Re: C as assembly language

VBDis wrote:
>
> Im Artikel 01-04-093, "Randall Hyde"
> <rhyde@transdimension.com> schreibt:
>
> >Guess it depends upon how you define "real" assembler.
>
> Right, I only had one point in mind, and forgot about all the other things,
> which an assembler is assumed to do :-(
>
> What I had in mind was, that an assembler is not normally allowed to insert
> instructions, modify the sequence of instructions, or the instructions
> themselves. At least this was the point of some hard core assembly programmers,
> which thought to know better how to write efficient code, and possibly also
> wrote self modifying code.


  Yes, but does it HAVE to be this way ?


> In short, when someone says "do it exactly this way", then an assembler should
> do it that way. In all other cases the assembler can provide optimizations and
> address calculations, macros and other handy stuff.


  We work a lot with PLD tools/compilers, and one attribute they have,
is control over KEEP / Collapse of logic nodes ( ie downstream LOGIC
minimise / optimise )
  Sometimes in PLD code you want to say "don't fiddle with this at all!"


  A similar option in Assemblers could benefit (all?) Assemblers, and
allow work on lower level optimise engines, as other threads have
discussed.


  - jg

Return to the comp.compilers page.
Search the comp.compilers archives again.