Blame mpn/generic/toom_eval_pm2.c

Packit 5c3484
/* mpn_toom_eval_pm2 -- Evaluate a polynomial in +2 and -2
Packit 5c3484
Packit 5c3484
   Contributed to the GNU project by Niels Möller and Marco Bodrato
Packit 5c3484
Packit 5c3484
   THE FUNCTION IN THIS FILE IS INTERNAL WITH A MUTABLE INTERFACE.  IT IS ONLY
Packit 5c3484
   SAFE TO REACH IT THROUGH DOCUMENTED INTERFACES.  IN FACT, IT IS ALMOST
Packit 5c3484
   GUARANTEED THAT IT WILL CHANGE OR DISAPPEAR IN A FUTURE GNU MP RELEASE.
Packit 5c3484
Packit 5c3484
Copyright 2009 Free Software Foundation, Inc.
Packit 5c3484
Packit 5c3484
This file is part of the GNU MP Library.
Packit 5c3484
Packit 5c3484
The GNU MP Library is free software; you can redistribute it and/or modify
Packit 5c3484
it under the terms of either:
Packit 5c3484
Packit 5c3484
  * the GNU Lesser General Public License as published by the Free
Packit 5c3484
    Software Foundation; either version 3 of the License, or (at your
Packit 5c3484
    option) any later version.
Packit 5c3484
Packit 5c3484
or
Packit 5c3484
Packit 5c3484
  * the GNU General Public License as published by the Free Software
Packit 5c3484
    Foundation; either version 2 of the License, or (at your option) any
Packit 5c3484
    later version.
Packit 5c3484
Packit 5c3484
or both in parallel, as here.
Packit 5c3484
Packit 5c3484
The GNU MP Library is distributed in the hope that it will be useful, but
Packit 5c3484
WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY
Packit 5c3484
or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License
Packit 5c3484
for more details.
Packit 5c3484
Packit 5c3484
You should have received copies of the GNU General Public License and the
Packit 5c3484
GNU Lesser General Public License along with the GNU MP Library.  If not,
Packit 5c3484
see https://www.gnu.org/licenses/.  */
Packit 5c3484
Packit 5c3484
#include "gmp.h"
Packit 5c3484
#include "gmp-impl.h"
Packit 5c3484
Packit 5c3484
/* DO_addlsh2(d,a,b,n,cy) computes cy,{d,n} <- {a,n} + 4*(cy,{b,n}), it
Packit 5c3484
   can be used as DO_addlsh2(d,a,d,n,d[n]), for accumulation on {d,n+1}. */
Packit 5c3484
#if HAVE_NATIVE_mpn_addlsh2_n
Packit 5c3484
#define DO_addlsh2(d, a, b, n, cy)	\
Packit 5c3484
do {					\
Packit 5c3484
  (cy) <<= 2;				\
Packit 5c3484
  (cy) += mpn_addlsh2_n(d, a, b, n);	\
Packit 5c3484
} while (0)
Packit 5c3484
#else
Packit 5c3484
#if HAVE_NATIVE_mpn_addlsh_n
Packit 5c3484
#define DO_addlsh2(d, a, b, n, cy)	\
Packit 5c3484
do {					\
Packit 5c3484
  (cy) <<= 2;				\
Packit 5c3484
  (cy) += mpn_addlsh_n(d, a, b, n, 2);	\
Packit 5c3484
} while (0)
Packit 5c3484
#else
Packit 5c3484
/* The following is not a general substitute for addlsh2.
