• Competitor rules

    Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.

Black Myth Wukong - Benchmark tool

If you're going to fall down the rabbit hole of the internet, i'm already there, its worth investigating a bit deeper, for some years now, since gamer gate there has been a war of power dynamics going on, of gate keeping.

It used to be that journalists would have the power to make or break a title, because public access to opinion on entertainment products would be through traditional print or curated online media, with that they could engage in their own form of social engendering, we don't like your politics you get a bad review, its a narcissistic power dynamic and its gate keeping.

Now with social media we can bypass journalists and get eachothers opinions, with that quite often things that don't pander to a given political view get through, think of all the games and movies that are in the public mind good but have a bad critics rating, or all the games and movies that are bad but have good critics ranting, you only need to go to places like Rotten Tomatoes to see 8/10 critics ratings and 2/10 audience ratings or vice versa.

Traditional gate keeping media are dying out, because we can now see them for what they are, they are angry and upset, they treat publishers who go around them with absolute distain, they are lashing out...


M1JZxn1.png
 
Last edited:
Classic paranoia mixed in with the rest of it.

Everything has a right to exist, because it does. You literally don’t get to decide.

Take it all leave it.
 
Getting back on topic.

I'm hitting a wall at around the 150 FPS mark with standard settings, and around 120 FPS with hardware ray tracing enabled.

The benchmark tool appears to run only 4 cores and 8 threads efficiently. It also appears to be using an earlier version of Unreal Engine 5.

Based on the benchmark, and having seen some trailers, it doesn't seem to have very advanced game mechanics.

Clearly more focussed on visual effects.


 
Getting back on topic.

I'm hitting a wall at around the 150 FPS mark with standard settings, and around 120 FPS with hardware ray tracing enabled.

The benchmark tool appears to run only 4 cores and 8 threads efficiently. It also appears to be using an earlier version of Unreal Engine 5.

Based on the benchmark, and having seen some trailers, it doesn't seem to have very advanced game mechanics.

Clearly more focussed on visual effects.


Yeah but thats like a 173p internal res right?
 
Last edited:
Yeah but thats like a 173p internal res right?
That's just so you understand it's likely completely stopped scaling.

The game also increases the internal image resolution with desktop resolution, so its actually 360p when at 4K.

It stops scaling at 720p actual on an RTX 3070.



I'm testing DX11 vs DX12 now, there seems to be a slight boost at higher resolutions whe using standard settings.


Edit.

Behaviour at 1080p when changing between TSR and DLAA. Not related to original point.
1080p actual with DLAA and TSR (performance goes up with CPU usage when using DLAA?)




Edit 2.

This is to do with GPU usage now, for context given the low resolution question, so nothing to do with original point.
720p actual with hardware ray tracing


720p DLAA with hardware ray tracing

Edit 3.

More strange stuff (edited).
Edit 4.

I’m not seeing the below now, so it could be that the settings weren’t applied properly.

If I change to 1080p actual with hardware ray tracing enabled, performance drops all the way down to 40 FPS.


Then if I change to 4K and use an internal resolution of 1080p (DLSS performance) with hardware ray tracing enabled, I'm still getting roughly the same 40 FPS.


However, changing to 4K and using an internal resolution of 720p (DLSS Ultra Peformance) with hardware ray tracing enabled, only results in roughly 60 FPS instead of the almost 100 FPS at an actual resolution of 720p.
 
Last edited:
Oh FFS let the CT crap go already.

My test results with 1440p, cinematic settings, motion blur off, FSR 70% and RT off.

7900X3D and RTX 4080

Avg FPS 72
Min FPS 58
Low 5th 64

7800X3D and 7900XT
Avg FPS 60
Min FPS 47
Low 5th 50

As usual RT tanks FPS and the graphical improvements are just not worth it IMHO. Though all in all it is not actually that impressive graphically, even with RT on. Maybe if I could turn off vignette and DoF.
 
7900xtx has decent performance right until you turn on Ray tracing then it dies


Look at the difference between the 3080/3090 cards and the 4070 cards here, what's going on? My 3090 is usually toe to toe more or less with 4070super/ 4070Ti Ti cards in most games (even with RT) so this massive win for the 4070 cards makes no sense at all.
 
Last edited:
1440P / Cinematic Quality / FG - DLSS - 75% / RT - Very High

Asus Strix 4090 - Mild OC

Average FPS: 138
Max FPS: 167
Min FPS: 114

K4RU9og.png
 
Last edited:
7900xtx has decent performance right until you turn on Ray tracing then it dies


Ray Tracing is already on without needing to further enable the "RT settings".
you're still using RT if you've disabled Full Ray tracing btw, UE5 uses RT at its core (Lumen) in software mode but in some games it is hybrid Lumen too. So that explains the performance on AMD if you're getting unexpected framerates. The Full RT setting is purely for path tracing on or off.

So as mrk says you are in fact turning on additioanl path tracing. Hence the results. So if the 7900XTX has decent performance, then that is in fact decent performance with UE5 Ray Tracing enabled.
 
Correct, Nvidia's naming scheme needs to be binned and redone, they've chopped and changed the labels a few times now, it shoudl be Ray Tracing, Path tracing. That's it, not Full Ray Tracing, not RTX (which is only to be used to detail the RTX card suite of features) or whatever else.
 
Last edited:
Back
Top Bottom