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
|
#pragma once
#include "intgemm_config.h"
#ifdef INTGEMM_COMPILER_SUPPORTS_AVX512VNNI
#include "avx512_gemm.h"
#include "types.h"
namespace intgemm {
struct AVX512VNNI_8bit : public AVX512_8bit {
template <typename Callback>
INTGEMM_AVX512VNNI static void Multiply(const int8_t *A, const int8_t *B, Index A_rows, Index width, Index B_cols, Callback callback) {
typedef __m512i Integer;
assert(width % sizeof(Integer) == 0);
assert(B_cols % 8 == 0);
assert(reinterpret_cast<uintptr_t>(A) % sizeof(Integer) == 0);
assert(reinterpret_cast<uintptr_t>(B) % sizeof(Integer) == 0);
auto callback_impl = callbacks::CallbackImpl<CPUType::AVX2, Callback>(callback);
const int simd_width = width / sizeof(Integer);
const Integer *B0_col = reinterpret_cast<const Integer*>(B);
Integer zeros = setzero_si<Integer>();
// Go over 8 columns of B at a time.
for (Index B0_colidx = 0; B0_colidx != B_cols; B0_col += 8 * simd_width, B0_colidx += 8) {
// Process one row of A at a time. Doesn't seem to be faster to do multiple rows of A at once.
for (Index A_rowidx = 0; A_rowidx < A_rows; ++A_rowidx) {
// Iterate over shared (inner) dimension.
const Integer *A_live = reinterpret_cast<const Integer *>(A + A_rowidx * width);
const Integer *A_end = A_live + simd_width;
const Integer *B_live = B0_col;
// TODO: separate first step.
Integer sum0 = zeros, sum1 = zeros, sum2 = zeros, sum3 = zeros, sum4 = zeros, sum5 = zeros, sum6 = zeros, sum7 = zeros;
for (; A_live != A_end; ++A_live, B_live += 8) {
Integer a = *A_live;
// Retrieve the conveniently consecutive values of B.
Integer b0 = *B_live;
Integer b1 = *(B_live + 1);
Integer b2 = *(B_live + 2);
Integer b3 = *(B_live + 3);
Integer b4 = *(B_live + 4);
Integer b5 = *(B_live + 5);
Integer b6 = *(B_live + 6);
Integer b7 = *(B_live + 7);
// Get a mask where a is negative.
__mmask64 neg_mask = _mm512_test_epi8_mask(a, _mm512_set1_epi8(-128));
Integer a_positive = _mm512_abs_epi8(a);
// Negate by subtracting from zero with a mask.
b0 = _mm512_mask_sub_epi8(b0, neg_mask, zeros, b0);
b1 = _mm512_mask_sub_epi8(b1, neg_mask, zeros, b1);
b2 = _mm512_mask_sub_epi8(b2, neg_mask, zeros, b2);
b3 = _mm512_mask_sub_epi8(b3, neg_mask, zeros, b3);
b4 = _mm512_mask_sub_epi8(b4, neg_mask, zeros, b4);
b5 = _mm512_mask_sub_epi8(b5, neg_mask, zeros, b5);
b6 = _mm512_mask_sub_epi8(b6, neg_mask, zeros, b6);
b7 = _mm512_mask_sub_epi8(b7, neg_mask, zeros, b7);
sum0 = _mm512_dpbusds_epi32(sum0, a_positive, b0);
sum1 = _mm512_dpbusds_epi32(sum1, a_positive, b1);
sum2 = _mm512_dpbusds_epi32(sum2, a_positive, b2);
sum3 = _mm512_dpbusds_epi32(sum3, a_positive, b3);
sum4 = _mm512_dpbusds_epi32(sum4, a_positive, b4);
sum5 = _mm512_dpbusds_epi32(sum5, a_positive, b5);
sum6 = _mm512_dpbusds_epi32(sum6, a_positive, b6);
sum7 = _mm512_dpbusds_epi32(sum7, a_positive, b7);
}
Integer pack0123 = Pack0123(sum0, sum1, sum2, sum3);
Integer pack4567 = Pack0123(sum4, sum5, sum6, sum7);
auto total = PermuteSummer(pack0123, pack4567);
callback_impl(total, callbacks::OutputBufferInfo(A_rowidx, B0_colidx, A_rows, B_cols));
}
}
}
template <typename Callback>
INTGEMM_AVX512VNNI static void Multiply8Shift(const uint8_t *A, const int8_t *B, Index A_rows, Index width, Index B_cols, Callback callback) {
typedef __m512i Integer;
assert(width % sizeof(Integer) == 0);
assert(B_cols % 8 == 0);
assert(reinterpret_cast<uintptr_t>(A) % sizeof(Integer) == 0);
assert(reinterpret_cast<uintptr_t>(B) % sizeof(Integer) == 0);
auto callback_impl = callbacks::CallbackImpl<CPUType::AVX2, Callback>(callback);
const int simd_width = width / sizeof(Integer);
const Integer *B0_col = reinterpret_cast<const Integer*>(B);
Integer zeros = setzero_si<Integer>();
// Go over 8 columns of B at a time.
