Why and where does it say that a job will be dispatched
to all available cores, because that is not the case.
I'm defining "available cores" as the ones that have IBM i entitlements.
IIRC, you indicated having IBM i entitlements for 3 cores out of 8 in the
socket. So active Job (threads) will be dispatched among the 3, according
to IBM i workload management. It may be possible that all threads are
dispatched to just 1 core on a mostly idle server. Much of this is
explained in Frank Soltis' books, but the concept is generally understood.
How do you know that only 1 core is being used?
My understanding is that vendors can limit their products to just 1 core by
using "workload groups", which is described in the Information Center. Is
that what you're seeing?