CUDA GPU passthrough with XenServer

Hi all,
I have a workstation with a Quadro K620, I’m interested in developing CUDA for multiple platforms (Win, Linux) and I would like to use virtual machines as developing environments. Does anyone know if the K620 supports GPU passthrough with XenServer and how can be enabled?
I’m not interested in sharing the GPU among different VM/users as vGPU/GRID would allow, I’m just interested in accessing full CUDA capabilities within a single virtual machine at a time.