for (Index B0_colidx = 0; B0_colidx != B_cols; B0_col += 8 * simd_width, B0_colidx += 8) {
// Process one row of A at a time. Doesn't seem to be faster to do multiple rows of A at once.
for (Index A_rowidx = 0; A_rowidx < A_rows; ++A_rowidx) {
// Iterate over shared (inner) dimension.
const Integer *A_live = reinterpret_cast<const Integer *>(A + A_rowidx * width);
const Integer *A_end = A_live + simd_width;
const Integer *B_live = B0_col;
// TODO: separate first step.
Integer sum0 = zeros, sum1 = zeros, sum2 = zeros, sum3 = zeros, sum4 = zeros, sum5 = zeros, sum6 = zeros, sum7 = zeros;
for (; A_live != A_end; ++A_live, B_live += 8) {
Integer a = *A_live;
//MultiplyAdd
sum0 = _mm512_dpbusds_epi32(sum0, a, *B_live);
sum1 = _mm512_dpbusds_epi32(sum1, a, *(B_live + 1));
sum2 = _mm512_dpbusds_epi32(sum2, a, *(B_live + 2));
sum3 = _mm512_dpbusds_epi32(sum3, a, *(B_live + 3));
sum4 = _mm512_dpbusds_epi32(sum4, a, *(B_live + 4));
sum5 = _mm512_dpbusds_epi32(sum5, a, *(B_live + 5));
sum6 = _mm512_dpbusds_epi32(sum6, a, *(B_live + 6));
sum7 = _mm512_dpbusds_epi32(sum7, a, *(B_live + 7));
}
Integer pack0123 = Pack0123(sum0, sum1, sum2, sum3);
Integer pack4567 = Pack0123(sum4, sum5, sum6, sum7);
auto total = PermuteSummer(pack0123, pack4567);
callback_impl(total, callbacks::OutputBufferInfo(A_rowidx, B0_colidx, A_rows, B_cols));
}
}
}
template <typename Callback>
INTGEMM_AVX512VNNI static void PrepareBias(const int8_t *B, Index width, Index B_cols, Callback callback) {
typedef __m512i Integer;
assert(width % sizeof(Integer) == 0);
assert(B_cols % 8 == 0);
assert(reinterpret_cast<uintptr_t>(B) % sizeof(Integer) == 0);
auto callback_impl = callbacks::CallbackImpl<CPUType::AVX2, Callback>(callback);
const int simd_width = width / sizeof(Integer);
const Integer *B0_col = reinterpret_cast<const Integer*>(B);
Integer zeros = setzero_si<Integer>();
const Integer a = set1_epi8<Integer>(1);
// Go over 8 columns of B at a time.
for (Index B0_colidx = 0; B0_colidx != B_cols; B0_col += 8 * simd_width, B0_colidx += 8) {
const Integer *B_live = B0_col; //In order to make the code look as much as possible as the above function
const Integer *B_end = B_live + simd_width*8;
// TODO: separate first step.
Integer sum0 = zeros, sum1 = zeros, sum2 = zeros, sum3 = zeros, sum4 = zeros, sum5 = zeros, sum6 = zeros, sum7 = zeros;
for (; B_live != B_end; B_live += 8) {
// Retrieve the conveniently consecutive values of B.
sum0 = _mm512_dpbusds_epi32(sum0, a, *B_live);
sum1 = _mm512_dpbusds_epi32(sum1, a, *(B_live + 1));
sum2 = _mm512_dpbusds_epi32(sum2, a, *(B_live + 2));
sum3 = _mm512_dpbusds_epi32(sum3, a, *(B_live + 3));
sum4 = _mm512_dpbusds_epi32(sum4, a, *(B_live + 4));
sum5 = _mm512_dpbusds_epi32(sum5, a, *(B_live + 5));
sum6 = _mm512_dpbusds_epi32(sum6, a, *(B_live + 6));
sum7 = _mm512_dpbusds_epi32(sum7, a, *(B_live + 7));
}
Integer pack0123 = Pack0123(sum0, sum1, sum2, sum3);
Integer pack4567 = Pack0123(sum4, sum5, sum6, sum7);
auto total = PermuteSummer(pack0123, pack4567);
callback_impl(total, callbacks::OutputBufferInfo(0, B0_colidx, 1, B_cols));
}
}
constexpr static const char *const kName = "8-bit AVX512VNNI";
static const CPUType kUses = CPUType::AVX512VNNI;
};
} // namespace intgemm
#endif
|
