Yes, convolution is a very common operation to offload to GPUs. However, the toolkit is only intended to help you call GPU functions safely from LabVIEW. The toolkit ships with a few pre-constructed wrappers to aid in generating your function wrappers for your own custom GPU functions.
You can construct a convolution using FFTs which the toolkit has wrappers for. Although this is likely not the most efficient implementation, it may be fast enough for your needs.
If you intend to use NVIDIA GPUs, you can google 'CUDA convolution' and you'll find a number of links on the subject.
Good luck!
