Markus Vieth

> Now I use the vcpkg toolchain to compile the PCL original source code, What do you mean exactly? Could you describe that in more detail, ideally with the commands...

For me (on Ubuntu Linux), setting the following CMake variables correctly sometimes helped CMake find cuda: `-DCMAKE_CUDA_COMPILER=`, also `CUDA_TOOLKIT_ROOT_DIR`, `CUDAToolkit_ROOT`

I think it could make sense to just parallelize the existing class, and not add another class for the parallel implementation. The only concern I would have is whether changing...

> Alright, this all sounds good. I'll run a quick benchmark today or tomorrow to see if it makes any difference (and personally I'm expecting it won't since statically sized...

Thanks! However I see a problem: the first 5 elements of `digest` are written to `sstream`. Previously, `digest` only had 5 elements (each 4 bytes), but in Boost 1.86 it...

Hi, thanks for the fix, I would like to ask two things: firstly, could you post a file where the problem occurs, so that I can understand the problem better?...

Thank you, the diff is much clearer now. And I think your solution of storing normals and averaging them per vertex makes sense. It would be good if we could...

Yeah, that might be a bit confusing. `CorrespondenceEstimationBase` is abstract (so not usable by itself) and the base class/parent class of four non-abstract correspondence estimation methods, including `CorrespondenceEstimation`. `CorrespondenceEstimation` is...

So in release mode, how are the results (the aligned cloud and the final transformation)? Do they make sense? Most users probably use some kind of filter (e.g. PassThrough or...

> The result in Release mode are meaningless. The pcd file is result of filter cropbox. Could you describe the crop-box filtering a bit more, perhaps post some code? Did...