Hi Robert,
Sure no problems.
Just to confirm, though, I notice that it states that the code is part
of GNU MP. That's absolutely no problem, but just wanted to make sure
that this is not because you based your code on existing code in GMP?
If so, it is very important to know that you based it on a version of
GMP that was also LGPL v2.1 or at the user's convenience any later
version. Alternatively, if the code is identical to code as it was in
GMP when licensed under v2.1 + and does not incorporate any
improvements that were added under the later license.
If the code is all entirely your own, then this issue doesn't arise, naturally.
I see it relies on many other mpz functions. I have to admit, I
haven't checked to what extent this is the "done thing" in the mpz
layer.
By the way, I have a gcd function which might be faster than the one
you use. When I get some time, hopefully later today or tomorrow, I'll
post that and we can do a comparison to see if it makes any
difference.
The algorithms are certainly of interest. I certainly get the
remainder trees and power residues. Nice!
Do you have a canonical reference on the Strong Lehmer test. As a
number theorist, I should have heard of it, but to my embarrassment, I
have not! {blush}
Bill.
2009/4/22 Robert Gerbicz <robert.gerb...@gmail.com>:
>
>
> 2009/4/22 Cactus <rieman...@googlemail.com>
>>
>>
>>
>> On Apr 22, 5:40 pm, gerrob <robert.gerb...@gmail.com> wrote:
>> > Hi!
>> >
>> > I've written a much faster code than the currently used in gmp/mpir.
>> > Million of bits numbers can be tested in about 1 sec, but slower by a
>> > factor of 2 up to about 1000 bits numbers on my pc. Tested a lot, it
>> > should be good.
>> > You can use it for whatever you want. I've sent it to Torbjorn
>> > Granlund (main developer of gmp project) also.
>> >
>> > See the uploaded file in the google group, it's name newperfpow.c
>>
>> Thank you for your contribution, which looks very interesting.
>>
>> MPIR is released under the LGPL v2.1 license whereas your code is LGPL
>> v3 licensed.
>>
>> Would you be willing to release it under a license that would allow
>> MPIR to incorporate it?
>>
>> Thank you for your interest, which is much appreciated.
>>
>> Brian Gladman
>>
>>
>
> OK, I've changed the license, is it good now?
>
> The idea of using power residues to cancel many exponents at power testing
> is an old method to avoid the Newton iteration. To do this we have to
> compute many remainders, for this I'm using a remainder tree to speedup this
> step. The strong Lehmer test is probably a new idea in this area.
>
> >
>
--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups
"mpir-devel" group.
To post to this group, send email to mpir-devel@googlegroups.com
To unsubscribe from this group, send email to
mpir-devel+unsubscr...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/mpir-devel?hl=en
-~----------~----~----~----~------~----~------~--~---
