Re: Optimizing Symbol#to_proc

On Friday 06 July 2007, Nobuyoshi Nakada wrote:
> Hi,
>
> At Sat, 7 Jul 2007 03:06:21 +0900,
>
> Sylvain Joyeux wrote in [ruby-core:11598]:
> > I don't what the "old Ruby version" was, but I tried to implement mine
> > and to my surprise my custom version is *much* faster. Does anybody
> > know why ?
> >
> > ========= 2. First naive pure-Ruby version ==============
> > class Symbol
> >     def to_proc
> >         proc { |x| x.send(self) }
> >     end
> > end
>
> VM would optimize the block code.
What do you mean by that ? It replaces x.send(self) by x.#{send} ?

> > ========= 3. Second pure-Ruby version ==============
> > class Symbol
> >     def to_proc
> >         @proc ||= eval <<-EOD
> >         proc { |x| x.#{self} }
> >         EOD
> >     end
> > end
>
> You added the cached Proc object.  It does get faster indeed.
Yes, I know that 3 is supposed to be faster than 2

The problem is the comparison with the built-in Ruby version ... The 1. is 
the one provided by default with Ruby, and it's much slower than the other 
two. Do you have any explanation ?
-- 
Sylvain Joyeux