pqc/external/flint-2.4.3/fmpz_poly/div_divconquer_recursive.c

109 lines
3.3 KiB
C

/*=============================================================================
This file is part of FLINT.
FLINT is free software; you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by
the Free Software Foundation; either version 2 of the License, or
(at your option) any later version.
FLINT is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
GNU General Public License for more details.
You should have received a copy of the GNU General Public License
along with FLINT; if not, write to the Free Software
Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA
=============================================================================*/
/******************************************************************************
Copyright (C) 2010 Sebastian Pancratz
******************************************************************************/
#include <stdlib.h>
#include <gmp.h>
#include "flint.h"
#include "fmpz.h"
#include "fmpz_vec.h"
#include "fmpz_poly.h"
#define FLINT_DIV_DIVCONQUER_CUTOFF 16
void
_fmpz_poly_div_divconquer_recursive(fmpz * Q, fmpz * temp,
const fmpz * A, const fmpz * B, slong lenB)
{
if (lenB <= FLINT_DIV_DIVCONQUER_CUTOFF)
{
_fmpz_poly_div_basecase(Q, temp, A, 2 * lenB - 1, B, lenB);
}
else
{
const slong n2 = lenB / 2;
const slong n1 = lenB - n2;
fmpz * q0 = Q;
fmpz * q1 = Q + n2;
/*
t is a vector of length lenB - 1, h points to the top n2 coeffs
of t; r1 is vector of length lenB >= 2 n1 - 1
*/
fmpz * t = temp;
fmpz * h = temp + (n1 - 1);
fmpz * r1 = temp + (lenB - 1);
/*
Set {q1, n1}, {r1, 2 n1 - 1} to the quotient and remainder of
{A + 2 n2, 2 n1 - 1} divided by {B + n2, n1}
*/
_fmpz_poly_divremlow_divconquer_recursive(q1, r1, A + 2 * n2, B + n2, n1);
_fmpz_vec_sub(r1, A + 2 * n2, r1, n1 - 1);
/*
Set the top n2 coeffs of t to the top n2 coeffs of the product of
{q1, n1} and {B, n2}; the bottom n1 - 1 coeffs may be arbitrary
For sufficiently large polynomials, computing the full product
using Kronecker segmentation is faster than computing the opposite
short product via Karatsuba
*/
_fmpz_poly_mul_KS(t, q1, n1, B, n2);
/*
If lenB is odd, set {h, n2} to {r1, n2} - {h, n2}, otherwise, to
{A + lenB - 1, 1} + {x * r1, n2} - {h, n2}
*/
if (lenB & WORD(1))
{
_fmpz_vec_sub(h, r1, h, n2);
}
else
{
_fmpz_vec_sub(h + 1, r1, h + 1, n2 - 1);
fmpz_neg(h, h);
fmpz_add(h, h, A + lenB - 1);
}
/*
Set t to h shifted to the right by n2 - 1, and set q0 to the
quotient of {t, 2 n2 - 1} and {B + n1, n2}
Note the bottom n2 - 1 coefficients of t are irrelevant
*/
t += (lenB & WORD(1));
_fmpz_poly_div_divconquer_recursive(q0, temp + lenB, t, B + n1, n2);
}
}