No hardware acceleration and for now limited to it seems 8gb ram allocated to the VM but for the sake of me being bored, it works.
Cool! But it’s not really useable as I understand, right?
If you have pen or touch input on your PC like I do, it’s useable but not fast.
Remeshing takes a long time, subdividing takes a long long time.
More ram means more total vertexes, but if I could get my GPU to provide HW acceleration it would be fast.
Maybe as the platform develops it might get better, or if you have a really really high end CPU (I’m running a Ryzen 3400g 4c8t 4ghz 32gb ram) it might be as fast as the New iPads maybe?
I’ll see if I can get more ram in the VM and HW acceleration to work.
It’s also a long process to get it working though, custom kernels and stuff, so it’s certainly not something I would recommend trying.
I am out
I would not install win 11 on my workstation anyway the next year. But if everything is accelerated it would be interesting with a RTX3090, i9 10900K and 64 Gig Ram. But my Cintiq has no touch, though.
You can just install an Android emulator like bluestack.
You can, but the overhead in the emulator means much worse performance than the near bare metal solution for the wsa.
At the moment there’s no vulkan API support, so once that’s running it will be interesting. Everything is CPU based right now and even that is able to handle a few million vertex sculpt.
It’s not cpu based, it supports OpenGL (not sure if it supports 100% of i thought).
Nomad simply uses OpenGL, Vulkan wouldn’t make much difference anyway.
If anything, the bottleneck is probably more on the CPU (overhead of VM? multithread? simd?, etc)
It’s definitely bottlenecking in the CPU as it’s using 0 GPU (even by nomads debug and log it’s using 0 GPU) it might be a configuration issue on my end as technically the subsystem should allow a vGPU with openGL and openCL support
I managed to get 16gb of ram Allocated and got stable 41M vertex models to sculpt with only 20gb system RAM used lol.
It doesn’t use the GPU.
The GPU is only used for rendering.
These logs don’t work, it always shows 0.
If your GPU is not used and if you have millions of face, it should run at <1fps.
CPU-optimized renderer are very rare (zbrush being the exception).