(Official?) Battlefield V with RTX benchmark

[Mira Yurizaki]

Maybe related maybe not. But something from the past to remember going forward:

 

 

[CarlBar]

30 minutes ago, M.Yurizaki said:

Maybe related maybe not. But something from the past to remember going forward:

 

 

 

Yeah only those of us who weren't complete idiots didn't fully expect this. We've seen this before whenever a new version of DX or some other improvement came out. 1 to 2 years from now when NVIDIA has finished optimizing their drivers for the 20 series platform and Game devs have got a better idea of how to get the best performance out of the tools they have we'll see 1440p 60hz RT being doable on a 2080Ti.

 

But no one with any memory of history excepted playable performance at anything but the lowest resolutions with it enabled on release day.

[Mira Yurizaki]

31 minutes ago, CarlBar said:

 

Yeah only those of us who weren't complete idiots didn't fully expect this. We've seen this before whenever a new version of DX or some other improvement came out. 1 to 2 years from now when NVIDIA has finished optimizing their drivers for the 20 series platform and Game devs have got a better idea of how to get the best performance out of the tools they have we'll see 1440p 60hz RT being doable on a 2080Ti.

 

But no one with any memory of history excepted playable performance at anything but the lowest resolutions with it enabled on release day.

I also found benchmarks of the same game, same settings, but a year later when benchmarking the GeForce 580 and Radeon HD 6970. There was a huge positive difference.

 

Though I'm too lazy to find it right now

[Drak3]

6 hours ago, CTR640 said:

Ye sorry but 1080p is ugly when you have experienced 1440p and up. Never gonna buy a €1400 GPU just to go back from 1440p to 1080p lmfao.

I regularly bounce between 1440p and 900p. It's rather easy.

[CarlBar]

Some interesting results out of this video. Seems that whilst peak frame rate drops a lot the lows aren't as far off the peak, they're still low enough to be concerning, but getting 60FPS consistent is going to possibly be a lot easier than we thought optimization wise.

 

 

[leadeater]

7 hours ago, The Benjamins said:

IDK, but it appears that it is bottlenecked by the GPU portion that handles ray tracing, which also points to the hardware being too new and needs work.

This is just DXR so no ray tracing at all, just up sampling on Tensor cores.

[leadeater]

7 hours ago, Morgan MLGman said:

This graph is a lot more interesting IMO. I mean, who's going to buy an RTX 2080Ti and play on a 1080p monitor? That's like buying a supercar and driving it with a 60MPH speed limiter of some sort.

Even at 1440p, the performance is barely over 40FPS with anything over DX12 + DXR Low enabled. At 4K, it's unplayable beyond that threshold. Note that this is a 1000$+ GPU that's marketed as a 4K GPU I don't remember Nvidia mentioning that it's only a 4K GPU if you don't use the ray-tracing features that they also market these cards with.

 

Edit: Dumb comment contained blow, was thinking DLSS lol.

Spoiler

 

One of the prevailing arguments for DXR is to upscale the image to 4K with internal render at 1080p, for higher FPS. That means you are still putting native resolution output to the monitor.

 

However this graph shows us something very interesting, 1080p DXR on any setting is worse than native 4K rendering to that's RIP to that DXR play book.

 

 

Edited November 15, 2018 by leadeater

[Princess Luna]

32 minutes ago, leadeater said:

.

So DXR is the whole "miracles" DLSS that would turn 1440p into 4k with no performance impact? I'm confused myself on what's going on... despite getting a RTX card myself I never really got into Ray Tracing I just wanted the rasterization performance haha, I also don't play battlefield so I have nothing to base on... but damn it's looking worse than I could've expected...

 

Feels odd knowing that somewhat half of my GPU is pretty much useless

[asus killer]

34 minutes ago, leadeater said:

This is just DXR so no ray tracing at all, just up sampling on Tensor cores.

"Included with this update will be the first release of DXR ray-traced reflections, which add lifelike, cutting-edge real-time reflections to surfaces and objects in Battlefield V’s epic 64-player multiplayer maps and cinematic War Stories. With DXR enabled, graphical fidelity, realism and immersion are elevated to previously unobtainable levels"

https://www.nvidia.com/en-us/geforce/news/battlefield-v-rtx-ray-tracing-out-now/

 

i'm not saying you're wrong but i'm confused.

[The Benjamins]

46 minutes ago, leadeater said:

This is just DXR so no ray tracing at all, just up sampling on Tensor cores.

DXR is ray tracing.

https://www.nvidia.com/en-us/geforce/news/battlefield-v-rtx-ray-tracing-out-now/

DXR is Direct X Ray tracing.

DLSS is the up scaling tech

 

[Sauron]

9 hours ago, voiha said:

DRX Medium : 64.5fps
DRX High: 66.4fps
DRX Ultra: 65.3fps

How is this possible ? Am I misunderstanding something ?

 

 

It's probably within the margin of error of the test. These scores are usually obtained by having the journalist run through a stage, trying to replicate the same movements as closely as possible every run. 2-3 FPS differences can generally be attributed to the journalist being a mere human. What's more likely happening is that there isn't much difference at all between medium/high/ultra DXR, or the card is hitting a specific bottleneck that prevents it from going higher even though parts of the RT cores can be used for higher detail levels.

