You are the one that is trying to use stupid comparisons when the guy that actually experienced both of them says that is not the case.
Math Fight!
I dunno, maybe Digital Foundry, which is the one that has seen both solutions on hand is actually the one that can make a good correlation instead of using a flawed model of looking at specs and thinking its all that matters, despite Nvidia generally trouncing all AMD solutions at "same" spec sheets.
Ok then when the RT performance is not 4x as better as the 2080Ti I expect you to come here and say paper performance has no meaning.
The raytracing they chose to use is probably one of the most expensive as well as minecraft (despite its looks) have a ton of interaction elements that would be active with raytracing if they were to be activated. Other games have much more controlled environments were the surfaces (and objects) were raytracing is active are much smaller (even then you will have performance problems).
news.xbox.com
That just says cutscenes are at 4k, in the same way that the Xbox One X has the cutscenes at 4k (but 30fps). And then we see something like Digital Foundry saying that there was similar perfomance to a 2080RTX. What we will see is a lower resolution with a pretty good scaler (which has improved a ton in the past gen)Xbox Series X: A Closer Look at the Technology Powering the Next Generation - Xbox Wire
A few months ago, we revealed Xbox Series X, our fastest, most powerful console ever, designed for a console generation that has you, the player, at its center. When it is released this holiday season, Xbox Series X will set a new bar for performance, speed and compatibility, all while allowing...
settings beyond ultra, cutscenes running at a fixed 4k60 (which cutscenes in gears 5 run at a capped framerate, thats normal) game running OVER 100fps, and they detail which settings are beyond ultra, or at least some of them.
That just says cutscenes are at 4k, in the same way that the Xbox One X has the cutscenes at 4k (but 30fps). And then we see something like Digital Foundry saying that there was similar perfomance to a 2080RTX. What we will see is a lower resolution with a pretty good scaler (which has improved a ton in the past gen)
Just a headsup of what jump you are talking when you say 4k100fps about because I am not sure you understand it:
4k 100 fps is *at best* 2 graphic cards generations away, and even then it makes no fucking sense because HDMI 2.1 only does 4k60fps. They were talking about getting to high refresh rates... for other resolutions.
Holy shit. I missed the 100fps remark yesterday.Xbox Series X: A Closer Look at the Technology Powering the Next Generation - Xbox Wire
A few months ago, we revealed Xbox Series X, our fastest, most powerful console ever, designed for a console generation that has you, the player, at its center. When it is released this holiday season, Xbox Series X will set a new bar for performance, speed and compatibility, all while allowing...
settings beyond ultra, cutscenes running at a fixed 4k60 (which cutscenes in gears 5 run at a capped framerate, thats normal) game running OVER 100fps, and they detail which settings are beyond ultra, or at least some of them.
and this is only done in a few weeks, so it hasnt even been optimized yet. So uh.... yeah..... 2080Ti does 60fps at 4k ultra in the gears benchmark, it runs around 100fps indoors and around 60 - 80 outdoors and in intense firefights, and this is doing better than that, with higher settings and no optimization. It bodes very well and makes comparisons to a 2080 a bit absurd. 2080 only does like 45fps at 4k ultra in gears 5 for reference.
Xbox Series X: A Closer Look at the Technology Powering the Next Generation - Xbox Wire
A few months ago, we revealed Xbox Series X, our fastest, most powerful console ever, designed for a console generation that has you, the player, at its center. When it is released this holiday season, Xbox Series X will set a new bar for performance, speed and compatibility, all while allowing...
settings beyond ultra, cutscenes running at a fixed 4k60 (which cutscenes in gears 5 run at a capped framerate, thats normal) game running OVER 100fps, and they detail which settings are beyond ultra, or at least some of them.
You are right on HDMI 2.1, I always get it wrong.
Yup, it doesn't state 4K 100fps, it just states 100fps, which could be any resolution really, maybe 1440p upscaled.
A cutscene is very different than real game performance, as many of the more complex calculations are turned off. The Xbox One X runs cutscenes at 4k30fps for instance...
