Author mark.dickinson
Recipients derekroconnor, eric.smith, isandler, mark.dickinson
Date 2010-04-04.09:59:11
SpamBayes Score 1.42126e-10
Marked as misclassified No
Message-id <1270375152.98.0.653333826323.issue8309@psf.upfronthosting.co.za>
In-reply-to
Content
> Python 3.1.2 -- 32 bit gives sin(2^60) = -0.7391806966492228

Interesting.  I get 

>>> sin(2.**60)
-0.83064921763725463

On both OS X 10.6.3 (64-bit build) and OS X 10.5.something (32-bit build).  But this is all down to the platform math library, of course.

I agree it's not ideal, but it's not really in Python's power to fix this, without adding a huge extra maintenance burden (from adding the necessary high-precision arithmetic to Python itself) or another external dependency.

You could always try to come up with a patch that adds a configure option for linking Python's math library into MPFR instead of the system math library.  (crlibm would also work, and would be faster;  it doesn't seem to be so well maintained these days though;  Pari/GP isn't such a great option since its floating-point model isn't so compatible with IEEE 754).
History
Date User Action Args
2010-04-04 09:59:13mark.dickinsonsetrecipients: + mark.dickinson, isandler, eric.smith, derekroconnor
2010-04-04 09:59:12mark.dickinsonsetmessageid: <1270375152.98.0.653333826323.issue8309@psf.upfronthosting.co.za>
2010-04-04 09:59:11mark.dickinsonlinkissue8309 messages
2010-04-04 09:59:11mark.dickinsoncreate