#15299: Incorrect results for analytic Sha due to low precision
-----------------------------------+----------------------------
Reporter: jdemeyer | Owner:
Type: defect | Status: needs_review
Priority: major | Milestone: sage-5.13
Component: elliptic curves | Resolution:
Keywords: | Merged in:
Authors: Jeroen Demeyer | Reviewers:
Report Upstream: N/A | Work issues:
Branch: | Commit:
Dependencies: | Stopgaps:
-----------------------------------+----------------------------
Comment (by cremona):
Replying to [comment:10 jdemeyer]:
> Corrected the error computation for `at1()`. I believe this is rigorous
now:
> {{{
> for n in xrange(1,k+1):
> term = (zpow * an[n])/n
> zpow *= z
> L += term
> # 8n+1 is the relative error in half-ulps to compute term.
> # For addition, multiplication, division, sqrt, this is
> # bounded by the number of operations. exp(x) multiplies the
> # relative error by abs(x) and adds 1 half-ulp. The relative
> # error for -2*pi/sqrtN is 3 half-ulps. Assuming that
> # 2*pi/sqrtN <= 2, the relative error for z is 7 half-ulps.
> # This implies a relative error of 8n-1 half-ulps for zpow.
> # Adding 2 for the computation of term gives:
> error += term.ulp()*(8*n+1) + L.ulp()
> }}}
I can see where this is in the code -- can you say how it affects any
doctest outputs? I am not a numerical analyst...
--
Ticket URL: <http://trac.sagemath.org/ticket/15299#comment:11>
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/groups/opt_out.
