NVIDIA Announces The GeForce RTX 40 Series With Much Better Ray-Tracing Performance

[theriddick]

It sounds amazing but I found the thing a little suspicious, it completely changed textures and objects. I don't think that was automatic, more like someone put much work in it.

I think they said it will be a mod pack that works inline with some ADA RTX stuff... So yes modders need to come in and do some work to create the mod pack. Which is fine IMO.

I also liked the massive improvements to Flight Simulator, that was impressive because original that game ran pretty badly.

[Anux]

If I understood that correct, the improvement comes from DLSS 3 which introduces motion interpolation, therefore the CPU has less frames to render.

Let's see how this looks in real games, I don't like the effect on modern TVs. Maybe this is done better, if they take some data from the game engine it could be pretty accurate.

[birdie]

Key points:

• DLSS 3.0 is fantastic though proprietary.
• Pricing is just bad
• Two 4080 SKUs with a different amount of shaders? Looks like NVIDIA decided to charge top dollar for what should have been RTX 4070 Ti. Let's see what RDNA 3.0 will bring because this is just ugly.
• I expect RDNA 3.0 to reach the RTRT performance of the RTX 30 series which again means NVIDIA will take the performance crown for heavy RT games for the next two years.
• Looks like we've reached the point where the laws of physics no longer allow to get more performance at the same power (envelope) which is really really sad.

[birdie]

If I understood that correct, the improvement comes from DLSS 3 which introduces motion interpolation, therefore the CPU has less frames to render.

Let's see how this looks in real games, I don't like the effect on modern TVs. Maybe this is done better, if they take some data from the game engine it could be pretty accurate.

It's not just motion interpolation, DLSS 3.0 renders almost full frames without using any classic GPU/CPU instructions which will result in an enormous performance increase/power savings. I'm not sure the tech will work for all games.

Why "almost", because in motion the image changes and there's new stuff (not seen before on the screen) to be rendered anyways using GPU/CPU but depending on how good DLSS 3.0 is, it will mean that you only have to render properly the percentage of the frame which has changed between DLSS interpolations. In games with little motion it means you can use just DLSS 3.0 instead of computing frames using GPU/CPU.

DLSS 3.0 looks like an alien tech - I've heard about it before but no one imagined anyone could make it work.

As much as people here hate NVIDIA (for reasons which I personally don't understand at all because they are absolutely illogical, let me repeat once again: NVIDIA does not owe anyone Linux support at all or open source drivers for Linux) the company has basically invented the modern game graphics pipeline starting with programmable shaders ending with RTX/RTRT.

[fong38]

As much as people here hate NVIDIA (for reasons which I personally don't understand at all because they are absolutely illogical, let me repeat once again: NVIDIA does not owe anyone Linux support at all or open source drivers for Linux) the company has basically invented the modern game graphics pipeline starting with programmable shaders ending with RTX/RTRT.

They don't but people also don't have to like their stance towards the OSS community, especially when they stand to gain so much from allowing contributions from external sources (see the recent massive vulkan draw call throughput improvements for the AMD and Intel mesa vulkan drivers by the same contributor for instance).

[WannaBeOCer]

Wow 1400 TOPS tensor performance compared to 320 from the RTX 3090 Ti. No wonder why DLSS 3 only supports the RTX 40 series. I remember just a few years ago Nvidia was demoing an AI rendering an urban city. Now they have an AI rendering frames in games.

https://youtu.be/ayPqjPekn7g

I wouldn’t be shocked if games in 2030 are rendered entirely by Tensor cores.

[pinguinpc]

As much as people here hate NVIDIA (for reasons which I personally don't understand at all because they are absolutely illogical, let me repeat once again: NVIDIA does not owe anyone Linux support at all or open source drivers for Linux) the company has basically invented the modern game graphics pipeline starting with programmable shaders ending with RTX/RTRT.

holy wars begins in 3.....2.......1......

