Le forum de XCAS

As always, thanks for this enlightening explanation Bernard! Yes, I will try to do benchmark of more polynomial systems (not just katsura9, as you say katsura>11 if this will show the power of giac!) to show the full power of giac. Anyway, now at least I understand. It seems though that in poly.lib...

You should also consider the giacsettings.threads value and also giacsettings.proba_epsilon (if probalistic algorithm are disabled it is much slower). In ipython %time gives both times: (cpu time is less with 1 thread but wall time is longer) In [5]: import giacpy In [7]: giacpy.giacsettings.threads...

Dans icas version 1.5.0-49 compilée moi meme: est ce que ca le fait avec tes versions aussi?
Frederic

No my previous example gives 0 under Xcas also.

Digits is avaible with giacpy see:

OK merci, en fait j'ai mis ce petit patch dans le package sage, il a deja eu une revue positive.

Fred

Bonjour Bernard, avec les sources 1.5.0-37 j'obtiens un résultat different d'avec une version de giac plus ancienne 1.4.9 pour une base de groebner de cyclic6 modulo p=2**31-1 (les termes B[24] et B[40] sont differents. p:=2**31-1; I:=[x0 + x1 + x2 + x3 + x4 + x5, x0*x1 + x1*x2 + x2*x3 + x3*x4 + x0*...

This is not giacpy related, I have a similar problem with n>=12: whatever is epsilon or Digits. n:=12:; expand((evalf(10**n,100)*x^2+x)/evalf(10**n,100)) (I put evalf with 100 digits to avoid global Digits or epsilon2zero limitations. Bernard, is there some epsilon2zero use in this division? or is i...

Yes but I think it is just because the windows dll is built with less dependency (may be lapack here?) so here is could encounter some accuract problem with [x0,x2] and not with [x2,x0]


Indeed it does the same in geogebra:

The problem is that ratnormal can output a symbol with sommet * or + So if the sommet is + you just have to do Q.op() >>> P=x**2+3*x+y >>> Q=P.ratnormal() >>> Q.sommet() '+' >>> Q.op() x**2,3*x,y >>> P2=x**2+3*x/2+y # warning 3/2*x is x in python2 but '3/2'*x or 3*x/2 are OK >>> Q2=P2.ratnormal() >>...

Thank you,
Yes they were errors. NB dimpase also mention that register gives a warning that will become an error in C++17

cf: https://trac.sagemath.org/ticket/26315#comment:19

Hi bernard,

curlbuild.h was removed from libcurl since 7.55, so giac's built is broken with recent versions of libcurl. Although it was replaced by system.h, it seems that removing this include is working and less version dependant.

Best
Frederic

Hi bernard,
on this topic,

https://trac.sagemath.org/ticket/26315#comment:14

dimpase suggest the following patch to build giac with clang 6.0.1


https://trac.sagemath.org/raw-attachmen ... g601.patch

could you tell us if you plan to include it in giac?
Best
Frederic

In fact my question was about icas. The one I have built for windows doesn't take ctrl-c when in nauty computation, and I was wondering if it was the same on linux.

Frederic

Dans la version que j'ai, return ; retourne tout de meme quelque chose, ca risque de perturber les gens. Je suis oblige de renvoyer quelque chose, peut-etre qu'on peut remplacer une sequence vide par autre chose, mais quoi? Si je fais un espace entre simple quotes return ' ' alors ca ne se voit pas...

J'ai essayé! L'editeur peut aussi servir au python car la police donne une vue plus globale du programme. Dans la version que j'ai, return ; retourne tout de meme quelque chose, ca risque de perturber les gens. est ce que: a=float(input("tutu")) est sensé marcher? (c'est sympa de pouvoir lancer un m...