>>13151522I have read all of Vol1, some of Vol2 mainly the part about arithmetical operations that was assigned to me in a complexity theory course, much of vol 3 out of interest and vol 4a only some so far. I also went through Knuth's 'Math prelim redux' in the new unreleased draft versions.
It is an excellent entertainment book that will blow your mind if you are interested in things like compiler optimization. For example his knew MMIX architecture you can run different competing branch predictions and experiment with custom pipelines, it's like the ultimate CPU theoretical test bench. MMIXAL is also much easier to write than MIXAL because there's 256 registers so the code is much shorter.
If you enjoy challenging puzzles, or math competitions, things like that, you should read it for entertainment because that's exactly why Knuth wrote that series, a series of challenging problems for computer scientists so instead of drinking a coffee on a sunday morning and doing the NY times crossword, you instead patch ancient algorithms, optimize computational arithmetic something that will be extremely useful for quantum virtual machines/simulators, and learn exactly how trees work and tricks to do O(1) operations on them. Even the shitty tape drive/external storage algorithms in Vol 3 are useful today because we all use external cloud storage over a network, and maybe you want the most efficient way to combine these together like Knuth did back when he combined tape drives in the 1960s.
I am also routinely surprised the quality of the physical books, I've never owned books made with such quality every thought and care like having the pages being able to open without folding and custom binding Knuth oversees. Totally beautiful typography and crystal clear. Knuth will explain the history of everything in modern CS in the books pointing to papers you can look up something invaluable when I had to do vol2 for school. 10/10 books if this kind of thing is your bag