On 15 November 2016 at 21:44, Jori Mäntysalo wrote:
> On Tue, 15 Nov 2016, John H Palmieri wrote:
>
>> Also, should we switch to a naive implementation of sigma: essentially
>> just
>> return sum(divisors(n))?
>
>
> The answer to questions like this is always the same, I guess:
>
> Add string-valu
On Tue, 15 Nov 2016, John H Palmieri wrote:
Also, should we switch to a naive implementation of sigma: essentially just
return sum(divisors(n))?
The answer to questions like this is always the same, I guess:
Add string-valued parameter 'algorithm' with None as default. For None use
some kind
I guess that it would be much better to also propose access to
1) the flint function
void fmpz_divisor_sigma ( fmpz_t res , const fmpz_t n , ulong k )
2) the pari function sigma
For example, pari timing is competitive in this range
$ sage -c "timeit('L = [sigma(n) for n in range(10**13, 10
On Tue, Nov 15, 2016 at 11:47 AM, John H Palmieri
wrote:
> Inspired by the ask.sagemath question
> https://ask.sagemath.org/question/35587/why-sigman-seems-not-so-performant-for-small-n/,
> I started looking at timings for the sigma function (sigma(n) = sum of the
> divisors of n, sigma(n, k) = su
Inspired by the ask.sagemath question
https://ask.sagemath.org/question/35587/why-sigman-seems-not-so-performant-for-small-n/,
I started looking at timings for the sigma function (sigma(n) = sum of the
divisors of n, sigma(n, k) = sum of the kth powers of the divisors of n).
On my computer, a