/*
Copyright (C) 2010 William Hart
This file is part of FLINT.
FLINT is free software: you can redistribute it and/or modify it under
the terms of the GNU Lesser General Public License (LGPL) as published
by the Free Software Foundation; either version 2.1 of the License, or
(at your option) any later version. See .
*/
#include
#include
#include
#include "flint.h"
#include "fmpz.h"
#include "fmpz_vec.h"
#include "fmpz_poly.h"
void _fmpz_poly_mulhigh_kara_recursive(fmpz * out, const fmpz * pol1,
const fmpz * pol2, fmpz * temp,
slong length);
/*
Multiplication using truncated karatsuba. Below length 7, classical
truncated multiplication is always theoretically faster, so we switch
to that as the basecase.
Above that we use the ordinary (left/right) karatsuba identity and
recursively do one full karatsuba multiplication and two truncated
karatsuba multiplications.
*/
void
_fmpz_poly_mulhigh_kara_recursive(fmpz * out, const fmpz * pol1,
const fmpz * pol2, fmpz * temp, slong length)
{
slong m1 = length / 2;
slong m2 = length - m1;
int odd = (length & 1);
if (length <= 6)
{
_fmpz_poly_mulhigh_classical(out, pol1, length, pol2, length,
length - 1);
return;
}
_fmpz_vec_add(out, pol1, pol1 + m1, m1);
if (odd)
fmpz_set(out + m1, pol1 + 2 * m1);
_fmpz_vec_add(out + m2, pol2, pol2 + m1, m1);
if (odd)
fmpz_set(out + m2 + m1, pol2 + 2 * m1);
_fmpz_poly_mulhigh_kara_recursive(temp, out, out + m2, temp + 2 * m2, m2);
_fmpz_poly_mul_karatsuba(out + 2 * m1, pol1 + m1, m2, pol2 + m1, m2);
fmpz_zero(out + 2 * m1 - 1);
_fmpz_poly_mulhigh_kara_recursive(out, pol1, pol2, temp + 2 * m2, m1);
_fmpz_vec_sub(temp + m2 - 1, temp + m2 - 1, out + m2 - 1, 2 * m1 - m2);
_fmpz_vec_sub(temp + m2 - 1, temp + m2 - 1, out + 2 * m1 + m2 - 1, m2);
_fmpz_vec_add(out + length - 1, out + length - 1, temp + m2 - 1, m2);
_fmpz_vec_zero(out, length - 1);
}
/* Assumes poly1 and poly2 are not length 0. */
void
_fmpz_poly_mulhigh_karatsuba_n(fmpz * res, const fmpz * poly1,
const fmpz * poly2, slong len)
{
fmpz *temp;
slong length, loglen = 0;
if (len == 1)
{
fmpz_mul(res, poly1, poly2);
return;
}
while ((WORD(1) << loglen) < len)
loglen++;
length = (WORD(1) << loglen);
temp = _fmpz_vec_init(2 * length);
_fmpz_poly_mulhigh_kara_recursive(res, poly1, poly2, temp, len);
_fmpz_vec_clear(temp, 2 * length);
}
void
fmpz_poly_mulhigh_karatsuba_n(fmpz_poly_t res,
const fmpz_poly_t poly1, const fmpz_poly_t poly2,
slong len)
{
slong lenr = poly1->length + poly2->length - 1;
int clear1 = 0, clear2 = 0;
fmpz *pol1, *pol2;
if (poly1->length == 0 || poly2->length == 0 || len - 1 >= lenr)
{
fmpz_poly_zero(res);
return;
}
if (poly1->length < len)
{
pol1 = (fmpz *) flint_calloc(len, sizeof(fmpz));
memcpy(pol1, poly1->coeffs, poly1->length * sizeof(fmpz));
clear1 = 1;
}
else
pol1 = poly1->coeffs;
if (poly2->length < len)
{
pol2 = (fmpz *) flint_calloc(len, sizeof(fmpz));
memcpy(pol2, poly2->coeffs, poly2->length * sizeof(fmpz));
clear2 = 1;
}
else
pol2 = poly2->coeffs;
if (res != poly1 && res != poly2)
{
fmpz_poly_fit_length(res, 2 * len - 1);
_fmpz_poly_mulhigh_karatsuba_n(res->coeffs, pol1, pol2, len);
_fmpz_poly_set_length(res, lenr);
}
else
{
fmpz_poly_t temp;
fmpz_poly_init2(temp, 2 * len - 1);
_fmpz_poly_mulhigh_karatsuba_n(temp->coeffs, pol1, pol2, len);
_fmpz_poly_set_length(temp, lenr);
fmpz_poly_swap(temp, res);
fmpz_poly_clear(temp);
}
if (clear1)
flint_free(pol1);
if (clear2)
flint_free(pol2);
}