Skip to content

Compile CompositeVectors filter with regular compiler

Kenneth Moreland requested to merge kmorel/vtk-m:composite-vectors-cuda into master

The CompositeVectors filter does not run any worklet of its own. It uses precompiled array manipulation and copies for its implementation.

It shouldn't matter if a device compiler is used. (It should be a quick compile.) But for some reason the nvcc compiler was choking on an auto. Rather than figure out why nvcc is barfing, I just stopped using it for this file.

Merge request reports