4 ms·

Adventures in JIT Compilation (2017)

• dang 6y ago
  Discussed a bit at the time: https://news.ycombinator.com/item?id=13938935 https://news.ycombinator.com/item?id=13938935
• moonchild 6y ago
  > since registers are specified in sub-byte nibbles, one needs very good memory and tons of experience to predict what goes where Nope, you just need to use octal instead of hex - https://web.archive.org/web/20041117160755if_/http://www.dabo.de:80/ccc99/www.camp.ccc.de/radio/help.txt https://web.archive.org/web/20041117160755if_/http://www.dab...
  • CalChris 6y ago
    Since x86 has 8-registers and x86_64 doubles that with a prefix (REX) byte, yeah, looking at that machine code in octal makes a lot more sense. It looks, dare I say it, kinda RISCy.
    • moonchild 6y ago
      Not to mention 16 more 256-bit simd registers.
• twhitmore 6y ago
  I would recommend building a tutorial around a source language that wasn't deliberately obtuse. I do see it makes the parser simple, but it reduces the audience to the small fraction that think a deliberately confusing language is useful. My opinion anyway.
• ficklepickle 6y ago
  I thoroughly enjoyed that. As someone who programs almost exclusively in high level languages, I learned a lot. It made brainfuck a bit more intelligible, too. It was actually a great choice of language for this exercise.