1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
|
/*
Copyright (C) 2013 Fredrik Johansson
This file is part of Arb.
Arb 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 <http://www.gnu.org/licenses/>.
*/
#include "fmpr.h"
/* like mpn_scan0, but takes an upper size */
static __inline__ flint_bitcnt_t
mpn_scan0b(mp_srcptr up, mp_size_t size, flint_bitcnt_t from_bit)
{
mp_limb_t t;
slong i, c;
i = from_bit / GMP_NUMB_BITS;
c = from_bit % FLINT_BITS;
t = ((~up[i]) >> c) << c;
while (t == 0)
{
i++;
if (i == size)
return size * FLINT_BITS;
else
t = ~up[i];
}
count_trailing_zeros(c, t);
return (i * FLINT_BITS) + c;
}
slong
_fmpr_set_round_mpn(slong * shift, fmpz_t man, mp_srcptr x, mp_size_t xn, int negative, slong prec, fmpr_rnd_t rnd)
{
slong bc, val, val_bits, val_limbs, ret;
int increment;
/* compute the total bit length of x */
count_leading_zeros(bc, x[xn - 1]);
bc = FLINT_BITS - bc;
bc += (xn - 1) * FLINT_BITS;
/* already odd */
if (x[0] & 1)
{
/* quick exit */
if (bc <= prec)
{
if (bc <= FLINT_BITS - 2)
{
mp_limb_t t = x[0];
_fmpz_demote(man);
*man = negative ? -t : t;
}
else
fmpz_set_mpn_large(man, x, xn, negative);
*shift = 0;
return FMPR_RESULT_EXACT;
}
else
{
val_limbs = val_bits = val = 0;
}
}
else
{
/* trailing zero bits: val = val_limbs * FLINT_BITS + val_bits */
val_limbs = 0;
while (x[val_limbs] == 0)
val_limbs++;
count_trailing_zeros(val_bits, x[val_limbs]);
val = val_bits + (val_limbs * FLINT_BITS);
}
/* no rounding necessary; just copy or shift to destination */
if (bc - val <= prec)
{
ret = FMPR_RESULT_EXACT;
increment = 0;
}
else
{
/* truncation */
if (!rounds_up(rnd, negative))
{
val = mpn_scan1(x, bc - prec);
increment = 0;
}
/* round to next higher odd mantissa */
else
{
val = mpn_scan0b(x, xn, bc - prec);
/* can overflow to next power of 2 */
if (val == bc)
{
fmpz_set_si(man, negative ? -1 : 1);
*shift = bc;
return prec - 1;
}
/* otherwise, we are cutting off at a zero bit, and
incrementing at that position will not cause carry
propagation below */
increment = 1;
}
val_limbs = val / FLINT_BITS;
val_bits = val % FLINT_BITS;
ret = prec - (bc - val);
}
/* the output mantissa is a small fmpz */
if (bc - val <= FLINT_BITS - 2)
{
mp_limb_t h;
if (val_limbs + 1 == xn || val_bits == 0)
h = x[val_limbs] >> val_bits;
else
h = (x[val_limbs] >> val_bits) | (x[val_limbs + 1] << (FLINT_BITS - val_bits));
h += increment;
_fmpz_demote(man);
*man = negative ? -h : h;
}
/* the output mantissa is an mpz */
else
{
mp_ptr dest;
slong res_limbs, res_alloc;
__mpz_struct * zptr = _fmpz_promote(man);
res_limbs = ((bc - val) + FLINT_BITS - 1) / FLINT_BITS;
/* todo: only allocate the required size, not the size before shifting */
res_alloc = xn - val_limbs;
if (zptr->_mp_alloc < res_alloc)
mpz_realloc2(zptr, res_alloc * FLINT_BITS);
dest = zptr->_mp_d;
/* right shift by val */
if (val_bits == 0)
flint_mpn_copyi(dest, x + val_limbs, res_limbs);
else
mpn_rshift(dest, x + val_limbs, xn - val_limbs, val_bits);
dest[0] += increment;
zptr->_mp_size = negative ? -res_limbs : res_limbs;
}
*shift = val;
return ret;
}
|
