Cryptography Reference
In-Depth Information
To save time in the computation, instead of the initialization required above,
the partial product
(
b
n
−
1
b
n
−
2
...b
0
)
B
·
a
0
is calculated in a loop and stored in
p
n
,p
n
−
1
,...,p
0
.
carry = 0;
av = *LSDPTR_L (a_l);
for (bptr_l = LSDPTR_L (b_l), pptr_l = LSDPTR_L (p_l);
bptr_l <= msdptrb_l; bptr_l++, pptr_l++)
{
*pptr_l = (USHORT)(carry = (ULONG)av * (ULONG)*bptr_l +
(ULONG)(USHORT)(carry >> BITPERDGT));
}
*pptr_l = (USHORT)(carry >> BITPERDGT);
Next follows the nested multiplication loop, beginning with the digit
a_l[2]
of
a_l
.
for (csptr_l = LSDPTR_L (p_l) + 1, aptr_l = LSDPTR_L (a_l) + 1;
aptr_l <= msdptra_l; csptr_l++, aptr_l++)
{
carry = 0;
av = *aptr_l;
for (bptr_l = LSDPTR_L (b_l), pptr_l = csptr_l;
bptr_l <= msdptrb_l; bptr_l++, pptr_l++)
{
*pptr_l = (USHORT)(carry = (ULONG)av * (ULONG)*bptr_l +
(ULONG)*pptr_l + (ULONG)(USHORT)(carry >> BITPERDGT));
}
*pptr_l = (USHORT)(carry >> BITPERDGT);
}
The largest possible length of the result is the sum of the numbers of digits of
a_l
and
b_l
. If the result has one digit fewer, this is determined by the macro
RMLDZRS_L
.
SETDIGITS_L (p_l, DIGITS_L (a_l) + DIGITS_L (b_l));
RMLDZRS_L (p_l);
If the result is larger than can be accommodated in a
CLINT
object, it is reduced,
and the error flag
OFL
is set to the value
E_CLINT_OFL
. Then the reduced result is
assigned to the object
pp_l
.
