Hi
I am implementing async clients.
I am trying to understand how many completion queues I should have on the client side
The options that I thought about are:
1. one in process
2. one per core in system
3. one per stub
In each case one thread to loop over it.
All the examples I saw, there was only one client (one stub), so I am not sure what is the way to go.
Thanks
Shiran