onednn-src 0.1.13

Source of oneAPI Deep Neural Network Library (oneDNN)
Documentation 
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

/*******************************************************************************
* Copyright 2019 Intel Corporation
*
* Licensed under the Apache License, Version 2.0 (the "License");
* you may not use this file except in compliance with the License.
* You may obtain a copy of the License at
*
*     http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*******************************************************************************/

#ifndef CPU_X64_JIT_UNI_X8S8S32X_1X1_CONV_KERNEL_HPP
#define CPU_X64_JIT_UNI_X8S8S32X_1X1_CONV_KERNEL_HPP

#include <memory>
#include "common/c_types_map.hpp"
#include "common/memory_tracking.hpp"

#include "cpu/x64/injectors/jit_uni_postops_injector.hpp"
#include "cpu/x64/jit_generator.hpp"
#include "cpu/x64/jit_primitive_conf.hpp"

namespace dnnl {
namespace impl {
namespace cpu {
namespace x64 {

template <cpu_isa_t isa, typename Vmm>
struct jit_uni_x8s8s32x_1x1_conv_kernel_vmm_t : public jit_generator_t {
    DECLARE_CPU_JIT_AUX_FUNCTIONS(jit_uni_x8s8s32x_1x1_conv_kernel_vmm_t)
    jit_uni_x8s8s32x_1x1_conv_kernel_vmm_t(const jit_1x1_conv_conf_t &ajcp,
            const primitive_attr_t &attr, const memory_desc_t &dst_md);

    int get_tail_size() { return jcp.oc_without_padding % jcp.oc_block; }

    jit_1x1_conv_conf_t jcp;
    const primitive_attr_t &attr_;

private:
    std::unique_ptr<injector::jit_uni_postops_injector_t<isa>>
            postops_injector_;

    enum {
        ker_max_reg_idx = 13,
    };
    const Xbyak::Reg64 reg_bcast_data = r8;
    const Xbyak::Reg64 reg_output_data = r9;
    const Xbyak::Reg64 reg_src_scales = r8;
    const Xbyak::Reg64 reg_wei_scales = r8;
    const Xbyak::Reg64 reg_scale_adjust = r8;
    const Xbyak::Reg64 reg_dst_scales = r12;
    const Xbyak::Reg64 reg_load_data = r10;
    const Xbyak::Reg64 reg_ptr_sum_scale = r10;
    const Xbyak::Reg64 reg_ptr_sum_zp = rdx;
    const Xbyak::Reg64 reg_reduce_loop_work = r11;
    const Xbyak::Reg64 reg_bias_data = r12;
    const Xbyak::Reg64 reg_comp_data = r12;
    const Xbyak::Reg64 reg_init_bcast = r13;
    const Xbyak::Reg64 reg_store_bcast = r13;
    const Xbyak::Reg64 reg_reduce_loop_iter = r13;
    const Xbyak::Reg64 aux_reg_bcast_data = r14;
    const Xbyak::Reg64 aux_reg_load_data = r15;
    const Xbyak::Reg64 aux_reg_saturation = r15;
    const Xbyak::Reg64 reg_reduce_pos_flag = rax;
    const Xbyak::Reg64 aux1_reg_bcast_data = rbx;
    const Xbyak::Reg64 reg_bcast_loop_work = rbx;
    const Xbyak::Reg64 reg_bcast_loop_iter = rdx;
    const Xbyak::Reg64 reg_load_loop_work = rsi;
    const Xbyak::Reg64 aux_reg_output_data = abi_not_param1;
    // zero-point computation
    const Xbyak::Reg64 reg_zp_compensation = aux_reg_load_data; // r15
    const Xbyak::Reg64 reg_src_zero_point = aux_reg_bcast_data; // r14
    const Xbyak::Reg64 reg_dst_zero_point = reg_src_zero_point;

    const Vmm vmm_tmp = Vmm(3);
    const Vmm vmm_one = Vmm(2);
    const Vmm vmm_zero = Vmm(1);
    const Vmm vmm_shift = Vmm(1);
    const Vmm vmm_bcast = Vmm(0);
    const Vmm vmm_saturation = Vmm(0);
    /* used during scale section of store_output */
    const Vmm vmm_scales = Vmm(1);
    const Vmm vmm_scales_tmp = Vmm(3); // Has dependency on `vmm_bias`.
    const Vmm vmm_dst_scales = Vmm(1);
    /* used during post_op sum section of store_output */
    const Vmm vmm_prev_dst = Vmm(1);
    /* used during bias section of store_output */
    const Vmm vmm_comp = Vmm(0); // only for signed input
    const Vmm vmm_bias = Vmm(3);
    /* zero-point */
    const Vmm vmm_zp = Vmm(1);
    const Vmm vmm_zp_comp = Vmm(2);

