I found this white paper from 2010
which includes this statement:
Care must be taken to assign interrupts correctly to both the physical core and its child hyper-thread.
Is this still true for current CPUs such as the Xeon E5-2697 and E5-2699? Are sibling threads equivalent or is one thread a real core (with better performance) and the other thread a hyperthreading core (with lower performance)? Reference: https://stackoverflow.com/questions/7274585/linux-find-out-hyper-threaded-core-id Linux find out Hyper-threaded core id - Stack Overflow
In other words, would assigning a hyperthreading core to servicing NIC interrupts result in lower performance?
My understanding is that hyperthreads are equivalent and that there is no performance gain (or loss) choosing one hyperthread over the other. When it comes to performance, the general answer is no, assigning a hyperthread to a particular interrupt shouldn't lower performance, but there are a lot of factors to be taken into account. For this, I bow to the documented wisdom.
Hope this helps,
In regard to your questions; Are sibling threads equivalent or is one thread a real Core and thread a hyperthreading core?
Threads are logical cores, and Hyper-Threading is the name of the technology. You can assign threads to NIC interrupts.
I am sending you a document from our Intel® Developer Zone. It describes the benefits and usage of Hyper-Threading Technology.
Introduction to Hyper-Threading Technology
Intel Customer Support