Different CPU numbers on dashboard vs VM vs dockers

[Fredrick]

Hi guys,

 

I'm trying to spread the load around and need some help understanding the numbering of my CPUs. I've got 2 Xeon X5650 processors. This is the view on the dashboard:

 

Note that CPU 24-27 does not have any load, and has never seen any load. Meanwhile CPU 12-15 is "missing". This load you can see here is with Handbrake running an encode with the following parameters:

 

 

As we are not seeing load on CPU 20-23, and instead seeing it on 16-18 I'm suspecting the dashboard view to be skewed, and that it should have been 0-23 instead of skipping 12-15.

 

Also, when I'm adding a VM this is the numbers I can choose from:

 

So my guess is that it's the dashboard view that is in fact "off". Any idea how I can fix this?

 

Also does this mean core pairing should be:

0-12

1-13

2-14

3-15

etc.

 

PS. Do you have Handbrake pinned to cores when its running with the SYS_NICE parameter, which I understand is "low priority"? Maybe I can just let it have access to all cores, and the OS will just prfioritize other tasks?

[bonienl]

You can try unRAID v6.4, it has some updates for this.

[Fredrick]

I'd prefer to stick with stable releases tbh

[Juventas]

I just built a dual-core Apollo Lake server.  unRAID shows "CPU 0" and "CPU 2".  There's no "CPU 1" and CPU 2 is always blank.  Is this a UI problem, or is unRAID only using one core?  Will this also possibly be fixed in 6.4?