Redhat 8.X does not see the Nvidia video card by default when I connect with vnc and xrdp

Hello, I have redhat 8.x installed on my server, nvidia drivers cuda installed on it, everything works normally, nvidia-sim cuda control everything is normal, but for some reason when I run an application when I connect with xrdp or vnc, it sees the default matrox graphics card of the server, while it should see my nvidia graphics card tesla V100 by default. What is the reason for this? How can I overcome this problem? How do I make the application see the Nvidia graphics card.