Metro Exodus Benchmark Performance, RTX & DLSS

[Mescalamba]

So, uh, how valuable are those user scores right now?

Exception for once. Bombing reviews isnt norm. Nor should be.

Beside if someone isnt idiot, they will figure where is truth too.. Using internet, especially today does require some common sense.

 

[mobiuus]

what is the differences between ultra and extreme graphic setting?
besides fps drop i can't notice any visible change...
perhaps higher antialiasing??

 

[Captain_Tom]

It's not just about Vram though. 580 is a faster card than 3GB 1060 in most things

There is no point in trying to reason with that guy lol. He thinks Infinity Fabric is a failure... and in fact he thinks he is so right that he doesn't see the folly in people who disagree with him in his signature.

 

[erixx]

I hope to see (somewhere) numbers for the 1080 Ti SLI vs the new topend card...

With ultra settings, 4K, and G-Sync, 60Hz, I get mostly 55 FPS.

 

[Frutika007]

Gamer's Nexus got 79fps with 1080p ultra(rtx high) while TPU got only 59fps on that same settings with RTX 2060. And people are saying TPU is showing more fps than actual and calling them nvidia shill?? LOL. It's the complete opposite. TPU is showing lower fps than what others actually got.

 

[Super XP]

Gamer's Nexus got 79fps with 1080p ultra(rtx high) while TPU got only 59fps on that same settings with RTX 2060. And people are saying TPU is showing more fps than actual and calling them nvidia shill?? LOL. It's the complete opposite. TPU is showing lower fps than what others actually got.View attachment 116567

Are test systems identical?

 

[erixx]

My FPS is just subjective, looking at value during gameplay. I just upped the settings to Extreme and FPS lowered to 40-50. Now it starts to feel uncomfortable. Ultra was nice (no waves of enemies).
I hear SLI calling. I found a used normal evga sc2 1080ti for 570 yourows, lol

 

[bug]

Are test systems identical?

TPU uses the i7-8700k and 16 GB of RAM, GN uses the i7-8086k and 32 GB of RAM.
The system doesn't matter much though, this is GPU bottlenecked scenario. The answer you're looking for is both sites have tested in-game footage which is obviously different. GN usually uses pre-baked benchmarks (and they say they will revert to that), they just made an exception this time because they were actually looking at pre-baked vs real game play differences. They found out the built-in is unrealistically heavy. That's ok in my book, but that wasn't @R4WN4K's point.

 

[EarthDog]

Are test systems identical?

test system dont make that much of a difference unless one of the test systems was a potato.

 

[jabbadap]

Steve does not use hairworks, W1z does.

 

[erixx]

I went to DX11 from DX12 and the fishes in far away river look much better from a tower (rephrasing, they looked bad in DX12 and look as should in DX11).
I got 58 fps.
Then under DX11 I lowered Shader level from 1.0 to 0.8 and now I have a stable 60 fps (v-sync and g-sync on) (1080ti light OC)
Then again back to DX12, Shader level 0.8 and it stays at fixed 60 FPS and fishes look good.... Whatever it might be (may it was the Extreme setting). I'll keep it a while.

 

[bug]

I went to DX11 from DX12 and the fishes in far away river look much better from a tower (rephrasing, they looked bad in DX12 and look as should in DX11).
I got 58 fps.
Then under DX11 I lowered Shader level from 1.0 to 0.8 and now I have a stable 60 fps (v-sync and g-sync on) (1080ti light OC)
Then again back to DX12, Shader level 0.8 and it stays at fixed 60 FPS and fishes look good.... Whatever it might be (may it was the Extreme setting). I'll keep it a while.

Ah, the lost art of looking at other setting besides presets. Respect, sir,

 

[erixx]

I finished totally seasick after playing several hours so tweaking is -still- a must, sir!

 

[Nkd]

are you using the builtin benchmark to run the test? I believe in game you get more but built in benchmark gave me 81fps at 1440p ultra without any RTX on and stuff on my 2080ti. and little over 60 fps with rtx high on 2080ti. That is with buit in benchmark though. The benchmark does seem to be on the heavy side and worst case.

