great job

[Juergen Bullinger]

Hi all,

I just downloaded two of the speed improvements from the swox homepage.
Namely Paul Zimmerman's mul_fft improvement and mpn_divexact. I expected
that it would improve my calculations based on mpz_powm by a few percent
or maybe even not at all. But in fact it more than doubled the speed. On
my Athlon 64 3200+ it took about 14 minutes to do a mpz_powm with
operands of about 31000 decimal digits size (for both the exponent and
the module) with GMP 4.1.4. Now it takes about 6 minutes to do such a
calculation.

Thank you. I am really looking forward for the next release of GMP :o)
Are there already schedules? On the GMP page
(http://www.swox.com/gmp/#FUTURE) I even read that "4.2 will probably
released in 2005".
If you need another beta tester, please contact me.

regards
Juergen
-- 
Juergen Bullinger <juergen.bullinger at gmx.net>