Just found this on reddit:
Looks good and could give us good numbers to compare performance.
6 Likes
And someone is already posting Pimax numbers to the leaderboard!
5 Likes
By the way. VoodooDE got 16.10 with a Pimax 8K. So not bad results for the 8KX in comparison, I guess!
8 Likes
Did a 2080Ti and a 9900K only get 14FPS on average?
The benchmark is made to kill your GPU and its not made to run smooth. It wont run smooth even on a Vive.
I used Normal FOV and PiTool 1.0, no PP.
With Large FOV + PP (basically double horizontal resolution) I was running the 8KX in Aerofly FS2 in 50fps in average. So yeah this benchmark doesn’t tell you anything…
3 Likes
I am not sure what is the purpose of this benchmark. Is it supposed to test the GPU? Or the headset? If the latter, it would be like calling 3DMark a monitor test and then concluding that with the higher resolution the frame rate get lower.
Plus it seems just to create an artificial overload and then putting low numbers, because, no GPU today can render keep up with the up.
The last point is “Rendered PPD”, which seems to be just the render resolution divided by angular resolution, and does not make much sense too. You can change the rendered res as you want (either in PiTool or in SteamVR) and thus produce basically arbitrary numbers for “Rendered PPD”.
1 Like
To me it looks like it’s just a comparitive test for your gpu hmd combination.
The numbers on their own dont seem to mean much.
I guess it tells you if you have a good overclock when you compare to others with the same gpu & hmd
If you get a low score in comparison to others, it might point out a problem ie… no oc or below average oc, bottleneck or settings?
5 Likes
Exactly.
It’s the first benchmark that applies a load to a given GPU/HMD combination and it eliminates SteamVR SS settings.
This allows for troubleshooting between comparable setups and overclocking.
I find it very useful.
5 Likes
Indeed some of the info is maybe not that useful. But it does say rendered PPD & not actual ppd. As most headsets do render at native panel res when at 100%
Although it didn’t seem to want to add my result on 5k+ @90hz normal fov pp on
I’m not sure if it utilises all of the cpu cores for people with high core counts but then again most vr games dont either so it’s probably more realistic for most experiences.
I think as long as people don’t try to cheat with doctoring results, then its probably a very useful tool.
The Info I took from it was I had good overclocks, didn’t seem to be bottlenecking my gpu in any way, & it made me glad I didn’t upgrade my 8700k.
2 Likes
Those results are pretty good.
But what’s the point of a benchmark that doesn’t give accurate results? What’s that used for?
Good question.
I wish VR games in general could implement a Benchmark mode instead. At least the games made for VR from start. Lets say Onward, Flyinside flight simulator, Beat Saber or whatever. That would at least give us some real life numbers.
1 Like
The tool is developed by the dev of cyubeVR and based on the engine that title uses. It’s the closest to reality benchmarking tool for VR we have at the moment.
2 Likes