Dying Light 2 Benchmark Test & Performance Analysis

[Assimilator]

Like W1zz explained pretty good, implementing DX12 (Ultimate) + RT. Nobody said the implementations are perfect. And it should not be forgotten that it is not a small task to reprogram the engine to use DX12 as well, and then RT on top. This is a lot of work.

Stop making excuses for Techland, they're not a resource-strapped indie developer. DX12 is not a new renderer and ray-tracing is not a new technology, having such a half-assed implementation of both is inexcusable in a game released in 2022. Do it right or don't do it at all.

 

[Kanan]

Stop making excuses for Techland, they're not a resource-strapped indie developer. DX12 is not a new renderer and ray-tracing is not a new technology, having such a half-assed implementation of both is inexcusable in a game released in 2022. Do it right or don't do it at all.

Nothing to do with that. I'm not trying to be a asshole, instead being a understanding human being. Aside simply using my knowledge and experience. Again: to implement DX12 + RT, is a lot of work. No wonder they didn't do it in time. I don't think you grasp the situation.

 

[Max(IT)]

I've read somewhere that because AMD is behind at raytracing, they're throwing money at devs to not use that many light sources and make them look too bad. If that was true, that's exactly what the end result will look like.

AMD RT performance are as usually poor. They are facing the consequences of that poor choice made with RDNA to skip RT entirely. Now they are one generation behind.

 

[Assimilator]

Nothing to do with that. I'm not trying to be a asshole, instead being a understanding human being. Aside simply using my knowledge and experience. Again: to implement DX12 + RT, is a lot of work. No wonder they didn't do it in time. I don't think you grasp the situation.

Expecting games to be finished on launch day makes me an asshole, huh?

 

[Max(IT)]

OLED 60 Hz is complete trash reactivity wise compared to my PC monitors, I see it everyday on my phone. The picture quality is excellent though.

Nothing to add about image quality: it is wonderful at 4K. I was just referring to people claiming playing 60Hz on an oled was better than 144 Hz on an IPS. I absolutely don’t agree.

Expecting games to be finished on launch day makes me an asshole, huh?

Unfortunately this seems to be the case with any game. I think at this point is wiser to buy games at least 6 months after release.

 

[Kanan]

Expecting games to be finished on launch day makes me an asshole, huh?

There is constructive criticism, fair assessments and then whatever you did.

 

[Footman]

Dying Light... more like Dying Shite.

@Assimilator now that's some funny shit!

 

[neatfeatguy]

AMD RT performance are as usually poor. They are facing the consequences of that poor choice made with RDNA to skip RT entirely. Now they are one generation behind.

I'm not sure if I feel bad for AMD that they can't handle RT as well as Nvidia.....

Or if I feel bad for Nvidia having dedicated cores for RT and they still suck at it......

Anyway, RT is still new in terms of implementing and GPUs having to handle it. I feel we still need to wait two more generations of GPUs to see what company can finally handle RT without having the fps crushed into the ground.

 

[Kanan]

AMD RT performance are as usually poor. They are facing the consequences of that poor choice made with RDNA to skip RT entirely. Now they are one generation behind.

They didn't skip it. They needed time to develop it for RDNA 2. Nvidia developed this in 10 years. That's why Nvidias is so much better. Including a way bigger RT core, not comparable to AMDs, which is tiny.

 

[Guwapo77]

Yes, OLED at 60hz is still more responsive than IPS at 120hz, it dawned on me just recently because I played a little bit on my previous 3440x1440 160hz screen (LG 34GN850-B) and felt 80FPS is kinda laggy compare to OLED at 60FPS

I'm crossing over to OLED as soon as a nice 27" panel becomes available sub $1,000. I have given up on "regular" panels. Thanks for further confirming what people like Coreteks has said for a couple years now.

 

[sector-z]

5800x is a little low end versus now a 12900k . I Hope to see the config updated soon.

 

[THU31]

"...added visual fidelity (from ray tracing) is limited". Not sure about that.

Digital Foundry video shows massive differences. It is a huge performance hit, but seems worth it in my opinion. After Metro Exodus EE, this might be the best RT implementation I have seen. It drastically changes the look of the game, instead of simply making shadows or reflections look better.

I will be starting my playthrough tomorrow. Will definitely use their optimized settings.

 

[Kanan]

"...added visual fidelity (from ray tracing) is limited". Not sure about that.

Digital Foundry video shows massive differences. It is a huge performance hit, but seems worth it in my opinion. After Metro Exodus EE, this might be the best RT implementation I have seen. It drastically changes the look of the game, instead of simply making shadows or reflections look better.

