An introduction to lockless algorithms

This is good related read: https://www.morganclaypool.com/doi/10.2200/S00962ED2V01Y2... (A Primer on Memory Consistency and Cache Coherency, Second Edition, currently free to download I believe)

I would also personally encourage readers to look elsewhere than the x86 QPI/UPI and HT stuff and toward an openly documented protocol with public documentation of its implementation. The Arm AMBA specs are available, as is their implementation in the Coherent Mesh Network CMN IP. Far more useful than the Intel/AMD alternatives.