[SciPy-User] nan handling by scipy.optimize

Micha michf@post.tau.ac...
Thu Jul 26 00:33:08 CDT 2012

On what computers are you and the student running? The are sometimes changes in hardware, especially with optimized code that uses sse and/or fused multiply add

Pauli Virtanen <pav@iki.fi> wrote:

>Alan G Isaac <alan.isaac <at> gmail.com> writes:
>> Has there been any recent change in nan handling in scipy.optimize?
>> I'm using a very recent EPD distribution, and a student
>> is using a release with Python 2.6.5.  I'm getting a
>> zero returned by brentq and bisect when (iiuc) the function
>> evaluation produces a nan, while he is not.
>> The search interval is bounded away from zero.
>There are no code changes in Scipy 0.8.0b1...v0.10.1 (which are
>the versions I guess apply for your issue, please clarify) for
>the `bisect` and `brentq` routines.
>Pauli Virtanen
>SciPy-User mailing list
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mail.scipy.org/pipermail/scipy-user/attachments/20120726/de4b9335/attachment.html 

More information about the SciPy-User mailing list