Client GPU at 100%, machine becomes unresponsive

[am312]

On my Windows client, task manager often shows nxplayer.bin GPU at 100% (see pic) and the machine gets very unresponsive.  Closing the NoMachine client brings GPU back near 0% and machine responds fine then.  This 100% GPU state can happen when the connected NoMachine session is idle.  The only change on the display then is the desktop’s tiny clock ticking.  Client CPU is approx 20%.  This condition does not always occur.  Sometimes  GPU is only about 60% and the client responds fine then.

client: Win 10 on Lenovo P53s laptop, GPU0 Intel UHD Graphics 620 (maxes out), GPU1 Nvidia Quadro P520 (stays at 0%)

server: CentOS 8 VM w/ICEWM virtual desktop on ESXi.  Hardware is HP gen10 server (no GPU) 24 core Xeon.  NoMachine 6.12.3_7.  Server CPU usage is low.

Why does does the GPU hop up to 100% for long periods?

Can I tell NoMachine to use my laptop’s GPU1 instead?  it appears to be just chilling there, waiting for some work.

Attachments:

1. 

   gpu.png
   

[fisherman]

Can you try to disable “multi-pass encoding” in session Display Settings of NoMachine Menu Panel. https://www.nomachine.com/DT04O00140#5.1

[am312]

That does appear to have fixed it so far.

It would be great if there was a more perm fix so that running the default client settings don’t make the machine unusable.

Thanks!

 

[Author]

Posts