Torbjorn Granlund t...@gmplib.org writes:
ni...@lysator.liu.se (Niels Möller) writes:
My bsqrt uses an iteration converging to a^{-1/2}, and broot uses an
iteration converging to a^{1/n - 1}. Both division free.
So binv_sqroot (from mpn/generic/perfpow.c and your bsqrt seem to
Hello,
I have a personal computer which is a PowerMac G5, so based on the
processor PowerPC 970MP.
Are you interested ?
Yours faithfully
Svyatoslav Covanov
___
gmp-devel mailing list
gmp-devel@gmplib.org
http://gmplib.org/mailman/listinfo/gmp-devel
ni...@lysator.liu.se (Niels Möller) writes:
Do you think we should have an advertised binv_sqrt function returning
a^{-1/2}? (And if so, should we have something analogous also for
euclidean square root? And for nth roots?)
Perhaps. A start would be to advertise it internally.
My
Torbjorn Granlund t...@gmplib.org writes:
With all respect for your functions, I don't think we should replace the
tested ones in perfpow. I think those should in he first place be
improved (and put in their own file).
I see. Do you want the functions in the separate files to compute
Torbjorn Granlund t...@gmplib.org writes:
OK. One might also want to consider which is the most useful function.
Computing x^{1/2} and x^{1/n} looks nice, at least in the manual ;-). And
we get there with a single mullo if we compute x^{-1/2} and x^{1/n-1}.
Which variants really are most