[birdie]

They don't but people also don't have to like their stance towards the OSS community, especially when they stand to gain so much from allowing contributions from external sources (see the recent massive vulkan draw call throughput improvements for the AMD and Intel mesa vulkan drivers by the same contributor for instance).

I'm fine with "not liking" NVIDIA. On the pages of these forums however you'll read "F you, NVIDIA" under every news piece related to the company.

I don't like a single company in the world BTW, and I certainly like neither AMD, nor Intel. They have enjoyed the x86 ISA duopoly for over two decades now. How this is totally fine and legal in this world I've no idea. Likewise NVIDIA and AMD have enjoyed the GPU market duopoly for two decades as well, only in the early 00s they fiercely competed in terms of pricing and performance and new graphics architectures came out a lot more frequently than nowadays. And around 4 years ago or so (Radeon VII) both companies coincidentally started copying one another pricing which, according to AMD fans, is totally fine as well. Everything AMD does has a justification, it's good, it's great, everything Intel and NVIDIA do is evil/anti-competitive/anti-consumer.

Speaking of the "the recent massive vulkan draw call throughput improvements" - I've no idea what you're talking about. Yes, AMD invented the proprietary Mantle Windows-only graphics API which was used as a basis for Vulkan/D3D 12 but NVIDIA swiftly adopted it for Pascal GPUs which had been designed long before Vulkan 1.0 was released and it worked beatififully on NVIDIA GPUs.

I don't know a single OSS development over the past two decades which NVIDIA has benefitted from aside from their super computer contracts but that's outside of OSS per se. Sorry. I know a ton of features NVIDIA has added to Linux however.

[birdie]

Here's another point of concern (shamefully stolen from TPU forums):

This will be the biggest difference between the flagship and the next fastest GPU in terms of SMX count that I can remember. The previous generations were like this:

Generation	Flagship SMX count	2nd tier GPU SMX count	Ratio	Comments
Kepler	15	12	1.25	GTX 780 Ti vs GTX 780
Maxwell	24	22	1.09	Titan X vs GTX 980 Ti
Pascal	30	28	1.07	Titan Xp vs GTX 1080 Ti
Turing	72	68	1.06	RTX Titan vs RTX 2080 Ti
Ampere	84	68	1.24	RTX 3090 Ti vs RTX 3080 10 GB
Ada	128	76	1.68	RTX 4090 vs RTX 4080 16 GB

One can easily see how the 4080 16 GB stands out as the runt and poor value.

Even if we go by die sizes for the actual 2nd tier full die, this generation is an outlier.

Generation	Flagship SMX count	Flagship Price	2nd tier GPU SMX count	2nd tier Price	SMX Ratio	Price Ratio	Comments
Kepler	15	699	8	330	1.88	2.12	GTX 780 Ti vs GTX 680
Maxwell	24	649	16	499	1.50	1.30	980 Ti vs GTX 980
Pascal	30	699	20	499	1.50	1.75	1080 Ti vs GTX 1080
Turing	72	1200	48	699	1.50	2.08	2080 Ti vs RTX 2080 Super
Ampere	84	1999	48	599	1.75	3.34	RTX 3090 Ti vs RTX 3070 Ti
Ada	128	1599	76	1199	1.68	1.33	RTX 4090 vs RTX 4080 16 GB

Now it looks better for the 4080 16 GB until you consider the price which is outside the historical norm for the lower tier GPUs. Only the GTX 980 was priced this close to the flagship, and that was an atypical generation in many ways.

I don't care too much because I've never bought cards of these tiers but it's still not looking good at all.

[kozman]

Boy do Jensen had a narcissistic photo for the video... "Look at me, I'm the best Taiwanese CEO, not that witch in the other company"...

Glad I am not the only one who looked at it and thought the exact same thing. At least Lisa smiles.

Do they teach this shitty "Pretension Face" in all limp dick "Business" schools? "Look into my nostrils as I *appear* to be way up here and I imagine you to be somewhere way down there." SMH.

Amending: A smile goes a long way.