Simultaneous Multithreading
How one physical core interleaves several thread contexts to fill idle issue slots, which state is replicated, which resources stay shared, and why SMT buys throughput instead of latency.
Multicore Organization
How several cores share LLC slices, a ring or mesh fabric, memory controllers and I/O inside one socket, and how that topology sets the latency and bandwidth a parallel program actually sees.
Heterogeneous Cores
How performance and efficiency cores share one ISA and one coherent domain, what differs inside them, and why the scheduler rather than the silicon decides whether the mix pays off.
Uncore Organization
What sits around the cores inside a socket: LLC slices, the ring or mesh, home agents, memory controllers, the I/O complex and the power unit that binds their clocks.