Please stop, you just picked a common multiplayer and claiemd it's industry standard 10 BVH depth, which is a blatant lie, ONCE MORE, NVIDIA doesn't use BVH depth to calculate Gigarays. Read their whitepaper for God's sake.
There's definitely missing context, but all of the direct statements we have say its at the very least running faster than a 2080 does, at higher settings than pc is even capable of running at. Maybe someone who actually uses twitter should tweet at him and see if we can get clarification or not.
They state their GigaRay figure is based off the combination of FP32 + INT/FP execution + RT cores, there is no mention whatsoever of your claimed BVH depth, which you are still failing to give us a link for in the first place, you can't even prove 10 depth is the industry standard.
Since I didn't really get a reply last time let me try and guess this "formula" of yours: you're basing this on the fact that there's only one RT core in Turing's SM but four TMUs in Navi's CU? What makes you think that the BVH testing h/w in RDNA2 is tied to each of four TMUs?
You're getting greedy with what to expect lol, the fact that it's even matching a 2080 is insane, don't push it to comparing against a 1200 dollar card.
Cutcenes are more demanding in Gears 5 than general gameplay. If, and that's a big if, the Series X is running the game at 4K 60 in cutscenes, then 4K 100fps would be very plausible.
A number of the PC settings have a level beyond ultra, called insane. I'd imagine they'd reference that if it was using those settings.
Considering it's running at a locked 4k60 on SX (which means 60 is the minimum framerate) I doubt it would run at only 100fps @1080p or that they would have any problems at all to get to 1080p 120.
Where is the depth level of 10 you claim?
WRONG, search Giga Ray and you will find dozens of references to it. Apparently you just skipped over like you always do, a sign of someone who doesn't do accurate math.
Not really. There is no sign of each of RDNA2 TMU containing a dedicated BVH unit inside them.
That's because this data is based on the same basics of the number of RT cores in each of said GPUs. Again, this means nothing really.
Because it's unknown at this point. You are the only one claiming to know it!
Because you picked a common multiplayer and ran with it.
Yes indeed it DOES NOT, which means they don't really use BVH depth as you claimed. You are the only one claiming to know a "formula" that NVIDIA doesn't even mention in the first place.
Basically you are dancing around the issue of flawed math .. NVIDIA doesn't use your formula to calculate anything, they use primary rays in a standard bunny test, with tera ops of ray tracing as their basis of math, their 10 Giga Rays figure is just that, marketing. You can't compare that to the figure Microsoft used (Intersection tests).
Exactly, there is also INT32 in the case of Turing.
He doesn't account for anything except his recently discovered "formula". He is too fixated on his claimed math, that he ignores basic rendering technical knowledge.
Basically you are dancing around the issue of flawed math .. NVIDIA doesn't use your formula to calculate anything, they use primary rays in a standard bunny test, with tera ops of ray tracing as their basis of math, their 10 Giga Rays figure is just that, marketing. You can't compare that to the figure Microsoft used (Intersection tests).
The only valid comparison points is the TF number because it's universal, Microsoft claimed 13 TF for the Series X, NVIDIA claimed 25TF for the 2080, hence why NVIDIA has more RT performance than RDNA2, which is showing in the Minecraft demo.
Glad we got that out of the way.
!!!!
That one is easy: NVIDIA takes the number of RT cores X GPU frequency , then divides the number by a factor to account for the role of RT cores in each frame.
Glad we got that out of the way.
!!!!
Please kindly show me how many NVIDIA intersection test are there in a Turing GPU?
That one is easy: NVIDIA takes the number of RT cores X GPU frequency , then divides the number by a factor to account for the role of RT cores in each frame.
So in the case of the RTX 2080: we have 46 RT cores x 1740MHz frequency, which would give you 80 TFLOPS of RT math, then they divide that number by a factor of 3.47 to arrive at the quoted 23 TF of traditional compute.
Microsoft did calculate their TF number and came up with the 13TF figure, NVIDIA did the same thing before them and came up with the 25TF figure, the burden is on Microsoft to prove their solution is faster, which they don't claim, ONLY you are claiming that.
And In real world workloads, the difference is obvious, clearly series x has less RT performance than a 2080 as shown in MineCraft.
