Although Muchnick wrote Advanced Compiler Design and Implementation in 1997, he lays out such an excellent overview of the compilation process and delves into such extreme detail with dozens of useful optimizations that it would be remiss for a serious low-level developer's library to forget this book.
Muchnick targeted Master's or PhD level Computer Scientists when he wrote this book, and I hadn't been introduced to some math concepts that he used, such as lattices. A reader attempting to find a useful sequence for reading this book might want to include some topic in advanced mathematics prior to reading, although Muchnick stays light on the proofs and overall isn't as taxing to read for non-mathematical audiences as Knuth.
All of his algorithms are written in ICAN, which is a pseudo-language he invented that looks remarkably similar to ML or Haskell with iterative qualities. The text explanations of what the algorithms do are usually strong enough that a reader familiar with data structures can follow along without referencing the gory details in the code.
All in all, I highly recommend Muchnick as a follow-up to Appel's Modern Compiler Implementation in ML, as the other text will bring a less advanced reader up to speed for this book.