AMD Radeon RDNA2 Navi (RX 6500, 6600, 6700, 6800, 6900 XT)

Possibly, that would explain the almost linear scaling between the 3080 and 3090 too.

You can calculate GI at a lower pixel resolution but it certainly requires more rays per pixel than mirror reflections. GI also lends itself better to stochastic multi-frame accumulation which reduces requirements a bit. In the end though I don't think reflections will be a worst case scenario for either AMD or Nvidia.

 

Says nothing about real time though.

 

Mirror-like reflections are not exactly a worst case scenario. Glossy reflections/refractions and GI on the other hand generate more challenging ray distributions.

 

https://docs.cryengine.com/pages/viewpage.action?pageId=25535599

How much image quality does hardware accelerated ray traced GI bring in comparison? At what performance?

 

Nice, but reminds me of next-gen console visuals.

 

Probably duo to budget increase, nowadays they are heavy in black while in most of the last decade it was a question if they will survive till next year.

 

Thats somewhat tangential isn’t it? Most developers have kicked off the RT era with reflections because the visual impact is most obvious and there aren’t very good alternatives. Cube maps suck and SSR can’t see anything off screen. GI has a few promising screen space and sparse world space solutions with RT being the most accurate but not necessarily the best choice.

 

The impact of the shape of the object is identical with the impact of "roughness", or the impact of more bounces. A skyscraper city of boxes is a lot more performing than a skyscraper city of cylinders or tori. And a city of perfect mirror cylinders can perform identical to a city of blurry glass boxes. Additionally, you'd have to factor in the quality of the BVH, but that is a much worst can of worms really (it's NP-hard).

In general, in raytracing, for performance, what matters is the divergence (or convergence) of directions, because that plays into predictability and locality of memory locations / access. You can practically use the metric of entropy (from compression) to measure ray redundancy, and that redundancy leads to better memory / data access. Try to compress a random data-set, good luck. So, if you pick the perfect rays, you always pay the full trip to memory, because they are truly unique, and perfectly distributed (from an information point of view) thus non-local and unpredictable.

Don't forget that texturing is also at times unpredictable, and an existing bandwidth consumer which can't really be eliminated; or you use cell-shading.

Only if you're in the totally scewed area of very little rays, or highly redundant rays will you find that TFLOPS are useful, because access is not the dominating part.

 

Anthony slips out "6800 is closer to 3080 than 3070"

 

Because it will be slower with RT.

 

I wouldnt make conclusions based on first batch of next gen games, these are almost universally optimized badly for Nvidia and AMD as well.

 

I like that AMD are doing demos for the graphics cards again, but I kind of miss the girl that ATI used to use.

Regards,
SB

 

Ruby.

 

Ruby has been cybernetically enhanced, haven't you heard?

 

Still have Dangerous Curves on my hdd somewhere I guess such demos won't fly in the current_year puritan socio-political climate

 

"We're proud of the great strides our company has made with RDNA2 and Zen3. Now sit back and enjoy our demo, Teenage Wank Fantasy III"

 

Didn't Nvidia just delete all their old demos that had "females" due to some sort SJW movement/backlash?

 

What is Vulkan Geekbench benchmark actually benching?

That was a mouthful.

EDIT: forgot to link the actual benchmark: https://browser.geekbench.com/v5/compute/1847356