[racket] Computer Language Benchmark Game

From: Eli Barzilay (eli at barzilay.org)
Date: Thu Feb 16 15:41:12 EST 2012

A few minutes ago, namekuseijin wrote:
> Looking at:
> http://shootout.alioth.debian.org/u64q/program.php?test=spectralnorm&lang=racket&id=3
> 
> I stumbled upon this:
> 
> ;; the big let improves performance by about 20%
> (let* ()
>   ... various defs...)
> 
> bizarre... why?  And is it still true?

(Maybe leftover from trying the code on the repl?)

> BTW, this:
> 
> http://shootout.alioth.debian.org/u64q/program.php?test=pidigits&lang=racket&id=2
> 
> seems to be measuring gmp performance rather than racket's...

Yes, it's unfortunate, but that's what practically everyone does.

BTW, my translation had some optimizations that made it at the top of
that benchmark for a long time.  It was later borrowed by a bunch of
other implementations.  (And mostly without attribution...  On a quick
glance, I see only one leftover comment in the ATS#2.)

BTW₂, the whole shootout thing is generally biased and almost
completely useless for drawing any kind of conclusions.  Really.
(I mean it in a stronger sense than usual benchmark disclaimers.)

-- 
          ((lambda (x) (x x)) (lambda (x) (x x)))          Eli Barzilay:
                    http://barzilay.org/                   Maze is Life!


Posted on the users mailing list.