Black Myth: Wukong PC benchmark tool available on Steam

Why is everybody testing at 50% resolution scale? Even the default 68 doesn't look good. Just tried 50% and would never play like that.
 
There is a side by side of RT on/off in this video, this is gonna be a tough sell and i'm an RT slut.



I remain convinced that RT is mostly a scam designed to sell graphics cards. It's largely fucking pointless in a majority of games, with even the ones that do effectively use it not really being that much different to justify its existence.
The thing that stands out the most is the quality of the shadows. It looks pretty bad with RT off. The shadows are blocky and kind of pixelated and tend to sort of flicker.

GGrscdK.png


This is just a screenshot, but they look significantly worse in motion. Still, the game looks very good with software Lumen anyway.
 
RT in this build of UE5 is not bad at all, compared to state of the art lumen it is around ~35% slower while looking better, at least on 40 series, 50 series will level the playing field more then RT is just gonna take over.
 
Last edited:
NO idea what i'm doing here as i dont game on my PC but wanted to benchmark... my monitors are only 1440 so no 4k res for me. I also noticed that even though i have full ray tracing enabled in settings options the readout at the end says that it is off? Not sure why... went back into settings again and it is set to on. I just tried to max everything to see what it would spit out. Reduced pic size by 50% hopefully i dont blow the page up with a huge image heh

1qCnyT0.jpeg
 
Too much nvidia cards on this topic. Where are my AMD Radeon bros? Will try this later tonight.

Serious Star Wars GIF by Disney+
4xMzDbm.jpeg

-No framegen


dmRhb1O.jpeg

-default settings with framegen

I can't seem to turn framegen off anymore though, it even enables when I select xess now; must be my older driver or something.
 
Last edited:
Based on a few minutes of testing, the DLSS resolution scale numbers don't seem to have much of an impact until the wording on the right changes.

For example:

62-89 - Quality. Changing the number within that range doesn't seem to impact FPS much.
Setting it to 90 changes to DLAA, which has a large impact.
Changing it to 61 sets it to Balanced, which has much more of an impact than anything between 62-89.

Based on that, it seems you might as well set it to the highest number within each setting.

Haven't done a visual comparison, and this is only based on a few runs, so I could be wrong.
 
Is this the first time we've gotten DLSS/FSR/etc implemented with a stepless slider instead of presets (Quality, Performance etc)? That's something I've been hoping would be implemented ever since DLSS was introduced. Ideally it would have gone beyond 1.0x for temporal supersampling (DLAA) etc.
 
NO idea what i'm doing here as i dont game on my PC but wanted to benchmark... my monitors are only 1440 so no 4k res for me. I also noticed that even though i have full ray tracing enabled in settings options the readout at the end says that it is off? Not sure why... went back into settings again and it is set to on. I just tried to max everything to see what it would spit out. Reduced pic size by 50% hopefully i dont blow the page up with a huge image heh

1qCnyT0.jpeg

Did you restart the tool after changing the Ray Tracing setting? It's required.
 
I don't know why but I find that the ray tracing + very high/cinematic settings, BUT with global illumination set to low, looks amazing. The scenes look a lot more atmospheric, darker from the tree canopy, grittier, almost horror-like. There must be less bounces of ray tracing or something but the effect is neat.

Try it.
 
lol come on. We are finally getting a next gen game that looks like a next gen game and taxes the GPUs like a next gen game.

If you are maxing out settings then of course its going to run poorly. RT is the only thing thats unoptimized and you can turn that off and still get Epic's own software Lumen for realtime dynamic GI, and great FPS.

With RT off and everything else maxed out, i was getting an average of 55 fps at 4k dlss performance (1080p) on a 3080. You NEVER play at max settings on PC. no one does. Im sure once i set them to high, i will easily get over 60 fps. PS5 will likely use a mix of high and medium and should look fine in the 30 fps mode. 60 fps might drop to 720p internal but thats just like every other next gen only game, and they dont look nearly as good as this.
Even without Raytracing, this game looks and runs great on my gaming laptop with high settings at 90% resolution with FSR. Even if it drops a lot with enemies and effects I can adjust it slightly and it would still looks great. I can even play with medium/high raytracing at 30 fps on medium/high settings. I'm surprised how well this game is running on my aged laptop.
 
I'm getting 120-140fps without RT at 1440p (with RT bellow). The game looks good in general, imo > looks better without RT. But to be honest this benchmark kinda sucks, since theres no action going on. Also, i noticed some stuttering 1 or 2 times, even with +7000gbps ssd.

WBwihs6.png
 
The thing that stands out the most is the quality of the shadows. It looks pretty bad with RT off. The shadows are blocky and kind of pixelated and tend to sort of flicker.

GGrscdK.png


This is just a screenshot, but they look significantly worse in motion. Still, the game looks very good with software Lumen anyway.

Heres a better difference

DMPlVsb.jpeg




Obviously the right side looks much better in terms of realism however will anyone actually notice playing the game? In cyberpunk at least it was very obvious since there are a ton of colored lights bouncing off surfaces which made each scene 10 times better, but in this case, dunno, very hard sell for me as well and I have a high end PC as well.
 
There is a side by side of RT on/off in this video, this is gonna be a tough sell and i'm an RT slut.



You can easily see how unstable the image is without RT. Shadows are square-ish and low-res, reflections are worse, and in some scenes RTGI looks better. You can easily spot the bad shadows (and that's on a compressed YT video).

Contrary to other games, RT in this game brings image stability. In motion, it can be really appreciated. Now if you are fine with square-ish low-quality shadows, then I guess it's fine. You can simply disable it. And yes, the RT effects are not as transformative as in Quake 2 RTX or Portal RTX (which are TRUE path-traced games). Still, I wouldn't call them useless.
 
