On Apr 17, 2:57 am, John Cremona <john.crem...@gmail.com> wrote:
> 2009/4/17 mabshoff <mabsh...@googlemail.com>:
<SNIP>
> >> File "/home/john/sage-3.4.1.rc3/devel/sage/sage/modular/dirichlet.py",
> >> line 1044:
> >> sage: e.kloosterman_sum_numerical()
> >> Expected:
> >> 7.21644966006e-16 + 1.73205080757*I
> >> Got:
> >> 5.55111512313e-16 + 1.73205080757*I
>
> > Hmm, is there any better input to pick for e so that the real value
> > isn't more or less zero? I can fix the test as numerical noise by
> > changing it to
>
> > '...e-16 + 1.73205080757*I'
>
> > but I would prefer to avoid that if a better value can be found.
>
> That would be a good idea, but I don't know enough about Kloosterman
> sums to help, sorry.
No problem, on Solaris/x86-64 32 bit I am seeing a slightly different
numerical noise value (this is rc2 upgraded to rc3):
sage -t -long devel/sage/sage/modular/dirichlet.py
**********************************************************************
File "/home/mabshoff/build-3.4.1.rc2/sage-3.4.1.rc2-fulvia-gcc-4.3.3/
devel/sage-main/sage/modular/dirichlet.py", line 10
44:
sage: e.kloosterman_sum_numerical()
Expected:
7.21644966006e-16 + 1.73205080757*I
Got:
6.66133814775e-16 + 1.73205080757*I
**********************************************************************
So I will go with the proposed solution for now.
> John
Cheers,
Michael
--~--~---------~--~----~------------~-------~--~----~
To post to this group, send email to sage-devel@googlegroups.com
To unsubscribe from this group, send email to
sage-devel-unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/sage-devel
URLs: http://www.sagemath.org
-~----------~----~----~----~------~----~------~--~---
