#18106: Maximum depth recursion in QQbar
---------------------------------+------------------------
       Reporter:  vdelecroix     |        Owner:
           Type:  defect         |       Status:  new
       Priority:  major          |    Milestone:  sage-6.6
      Component:  number fields  |   Resolution:
       Keywords:  sd66           |    Merged in:
        Authors:                 |    Reviewers:
Report Upstream:  N/A            |  Work issues:
         Branch:                 |       Commit:
   Dependencies:                 |     Stopgaps:
---------------------------------+------------------------

Comment (by vdelecroix):

 Replying to [comment:2 gagern]:
 > On the other hand, we could also try addressing the source of this deep
 recursion. The way I see it, that's because addition is left associative,
 so that cyclotomic polynomial will be a very deep but thin binary tree. If
 we had a representation which describes a sum (and perhaps also a product)
 of an arbitrary number of algebraic numbers using a single descriptor, the
 data structure would become much more shallow.

 Looks like a good idea to have a polynomial descriptor for one (or
 several?) algebraic numbers. It might even be used to get faster and more
 precise interval evaluations.

 > As a third solution, we might set up our own evaluation machinery for
 these trees, with our own stack instead of Python recursion. I haven't yet
 worked out all the details, but if this sounds interesting I might write
 some code to see how this approach feels.

 Looks reasonable to do it without recursion. We might obtain a good speed
 up.

 > The way I see it, since the backtrace is about `_interval_fast` and
 `_more_precision`, all of this is happening before exact computation is
 triggered, right?

 Right!

 Do you find reasonable to open two tickets:
  - one for polynomial descriptor in one variable
  - one for evaluation without recursion

 Vincent

--
Ticket URL: <http://trac.sagemath.org/ticket/18106#comment:3>
Sage <http://www.sagemath.org>
Sage: Creating a Viable Open Source Alternative to Magma, Maple, Mathematica, 
and MATLAB

-- 
You received this message because you are subscribed to the Google Groups 
"sage-trac" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
To post to this group, send email to [email protected].
Visit this group at http://groups.google.com/group/sage-trac.
For more options, visit https://groups.google.com/d/optout.

Reply via email to