On Jan 15, 9:23 pm, Robert Bradshaw <[EMAIL PROTECTED]> wrote: > I was under the impression that MPFR was supposed to give identical > answers across all platforms (unlike, say, RDF). I was wondering if > any experts in the area could explain the numerical noise in cases like > > http://sagetrac.org/sage_trac/ticket/1790 > > Is this something to be worried about, or are MPFR results not truly > platform independent?
Computations done only with MPFR/MPFI should definitely be platform independent. However, p.roots(ring=CIF) is done by first finding the roots numerically (using either numpy or Pari, depending on the precision), and then verifying/refining the roots using MPFI. The initial numpy or Pari roots may differ across platforms, so the MPFI refinement may also give different results. (I believe that the results are always correct, though, in that the returned intervals do always include the real root.) I don't know how to make the results platform-independent without writing a lot of code (and possibly slowing down root-finding significantly). Carl --~--~---------~--~----~------------~-------~--~----~ To post to this group, send email to sage-devel@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/sage-devel URLs: http://sage.scipy.org/sage/ and http://modular.math.washington.edu/sage/ -~----------~----~----~----~------~----~------~--~---