Why is everybody testing at 50% resolution scale? Even the default 68 doesn't look good. Just tried 50% and would never play like that.

In 4k 50% DLSS (so performance setting) is completely fine, other upscalers obviously don't compare.

Shadows and reflections are the biggest difference here:

H4BsL5g.jpeg
gQg7CpS.jpeg
WKRgkwM.png


Without body of water to reflect stuff game looks almost the same, not bad for over 2x better performance:

MfNog3b.jpeg
RIN6J10.jpeg
iyPqs0b.jpeg


I think they aren't using Virtual Shadow Maps here because usually UE5 have better shadows than that, maybe they do that to make differences look bigger. Nvidia money can't go to waste.

Compared to games that have RTGI vs. no GI at all without RT I think using hardware RT in UE5 is kind of waste.
 
Last edited:
Tweaked some settings. RT Off/Low with Recommended Settings @ 2k w/DLSS On is where it's at for now I think

r3Raixc.jpeg
 
Last edited:
Cinematic, 3840 x2160p
HKOJUJY.jpeg



Recommended, 2560x1440p
XnQnp2d.jpeg


Refund's imminent @ Steam:pie_grinning_sweat:

It's hard to believe how much slower Ampere is in this game in comparison to ADA Lovelace. Even in Alan Wake 2 (PT), my RTX4080S was only 30% faster compared to the RTX3090, but I would never believe it could double the results of the RTX3090, yet that's exactly the case (22fps vs 50fps with the same settings). It seems to me this game is only optimized for ADA GPUs, because no one can explain such massive difference between the RTX3090 and the RTX4080S.


4080S-4K.jpg
 
Last edited:
I'll probably slowly bump up the super resolution until it's a little over 60fps and then just lock it. Wonder why they don't give you Fullscreen Exclusive Display Mode.

PGald7L.jpeg
 
Uh? To me it just shows the typical improvements Ada has over Ampere with RT.

Nope, in normal RT there usually is very little difference. With path tracing Ada shows bigger differences but not to this extend:

AW2, 4070 is 14% better

UajJmC5.jpeg


CP2077, 4070 is 20% better

ficmyB2.jpeg


BMW (this time in 1080p), 4070 is 42% better

zwAKIuR.jpeg


And this is normal RT average between those GPUs from TPU:

relative-performance-rt-1920-1080.png
 
This game is well optimized like other games that came from Korea such as Lies of P. Western developers and some Japanese (not you FromSoftware, wink wink) have a lot to learn.


N2h1tGa.png
 
I wonder if RT set to Low just adds RT shadows while medium adds RT reflections and Very high the full hardware GI.

If RT shadows are the biggest win, id rather just have them on.
 
Which one is better? Tsr or FSR? Or DLSS.

My 3070 doesn't support frame generation for DLSS but it does support Fsr and Tsr. Obvious choice is not DLSS but I see everyone like DLSS the most. I never play game on PC so don't know anything about those options.

I turn on frame generation I could go "Very High" setting or even "Cinematic" with ~70-80fps on Fsr or Tsr.
 
Last edited:
Which one is better? Tsr or FSR? Or DLSS.

My 3070 doesn't support frame generation for DLSS but it does support Fsr and Tsr. Obvious choice is not DLSS but I see everyone seem like prefer DLSS the most. I never play game on PC so don't know anything about those options.

I turn on frame generation I could go "Very High" setting or even "Cinematic" with ~70-80fps on Fsr or Tsr.

I didn't know game has FSR3 frame gen, thanks to you I know now! The best thing is that it works with TSR and FSR but it's not usable with DLSS.

Fucking Nvidia, they know that FSR3.1 FG combined with DLSS upscaling makes their DLSS3 frame gen irrelevant (it also uses much less vram) so they started to block this combo in sponsored games...
 
Well, I'm happy with my results.

I won't expect to get near this performance on PS5 so, guess I'll eventually play the game on PC.

Can't go higher than RT low with cinematic settings, otherwise FPS takes a dive. But I didn't built my PC for RT anyway, so I don't honestly care.

1440p, Frame Gen ON:

O6WTWuf.jpeg


1440p, Frame Gen OFF:

bxur3hm.jpeg


1440p, Frame Gen ON, RT LOW:


0XvrjSV.jpeg
 
Last edited:
This game is well optimized like other games that came from Korea such as Lies of P. Western developers and some Japanese (not you FromSoftware, wink wink) have a lot to learn.


N2h1tGa.png

frame gen on at 37% screen res and you call that optimized? Not sure if serious
 
I didn't know game has FSR3 frame gen, thanks to you I know now! The best thing is that it works with TSR and FSR but it's not usable with DLSS.

Fucking Nvidia, they know that FSR3.1 FG combined with DLSS upscaling makes their DLSS3 frame gen irrelevant (it also uses much less vram) so they started to block this combo in sponsored games...

It worked like a charm , Cinematic with 100% scaling for >70fps. But I see people complain about the input lag.

Not sure should I turn frame gen on since I can't play the game yet :lollipop_neutral:
 
Last edited:
Is anyone able to get this to run at 8k?
For me it just stops on the loading screen after I click on run.

Oh wait, it's actually running… my system was just struggling hard lol. Results incoming…
 
Last edited:
99FPS at 4K with Ray Tracing at Max with Frame Generation (63FPS without Frame Generation).

dAb2n65.jpg


122FPS at 4K with No Ray Tracing.


wtxHC4t.jpg


Not bad at all.
Bad is that amount of tray icons, my dude. wth! A special red flag for that "blue M".
 
Last edited:
Top Bottom