GPU Ray Tracing Performance Comparisons [2021-2022]

Nah, scenes with RT reflections and high amount of RT lights lean more towards the 50% gap. This is only in the new edition.
From the benchmarks I've seen of Cyberpunk/Control, the performance is at least 50% better on the 3080 when everything is on Ultra. So if performance only dropped heavily in select scenes, it would be an improvement. I guess we will have to wait until DF can do a performance analysis of the whole title.
 
Very interesting. Was expecting the opposite
This would have been unexpected for RT heavy game with per pixel RayTracing.
But in this case it's likely that BVH building takes the most time and its cost is fixed and independent of resolution. If they are also doing something like RTXGI for GI, then ray tracing portion is also non depend on resolution. So increasing resolution would make raster portion of frame larger and RT portion of frame smaller, hence the scaling you see in the gamegpu results.

Just checked numbers and yep, that's the case, rasterization is 65% faster in 1080p, RT workload takes 2.32 ms.
In 4K, rasterization only mode is just 20% faster, and RT takes just 1.9 ms. That's for 3090

On 6900 XT, RT portion takes 4.47 ms in 1080p and 3.9 ms in 4K.

Probably RT is slightly limited by CPU in 1080p (this would explain slightly higher RT cost in 1080p)
 
Last edited:
This would have been unexpected for RT heavy game with per pixel RayTracing.
But in this case it's likely that BVH building takes the most time and its cost is fixed and independent of resolution. If they are also doing something like RTXGI for GI, then ray tracing portion is also non depend on resolution. So increasing resolution would make raster portion of frame larger and RT portion of frame smaller, hence the scaling you see in the gamegpu results.

Just checked numbers and yep, that's the case, rasterization is 65% faster in 1080p, RT workload takes 2.32 ms.
In 4K, rasterization only mode is just 20% faster, and RT takes just 1.9 ms. That's for 3090

On 6900 XT, RT portion takes 4.47 ms in 1080p and 3.9 ms in 4K.

Probably RT is slightly limited by CPU in 1080p (this would explain slightly higher RT cost in 1080p)
Pretty neat. This is a great response.
I guess my only question is why would the RT costs be higher in 1080p and less in 4K for both systems. I just don't understand that part. Shouldn't RT costs be higher with higher resolution? I'm a bit lost here on this even though I understand that that RT costs should be fixed let's say; I still don't understand why we are seeing the discrepancy in RT numbers between 1080p and 4K on the same card.
 
Yeah we have so little real data about what part of the system does what, and even when we do have data it will only tell us something about that particular engine, and then with the PC game optimized drivers it’s even harder sometimes.
 
Resident Evil 8 Ray Tracing tests:

1-The game has two RT effects, GI and Reflections.
2-The game is optimized for RDNA2, which translates to a limited application of RT reflections, RT GI is quite good though but still limited.
3-RDNA2 GPUs lose 29% of performance running RT, NVIDIA GPUs lose 17%.
4-Generally speaking a 3090 is 15% to 25% faster than 6900XT @4K depending on the scene.

https://www.computerbase.de/2021-05...mm-resident-evil-village-3840-2160-raytracing
https://wccftech.com/resident-evil-village-ray-tracing-impact-explored/
https://www.guru3d.com/articles_pag..._graphics_performance_benchmark_review,7.html
 
I wonder how much you can optimize RT for Rdna2, or Ampere, on PC where direct x/rt is limiting what you can do anyway... Invisible RT is not an optimisation...
 
Back
Top