Coming to think about it, you are right!
At the end benchmarked are debugging tools. From switching times of single transistors and grey-grey switch times of pixels and microbebchnarks to benchmarking supercomputer and distributed computing systems.

They tell you if there is a problem within their view range and can tell you what should or likely won’t not improve your performance.

Thank you, that is a valuable insight.

K

so it's a debugging tool not a benchmarker

noted

I am a developer. But I am not involved in this benchmark, nor do I do VR or game development.

What you want is not a VR benchmark, it is a system benchmark for physics heavy games used through VR. Which is valuable, but only to physics heavy games in VR.
You can visit the wreck of the Titanic in VR. That is not going to be physics heavy.

A VR benchmark compares the relative performance of different GPU and VR headgear combination. Stuff like you finding the Oculus software eating performance. That is the point of it.

As you put it: Yet there are many games in VR that absolutely will stress the cpu, and the software itself does. Obviously games do more than just pass a scene to be rendered to the GPU. But that is independent of the VR.

However, none of them will ever be more cpu than gpu intensive.
Counter-Example: A chess game using an engine like Deep Blue, in VR.

Perhaps calculate the physics of the blocks being demolished to a severe degree. Or any other physics based mechanic, a la 3dmark's cpu run tests. Lots of medium/high graphical fidelity objects moving in relatively indeterminate fashions. (or predetermined, but physics-based, noodly).

But I'm not a programmer, simply a benchmarker and neither are my job I'm afraid :)

If I wanted to run a benchmark I would come up with some way to make boneworks consistent between runs- that game is cpu heavy.

Apologies if you have some stake in the game and I am unsure if you were the dev or not, if you are then you should really just ignore me. If you aren't then it's a voxel styled cubes game how can it possibly be a benchmark for anything unless it has rtx

Ok, so what should the CPU do while running this benchmark?

I'm not being held back by my CPU, because I have almost no % usage. I would be GPU bound. Yet there are many games in VR that absolutely will stress the cpu, and the software itself does. However, none of them will ever be more cpu than gpu intensive. The problem is it doesn't do both, which is sad.

And no I don't have a trash cpu or something it doesn't really get much better than this in games that will not use more than 6-8 cores. 5.1ghz allcore being used .4% with a bad gpu would be even worse if I had a much higher consistency with the cpu. There is no contrary to VR. You are right, it is very GPU based. But to say this benchmark proves anything is like saying CLEARLY english is important to IQ, because only english speakers pass my english IQ test.

>as this VR benchmark proves.

uhhhh....

that's because this VR benchmark is bad

This is supposed to be a benchmark of VR, not a benchmark of this game. That was the point.

I I am not sure I understand the problem.
You want the CPU included in the VR benchmark, but what does that say about the VR performance of the setup? VR performance by itself is clearly not CPU-bound , as this VR benchmark proves.

A benchmark of a specific game will depend on the CPU. If your CPU is the limit, getting an ever so good VR setup will not help you at all. On the contrary, someone with a better CPU but a worse GPU may run circles around you. Deciding to buy VR GPU A or B and headset C or D is not helping me when I see that a benchmark on your computer says "they all deliver about the same performance" if you are held back by your CPU.


What am I missing here?