Let's take a look at some useful aspects of the Black Myth: Wukong benchmarking tool

Unlike Ed, I wasn't considered important enough, or young enough or handsome enough, to receive a Black Myth: Wukong review code, which left me with hopes of doing some hardware performance reviews with the recently released benchmarking tool. As developers Game Science themselves admit, it “may not fully represent the actual gameplay experience and final performance at the time the game launches.” Monkey nuts.

The tool allows you to tinker with individual graphics settings, as well as flashier tricks like ray tracing and both DLSS 3 and FSR 3.1 frame-rendering varieties. But considering Wukong is a deep-level Soulslike, the benchmarking process is pretty sedate, mostly a quiet river with only brief periods of idle NPCs to be seen. There's no way to judge how a boss fight's storm of particle effects will strain a GPU, for example, or what would happen if you tried to play from an old hard drive instead of ignoring the SSD requirement in the system requirements.

But I think there's value to be had from this peaceful forest traversal. For one thing, it's encouraging for Wukong's chances on the Steam Deck: I got 65fps on the Valve handheld, with the Low preset and FSR boost set to 66%. Even if the car isn't perfectly 1:1 with the full game's launch performance, the Deck would have to be extremely off-track to prevent it from having a smooth time.

Black Myth: The graphics settings menu for Wukong's benchmark tool.

Image credit: Rock Paper Gun/Game Science

More generally, the tool can also give you an idea of ​​which specific settings are worth lowering if you want smoother performance than the presets offer. The benchmark might suggest, for example, that you can get big frame rate gains by lowering something, even if it's not the best display of visual effects. is If it's heavily represented (shadows, textures, reflections, etc.) you can make a good bet that lowering those same settings will help speed up the full game.

To find out which options claimed the biggest performance hit, I started with an average of 36fps – that's on the RTX 4060, at 1080p, using the highest Cinematic preset with DLSS at 75% – and re-ran the benchmark after tweaking each one individually. Here are the results:

Adjustment Average performance
Motion blur, Strong to Off 36 frames/sec
Super-resolution sampling, from DLSS to TSR 32 frames per second
Super resolution sampling, from DLSS to FSR 34 frames per second
Creating a frame, from Closed to Open 60fps
Full ray tracing, Off to Medium 38fps
Picture quality, Cinematic to Medium 37fps
Anti-aliasing quality, Cinematic to Medium 37fps
Post-effects quality, Cinematic to Medium 37fps
Shadow quality, Cinematic to Medium 51 frames per second
Texture quality. Cinematic to Medium 36 frames/sec
Visual effects quality, Cinematic to Medium 39fps
Hair quality, Cinematic to Medium 38fps
Vegetation quality, Cinematic to Medium 40fps
Global illumination quality, Cinematic to Medium 45fps
Reflection quality, Cinematic to Medium 37fps

If we just stick to quality settings for a moment, we can see that shadows, foliage, and global illumination are the biggest framerate hogs and therefore likely candidates for lowering in the full game. With all three on Medium at the same time, with the rest on Cinematic, my RTX 4060 managed a much smoother 75fps average. Of course, that exact number may increase or decrease slightly in Wukong, but it’s very likely that these three settings will still be the biggest drag on your system.

As we get into the more high-end stuff, DLSS (once again) seems like the best upscaling option, but if you don't have a compatible RTX graphics card for it, neither TSR nor FSR are particularly ugly alternatives. There's also the slightly trickier way to enable ray tracing. to develop Performance based on the Cinematic preset; normally the RT effects are the most GPU-heavy in the menu. Since Wukong is an Unreal Engine 5 game, it uses Epic's Lumen technology for its highest shadows and overall lighting settings, and this can be just as hard as regular ray tracing on your GPU. Maybe even harder. Choosing full ray tracing (or path tracing as it's often called) seems to replace these Lumen effects, hence the small performance hit.

A screenshot of the Black Myth: Wukong benchmarking tool in action shows a wooden bridge over a calm river.

Image credit: Rock Paper Gun/Game Science

So, the Black Myth: Wukong benchmark: it's not as pointless as it seems. I still don't think solid 60fps in the vehicle equates to solid 60fps in the final game, but when the latter releases on August 20, you can at least go into it knowing which settings to cut if performance becomes wonky.

Leave a Comment