto-craze, they were pure compute tasks and AMD's GPU's were built better for that.

Alternatives to DLSS exist.
Radeon Image Sharpening for example.

@Pemalite   
I got a question if you dont mind.
But what would bottleneck a gpu more when talking vram, its final bandwidth or it the amount of vram?
Obviously we want faster and more, but wich to prioritize?

I was mainly asking because of the 3070 and 3080 have the same amount but different bandwidth's. So im not sure if the 3080 would reach some kind of limit in some cases, where the 3070 would not as it would get there anyways.