    constexpr static int simd_w = isa == avx2 ? 8 : 4;
    constexpr static int reg64_size = sizeof(int64_t);
    constexpr static int bcast_loop_work_off = 0;
    constexpr static int reg_bias_data_off = 1 * reg64_size;
    constexpr static int reg_bcast_data_off = 2 * reg64_size;
    constexpr static int reg_load_data_off = 3 * reg64_size;
    constexpr static int reg_src_scales_off = 4 * reg64_size;
    constexpr static int reg_wei_scales_off = 5 * reg64_size;
    constexpr static int reg_dst_scales_off = 6 * reg64_size;
    constexpr static int reg_bcast_loop_iter_off = 7 * reg64_size;
    constexpr static int reg_comp_data_off = 8 * reg64_size;
    constexpr static int reg_zp_compensation_off = 9 * reg64_size;
    constexpr static int reg_src_zero_point_off = 10 * reg64_size;
    constexpr static int reg_dst_zero_point_off = 11 * reg64_size;
    constexpr static int stack_space_needed = 12 * reg64_size;

    int vreg_accum_idx(
            const int load_loop_blk, const int i_load, const int i_ur);
    Vmm vreg_accum(const int load_loop_blk, const int i_load, const int i_ur);
    int output_ptr(const int i_load, const int i_ur);
    void bcast_loop(int load_loop_blk);
    void apply_sum(const int ur, const int load_loop_blk,
            const bool mask_flag_in, const float *p_sum_scale,
            const int32_t *p_sum_zp);
    void apply_postops(const int ur, const int load_loop_blk,
            const bool mask_flag_in, const float *p_sum_scale,
            const int32_t *p_sum_zp);
    void reduce_loop(int load_loop_blk, int ur, bool wraparound);

    void generate() override;
    void cvt2ps(data_type_t type_in, const Vmm &vmm_in, const Xbyak::Reg64 &reg,
            int offset, int load_size);
};

template <cpu_isa_t isa>
struct jit_uni_x8s8s32x_1x1_conv_kernel_t {

    jit_uni_x8s8s32x_1x1_conv_kernel_t(const jit_1x1_conv_conf_t &ajcp,
            const primitive_attr_t &attr, const memory_desc_t &dst_md)
        : kernel_(nullptr) {

        switch (isa) {
            case avx2:
                kernel_ = utils::make_unique<jit_avx2_x8s8s32x_1x1_conv_kernel>(
                        ajcp, attr, dst_md);
                return;
            case sse41:
                kernel_ = utils::make_unique<
                        jit_sse41_x8s8s32x_1x1_conv_kernel>(ajcp, attr, dst_md);
                return;
            default: assert(!"Current ISA is not supported!");
        }
    }

    status_t create_kernel() {
        if (kernel_) return kernel_->create_kernel();
        return status::out_of_memory;
    }

    ~jit_uni_x8s8s32x_1x1_conv_kernel_t() = default;

    void operator()(const jit_1x1_conv_args_t *p) const { (*kernel_)(p); }

    static status_t init_conf(jit_1x1_conv_conf_t &jcp,
            const convolution_desc_t &cd, const memory_desc_wrapper &src_d,
            const memory_desc_wrapper &weights_d,
            const memory_desc_wrapper &dst_d, const memory_desc_wrapper &bias_d,
            primitive_attr_t &attr, int nthreads, bool reduce_src);

    static void init_scratchpad(memory_tracking::registrar_t &scratchpad,
            const jit_1x1_conv_conf_t &jcp, const primitive_attr_t &attr);

    using jit_sse41_x8s8s32x_1x1_conv_kernel
            = jit_uni_x8s8s32x_1x1_conv_kernel_vmm_t<sse41, Xbyak::Xmm>;
    using jit_avx2_x8s8s32x_1x1_conv_kernel
            = jit_uni_x8s8s32x_1x1_conv_kernel_vmm_t<avx2, Xbyak::Ymm>;

    constexpr static int simd_w = isa == avx2 ? 8 : 4;

private:
    DNNL_DISALLOW_COPY_AND_ASSIGN(jit_uni_x8s8s32x_1x1_conv_kernel_t);
    std::unique_ptr<jit_generator_t> kernel_;
};

} // namespace x64
} // namespace cpu
} // namespace impl
} // namespace dnnl

#endif