A Programmer's View of Computer Architecture: With Examples from the MIPS RISC Architecture

I've used this book for several years to teach an undergrad course introducing CS majors to assembly language and computer representation of data. The authors have chosen an interesting way to ease the transition from high-level language to assembly: they use several successively more realistic versions of the same (ultimately MIPS) assembly language, all of which run on a simulator provided with the book. The first models a memory-to-memory machine, with typed variables and no registers, allowing students to learn about the minimal arithmetic and control operations (including a limited form of procedure calling) of assembly language without worrying about other concerns. In this context they spend two chapters on integer, floating-point, and character representation. In Chap. 7 they introduce memory addresses, using an array-like syntax familiar to high-level-language programmers, and show how to implement simple data structures. In Chap. 8 they introduce registers and type-specific operations thereon, pointing out that in a load/store architecture like MIPS, all arithmetic actually works on registers. Chap. 9 treats procedures more fully. This constitutes a minimal course; the remaining five chapters can be used as time allows. Chap. 10 discusses assemblers, machine code format, and the "true" MIPS assembly language; chap. 11 discusses I/O, chap. 12 interrupts and exceptions; chap. 13 performance; and chap. 14 other approaches to computer architecture. I switched to this book when I found Hennessy & Patterson too advanced for my students, and it has served me well. Students are sometimes a little confused about which version of the assembly language we're using at the moment, and I wish the author of the simulator had put in a three-way choice rather than accepting all three languages at once, but I still think the approach works better than throwing the kids in the deep end.