This book was recommended to me by one of my colleagues and I must say I'm very dissatisfied by the book's organisation and approach. It's important to take the title literally: "A Quantitative Approach". Sometimes it feels more like you're reading a database of SPECint benchmark scores for a multitude of niche memory configurations than a "Computer Architecture Bible". I don't feel like I've actually learnt anything substantial in the ~300 pages I've read, and it does little to hold the interest of even a patient reader. I often find myself being diverted to one of the Appendices (which often feel like better versions of entire chapters) and there are more spelling/grammatical errors than I'd expect.
This book is the absolute guide to the inner realm of CPU architectures and in helping anyone also to fully grasp the notions of parallelism in modern Computer architectures at both CPU & GPU implementations. It is also an excellent guide to the rising omnipotent Open RISC-V architecture. John