[Shootout-list] Mandelbrot Set for CMUCL and SBCL 2nd try

Sebastien Loisel Sebastien Loisel <sloisel@gmail.com>
Mon, 28 Mar 2005 14:39:10 -0800


Reproducing the exact floating point behavior of a C program in
another language isn't reasonable. The image comparison should have a
threshold.

Sebastien Loisel


On Sun, 27 Mar 2005 14:47:01 -0500, Yannick Gingras
<ygingras@ygingras.net> wrote:
> On March 27, 2005 02:26 pm, Isaac Gouy wrote:
> > Sorry we need to add an example output file for this benchmark.
> >
> > Quickest approach - use the Python or Lua or Perl or ... program to
> > generate example output and diff against that.
> 
> Yes, I'm really close to the C output but I still get a few bits here
> and there that are not the same.  If I look at the pictures they
> really look the same.  About 99.99% of the bits are the same.  It
> probably is some floating point errors creeping in...
> 
>  : \
> 
> Should my output be exactly the same ?  There is also a few
> implementations skipping the last byte if the size is not a multiple
> of 8, should I do so too ?
> 
>  ; )
> 
> --
> Yannick Gingras         Math Student, UQAM
> Join FSF as an Associate Member at:
>   http://member.fsf.org/join?referrer=2327
> 
> 
>