Re: Writing fast compilers...

[From comp.arch -John]


davidsen@crdos1.crd.ge.COM (Wm E Davidsen Jr) writes:


> An overstatement, I think. Optimizing compilers are not going to
>produce as much improvement on well written code as they will on student
>first attempts, because the common subexpressions have be done in
>source, the invariant stuff is out of loops, the loops are unrolled, the
>register type has been used here and there... all by a one pass code
>generator called a programmer.


The only program i ever generated in one pass did little more than
print out `Hello world!', and even then i cheated by looking
up the hard parts :-).


About doing optimizations at the source level ... yes, but. Not all
common subexpressions are available at the source level, and many of
those that are tend to turn a readable program into a morass of
pointer arithmetic and mysterious extra variables. excessive loop
unrolling is maybe attractive for people who get paid by the line,
hardly for the maintenance programmer. Using the `register' storage
class specifier (in languages where such an anachronism is available)
to achieve speedups is hardly worthwhile with modern compilers. they
tend to ignore it.


The gist is, a reasonable amount of optimization seems worthwhile
even for `well written code'.
--
Alexander Vrchoticky | alex@vmars.tuwien.ac.at
TU Vienna, CS/Real-Time Systems | +43/222/58801-8168
--

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