How to display TX2 screen on laptop?

Autonomous Machines Jetson & Embedded Systems Jetson TX2
1

I have been searching on how to display tx2 screen on my laptop. I thought that I can display with using Wi-Fi but I have no idea. Jetson will be at least 30 metres away.

I’m really bad at setting IPs and DHCP.

Thanks.

2

Do you need text [console] or GUI[Graphical User Interface] displayed at your laptop?
For former you could use ssh and for the latter you could use some remote desktop application.
Whatever you use you will need to distinguish the ip address of the Jetson, that you can do executing:

ifconfig

Once you got it determined you can connect over ssh to it or with graphical client somehow from your laptop

3

I need GUI. For example, I want to see CUDA Samples(executed samples) from my laptop. I have found this but didn’t understand clearly.

4

does your laptop have GPU ?

5

Yes, GTX 1050 Ti. I installed CUDA and CUDNN.

6

as far as I remember there are two ways to get cuda samples running and shown remotely at laptop screen.
Scenario I: You execute cuda samples from your laptop at tx2 jetson and your local gpu processes execution of samples.
Scenario II: Samples are executed at iGPU of jetson and somehow streamed to laptop.
Perhaps the utilized scenario will depend on software application that you will be using as remote desktop sharing medium. To a certain extent ssh -X providers GUI feature. If you can get the same cuda version and cudnn version at laptop that you have at jetson that, in my opinion will reduce possible issues. However, you will have to determine if ssh -X provides enough capabilities or you will need to use more sophisticated solution, in my opinion.

7

@Andrey1984 is correct, but to extend the knowledge: “ssh -Y” or “ssh -X” will forward X events to the laptop…then the CUDA and GPU to use must be on the laptop. The program running on the TX2 would have to be designed for the CUDA architecture of the laptop…not the TX2.

If you want the CUDA and other NN software to actually run on the TX2, then you need to have the X events interpreted on the TX2. You could run an X login on the TX2 using a real X server, or you could use a virtual server on the TX2. In the case of a virtual server you won’t need a display, but a software client running on the laptop could interact with that desktop as if it is local to the laptop…though events would process on the TX2 (and thus so would any CUDA). I do not have any advise on specific virtual desktops.