I will be starting my playthrough tomorrow. Will definitely use their optimized settings.

I have to say, RT on in this game looks much more realistic, just judging by these pics.

 

[mama]

I'm crossing over to OLED as soon as a nice 27" panel becomes available sub $1,000. I have given up on "regular" panels. Thanks for further confirming what people like Coreteks has said for a couple years now.

What's the Sony OLED like?

 

[W1zzard]

"...added visual fidelity (from ray tracing) is limited". Not sure about that.

Digital Foundry video shows massive differences. It is a huge performance hit, but seems worth it in my opinion. After Metro Exodus EE, this might be the best RT implementation I have seen. It drastically changes the look of the game, instead of simply making shadows or reflections look better.

I will be starting my playthrough tomorrow. Will definitely use their optimized settings.

Looks just "different" to me, make sure to toggle between RT often and let us know how you like it

 

[Kanan]

To me the non-RT version just looks broken. Light everywhere, corners, corridors, it makes no sense to me, in a realistic way. I think they concentrated a bit too much on making the RT version here, and disregarded the normal version. IMO, I would never play this game without RT on.

edit: again just judging by those pics.

 

[THU31]

To me the non-RT version just looks broken. Light everywhere, corners, corridors, it makes no sense to me, in a realistic way. I think they concentrated a bit too much on making the RT version here, and disregarded the normal version. IMO, I would never play this game without RT on.

edit: again just judging by those pics.

I do agree with that. Devs are amazing at faking things these days. But without RT this game looks like from the X360/PS3 era in some aspects. Most AAA games look way better without using RT.

Also, textures are bad (ones like posters and stuff). My friend has taken a lot of screenshots, he is obsessed with texture quality. 43 GB install size kind of explains that, so does the DX11 VRAM usage.

 

[lightning70]

On the RTX 3090, only 30-35 FPS with RT enabled in 4k. Pretty little for this card, isn't it?

 

[Kanan]

On the RTX 3090, only 30-35 FPS with RT enabled in 4k. Pretty little for this card, isn't it?

DLSS is your best friend.

 

[Scircura]

FPS Analysis on page 5 shows the God of War - Relative GPU Performance graphic.

Thanks for another analysis, W1zzard.

 

[wolf]

The "AMD optimized" titles have way better performance for RT in comparison to Nvidia, often same performance-loss, reviews are on TPU.

Not quite... I used various results to compare this for myself, the FPS you get at the end (RT on) might seem similar, in part because RDNA2 cards tend to start with higher rasterization performance in AMD optimized titles, but the RT cost in terms of milliseconds added to each frames render time is inarguably higher on an RDNA2 card vs it's Ampere competitor.

Generally speaking, they have better RT-on performance compared to "Nvidia optimized" games because RDNA2 is considerably weaker at RT, which isn't really in debate, so they severely cut back the amount of effects and the quality of the RT effects themselves. It's just about the only play they can use this generation (outside of HD texture packs) if they intend to use RT at all, or they'd pants their own cards.

Really enjoying DL2 so far, another powerhouse of an RT showcase and gorgeous to behold so often when playing. My only wish would be infinite bounce RTGI rather than single, but the dark aesthetic certainly works.

 

[nguyen]

Looks just "different" to me, make sure to toggle between RT often and let us know how you like it

Sometimes it's better to turn off the FPS counter and enjoy the game and all its graphical glory .

BTW looks like force enabling Resizeable Bar in Dying Light 2 boost FPS by 6% for 3090 from my testing. Now averaging 68FPS outdoor and 80FPS indoor, very playable FPS to me.

 

[W1zzard]

FPS Analysis on page 5 shows the God of War - Relative GPU Performance graphic.

Fixed, was just the wrong title

 

[Kanan]

Not quite

Very quite, as long as I used the word "often" what I described is perfectly sound, and in accordance with the many professional reviews I've read about it. I know where you're coming from, so I'm not concerned. Keep adding laugh smilies to my postings.

 

[wolf]

Very quite, as long as I used the word "often" what I described is perfectly sound, and in accordance with the many professional reviews I've read about it. I know where you're coming from, so I'm not concerned.

An interesting way to appear to dismantle my point without really addressing it. So where do you believe I am coming from? I've extrapolated numbers from professional reviews, feel free to share others and I'll happily do analysis of them, where possible, to show the true cost and performance loss for enabling RT.

And I don't believe your reasoning that its "often same performance-loss", it's my understanding the performance loss is always higher, relative to the performance either had before enabling RT. I am always willing to learn however and if you present evidence showing that the performance loss is often the same, I'd love to see it. Why would anyone need to be concerned? I am fascinated by technology and strive to learn and understand it better every day.