Dnia Sun, 31 Jan 2010 09:48:47 -0700, Robert Dodier napisa?(a):
> On 1/30/10, Adam Majewski <adammaj1 at o2.pl> wrote:
>
>> Thx for all answers.
>> Here is a result :
>> http://fraktal.republika.pl/eigensolve.html
>
> Terrific, keep up the good work.
>
> I wonder why you are calling the external program eigensolve. Is it
> faster than the corresponding Maxima function or Maxima has some bug or
> what??
>
> best
>
> Robert Dodier
Thank you for a nice words, Maxima and help.
It's a complex problem. (:-)
First because period is any positive integer so there is no upper limit
and only hardware/software limitations.
Thus this could be treated as an numerical benchmark.
I do not know which program is the best.
I have checked some :
Maxima ( allroots, bfallroots, polyroots)
mpsolve
eigensolve
octave
Web Solver
Results are on my page : fraktal.republika.pl
Now I know that on my computer Maxima fails with polynomial
for period 14 ( degree 8127 ) - computing this polynomial ( before
computing its roots)
The comparison is hard because :
- some programs were done under different OS : win32 and Linux 64 bit
- last algorithm is optimised
- (I hope that) new programs are better,
- I do not know above programs in the same level
so my scripts are not using all power of programs
I'm intresting in compiling Maxima from sources with maximum memory.
(Now I have SBCL on 64 bit and my maxima is from packages by Istvan
Blahota :
http://zeus.nyf.hu/~blahota/
It would be great if some notes about it one could find in doc.
I have found sth like :
http://fraktal.republika.pl/doc/maxima%20error%20codes.txt
Maxima is great, Lisp also (:-)
Best regards
Adam