Should you use the video output of the GPU?

[dannieboiz]

When using DS GPU version, are you also using the GPU's video out or dedicate it all to DS?

I have 3 monitors that I need to drive,

 

[TBurt]

When using DS GPU version, are you also using the GPU's video out or dedicate it all to DS?

I have 3 monitors that I need to drive,

I've done both ways. I have an AMD video card so I thought I might as well use and leave the Nvidia dedicated to Deep Stack. But before that, I drove my monitor and DS off the same GPU with no noticeable slowdowns or stability issues. One thing I would not try is using an NVidia GPU for Deep Stack and encoding/reencoding as it will starve DS to the point you will get AI has timed out errors. At least I did with my low-end GTX 1060 with only 3GB of memory. Use Intel hardware acceleration if your CPU supports it, otherwise just stick with a regular CPU.

 

[sebastiantombs]

I used a GTX970 for DS and video for quite a while with no noticeable delay in either. Now I'm using Intel for video, new mobo with onboard video. No significant change with DS detection times.

 

[dannieboiz]

Thanks, I went ahead and drove 2x monitor with the GPU and the other using the onboard output so far it's working.

Only thing for me is when DS run it also increases my CPU a little from python