|
|
nemesis wrote:
> Still, C is a very preferred target for Scheme fast code generators.
> Many translation techniques have matured and proven very useful in using
> C as intermediate language.
I've seen translators where the outputted C code is basically one function
per high-level source file, and the body of the function is full of gotos
and switches and all kinds of stuff like that. Again, completely unreadable,
but that isn't really the point. I've seen where it (for example) generates
C functions with >65536 labels in a file, breaking the compiler.
I would think if you want to optimize *that* sort of thing, you'd be better
off with a lower-level IM like JVM or CIL.
> Yes, it will never be as fast as C obfuscated and cryptic code written
> by hand by a hacker, but the point is not that: it is to write very
> high level code that still performs quite favorably to low-level C as
> assembly.
Sure. I guess taking advantage of the more complex chips these days can be
done much more easily by writing C code than by trying to generate good
assembly code. Perhaps I just haven't updated my prejudices since CPUs have
been simple enough to program by hand. :-)
--
Darren New, San Diego CA, USA (PST)
There's no CD like OCD, there's no CD I knoooow!
Post a reply to this message
|
|