P2P communication between GPU and FPGA

Could you share which specific GPU and FPGA are being used and the driver/toolkit versions.
This information may help to trivially reject capability concerns.