[QUESTION] Converting openvdb grid to nanovdb::cuda::DeviceBuffer

[Ayamur0]

I am trying to convert a openvdb::FloatGrid to a nanovdb grid on the GPU.
From the examples I gathered that I should use a nanovdb::cuda::DeviceBuffer and then the .deviceUpload(stream, sync) function.

But I am struggling with converting the openvdb grid into a nanovdb::GridHandle<nanovdb::HostBuffer>.

Converting it to a HostBuffer works fine like this:
auto handle = nanovdb::tools::openToNanoVDB(vdbGrid);

But when I try to convert it to a cuda::DeviceBuffer, with either of these two calls:
auto d_handle = nanovdb::tools::openToNanoVDB<nanovdb::cuda::DeviceBuffer>(vdbGrid);
auto d_handle = nanovdb::tools::createNanoGrid<openvdb::FloatGrid, float, nanovdb::cuda::DeviceBuffer>(*vdbGrid);

I keep getting a LNK2019 error when compiling:

unresolved external symbol "public: __cdecl nanovdb::GridHandle<class 
nanovdb::cuda::DeviceBuffer>::GridHandle<class nanovdb::cuda::DeviceBuffer><class 
nanovdb::cuda::DeviceBuffer,0>(class nanovdb::cuda::DeviceBuffer &&)" (??$?0VDeviceBuffer@cuda@nanovdb@@$0A@@?
$GridHandle@VDeviceBuffer@cuda@nanovdb@@@nanovdb@@QEAA@$$QEAVDeviceBuffer@cuda@1@@Z) 
referenced in function "private: class nanovdb::GridHandle<class nanovdb::cuda::DeviceBuffer> __cdecl 
nanovdb::tools::CreateNanoGrid<class openvdb::v12_0::Grid<class openvdb::v12_0::tree::Tree<class 
openvdb::v12_0::tree::RootNode<class openvdb::v12_0::tree::InternalNode<class 
openvdb::v12_0::tree::InternalNode<class openvdb::v12_0::tree::LeafNode<float,3>,4>,5> > > > 
>::initHandle<float,class nanovdb::cuda::DeviceBuffer>(class nanovdb::cuda::DeviceBuffer const &)" (??
$initHandle@MVDeviceBuffer@cuda@nanovdb@@@?$CreateNanoGrid@V?$Grid@V?$Tree@V?$RootNode@V?
$InternalNode@V?$InternalNode@V?
$LeafNode@M$02@tree@v12_0@openvdb@@$03@tree@v12_0@openvdb@@$04@tree@v12_0@openvdb@
@@tree@v12_0@openvdb@@@tree@v12_0@openvdb@@@v12_0@openvdb@@@tools@nanovdb@@AEAA?
AV?$GridHandle@VDeviceBuffer@cuda@nanovdb@@@2@AEBVDeviceBuffer@cuda@2@@Z)

I am using the newest version from vcpkg (openvdb 12.0.0#1) and have both NANOVDB_USE_CUDA and NANOVDB_USE_OPENVDB macros defined.
The Examples from the documentation also compile perfectly fine (loading a grid from a file like this:
auto handle = nanovdb::io::readGrid<nanovdb::CudaDeviceBuffer>("data/sphere.nvdb");)

I don't know what I am doing wrong. Is there a different way to convert the openvdb grid?

[Ayamur0]

Solved: Seems like CudaDeviceBuffer can only be used with the Cuda compiler not the normal Cpp compiler.