/* Copyright (C) 2008, 2009 William Hart Copyright (C) 2010 Sebastian Pancratz Copyright (C) 2012 Fredrik Johansson 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 "flint.h" #include "fmpz.h" #include "fmpz_vec.h" #include "fmpz_poly.h" static int __fmpz_poly_divrem_divconquer(fmpz * Q, fmpz * R, const fmpz * A, slong lenA, const fmpz * B, slong lenB, int exact) { if (lenA < 2 * lenB - 1) { /* Convert unbalanced division into a 2 n1 - 1 by n1 division */ const slong n1 = lenA - lenB + 1; const slong n2 = lenB - n1; const fmpz * p1 = A + n2; const fmpz * d1 = B + n2; const fmpz * d2 = B; fmpz * W = _fmpz_vec_init((2 * n1 - 1) + lenB - 1); fmpz * d1q1 = R + n2; fmpz * d2q1 = W + (2 * n1 - 1); if (!_fmpz_poly_divrem_divconquer_recursive(Q, d1q1, W, p1, d1, n1, exact)) { _fmpz_vec_clear(W, (2 * n1 - 1) + lenB - 1); return 0; } /* Compute d2q1 = Q d2, of length lenB - 1 */ if (n1 >= n2) _fmpz_poly_mul(d2q1, Q, n1, d2, n2); else _fmpz_poly_mul(d2q1, d2, n2, Q, n1); /* Compute BQ = d1q1 * x^n1 + d2q1, of length lenB - 1; then compute R = A - BQ */ _fmpz_vec_swap(R, d2q1, n2); _fmpz_vec_add(R + n2, R + n2, d2q1 + n2, n1 - 1); _fmpz_vec_sub(R, A, R, lenA); _fmpz_vec_clear(W, (2 * n1 - 1) + lenB - 1); } else /* lenA = 2 * lenB - 1 */ { fmpz * W = _fmpz_vec_init(lenA); if (!_fmpz_poly_divrem_divconquer_recursive(Q, R, W, A, B, lenB, exact)) { _fmpz_vec_clear(W, lenA); return 0; } _fmpz_vec_sub(R, A, R, lenA); _fmpz_vec_clear(W, lenA); } return 1; } int _fmpz_poly_divrem_divconquer(fmpz *Q, fmpz *R, const fmpz *A, slong lenA, const fmpz *B, slong lenB, int exact) { if (lenA <= 2 * lenB - 1) { if (!__fmpz_poly_divrem_divconquer(Q, R, A, lenA, B, lenB, exact)) return 0; } else /* lenA > 2 * lenB - 1 */ { slong shift, n = 2 * lenB - 1; fmpz *QB, *W; _fmpz_vec_set(R, A, lenA); W = _fmpz_vec_init(2 * n); QB = W + n; while (lenA >= n) { shift = lenA - n; if (!_fmpz_poly_divrem_divconquer_recursive(Q + shift, QB, W, R + shift, B, lenB, exact)) { _fmpz_vec_clear(W, 2 * n); return 0; } _fmpz_vec_sub(R + shift, R + shift, QB, n); lenA -= lenB; } if (lenA >= lenB) { if (!__fmpz_poly_divrem_divconquer(Q, W, R, lenA, B, lenB, exact)) { _fmpz_vec_clear(W, 2 * n); return 0; } _fmpz_vec_swap(W, R, lenA); } _fmpz_vec_clear(W, 2 * n); } return 1; } void fmpz_poly_divrem_divconquer(fmpz_poly_t Q, fmpz_poly_t R, const fmpz_poly_t A, const fmpz_poly_t B) { const slong lenA = A->length; const slong lenB = B->length; fmpz_poly_t tQ, tR; fmpz *q, *r; if (lenB == 0) { flint_printf("Exception (fmpz_poly_divrem_divconquer). Division by zero.\n"); flint_abort(); } if (lenA < lenB) { fmpz_poly_set(R, A); fmpz_poly_zero(Q); return; } if (Q == A || Q == B) { fmpz_poly_init2(tQ, lenA - lenB + 1); q = tQ->coeffs; } else { fmpz_poly_fit_length(Q, lenA - lenB + 1); q = Q->coeffs; } if (R == A || R == B) { fmpz_poly_init2(tR, lenA); r = tR->coeffs; } else { fmpz_poly_fit_length(R, lenA); r = R->coeffs; } _fmpz_poly_divrem_divconquer(q, r, A->coeffs, lenA, B->coeffs, lenB, 0); if (Q == A || Q == B) { _fmpz_poly_set_length(tQ, lenA - lenB + 1); fmpz_poly_swap(tQ, Q); fmpz_poly_clear(tQ); } else _fmpz_poly_set_length(Q, lenA - lenB + 1); if (R == A || R == B) { _fmpz_poly_set_length(tR, lenA); fmpz_poly_swap(tR, R); fmpz_poly_clear(tR); } else _fmpz_poly_set_length(R, lenA); _fmpz_poly_normalise(Q); _fmpz_poly_normalise(R); }