Robert Sulej

Results 39 comments of Robert Sulej

Oh, wait a second! I just checked again: Wed Sep 4 08:41:42 2019 +-----------------------------------------------------------------------------+ | NVIDIA-SMI 430.40 Driver Version: 418.67 CUDA Version: 10.1 | |-------------------------------+----------------------+----------------------+ That one should work, I'll...

Ok, it seems that the problem is with the OptiX library. I am discussing this issue [here](https://devtalk.nvidia.com/default/topic/1062573/optix/optix-6-ptx-parse-error-internal-error-while-parsing-builtin-declarations/?offset=3#5380687). If there is a workaround I'll fix it with the next release. If...

Yup, set/update defenitively to be changed. I'll improve on that as soon as I finish current project.

Thanks for your replies! You are right, I checked again and `return x.to(torch.int32)` gives me correct results. Also `return x.to(torch.int32).to(torch.float32)` correctly returns integers casted to floats. Only the code from...

Closing, should be solved in the new releases.

I came back to check if/how Colab is working. They still don't have the recent drivers, but with a compatible plotoptix release everything seems ok. Here is the [working notebook](https://colab.research.google.com/drive/1RSVKYSZl1NjfttBwyuoehaEkNWnXwy6t?usp=sharing)....

Closing since there was no activity over long time.

There is a [sample notebook](https://github.com/rnd-team-dev/plotoptix/blob/master/examples/1_basics/4_materials_configuration.ipynb); cell 4. has a few examples on how to use `make_material()` function. Function is documented [here](https://plotoptix.rnd.team/materials.html#plotoptix.materials.make_material), and [here](https://plotoptix.rnd.team/enums.html#plotoptix.enums.MaterialType) are available shaders. For a better performance...

Ah.. there was a message from @skiler07 that I didn't manage to answer. Yes, materials setup is still ugly in plotoptix. And right, now it is not possible to use...

Yes, that is on the list. I'll work on all the materials related issues after the noise-balanced work distribution and support for the micromesh displacement. Work distribution is practicaly ready,...