80 bit precision ?

[km]

Hi all,

does python currently support 80 bit precision Floating Point Unit ?

regards,
KM

[Sébastien Boisgérault]

Re: 80 bit precision ?


Not in the core language or the std library.

However, if you are insterested in high-precision
computations, gmpy may be useful:

http://sourceforge.net/projects/gmpy/

To be honest, I have never used it ;). A review
would be appreciated.

Regards,

SB

[Paul Rubin]

Re: 80 bit precision ?

"Sébastien Boisgérault" <Sebastien.Bois gerault@gmail.c om> writes:[color=blue]
> http://sourceforge.net/projects/gmpy/
>
> To be honest, I have never used it ;). A review would be
> appreciated.[/color]

I've used it for large integer calculations and it's great, around 5x
faster than Python longs if I remember right. I did hit a couple bugs
that have been fixed since then, and there are a few interface quirks.

[chris]

Re: 80 bit precision ?


"km" <km@mrna.tn.nic .in> wrote in message
news:mailman.50 .1116011033.293 28.python-list@python.org ...[color=blue]
> Hi all,
>
> does python currently support 80 bit precision Floating Point Unit ?
>
> regards,
> KM[/color]

I've just been using the decimal module for Python 2.3 (native in Python
2.4) at 100 digit precision to solve some very high order polynomials. Slow
but worked well for me. Also used a rational.py module ported to Py 2.3.
Perhaps you can do your calcs to the precision you want with a combination
of decimal and rational arithmetic.

bwaha