[Discuss] Test for positive gives wacko results for underflowed values for the gcc compiler suite (including the gcc compiler, and g77 and gfortran compilers)

[Alan W. Irwin]

On 2007-06-25 14:19-0700 Murray Strome wrote:

> I don't really know the answer to this one, but just as a guess, is it 
> possible that Python would automatically use a larger word size for 
> calculating these things (e.g. 64 bit or 128 bit floating point as opposed to 
> 32bit)? Might that be the difference?

That's an interesting question.

According to python documentation (Python Essential reference p. 23), python
uses native double precision (64-bit floating point), i.e., it is identical
to the C double type.  Fortran real*8 (for g77 and gfortran) is also
identical to the C double type.  Thus, you should expect to get the same
answers from g77/gfortran, gcc, and python for these test cases, but you
don't.

Alan
__________________________
Alan W. Irwin

Astronomical research affiliation with Department of Physics and Astronomy,
University of Victoria (astrowww.phys.uvic.ca).

Programming affiliations with the FreeEOS equation-of-state implementation
for stellar interiors (freeeos.sf.net); PLplot scientific plotting software
package (plplot.org); the libLASi project (unifont.org/lasi); the Loads of
Linux Links project (loll.sf.net); and the Linux Brochure Project
(lbproject.sf.net).
__________________________

Linux-powered Science
__________________________