Review: Metro Exodus: PC graphics performance analysis

Erm, by this logic, you might as well ignore all the benches because they never test the whole game.

The point is, and trust me, the benchmark runs a lot worse than actual gameplay. TPU confirm that. GN confirm that. Even HH says that the built in bench stresses the system more than gameplay.

 

Unless the bench is based on a different game, I would think that would be the best way to test performance.

Unless you want to play half the game where the frames are more favourable, and then come back and play the other parts with better hardware or by dropping resolution.

 

but of course a bench is best to test, because it'll provide the most consistent results, but in this case, and in some other games, the bench doesn't reflect the in game performance.

Why do I have to repeat myself?

Nobody is going to play through half of any game to get "accurate" results, nobody has time for that. Benches are a quick way to measure performance.

 

Its interesting how much the results rankings for FPS changed when TPU benched it with an 8700K.

 

DLSS doesn't seem to work in the comparison shot from Hilbert, it's way too blurry and looks like DLSS input/raw upscaled footage.

 

Guess you missed the discussion. TPU benched actual gameplay while G3D used the tougher ingame benchmark tool. Also there is an i9-9900k to compare results vs the 5960x. Yes, the 8700k and 9900k should produce better results at 1080p, but the higher you go up in res, it narrows down (although number of cores seems to matter, if less than 6).

https://www.guru3d.com/articles-pages/metro-exodus-pc-graphics-performance-benchmarks,4.html

 

Curious that Nvidia recommended TechPowerUp use driver 417.22 for their analysis. Is it possible there is a bug in subsequent drivers?

 

Really great article Hilbert! Thank you for posting it.

 

Love the sliders @Hilbert Hagedoorn also great performance review !
As expected like any metro non-redux in the past this is an fps eater and gpu bully ! but tweaking settings etc i think a gtx 1060 / 570-580-590 or so can hit those 60 fps ((1080p)) , i mean it is not like a rule you have to max out the sliders ! On the other hand by the time this hit a sale on what ever store for say 5~10 usd/eur .... most likely our computers will be up to snuff to max everything out and hit 60+ maybe on 1440p ... except people with a 2080ti ... you guys are ready to go !

 

Metro has always been a beast on hardware when it launches. I remember back in the day when I had a 2600K with a GTX570 and the first Metro came out... I was getting 20fps in the benchmarks and my setup was definitely top of the line at that time, with only the 2700K and the GTX580 being faster. I remember looking at the benchmark results and through there is no ways my pc can be that bad. Lol! I jumped in my car and went and bought another GTX570 to run SLi and it made quite a big difference. Atleast now I could play above 30fps. Lol!
Anyway the game looks good for the most part as any Metro normally looks. However the RTX feature to me personally looks like just a SweetFX colour mod slapped on top of it. There is only one spot on the ice in the first image where I can see a bit more reflection, but the other images looks like just a colour change???

 

Oh and don’t talk about DLSS... That actually looks like you turned the graphics down as everything is a blurry mess.

 

imagine this played from a competitive gamer but wants dxr ....dxr on dlss on ! then proceeds to max out brightness so is not as dark !

 

hmmm, so, they implented DXR lightning..
Am i the only one that thinks that they didnt bother to make nice NON DXR Lightning?
even in Metro 2033 the lightning looked better then the NON DXR in metro..
talking about the indoor scenes in the sand dunes for example...

example:


VS




ummmmmm....

So they went with DXR and then sayd to themselves, well f*#k it thats enough, whos gonna use an non RTX Card anyway..


these are just my thoughts, like i sayd teh lightning looks terrible without DXR/RTX...
Way to goo, force people to buy the shitty RTX Cards...

 

I've noticed this is the only review where results are different from other tests. Am I missing something?

https://www.techpowerup.com/reviews/Performance_Analysis/Metro_Exodus/6.html
https://www.computerbase.de/2019-02...u-savegame-directx-11-vs-directx-12-3840-2160
http://www.pcgameshardware.de/Metro...ik-Test-Raytracing-DLSS-DirectX-12-1275286/2/

 

Why reviewers all over the world didn't wait for amd optimized driver once again? 19.2.2 was the way to go. Now what? Amd forced to be weaker than intended?

 

One day later is too late

First allways gives more money (click views)

 

Another game another reason why paying a premium for RTX is dumb.

 

Altough it's a clearly NV supported title, AMD blasts their NV counterparts. Nicely done.

 

Thanks for not taking personal and clarifying that. I have seen you by the years re-testing titles with newer drivers. It is just extremely interesting if not all, but the most of times, it is AMD/ATI who provides drivers later or just upon launch, as a result of not allowing 100% accurate numbers.

I wonder if it is a random thing from their shake or it has something to do with which vendor sponsoring each game exclusively.