[bitsandpieces]

And AMD expressed wishes to bring DXR to the midrange, good luck with that 

[mr moose]

4 hours ago, M.Yurizaki said:

Maybe related maybe not. But something from the past to remember going forward:

 

 

Yep, I don't know why anyone bothered to push forward with AA, clearly it was a failure of a technology that should never have been pursued. oh wait...  

[aezakmi]

benchmarking in 1920x1080 with THAT card? they gotta be fucking kidding, even my 14 year old monitor has a higher resolution.

 

no sane person would spend $2000 on a card to play on 1080p, I think whoever owns a 2080Ti has AT LEAST a 2160p screen.

 

 

[leadeater]

2 hours ago, asus killer said:

"Included with this update will be the first release of DXR ray-traced reflections, which add lifelike, cutting-edge real-time reflections to surfaces and objects in Battlefield V’s epic 64-player multiplayer maps and cinematic War Stories. With DXR enabled, graphical fidelity, realism and immersion are elevated to previously unobtainable levels"

https://www.nvidia.com/en-us/geforce/news/battlefield-v-rtx-ray-tracing-out-now/

 

i'm not saying you're wrong but i'm confused.

No I was confused lol, was thinking DLSS. Need more sleep at night haha.

 

1 hour ago, The Benjamins said:

DXR is ray tracing.

https://www.nvidia.com/en-us/geforce/news/battlefield-v-rtx-ray-tracing-out-now/

DXR is Direct X Ray tracing.

DLSS is the up scaling tech

See above, 

[pas008]

1 hour ago, aezakmi said:

benchmarking in 1920x1080 with THAT card? they gotta be fucking kidding, even my 14 year old monitor has a higher resolution.

 

no sane person would spend $2000 on a card to play on 1080p, I think whoever owns a 2080Ti has AT LEAST a 2160p screen.

 

 

 

resolution isnt the only way to improve visuals if you actually paid attention and compared what rt really does visually

 

[williamcll]

Didn't they marketed it as 108060fps ray tracing OR 4k60fps back in the original presentation?

[Mira Yurizaki]

2 hours ago, mr moose said:

Yep, I don't know why anyone bothered to push forward with AA, clearly it was a failure of a technology that should never have been pursued. oh wait...  

I don't know how to react to this :c

[Maticks]

thats also a benchmark folks, we all know real games will be worse than this.

at 40fps a FPS becomes more a turn based game. ouchy. i am assuming its a driver issue not what these cards are capable of.

That would really kill RTX before it got of the ground. 

I would expect 65-70FPS be the spot they want to hit given its fast enough to feel fluid obviously not 100FPS fluid but smooth enough.

 

If this is what we expect 40FPS ouch, no one will be turning on ray tracing.

[Mira Yurizaki]

Well in light of @mr moose's comment, I did want to see what was a factor in DirectX 11 performance vs. DirectX 10 since I've pulled up more benchmarks that paint a better picture. The only correlation I can find was one of DirectX 11's most touted feature, tessellation, could wreck havoc with performance.

 

 

 

Though to be fair, AMD's tessellator wasn't as good as NVIDIA's at the time. And some games that did use tessellation, like Deus Ex: Human Revolution, did so only on fewer elements so the performance impact wasn't that bad at all. Also, NVIDIA did make special hardware for Tessellation they called the Polymorph Engine.

 

Ah well, I don't know what I'm trying to get at here other than we've seen this before in a different feature.

[Speed Weed]

RTX graphics cards are still expensive. 

[mr moose]

1 hour ago, M.Yurizaki said:

I don't know how to react to this :c

My point was more just that AA in the early days caused a big drop in FPS.  Now if people are claiming that because RTX has such a massive hit to FPS that it is a failure of a technology,  then shouldn't the same logic should apply to AA? Did these people honestly think AA was going to fail because it was perfect out of the gate?

[Mira Yurizaki]

2 minutes ago, mr moose said:

My point was more just that AA in the early days caused a big drop in FPS.  Now if people are claiming that because RTX has such a massive hit to FPS that it is a failure of a technology,  then shouldn't the same logic should apply to AA? Did these people honestly think AA was going to fail because it was perfect out of the gate?

Well, AA was used, but the graphs were also showing the difference between DX10 and DX11.

 

That's why I also went out and found some other benchmarks.

[mr moose]

2 minutes ago, M.Yurizaki said:

Well, AA was used, but the graphs were also showing the difference between DX10 and DX11.

 

That's why I also went out and found some other benchmarks.

I definitely wasn't trying to say there was some intrinsic error in your post, quite the contrary, I found it a very relevant look at how new GPU technology becomes a thing and how well it performs initially versus better optimizations and uptake. 

[Mira Yurizaki]

7 minutes ago, mr moose said:

I definitely wasn't trying to say there was some intrinsic error in your post, quite the contrary, I found it a very relevant look at how new GPU technology becomes a thing and how well it performs initially versus better optimizations and uptake. 

Literally at this point, the only complaint I'll accept is NVIDIA's charging too much for this tech. Which I think they are.

 

Everything else is just repeating history.