Skip to content
Newer
Older
100644 115 lines (95 sloc) 2.81 KB
e549ccf added libtommath-0.32
Tom St Denis authored
1 #include <tommath.h>
2 #ifdef BN_FAST_S_MP_SQR_C
57354e1 added libtommath-0.12
Tom St Denis authored
3 /* LibTomMath, multiple-precision integer library -- Tom St Denis
4 *
c1da6aa added libtommath-0.25
Tom St Denis authored
5 * LibTomMath is a library that provides multiple-precision
57354e1 added libtommath-0.12
Tom St Denis authored
6 * integer arithmetic as well as number theoretic functionality.
7 *
c1da6aa added libtommath-0.25
Tom St Denis authored
8 * The library was designed directly after the MPI library by
57354e1 added libtommath-0.12
Tom St Denis authored
9 * Michael Fromberger but has been written from scratch with
10 * additional optimizations in place.
11 *
12 * The library is free for all purposes without any express
13 * guarantee it works.
14 *
333aebc added libtommath-0.41
Tom St Denis authored
15 * Tom St Denis, tomstdenis@gmail.com, http://libtom.org
57354e1 added libtommath-0.12
Tom St Denis authored
16 */
17
e549ccf added libtommath-0.32
Tom St Denis authored
18 /* the jist of squaring...
fdfa2f4 added libtommath-0.35
Tom St Denis authored
19 * you do like mult except the offset of the tmpx [one that
20 * starts closer to zero] can't equal the offset of tmpy.
21 * So basically you set up iy like before then you min it with
22 * (ty-tx) so that it never happens. You double all those
23 * you add in the inner loop
e549ccf added libtommath-0.32
Tom St Denis authored
24
25 After that loop you do the squares and add them in.
26 */
27
350578d added libtommath-0.30
Tom St Denis authored
28 int fast_s_mp_sqr (mp_int * a, mp_int * b)
57354e1 added libtommath-0.12
Tom St Denis authored
29 {
e549ccf added libtommath-0.32
Tom St Denis authored
30 int olduse, res, pa, ix, iz;
31 mp_digit W[MP_WARRAY], *tmpx;
32 mp_word W1;
33
34 /* grow the destination as required */
35 pa = a->used + a->used;
36 if (b->alloc < pa) {
37 if ((res = mp_grow (b, pa)) != MP_OKAY) {
57354e1 added libtommath-0.12
Tom St Denis authored
38 return res;
39 }
40 }
41
e549ccf added libtommath-0.32
Tom St Denis authored
42 /* number of output digits to produce */
43 W1 = 0;
4b7111d added libtommath-0.33
Tom St Denis authored
44 for (ix = 0; ix < pa; ix++) {
e549ccf added libtommath-0.32
Tom St Denis authored
45 int tx, ty, iy;
46 mp_word _W;
47 mp_digit *tmpy;
48
49 /* clear counter */
50 _W = 0;
51
52 /* get offsets into the two bignums */
53 ty = MIN(a->used-1, ix);
54 tx = ix - ty;
55
56 /* setup temp aliases */
57 tmpx = a->dp + tx;
58 tmpy = a->dp + ty;
59
fdfa2f4 added libtommath-0.35
Tom St Denis authored
60 /* this is the number of times the loop will iterrate, essentially
e549ccf added libtommath-0.32
Tom St Denis authored
61 while (tx++ < a->used && ty-- >= 0) { ... }
62 */
63 iy = MIN(a->used-tx, ty+1);
64
65 /* now for squaring tx can never equal ty
66 * we halve the distance since they approach at a rate of 2x
67 * and we have to round because odd cases need to be executed
68 */
69 iy = MIN(iy, (ty-tx+1)>>1);
70
71 /* execute loop */
72 for (iz = 0; iz < iy; iz++) {
73 _W += ((mp_word)*tmpx++)*((mp_word)*tmpy--);
57354e1 added libtommath-0.12
Tom St Denis authored
74 }
e549ccf added libtommath-0.32
Tom St Denis authored
75
76 /* double the inner product and add carry */
77 _W = _W + _W + W1;
78
79 /* even columns have the square term in them */
80 if ((ix&1) == 0) {
81 _W += ((mp_word)a->dp[ix>>1])*((mp_word)a->dp[ix>>1]);
82 }
83
84 /* store it */
fdfa2f4 added libtommath-0.35
Tom St Denis authored
85 W[ix] = (mp_digit)(_W & MP_MASK);
e549ccf added libtommath-0.32
Tom St Denis authored
86
87 /* make next carry */
88 W1 = _W >> ((mp_word)DIGIT_BIT);
57354e1 added libtommath-0.12
Tom St Denis authored
89 }
90
91 /* setup dest */
0ef44ce added libtommath-0.18
Tom St Denis authored
92 olduse = b->used;
e549ccf added libtommath-0.32
Tom St Denis authored
93 b->used = a->used+a->used;
57354e1 added libtommath-0.12
Tom St Denis authored
94
e549ccf added libtommath-0.32
Tom St Denis authored
95 {
96 mp_digit *tmpb;
0ef44ce added libtommath-0.18
Tom St Denis authored
97 tmpb = b->dp;
e549ccf added libtommath-0.32
Tom St Denis authored
98 for (ix = 0; ix < pa; ix++) {
99 *tmpb++ = W[ix] & MP_MASK;
b66471f added libtommath-0.13
Tom St Denis authored
100 }
101
e549ccf added libtommath-0.32
Tom St Denis authored
102 /* clear unused digits [that existed in the old copy of c] */
b66471f added libtommath-0.13
Tom St Denis authored
103 for (; ix < olduse; ix++) {
104 *tmpb++ = 0;
105 }
106 }
57354e1 added libtommath-0.12
Tom St Denis authored
107 mp_clamp (b);
108 return MP_OKAY;
109 }
e549ccf added libtommath-0.32
Tom St Denis authored
110 #endif
9294e79 added libtommath-0.36
Tom St Denis authored
111
112 /* $Source$ */
113 /* $Revision$ */
114 /* $Date$ */
Something went wrong with that request. Please try again.