[Shootout-list] Regexmatch woes with Haskell
Isaac Gouy
igouy2@yahoo.com
Tue, 8 Mar 2005 10:25:32 -0800 (PST)
imo We already have a solution for this issue, and we've already
started working on it. The solution is to define benchmarks that do
real-work rather than make-work.
We'll be able to define a fine replacement for regexmatch based on
FASTA files. That will happen. It hasn't happened yet because we're in
the process of replacing some of the other benchmarks.
--- Brent Fulgham <bfulg@pacbell.net> wrote:
>
> --- Einar Karttunen <ekarttun@cs.helsinki.fi> wrote:
>
> > Implementing regexmatch correctly in Haskell seems
> > quite problematic.
> [ ... ]
> > The GHC optimizer looks at that and generates
> > code for "print x" - as the thing inside the loop is
> > not used for anything and has no side-effects.
>
> I've had this issue bounching around in my head for
> some time. I don't intend for the benchmarks to
> become
> confounding, arcane messes just to require the
> language to perform the right number of steps.
>
> I wonder if it would be better to allow the compiler
> to perform any optimizations it wants, and just use
> much larger input files?
__________________________________
Celebrate Yahoo!'s 10th Birthday!
Yahoo! Netrospective: 100 Moments of the Web
http://birthday.yahoo.com/netrospective/