Re: The boogus: Shootout website

On May 20, 11:05 pm, Razi<fg...@[EMAIL PROTECTED]
> wrote:
> On Mon, 5 May 2008 08:27:23 -0700 (PDT), Isaac Gouy <igo...@[EMAIL PROTECTED]
>
> wrote:
>
> >The FAQ gives instructions on how to contribute better programs
>
> >http://shootout.alioth.debian.org/gp4/faq.php#play
>
> Well, I have been around but wasting my time on this bogusshootout
> site.
>
> I submitted a twice faster "partialsums" but it was put in
> "interesting alternative" section, even though it follows all
> benchmark rules.
>
> http://shootout.alioth.debian.org/gp4/benchmark.php?test=partialsums&...
>
> Also, more seriously, in the "revcomp" benchmark, both java versions,
> that like top C and D programs, use byte arrays (instead of create
> String objects) are put in "alternative" sections.
>
> http://shootout.alioth.debian.org/gp4/benchmark.php?test=revcomp&lang...


As you already know those revcomp Java programs are doing 8k block
reads - contribute a program that reads 82 chars at a time.


> the "alternative" sections is not used in the overall ranking, so it's
> total waste.  The excuse given is that all programs must read data
> line-by-line. Just to humor the guy, I "fixed" it so that it reads
> bytes line-by-line (and is still twice faster than the version that
> creates String objects). Gouy totally ignored the "fix" and apparently
> has no interest in using the twice faster version.


Your revcomp program is waiting safely, along with other peoples
programs - I will deal with it when I find time.


> Also, remember the bogus excuses he gave for not including Jet, even
> after Excelsior gave him permission in email.
>
> Conclusion: The site seems to have an hardcore agenda. It's an ok site
> but just remember that they have an agenda. They probably are getting
> paid to promote some languages and/or products/compilers.


Stop making things up.

Our hardcore agenda is to show programs in many programming languages.

We work on the benchmarks game for free - we don't get money, we don't
get stuff.



> For now I am done wasting time on it, I think, especially since anything
faster
> that changes the ranking is ignored or put in "alternative."


Your style of name calling might give people the impression that
whatever you contributed to the benchmarks game was worthless - but
that would be wrong.

I congratulate you on searching out the slowdowns in the Java programs
and trying to find workable alternatives.