NVIDIA Announces Turing-based Quadro RTX 8000, Quadro RTX 6000 and Quadro RTX 5000

[cucker tarlson]

Again it was very silly debate, and you still get them wrong way a round. Without tensors it could have been 617 x 6 cuda core SMs on 3072cc card and 768 x 6 cuda core SMs on 4608cc card(was not possible other way around). So it's not huge it's very tiny SMs.

I get it now,but it's very stupid.

 

[FordGT90Concept]

Judging by the picture, less than half of the die real-estate is useful for gaming. If everything else were equal (clock speeds, shader architecture, etc.), Pascal would be a better gaming card.

NVIDIA has to invent new software to use the hardware features in order to justify the waste of all that silicon in the gaming segment. It's PhysX all over but instead of "buy a second GeForce for PhysX" it's going to be "buy a GeForce RTX for ray tracing!"

 

[cucker tarlson]

Judging by the picture, less than half of the die real-estate is useful for gaming. If everything else were equal (clock speeds, shader architecture, etc.), Pascal would be a better gaming card.

NVIDIA has to invent new software to use the hardware features in order to justify the waste of all that silicon on them in the gaming segment. It's PhysX all over; an argument NVIDIA can't win.

you mean microsoft's DirectX R or Vulkan RTX ?

Radeon RXT coming 2019

 

[jabbadap]

I get it now,but it's very stupid.

Yeah agree, It's very stupid and even impossible when adding tensor cores on the equation. But it's only way to get 3702 and 4608 to have common divisor.

 

[FordGT90Concept]

you mean microsoft's DirectX R or Vulkan RTX ?

Radeon RXT coming 2019

Or not, because graphics cards need to be about eight times more powerful to realistically do real-time raytracing. By the time hardware is eight times more powerful to do it, the bar will move in terms of scene complexity so they'll need eight times more powerful hardware after they got eight times more powerful hardware.

Lighting isn't something game developers want to budget a lot of compute time for.

Real time ray tracing has been a carrot dangling from a stick for decades now and there's no sign of that changing.

 

[cucker tarlson]

That's why they utilize specified RT hardware like RT cores accelerated by tensor cores, that'll do the job better than just adding more fp32 cores.Plus I don't think RT is that compicated after all, I've seen people say it not, it's just compute heavy.

Yeah agree, It's very stupid and even impossible when adding tensor cores on the equation. But it's only way to get 3702 and 4608 to have common divisor.

 

[FordGT90Concept]

All that's different with RTX is approximation. Even with cutting corners, fairly simple scenes still render at less than 1 fps on a $3000 card. Unless you're making a game like Myst, that's worthless.

I'll let you in on a secret: the numbers were more or less the same two decades ago. Only difference is that scene is more complex now and still unattainable because the goal post keeps moving.


This is a pretty big deal for the digital animation industry because instead of buying racks of servers, they can do the same thing with a single rack of graphics cards. Not useful for consumers.

 

[cucker tarlson]

I can't wait to play metro exodus at half fps then.

 

[FordGT90Concept]

Much, much worse than half. Traditional techniques look a lot better than RTRT forced to run at 15 fps.

 

[cucker tarlson]

Well, we'll know more on monday.

 

[Octopuss]

All this tech fluff is all nice and fancy, but what new releases of cards really turn me on are the new features available NOW and in ANY game. Like, for example, utilizing all these fancy Tensor cores to make next level post-process AA that's light on GPU but smooths edges better than crappy FXAA they give us in NV CP. Like, freaking at least give us option to switch between FXAA, MLAA and SMAA ffs. Just give us something new we can use now, not something we might hypothetically see being used in games in 5 years time...

It's why Pascal was such boring release for me. Sure, it was fast, but other than that, it brought basically nothing new and exciting to the end user. New post process AA modes would be a nice start, just like it is that post process image thing they released some time ago to change game colors, sharpness, tone and so on. That's cool, but you need to use stupid NVIDIA Experience to have it which sucks. So, that's another problem thanks to archaic original NV CP. Anyway, I'm rambling again, give us more features for today so we can easier wait for the features of tomorrow... I usually buy new graphic cards because of these features even when I don't really have to buy new one totally out of curiosity, not for what I might use it in 5 years time, maybe.

Perhaps I am missing something, but isn't this about development cards and not gaming cards?

 

[cucker tarlson]

Perhaps I am missing something, but isn't this about development cards and not gaming cards?

He paid so much for his card he's got a right ot rant about anything,whether it makes sense or not.

 

[Anymal]

He paid so much for his card he's got a right ot rant about anything,whether it makes sense or not.

He is always waiting an idiot to call him amd fanboy, but he IS that idiot amd fanboy who paid 790eur for top gaming gpu from not so beloved nvidia

 

[efikkan]

Or not, because graphics cards need to be about eight times more powerful to realistically do real-time raytracing. By the time the time hardware is eight times more powerful to do it, the bar will move in terms of scene complexity so they'll need eight times more powerful hardware after they got eight times more powerful hardware.

Lighting isn't something game developers want to budget a lot of compute time for.

Real time ray tracing has been a carrot dangling from a stick for decades now and there's no sign of that changing.

As I said earlier too, I think it will take multiple iterations before this becomes useful in gaming. But even then, let's say eight times as fast as you suggest, it will still have to "fake it" by doing a low sample raytracing and either blur or "denoise" the result. Full-scene raytracing per pixel would require something about a thousand times faster than this. By studying the videos from Nvidia, it becomes apparent to me that they are pretty cleverly crafted to look impressive. Once you move to typical game scenes with more objects, high-contrast grainy textures, and large scenes with distant lighting, etc.

Using low sample raytracing might be enough for many situations, especially if you don't need water or glass reflections, and just having more realistic soft shadows and color reflections will do a lot for realism (or cool effects). But that "denoising" thing is a gimmick, it will certainly create a lot of artifacts, it will struggle a lot in various situations, especially with fast animations and challenging textures.