Packit 5c3484
   It is correct if d == b, but it is not if d == a.  */
Packit 5c3484
#define DO_addlsh2(d, a, b, n, cy)	\
Packit 5c3484
do {					\
Packit 5c3484
  (cy) <<= 2;				\
Packit 5c3484
  (cy) += mpn_lshift(d, b, n, 2);	\
Packit 5c3484
  (cy) += mpn_add_n(d, d, a, n);	\
Packit 5c3484
} while (0)
Packit 5c3484
#endif
Packit 5c3484
#endif
Packit 5c3484
Packit 5c3484
/* Evaluates a polynomial of degree 2 < k < GMP_NUMB_BITS, in the
Packit 5c3484
   points +2 and -2. */
Packit 5c3484
int
Packit 5c3484
mpn_toom_eval_pm2 (mp_ptr xp2, mp_ptr xm2, unsigned k,
Packit 5c3484
		   mp_srcptr xp, mp_size_t n, mp_size_t hn, mp_ptr tp)
Packit 5c3484
{
Packit 5c3484
  int i;
Packit 5c3484
  int neg;
Packit 5c3484
  mp_limb_t cy;
Packit 5c3484
Packit 5c3484
  ASSERT (k >= 3);
Packit 5c3484
  ASSERT (k < GMP_NUMB_BITS);
Packit 5c3484
Packit 5c3484
  ASSERT (hn > 0);
Packit 5c3484
  ASSERT (hn <= n);
Packit 5c3484
Packit 5c3484
  /* The degree k is also the number of full-size coefficients, so
Packit 5c3484
   * that last coefficient, of size hn, starts at xp + k*n. */
Packit 5c3484
Packit 5c3484
  cy = 0;
Packit 5c3484
  DO_addlsh2 (xp2, xp + (k-2) * n, xp + k * n, hn, cy);
Packit 5c3484
  if (hn != n)
Packit 5c3484
    cy = mpn_add_1 (xp2 + hn, xp + (k-2) * n + hn, n - hn, cy);
Packit 5c3484
  for (i = k - 4; i >= 0; i -= 2)
Packit 5c3484
    DO_addlsh2 (xp2, xp + i * n, xp2, n, cy);
Packit 5c3484
  xp2[n] = cy;
Packit 5c3484
Packit 5c3484
  k--;
Packit 5c3484
Packit 5c3484
  cy = 0;
Packit 5c3484
  DO_addlsh2 (tp, xp + (k-2) * n, xp + k * n, n, cy);
Packit 5c3484
  for (i = k - 4; i >= 0; i -= 2)
Packit 5c3484
    DO_addlsh2 (tp, xp + i * n, tp, n, cy);
Packit 5c3484
  tp[n] = cy;
Packit 5c3484
Packit 5c3484
  if (k & 1)
Packit 5c3484
    ASSERT_NOCARRY(mpn_lshift (tp , tp , n + 1, 1));
Packit 5c3484
  else
Packit 5c3484
    ASSERT_NOCARRY(mpn_lshift (xp2, xp2, n + 1, 1));
Packit 5c3484
Packit 5c3484
  neg = (mpn_cmp (xp2, tp, n + 1) < 0) ? ~0 : 0;
Packit 5c3484
Packit 5c3484
#if HAVE_NATIVE_mpn_add_n_sub_n
Packit 5c3484
  if (neg)
Packit 5c3484
    mpn_add_n_sub_n (xp2, xm2, tp, xp2, n + 1);
Packit 5c3484
  else
Packit 5c3484
    mpn_add_n_sub_n (xp2, xm2, xp2, tp, n + 1);
Packit 5c3484
#else /* !HAVE_NATIVE_mpn_add_n_sub_n */
Packit 5c3484
  if (neg)
Packit 5c3484
    mpn_sub_n (xm2, tp, xp2, n + 1);
Packit 5c3484
  else
Packit 5c3484
    mpn_sub_n (xm2, xp2, tp, n + 1);
Packit 5c3484
Packit 5c3484
  mpn_add_n (xp2, xp2, tp, n + 1);
Packit 5c3484
#endif /* !HAVE_NATIVE_mpn_add_n_sub_n */
Packit 5c3484
Packit 5c3484
  ASSERT (xp2[n] < (1<<(k+2))-1);
Packit 5c3484
  ASSERT (xm2[n] < ((1<<(k+3))-1 - (1^k&1))/3);
Packit 5c3484
Packit 5c3484
  neg ^= ((k & 1) - 1);
Packit 5c3484
Packit 5c3484
  return neg;
Packit 5c3484
}
Packit 5c3484
Packit 5c3484
#undef DO_addlsh2