• Competitor rules

    Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.

AMD Navi 23 ‘NVIDIA Killer’ GPU Rumored to Support Hardware Ray Tracing, Coming Next Year

Status
Not open for further replies.
I couldn't go back to AMD now, as theres just too much crap in the CP for me, its just bloatware.

I liked this one, Crimson ?

DH-012-07.png

then just install the new drivers and use the older control panel.
 
I did express concerns about performance of Ray Tracing when that developer said you could either have Ray Tracing or high FPS.

Sony did say they had some secret sauce for Ray Tracing, it might have been only market speak, but I wonder what all this means for AMD's Ray Tracing performance on the PC?

Annoys me a bit as I've been playing around with Quake 2 RTX a lot and seeing the potential - but seems AMD are just making some optimisations to existing hardware to be able to run it a bit rather than putting the effort this round into proper RT acceleration.
 
Annoys me a bit as I've been playing around with Quake 2 RTX a lot and seeing the potential - but seems AMD are just making some optimisations to existing hardware to be able to run it a bit rather than putting the effort this round into proper RT acceleration.


There is no such thing as improper RT acceleration. There are just more 'Optimized" ways of doing it. Finding new innovative ways of doing something is what drives technology forward. Console developers are not going to use large expensive chips to drive an "unoptimised" technology, so that technology will not become mainstream, just a normal technology feature.

What AMD are doing is normalizing Ray Tracing.
 
Annoys me a bit as I've been playing around with Quake 2 RTX a lot and seeing the potential - but seems AMD are just making some optimisations to existing hardware to be able to run it a bit rather than putting the effort this round into proper RT acceleration.

Yet, looking at the Patents there seems to be a definite hardware element to AMD's Ray Tracing solution. So maybe there is something lacking in Sony's solution. The GPU has been custom built for them and it's their software. Maybe DXR 1.1 on the PC/xbox side will be better.

I continue to hope despite my concerns :)
 
Yet, looking at the Patents there seems to be a definite hardware element to AMD's Ray Tracing solution. So maybe there is something lacking in Sony's solution. The GPU has been custom built for them and it's their software. Maybe DXR 1.1 on the PC/xbox side will be better.

I continue to hope despite my concerns :)

AMD added an instruction set and some sort of queueing or algorithm to allocate Ray Tracing tasks to shaders that would otherwise be idle and wasted, its an innovative efficient use of the hardware that is right there in the GPU anyway, there is no need for adding dedicated chunks of hardware.
 
What AMD are doing is normalizing Ray Tracing.

What do you mean by this statement?

What are AMD normalising? They have created their own solution just like Nvidia. They are using DXR, just like Nvidia. AMD's isn't using Microsoft's software fallback layer, they have their own solution that is called by in by DXR as needed by the game.

Console developers are not going to use large expensive chips to drive an "unoptimised" technology, .

I don't know about Sony, but on the Xbox/PC gaming developers are going to be designing games to run in DXR.
 
Yet, looking at the Patents there seems to be a definite hardware element to AMD's Ray Tracing solution. So maybe there is something lacking in Sony's solution. The GPU has been custom built for them and it's their software. Maybe DXR 1.1 on the PC/xbox side will be better.

I continue to hope despite my concerns :)

AMD has hardware elements but the information I've seen so far doesn't seem to indicate they will approach it like nVidia - instead just make it so that their existing architectures can do some degree of ray tracing on the side using existing shaders so to speak but that is a far inferior solution to what nVidia are doing as even best case reusing existing shader functionality like that is far less efficient than the kind of approaches nVidia are using in hardware and means on AMD it is gimped to handling a small number of features in a scene rather than being able to move towards most or all of the scene being rendered that way.
 
I alluded to it in the rest of what i wrote. RTX cards are expensive because the GPU is large, a good chunk of space is taken up but the dedicated Ray Tracing portion.

AMD GPU's do not have that dedicated Ray Tracing portion on the die, instead they added a Ray Tracing Instruction to the GPU's existing shaders and built an algorithmic or prediction engine of some sort that can allocate Ray Tracing tasks to the existing shaders where they would otherwise be idle and wasted, which a lot of the time they are.

With that AMD are Ray Tracing without dedicated RT cores.

AMD's RDNA 2 GPU's Like Turing and Ampere will run Ray Tracing in DXR, the software or API between AMD, Nvidia and even the consoles are the same thing, what's different is how AMD and Nvidia achieve the same thing at the hardware level.
 
