File: [local] / OpenXM_contrib / gmp / mpn / generic / Attic / mul_basecase.c (download)
Revision 1.1.1.2 (vendor branch), Mon Aug 25 16:06:20 2003 UTC (21 years ago) by ohara
Branch: GMP
CVS Tags: VERSION_4_1_2, RELEASE_1_2_3, RELEASE_1_2_2_KNOPPIX_b, RELEASE_1_2_2_KNOPPIX Changes since 1.1.1.1: +73 -38
lines
Import gmp 4.1.2
|
/* mpn_mul_basecase -- Internal routine to multiply two natural numbers
of length m and n.
THIS IS AN INTERNAL FUNCTION WITH A MUTABLE INTERFACE. IT IS ONLY
SAFE TO REACH THIS FUNCTION THROUGH DOCUMENTED INTERFACES.
Copyright 1991, 1992, 1993, 1994, 1996, 1997, 2000, 2001, 2002 Free Software
Foundation, Inc.
This file is part of the GNU MP Library.
The GNU MP Library is free software; you can redistribute it and/or modify
it under the terms of the GNU Lesser General Public License as published by
the Free Software Foundation; either version 2.1 of the License, or (at your
option) any later version.
The GNU MP Library 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 Lesser General Public
License for more details.
You should have received a copy of the GNU Lesser General Public License
along with the GNU MP Library; see the file COPYING.LIB. If not, write to
the Free Software Foundation, Inc., 59 Temple Place - Suite 330, Boston,
MA 02111-1307, USA. */
#include "gmp.h"
#include "gmp-impl.h"
/* Multiply {up,usize} by {vp,vsize} and write the result to
{prodp,usize+vsize}. Must have usize>=vsize.
Note that prodp gets usize+vsize limbs stored, even if the actual result
only needs usize+vsize-1.
There's no good reason to call here with vsize>=MUL_KARATSUBA_THRESHOLD.
Currently this is allowed, but it might not be in the future.
This is the most critical code for multiplication. All multiplies rely
on this, both small and huge. Small ones arrive here immediately, huge
ones arrive here as this is the base case for Karatsuba's recursive
algorithm. */
void
mpn_mul_basecase (mp_ptr rp,
mp_srcptr up, mp_size_t un,
mp_srcptr vp, mp_size_t vn)
{
ASSERT (un >= vn);
ASSERT (vn >= 1);
ASSERT (! MPN_OVERLAP_P (rp, un+vn, up, un));
ASSERT (! MPN_OVERLAP_P (rp, un+vn, vp, vn));
/* We first multiply by the low order limb (or depending on optional function
availability, limbs). This result can be stored, not added, to rp. We
also avoid a loop for zeroing this way. */
#if HAVE_NATIVE_mpn_mul_2
if (vn >= 2)
{
rp[un + 1] = mpn_mul_2 (rp, up, un, vp);
rp += 2, vp += 2, vn -= 2;
}
else
{
rp[un] = mpn_mul_1 (rp, up, un, vp[0]);
return;
}
#else
rp[un] = mpn_mul_1 (rp, up, un, vp[0]);
rp += 1, vp += 1, vn -= 1;
#endif
/* Now accumulate the product of up[] and the next low-order limb (or
depending on optional function availability, limbs) from vp[0]. */
#define MAX_LEFT MP_SIZE_T_MAX
#if HAVE_NATIVE_mpn_addmul_4
while (vn >= 4)
{
rp[un + 4 - 1] = mpn_addmul_4 (rp, up, un, vp);
rp += 4, vp += 4, vn -= 4;
}
#undef MAX_LEFT
#define MAX_LEFT 3
#endif
#if HAVE_NATIVE_mpn_addmul_3
while (vn >= 3)
{
rp[un + 3 - 1] = mpn_addmul_3 (rp, up, un, vp);
rp += 3, vp += 3, vn -= 3;
if (MAX_LEFT - 3 <= 3)
break;
}
#undef MAX_LEFT
#define MAX_LEFT 2
#endif
#if HAVE_NATIVE_mpn_addmul_2
while (vn >= 2)
{
rp[un + 2 - 1] = mpn_addmul_2 (rp, up, un, vp);
rp += 2, vp += 2, vn -= 2;
if (MAX_LEFT - 2 <= 2)
break;
}
#undef MAX_LEFT
#define MAX_LEFT 1
#endif
while (vn >= 1)
{
rp[un] = mpn_addmul_1 (rp, up, un, vp[0]);
rp += 1, vp += 1, vn -= 1;
if (MAX_LEFT - 1 <= 1)
break;
}
}