Hacker News Reader

Comment by ajross | Hacker News Reader

Honestly? Learn old-style 8086 assembly, then 386 code. The new stuff may be architecturally simpler in many ways, but it sits as an edge case on top of a very thick historical stack that seems completely insane if you look at it a priori.

But the early CPUs were actually quite simple! At the time "CISC" was a good thing because it provided straightforward mechanisms for expressing things (e.g. "push", "call", load a struct field with an offset...) that real world programmers needed to do all the time.

View on HN

Most 64 bit processors only use 48 or 52 bits for physical memory addressing. So you might have a 64 bit virtual address, but that gets crunched down to your real physical memory which might only need e.g. 36 bits for 64GB RAM.

You rarely actually use 64 bit integers. So much of the processing is on strings which use 8 or 16 bits.

Furthermore, your Intel x64 processor still contains the features first introduced with the first IBM PC using the 8088 which in turn was conceptually compatible with the 8008 and 8080. When you try to understand a modern Intel or AMD CPU, you need to be aware of the legacy of 40+ years of backward compatibility. Once you understand the historical decisions, SSE, AVX architectures start making more sense.

I think segmented memory really hampered my ability to learn 8086 assembly well. It really made a mess of just about everything.

It... kinda didn't though. I mean, yes, it seems complicated when you look back and ask "why can't it just have been a flat space", etc... But the same point persists: you have to look at what came before.

Take a look at the complexities of 8080/Z80 or 6502 addressing modes, or the kind of tricks "big" 16 bit architectures like the PDP/11-70 were playing to stretch addressible memory. The 8086 was a breath of fresh air! Your code could be separate from your stack and from your heap, and all three could be a full 64k without any crazy bank switching or copying! And all you had to do was set up 4 segment selectors and then ignore them. And everything you were used to running did so in a clean, unconstrained environment.

Really, read that 8086 datasheet again (it's like 8 pages), it was great stuff at the time.

Never thought in this way in those years. But work is like this. Great perspective.

It also carries over to OS implementations. For example, many OS designed for the 386 failed to implement demand paging for which the CPU had some great support. I believe Windows continued using segmented addressing due to the backward compatibility with earlier releases of the OS.

Reply on news.ycombinator.com