| 1 |
On Wed, 16 Apr 2014 19:13:54 +1200 |
| 2 |
François Bissey <fbissey@××××××××××××.nz> wrote: |
| 3 |
|
| 4 |
> On 2014-04-16 19:06, François Bissey wrote: |
| 5 |
> > On 2014-04-16 18:25, Thomas Kahle wrote: |
| 6 |
> >> On 16/04/14 08:15, grozin@g.o wrote: |
| 7 |
> >>> On Sun, 13 Apr 2014, hasufell wrote: |
| 8 |
> >>>> I'm looking for a C lib that can do basic polynomial arithmetic |
| 9 |
> >>>> as well |
| 10 |
> >>>> as algebraic stuff like modular inverse. |
| 11 |
> >>> Maybe GiNaC. It definitely can do polynomials. |
| 12 |
|
| 13 |
I wouldn't recommend using GiNaC if you know which algebraic domain you |
| 14 |
are working with. Once the domain is known, there are special algorithms |
| 15 |
you can take advantage of for arithmetic. GiNaC's generic |
| 16 |
implementation is faster only if you have really sparse polynomials. |
| 17 |
|
| 18 |
> >>> Another possibility is pari. It can be used as a library, not |
| 19 |
> >>> only via |
| 20 |
> >>> gp. It is highly optimized and efficient. |
| 21 |
> >> |
| 22 |
> >> I'm pretty sure that this can be done in sage. You could see |
| 23 |
> >> which library they use under the hood. Taking this to |
| 24 |
> >> sci-mailing list, there your question gets a lot more attention. |
| 25 |
> >> Definitely can be done in sage. What they use under the hood |
| 26 |
> >> depends |
| 27 |
> > on the ring the polynomial are defined over. But I think for most |
| 28 |
> > applications singular is used. |
| 29 |
|
| 30 |
As Thomas said, Sage picks the underlying library depending on the |
| 31 |
coefficient ring, number of variables, etc. Which domain do you want to |
| 32 |
work with? |
| 33 |
|
| 34 |
NTL [1] and FLINT [2] are good options for univariate polynomials over |
| 35 |
various base rings for instance. |
| 36 |
|
| 37 |
[1] http://www.shoup.net/ntl/ |
| 38 |
|
| 39 |
[2] http://flintlib.org/ |
| 40 |
|
| 41 |
|
| 42 |
Cheers, |
| 43 |
Burcin |
Archives