Nikolaos Kyriazis

Nothing wrong with it. This is what I'm referring to. How is one supposed to easily link against a built libshaderc from another cmake project without a target export of...

It does not seem to be a memory outage issue, since some iterations do go through, and the GPU monitor shows enough memory available.

```bash COMPUTE_PROFILE=0 cuda-memcheck python debug.py ========= CUDA-MEMCHECK ========= Internal Memcheck Error: Initialization failed ... ``` :( There are several obstacles in sharing the offending example. I was hoping that the...

I've put together a private repo that replicates the problem, to which I've added you. You'll find a commented out inclusion of normals, which, if enabled, crashes.

This is reproducible for me in both Ubuntu and Windows 10, although in Windows 10 it is ~4X slower.

I upgraded to the latest version, but I'll check again. Could you share a bit about your specs, pip freeze, driver version, etc?

I've tested it in colab and another machine and it runs OK. So it's an issue with my machine. It's disheartening that it is reproducible in both OSes :(

Perhaps a [H/W issue](https://wccftech.com/nvidia-titan-v-error/)?

Still no luck. I still can't run cuda-memcheck. I made a script out of the shared notebook and run nvprof with it. I'm attaching the [output](https://github.com/BachiLi/redner/files/4466908/debug.txt). There's a segfault as...

I managed to run `cuda-memcheck`, but, even for the smallest of problems, I'm getting `out of host memory`. I Ubuntu one needs to set `CUDA_MEMCHECK_PATCH_MODULES=1`. The error I'm getting is...