File: bitread.h

package info (click to toggle)
wham-align 0.1.5-8
  • links: PTS, VCS
  • area: main
  • in suites: bookworm, bullseye, forky, sid, trixie
  • size: 892 kB
  • sloc: cpp: 8,769; sh: 76; makefile: 52
file content (223 lines) | stat: -rw-r--r-- 8,100 bytes parent folder | download
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
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
#ifndef BITREAD_H_
#define BITREAD_H_

/**
 *    WHAM - high-throughput sequence aligner
 *    Copyright (C) 2011  WHAM Group, University of Wisconsin
 *
 *    This program is free software: you can redistribute it and/or modify
 *    it under the terms of the GNU General Public License as published by
 *    the Free Software Foundation, either version 3 of the License, or
 *    (at your option) any later version.
 *
 *    This program 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 General Public License for more details.
 *
 *    You should have received a copy of the GNU General Public License
 *    along with this program.  If not, see <http://www.gnu.org/licenses/>.
 */

/* $Id: bitread.h 162 2012-07-31 09:14:04Z yinan $ */

#include "lib.h"

#define BITS_PER_BASE 3
#define BITS_PER_BASE_LL 3LL
#define BITS_PER_BYTE 8
#define BITS_PER_WORD 32
#define BITS_PER_LONGWORD 64
#define BITS_LONGWORD_SHIFT 6
#define BITS_LONGWORD_MASK	0x000000000000003f
#define BITS_TWO_LONGWORD 128
#define BITS_THREE_LONGWORD 192
#define BITS_FOUR_LONGWORD 256
#define BITS_FIVE_LONGWORD 320
#define WORDS_PER_READ 6

class BitRead {
public:

  /*
   *	BitRead::extract
   *	extract a subsequence from the sequence. The subsequence
   *	starts from offset-th character of the sequence and has len
   *	characters.
   *
   *	This implementation is highly optimized for modern CPUs. All IF
   *	statements are implemented by arithmetic operations to eliminate the
   *	branch. It support up to 256-bit sequence (85 bases) that can be
   *	fitted into four 64-bit integers.
   */

  inline static void extract(int64 * a, int64 * b, int64 offset, int len) {
    int shift = ((offset) + (len)) & BITS_LONGWORD_MASK;
    int index = ((offset) + (len)) >> BITS_LONGWORD_SHIFT;
    int64 mask = 0LL - (shift != 0);
    int cshift = BITS_PER_LONGWORD - shift;

    b[0] = (mask & (a[index - 5] >> cshift)) | (a[index - 6] << shift);
    b[1] = (mask & (a[index - 4] >> cshift)) | (a[index - 5] << shift);
    b[2] = (mask & (a[index - 3] >> cshift)) | (a[index - 4] << shift);
    b[3] = (mask & (a[index - 2] >> cshift)) | (a[index - 3] << shift);
    b[4] = (mask & (a[index - 1] >> cshift)) | (a[index - 2] << shift);
    b[5] = (mask & (a[index] >> cshift)) | (a[index - 1] << shift);

    shift = BITS_PER_LONGWORD - ((len) & BITS_LONGWORD_MASK);
    b[0] = (0LL - (len > BITS_FIVE_LONGWORD)) & (b[0] << shift >> shift);
    cshift = (0LL - (len < BITS_FIVE_LONGWORD)) & shift;
    b[1] = b[1] << cshift >> cshift;
    b[1] = (0LL - ((len) > BITS_FOUR_LONGWORD)) & b[1];
    cshift = (0LL - (len < BITS_FOUR_LONGWORD)) & shift;
    b[2] = b[2] << cshift >> cshift;
    b[2] = (0LL - ((len) > BITS_THREE_LONGWORD)) & b[2];
    cshift = (0LL - (len < BITS_THREE_LONGWORD)) & shift;
    b[3] = b[3] << cshift >> cshift;

    b[3] = (0LL - ((len) > BITS_TWO_LONGWORD)) & b[3];
    cshift = (0LL - (len < BITS_TWO_LONGWORD)) & shift;
    b[4] = b[4] << cshift >> cshift;
    b[4] = (0LL - ((len) > BITS_PER_LONGWORD)) & b[4];
    cshift = (0LL - (len < BITS_PER_LONGWORD)) & shift;
    b[5] = b[5] << cshift >> cshift;
  }

  inline static void genHeadMask(int64 * mask, int l) {
    int64 tmp[WORDS_PER_READ];
    for (int i = 0; i < WORDS_PER_READ; i++)
      tmp[i] = -1;
    removeHead2(tmp, mask, l);
  }

  /*
   *	BitRead::removeHead
   *	This function is used to extract rightmost portion of sequence a
   *	into sequence b. The extracted portion has l bits.
   *
   *	This implementation is highly optimized for modern CPUs. All IF
   *	statements are implemented by arithmetic operations to eliminate the
   *	branch. It support up to 256-bit sequence (85 bases) that can be
   *	fitted into four 64-bit integers.
   */

  inline static void removeHead(int64 * a, int64 * b, int64 * mask) {
    b[0] = a[0] & mask[0];
    b[1] = a[1] & mask[1];
    b[2] = a[2] & mask[2];
    b[3] = a[3] & mask[3];
    b[4] = a[4] & mask[4];
    b[5] = a[5] & mask[5];
  }

  inline static void removeHead2(int64 * a, int64 * b, int l) {
    int shift;
    int lshift = BITS_PER_LONGWORD - ((l) & BITS_LONGWORD_MASK);