 

[W1zzard]

are you using the builtin benchmark to run the test? I believe in game you get more but built in benchmark gave me 81fps at 1440p ultra without any RTX on and stuff on my 2080ti. and little over 60 fps with rtx high on 2080ti. That is with buit in benchmark though. The benchmark does seem to be on the heavy side and worst case.

As mentioned several times before, I do not use the in-game benchmark, I'm using actual gameplay (in open world)

 

[Mescalamba]

As mentioned several times before, I do not use the in-game benchmark, I'm using actual gameplay (in open world)

I know you probably dont need confirmation, but what you doing is best. Benchmark in games are often not very telling in terms of how actual gameplay looks like.

 

[bug]

I know you probably dont need confirmation, but what you doing is best. Benchmark in games are often not very telling in terms of how actual gameplay looks like.

Built-in benchmarks aren't meant to be that telling (one could argue recorded bits aren't either). The point about built-in benchmarks is they're repeatable.

 

[EarthDog]

Built-in benchmarks aren't meant to be that telling (one could argue recorded bits aren't either). The point about built-in benchmarks is they're repeatable.

People have lost site of the fact that integrated benchmarks aren't direct analogs to game play anyway. What they do provide is an empirical method of testing where results are 100% repeatable and show relative performance between a group of cards. It shouldn't be translated directly to FPS in-game, regardless.

 

[bug]

People have lost site of the fact that integrated benchmarks aren't direct analogs to game play anyway. What they do provide is an empirical method of testing where results are 100% repeatable and show relative performance between a group of cards. It shouldn't be translated directly to FPS in-game, regardless.

I'd say built-in benchmarks are representative of gameplay. It's just that they tend to be a worst case scenario, otherwise the developer would stand accused they try to make it seem their game performs better than it actually does. I don't have a problem with that, but people don't have all that in mind. people tend to just look at numbers. If they don't like the numbers, it's either the reviewer's fault (somehow) or the other camp has paid to gimp the game for their video cards. It's so predictable it's not even boring anymore.

 

[EarthDog]

I think it gives a good idea, in general... like close enough...horsehoes/hand grenades style, but the real takeaway is the relative distance between cards for each benchmark.

It's so predictable it's not even boring anymore.

It isnt always like that... it depends on where (what forum) you are lay your head.

 

[bug]

It isnt always like that... it depends on where (what forum) you are lay your head.

I admit I don't frequent that many forums. But the ones I do are like that.

 

[megaclite]

Also Guru3D have Radeons doing MUCH better, I'm so confused:

Because TPU fake

 

[londiste]

Interview with 4A guys has some nice details and thoughts:
Tech Interview: Metro Exodus, ray tracing and the 4A Engine's open world upgrades (EuroGamer Digital Foundry)

Ray tracing has two quality settings: high and ultra. Ultra setting traces up to one ray per pixel, with all the denoising and accumulation running in full. The high setting traces up to 0.5 rays per pixel, essentially in a checkerboard pattern, and one of the denoising passes runs as checkerboard. We recommend high for the best balance between image quality and performance, but please note that we are still experimenting a lot, so this information is valid only at the time of writing.

 

[bug]

Interview with 4A guys has some nice details and thoughts:
Tech Interview: Metro Exodus, ray tracing and the 4A Engine's open world upgrades (EuroGamer Digital Foundry)

This pretty much sums up the whole discussion about RTX: it's in its infancy. If you want to be an early adopter (and can afford the asking price), go ahead and show your support. If you like things ironed out for you, then wait.
But no, this level of common sense cannot be expected on the Internet. We need to go all sjw and point out how companies are evil and such instead.

 

[dirtyferret]

People have lost site of the fact that integrated benchmarks aren't direct analogs to game play anyway. What they do provide is an empirical method of testing where results are 100% repeatable and show relative performance between a group of cards. It shouldn't be translated directly to FPS in-game, regardless.

+1
It's similar to all those "which CPU should I buy" threads. Should I get the CPU that's faster in synthetic benchmarks or the one that's just as fast or faster in virtually every real world scenario? I don't know, are you building a PC that only runs sythetic benchmarks?