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
|
#pragma once
#include <library/cpp/digest/argonish/internal/rotations/rotations_sse2.h>
namespace NArgonish {
static inline void BlamkaG1SSE2(
__m128i& a0, __m128i& a1, __m128i& b0, __m128i& b1,
__m128i& c0, __m128i& c1, __m128i& d0, __m128i& d1) {
__m128i ml = _mm_mul_epu32(a0, b0);
ml = _mm_add_epi64(ml, ml);
a0 = _mm_add_epi64(a0, _mm_add_epi64(b0, ml));
ml = _mm_mul_epu32(a1, b1);
ml = _mm_add_epi64(ml, ml);
a1 = _mm_add_epi64(a1, _mm_add_epi64(b1, ml));
d0 = _mm_xor_si128(d0, a0);
d1 = _mm_xor_si128(d1, a1);
d0 = Rotr32(d0);
d1 = Rotr32(d1);
ml = _mm_mul_epu32(c0, d0);
ml = _mm_add_epi64(ml, ml);
c0 = _mm_add_epi64(c0, _mm_add_epi64(d0, ml));
ml = _mm_mul_epu32(c1, d1);
ml = _mm_add_epi64(ml, ml);
c1 = _mm_add_epi64(c1, _mm_add_epi64(ml, d1));
b0 = _mm_xor_si128(b0, c0);
b1 = _mm_xor_si128(b1, c1);
b0 = Rotr24(b0);
b1 = Rotr24(b1);
}
static inline void BlamkaG2SSE2(
__m128i& a0, __m128i& a1, __m128i& b0, __m128i& b1,
__m128i& c0, __m128i& c1, __m128i& d0, __m128i& d1) {
__m128i ml = _mm_mul_epu32(a0, b0);
ml = _mm_add_epi64(ml, ml);
a0 = _mm_add_epi64(a0, _mm_add_epi64(b0, ml));
ml = _mm_mul_epu32(a1, b1);
ml = _mm_add_epi64(ml, ml);
a1 = _mm_add_epi64(a1, _mm_add_epi64(b1, ml));
d0 = _mm_xor_si128(d0, a0);
d1 = _mm_xor_si128(d1, a1);
d0 = Rotr16(d0);
d1 = Rotr16(d1);
ml = _mm_mul_epu32(c0, d0);
ml = _mm_add_epi64(ml, ml);
c0 = _mm_add_epi64(c0, _mm_add_epi64(d0, ml));
ml = _mm_mul_epu32(c1, d1);
ml = _mm_add_epi64(ml, ml);
c1 = _mm_add_epi64(c1, _mm_add_epi64(ml, d1));
b0 = _mm_xor_si128(b0, c0);
b1 = _mm_xor_si128(b1, c1);
b0 = Rotr63(b0);
b1 = Rotr63(b1);
}
static inline void DiagonalizeSSE2(
__m128i& b0, __m128i& b1, __m128i& c0, __m128i& c1, __m128i& d0, __m128i& d1) {
__m128i tmp0 = d0;
__m128i tmp1 = b0;
d0 = c0;
c0 = c1;
c1 = d0;
d0 = _mm_unpackhi_epi64(d1, _mm_unpacklo_epi64(tmp0, tmp0));
d1 = _mm_unpackhi_epi64(tmp0, _mm_unpacklo_epi64(d1, d1));
b0 = _mm_unpackhi_epi64(b0, _mm_unpacklo_epi64(b1, b1));
b1 = _mm_unpackhi_epi64(b1, _mm_unpacklo_epi64(tmp1, tmp1));
}
static inline void UndiagonalizeSSE2(
__m128i& b0, __m128i& b1, __m128i& c0, __m128i& c1, __m128i& d0, __m128i& d1) {
__m128i tmp0 = c0;
c0 = c1;
c1 = tmp0;
tmp0 = b0;
__m128i tmp1 = d0;
b0 = _mm_unpackhi_epi64(b1, _mm_unpacklo_epi64(b0, b0));
b1 = _mm_unpackhi_epi64(tmp0, _mm_unpacklo_epi64(b1, b1));
d0 = _mm_unpackhi_epi64(d0, _mm_unpacklo_epi64(d1, d1));
d1 = _mm_unpackhi_epi64(d1, _mm_unpacklo_epi64(tmp1, tmp1));
}
}
|
