Hi Waldek,

thank you for taking a look at this issue. A factor 50 speedup looks
promising. However, I do not think it is enough, if I can tweak the
factorizer to be compiled at higher optimization level. There would be
at least need to exactly explain how anyone can easily do this.

Nevertheless, I think an interface to other existing libraries would be
a good idea.

> Looking again at the problem it is not clear what is the
> intent.

I had posted the original intend here.
https://groups.google.com/g/fricas-devel/c/jmEWZvM5Jdk/m/1j33ik34AAAJ

https://www.mail-archive.com/[email protected]/msg13091.html

I did this some time ago so I guess the code attached there is not
up-to-date. But you asked about the intend.

What I have posted at the beginning of the "slow factorization in finite
field" thread is connected to it. I experimented with my implementation
of a dynamic algebraically closed field and wondered about long running
times. The polynomials

p3 := x^3 - 29
p5 := x^5 - 29
p7 := x^7 - 29

are somewhat made up, because I wanted to test whether my code can
recognize zeros in an algebraic setup.

In fact, a dynamically algebraic closure is like AlgebraicNumber with
the property that the zero test really finds zeros. For example, it can
decide whether sqrt(2)*sqrt(3)-sqrt(6) is true or false. That example is
too easy, but you get the point. Note that is could also be that by
independently creating sqrt(2), sqrt(3) and sqrt(6), it coudl also be
that for consistency, we would have

zero?(sqrt(2)*sqrt(3) - sqrt(6))  is false
and
zero?(sqrt(2)*sqrt(3) + sqrt(6))  is true.

Internally, the dynamic algebraic closure relies on a projection into a
finite field (with enough roots). Clearly with the dynamic extension of
the field one also must sometimes extend the underlying finite field.
And there I need to factor polynomials in finite fields.

If this is too vague, I can try to single out the respective code that I
have and make it public.

Ralf

-- 
You received this message because you are subscribed to the Google Groups 
"FriCAS - computer algebra system" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
To view this discussion on the web visit 
https://groups.google.com/d/msgid/fricas-devel/21ef5a56-bf0b-88d7-77e7-43bd9e8c9b7f%40hemmecke.org.

Reply via email to