I realise I can PIN docker CPUs but I was just wondering what the docker sees when I don't pin any? Specifically I have a coffee lake xeon which I want to use the plex hardware decoding with, I know there's seperate setup for that, but do you need CPU pinning? The reason I ask is I actually want to give plex pretty much everything the server has, I'm wondering if it's better to pin lots of cores and ht to it or just let it run (if running gives it all the cores).

 

The other thing I was thinking was pinning means you'll need all the cores/ht you pin to be available to do a single cycle, which is why I was thinking perhaps not pinning is better. 

 

But if not pinning gives you just say one core then obviously I'll need to pin.

 

Thanks in advance.

Edited December 3, 20187 yr by dgwharrison

Not pinning gives the container access to all cores that are not isolated

Sent via telekinesis

thanks @Squid