AMD added an instruction set and some sort of queueing or algorithm to allocate Ray Tracing tasks to shaders that would otherwise be idle and wasted, its an innovative efficient use of the hardware that is right there in the GPU anyway, there is no need for adding dedicated chunks of hardware.

But, AMD's solution still needs extra hardware, it doesn't just use existing GPU hardware to do all the Ray Tracing, there is extra hardware needed to do the BVH calculations that can be bypassed if not needed.
 
But, AMD's solution still needs extra hardware, it doesn't just use existing GPU hardware to do all the Ray Tracing, there is extra hardware needed to do the BVH calculations that can be bypassed if not needed.

it doesn't just use existing GPU hardware to do all the Ray Tracing

This needs explaining...
 
AMD has hardware elements but the information I've seen so far doesn't seem to indicate they will approach it like nVidia - instead just make it so that their existing architectures can do some degree of ray tracing on the side using existing shaders so to speak but that is a far inferior solution to what nVidia are doing as even best case reusing existing shader functionality like that is far less efficient than the kind of approaches nVidia are using in hardware and means on AMD it is gimped to handling a small number of features in a scene rather than being able to move towards most or all of the scene being rendered that way.

Yes, AMD's solution does have hardware elements.

Whether it's an inferior solution or not remains to be seen. It certainly looks like a neater solution.
 
Yes, AMD's solution does have hardware elements.

Whether it's an inferior solution or not remains to be seen. It certainly looks like a neater solution.

Only looks neater IMO as it integrates more with existing functionality but that isn't actually a good thing as you are penalised ultimately in performance even though it brings a greater level of performance than currently doing it on shaders allows for.
 
I alluded to it in the rest of what i wrote. RTX cards are expensive because the GPU is large, a good chunk of space is taken up but the dedicated Ray Tracing portion.

AMD GPU's do not have that dedicated Ray Tracing portion on the die, instead they added a Ray Tracing Instruction to the GPU's existing shaders and built an algorithmic or prediction engine of some sort that can allocate Ray Tracing tasks to the existing shaders where they would otherwise be idle and wasted, which a lot of the time they are.

With that AMD are Ray Tracing without dedicated RT cores.

AMD's RDNA 2 GPU's Like Turing and Ampere will run Ray Tracing in DXR, the software or API between AMD, Nvidia and even the consoles are the same thing, what's different is how AMD and Nvidia achieve the same thing at the hardware level.

This needs explaining...

Answering both of these together :)

It's not that large of chunk, RT cores use around 6% of the die space according to D.P. Tensor Cores and RT cores together take up around 15%.

AMD cards will have dedicated hardware for Ray Tracing. Not as much as Nvidia, but it isn't all done in software. It would be far too slow to do without some kind of hardware dedicated to Ray Tracing calculations.
 


These are the games that will have ray tracing. We already knew it wouldn't be brute force ray tracing. To think that was a bit of an extreme.
However, it's still offering ray tracing. So the question we need to consider? Will the average gamer, those who want to play games, pay over a grand for just a card offering subtle difference in RT (where you need to screenshoot and blow up)? Or pay the console price and get the immersion they expect from next gen consoles.

At the end of the day, not everyone is brand loyal when it comes to graphics card.
 
Answering both of these together :)

It's not that large of chunk, RT cores use around 6% of the die space according to D.P. Tensor Cores and RT cores together take up around 15%.

AMD cards will have dedicated hardware for Ray Tracing. Not as much as Nvidia, but it isn't all done in software. It would be far too slow to do without some kind of hardware dedicated to Ray Tracing calculations.

AMD's solution allows for a hybrid approach for more optimal results but much of the work they are doing hardware wise is opening up the ability to repurpose existing hardware functionality for ray tracing purposes without the bottlenecks you have in Pascal, etc. but ultimately it falls short of the efficiency of dedicated RT hardware by some margin.
 
Only looks neater IMO as it integrates more with existing functionality but that isn't actually a good thing as you are penalised ultimately in performance even though it brings a greater level of performance than currently doing it on shaders allows for.

The performance of this solution if purley speculation on your part presented as fact, Nvidia's solution is 1:2 if not 1:1 resolution but the performance hit is 40%, we can't know yet but there are apparent slides that Ray Tracing on RDNA 2 GPU's has a performance hit of ~10%.

There is no reason why the resolution must be at 1:1 all the time or at all, not unless you want perfectly mirrored reflection's which in my view have in these RTX games looked completely ridiculous.
 
Status
Not open for further replies.
Back
Top Bottom