mthkvv

160 Reputation

3 Badges

6 years, 130 days

MaplePrimes Activity


These are replies submitted by mthkvv

@dohashi 

No, 64-bit. And I didn't set any limits

Hi all,

Are there any developers of modp1 here?

Thank you

@Carl Love 

> I don't understand "unit" or the significance of theta, which doesn't otherwise appear in the algorithm.

It's really doesn't matter. You can just read that Input of norm - two polynomials f(x) and beta(x) (that appears in line 1). In my doc XGCD.mw - ff(t) and gg(t) respectively. I combined both functions into one XGCD in my code.

Hi all,

Any ideas?

@vv 

It would be great)

@vv 

I think so. I would like to know more precisely

@Carl Love 

"You can speed up the coefficient arithmetic by exploiting the fact that the modulus is a power of two. Thus, the mod operation is just a bit truncation."

I doubt that still not included in GMP.

@Carl Love 

So I can't get it from Maple? And should use C++ or Java?

@Carl Love 

No, this isn't a solution. I'm looking in a way of Toom-Cook and Schonhage-Strassen algorithms. Do you familiar with them?

@Carl Love 

Yes. More precisely - lifting of curve in p-adic ring with N precision. See Ch. 3 in Vercauteren's thesis

@vv 

Thank you! Maybe some another way?

@Carl Love 

Thank you.

pols.txt

@Carl Love 

Now version for m=1031:

ex_n.mw

@vv 

It's a part of different algorithms of elliptic curves point counting. N - is a so called precision (or accuracy) of lifting of polynom. You can read about this:

https://www.semanticscholar.org/paper/Computing-Zeta-Functions-of-Curves-over-Finite-Vercauteren/c945c98267db064b272c87a885fc5eeb764b0b2d

It's equivalent to Rem(a*a,f_t,t) mod 2^N. I need to speed up this simple code line. Any ideas?

PS. And Rem(a*b,f_t,t) mod 2^N, where a<>b

@Carl Love 

I verified. modpol working well with prime-power numbers. How I can to speed up this? Maybe with modp1, GF or domains?

1 2 3 4 5 Page 3 of 5