What this part is for¶
Two machines to get working, and a script that says what each one can currently run.
The book uses three targets on two machines, and the two machines are not equally urgent. The
computer you are reading this on runs bare and xv6 under QEMU, needs one cross compiler, and
covers everything up to Part V. The board is a second machine — the only place a
timing may be measured — and nothing needs it until ch24.
So the two chapters are in that order, and they are separate on purpose: you can finish ch00 this afternoon on a laptop, and do ch01 whenever the hardware turns up. The board chapter is the next one in the book, but twenty-two chapters sit between finishing the first and needing the second.
What it leaves out¶
The decision about what to buy. That is Appendix H, because it is read once, before you own anything, and then only to settle an argument with a spec sheet. This part assumes the machines exist.
Any reason for the arrangement. The preface argues for three targets on two machines and for the instruction sets not matching. These chapters take that as settled and get it working.
Where to start¶
ch00, on whatever you are reading this on. It ends with one program compiled for two targets and run on both, agreeing about every answer, which is how you know the setup is real rather than merely installed.
Then ch01 when the board arrives — or later. Nothing before Part V depends on it, and ch00’s script will tell you at any point which targets your machines can currently run.
Which machine, and what it cannot tell you¶
Both, which is what makes this a part rather than a chapter. The emulated targets answer questions about what a program does, and nothing here or in Parts I to IV is ever timed: QEMU models no cache, no branch predictor and no memory latency, so a duration measured inside it describes the laptop, and how the emulator translates instructions, rather than the machine being emulated. The board answers questions about cost, and is the only place in the book a timing may be taken.
Neither can tell you yet whether any of it is worth measuring. Setting a machine up proves it runs; it does not prove a number coming out of it means anything, and ch24 is the chapter that takes that apart.
Where this leaves you¶
Two machines that work, and a script that says so. You will not have learned anything about computers yet. You will have the apparatus the rest of the book measures with, and a way to check at any point that it is still telling the truth.