    shift = lshift;
    b[0] = ((int64) 0 - (l > BITS_FIVE_LONGWORD)) & (a[0] << shift >> shift);
    shift = ((int64) 0 - (l < BITS_FIVE_LONGWORD)) & lshift;
    b[1] = a[1] << shift >> shift;
    shift = lshift;
    b[1] = ((int64) 0 - ((l) > BITS_FOUR_LONGWORD)) & b[1];
    shift = ((int64) 0 - (l < BITS_FOUR_LONGWORD)) & lshift;
    b[2] = a[2] << shift >> shift;
    shift = lshift;
    b[2] = ((int64) 0 - ((l) > BITS_THREE_LONGWORD)) & b[2];
    shift = ((int64) 0 - ((l) < BITS_THREE_LONGWORD)) & lshift;
    b[3] = a[3] << shift >> shift;

    shift = lshift;
    b[3] = ((int64) 0 - ((l) > BITS_TWO_LONGWORD)) & b[3];
    shift = ((int64) 0 - ((l) < BITS_TWO_LONGWORD)) & lshift;
    b[4] = a[4] << shift >> shift;
    shift = lshift;
    b[4] = ((int64) 0 - ((l) > BITS_PER_LONGWORD)) & b[4];
    shift = ((int64) 0 - ((l) < BITS_PER_LONGWORD)) & lshift;
    b[5] = a[5] << shift >> shift;
  }

  /*
   *	BitRead::removeInterval
   *	This function is used to remove an interval segment from sequence a,
   *	and extract others into sequence b. The interval starts from x-th
   *	bit in the sequence and has len bits.
   *
   *	This implementation is highly optimized for modern CPUs. All IF
   *	statements are implemented by arithmetic operations to eliminate the
   *	branch. It support up to 256-bit sequence (85 bases) that can be
   *	fitted into four 64-bit integers.
   *
   *	This implementation requires that the sequence a is stored in the
   *	rightmost of seven 64-bit integers. All unoccupied bits in the seven
   *	64-bit integers are set to be 0s.
   */

  inline static void removeInterval(int64 * a, int64 * b, uint32 x, int p) {
    int index1 = WORDS_PER_READ - 1 - ((x) >> BITS_LONGWORD_SHIFT);
    int index2 = WORDS_PER_READ - 1 - (((x) + (p)) >> BITS_LONGWORD_SHIFT);
    int shift = (x) & BITS_LONGWORD_MASK;
    int shift2 = ((x) + (p)) & BITS_LONGWORD_MASK;
    int poffset = (p) >> BITS_LONGWORD_SHIFT;
    int pshift = (p) & BITS_LONGWORD_MASK;
    int64 mask = ((int64) 0 - (shift != 0))
        & ((int64) -1 >> (BITS_PER_LONGWORD - shift));
    int64 mask2 = (int64) 0 - (pshift != 0);
    int index3 = index1 - poffset;
    int rpshift = BITS_PER_LONGWORD - pshift;

    b[5] = a[5];
    b[4] = a[4];
    b[3] = a[3];
    b[2] = a[2];
    b[1] = a[1];
    b[index1] = (a[index1] & mask) | (a[index2] >> shift2 << shift);
    b[index1] |= ((int64) 0 - (shift2 > shift)) & ((a[index2 - 1] << rpshift));

    b[index1 - 1] = (a[index3 - 1] >> pshift)
        | (mask2 & (a[index3 - 2] << rpshift));
    b[index1 - 2] = (a[index3 - 2] >> pshift)
        | (mask2 & (a[index3 - 3] << rpshift));
    b[index1 - 3] = (a[index3 - 3] >> pshift)
        | (mask2 & (a[index3 - 4] << rpshift));
    b[index1 - 4] = (a[index3 - 4] >> pshift)
        | (mask2 & (a[index3 - 5] << rpshift));
    b[index1 - 5] = (a[index3 - 5] >> pshift)
        | (mask2 & (a[index3 - 6] << rpshift));
  }

  inline static void copy(int64 * a, int64 * b) {
    b[5] = a[5];
    b[4] = a[4];
    b[3] = a[3];
    b[2] = a[2];
    b[1] = a[1];
    b[0] = a[0];
  }

  inline static bool compare(int64 * a, int64 * b) {
    return (a[0] == b[0]) && (a[1] == b[1]) && (a[2] == b[2]) && (a[3] == b[3])
        && (a[4] == b[4]) && (a[5] == b[5]);
  }

  inline static void leftShift(int64 * a, int64 * b) {
    b[0] = (a[0] << 3) | (a[1] >> 61);
    b[1] = (a[1] << 3) | (a[2] >> 61);
    b[2] = (a[2] << 3) | (a[3] >> 61);
    b[3] = (a[3] << 3) | (a[4] >> 61);
    b[4] = (a[4] << 3) | (a[5] >> 61);
    b[5] = (a[5] << 3);
  }

  inline static void rightShift(int64 * a, int64 * b) {
    b[0] = (a[0] >> 3);
    b[1] = (a[1] >> 3) | (a[0] << 61);
    b[2] = (a[2] >> 3) | (a[1] << 61);
    b[3] = (a[3] >> 3) | (a[2] << 61);
    b[4] = (a[4] >> 3) | (a[3] << 61);
    b[5] = (a[5] >> 3) | (a[4] << 61);
  }

};

#endif /* BITREAD_H_ */