From: Robert Klemme <bob.news@...>
Date: 2003-02-27T19:19:01+09:00
Subject: Re: Ruby Compile-time optimization


"Daniel Carrera" <dcarrera@math.umd.edu> schrieb im Newsbeitrag
news:20030227072950.GD17232@math.umd.edu...
> On Thu, Feb 27, 2003 at 04:18:24PM +0900, Robert wrote:
> > ^Hi,
> >
> > did you do this with a loop body that does really do something?  I
mean,
> > your example is a bit artificial and optimizing this scenario would
have not
> > much impact for real code, because nobody would repeat a constant
operation
> > in a loop exept for performance measurements.
> >
> > You could simply sum up a sequence of random numbers or something
similar.
> >
> > Kind regards
> >
> >     robert
>
> Correct me if I'm wrong, I very well might be.  I think that my example
> suggests that Ruby's loops and assignments are not more expensive than
> Perl's, but Ruby's Fixnum#+ is.  We might infer that this is true for
> other arithmetic operations as well.  Am I right?

When it comes to performance measuring is much better than infering,
because experience shows that out intuition about bottlenecks often betrays
us.

> This is only the most miniscule example of optimization, I know.  But I
> wonder if we can use traditional optimization techniques to (like
constant
> folding) to speed up Ruby without significantly altering Ruby's
> implementation.

I doubt the "significant" for constant folding: Does constant folding
really help in many cases?  If I use a constant value I will typically
assign it to a constant variable and reuse that - at least I think this is
a typical way to proceed.

But OTOH improving performance of "+" will surely help a lot.

> It's just a thought.  I don't really know much about the subject.

Me either. :-)

    robert