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
|
/*
* Copyright 2019 Xilinx Inc.
*
* 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.
*/
#include <cmath>
#include <fstream>
#include <future>
#include <iostream>
#include <random>
#include <thread>
#include <vitis/ai/performance_test.hpp>
#include "vitis/ai/parse_value.hpp"
#include "vitis/ai/profiling.hpp"
#include "xir/sfm_controller.hpp"
using xir::SfmController;
using namespace std;
DEF_ENV_PARAM(DEBUG_TEST, "0");
static void softmax_c(const int8_t* input, float scale, unsigned int cls,
float* output) {
float sum = 0.f;
for (unsigned int i = 0; i < cls; ++i) {
output[i] = exp(input[i] * scale);
sum += output[i];
}
for (unsigned int i = 0; i < cls; ++i) output[i] /= sum;
}
static void softmax_c(const int8_t* input, float scale, unsigned int cls,
unsigned int group, float* output) {
for (unsigned int i = 0; i < group; ++i) {
softmax_c(input, scale, cls, output);
input += cls;
output += cls;
}
}
static vector<int8_t> random_vector_char(size_t sz) {
static std::mt19937 rng(100);
static std::uniform_int_distribution<char> dist;
auto ret = vector<int8_t>(sz);
for (auto i = 0u; i < ret.size(); ++i) {
ret[i] = dist(rng);
}
return ret;
}
// static vector<vector<int8_t>> random_vector_char(int num_of_threads, int cls,
// int group) {
// auto ret = vector<vector<int8_t>>((size_t)num_of_threads);
// for (auto& r : ret) {
// r = random_vector_char((size_t)(cls * group));
// }
// return ret;
// }
static vector<float> generate_ref(const int8_t* input, float scale, int cls,
int group) {
auto ret = vector<float>((size_t)(cls * group));
softmax_c(input, scale, cls, group, &ret[0]);
return ret;
}
static void compare(int cls, int group, const signed char* input,
const float* output1, const float* output2) {
int error = 0;
for (auto g = 0; g < group; ++g) {
for (auto i = 0; i < cls; ++i) {
auto idx = g * cls + i;
auto diff = output1[idx] - output2[idx];
if ((diff != 0.0 && std::abs(diff) > 0.001)) {
error = error + 1;
}
}
}
CHECK_EQ(error, 0);
}
class MyPerformanceTestRunner : public vitis::ai::PerformanceTestRunner {
public:
explicit MyPerformanceTestRunner(float scale, int cls, int group);
virtual ~MyPerformanceTestRunner() = default;
MyPerformanceTestRunner(const PerformanceTestRunner& other) = delete;
MyPerformanceTestRunner& operator=(const PerformanceTestRunner& rhs) = delete;
public:
virtual void step(size_t idx, int thread_id) override;
virtual size_t get_result() override;
private:
float scale_;
int cls_;
int group_;
std::shared_ptr<xir::SfmController> sfm_;
vector<int8_t> input_;
vector<float> output_;
size_t result_ = 0;
};
MyPerformanceTestRunner::MyPerformanceTestRunner(float scale, int cls,
int group)
: scale_{scale}, //
cls_{cls}, //
group_{group}, //
sfm_{xir::SfmController::get_instance()}, //
input_{random_vector_char(cls_ * group_)}, //
output_{generate_ref(&input_[0], scale, cls, group)} //
{}
void MyPerformanceTestRunner::step(size_t idx, int thread_id) {
vector<float> output1(cls_ * group_);
sfm_->run(&input_[0], scale_, cls_, group_, &output1[0]);
compare(cls_, group_, &input_[0], &output1[0], &output_[0]);
result_ = result_ + 1u;
return;
}
size_t MyPerformanceTestRunner::get_result() {
LOG(INFO) << "result_ " << result_ << " " //
;
return result_;
}
int main(int argc, char* argv[]) {
if (argc < 5) {
cout << "usage: " << argv[0] << "<num_of_threads> <fixpos> <cls> <group>"
<< endl;
return 0;
}
int num_of_threads = stoi(argv[1]);
int fixpos = stoi(argv[2]);
int cls = stoi(argv[3]);
int group = stoi(argv[4]);
float scale = std::exp2f(-1.0f * (float)fixpos);
auto runners = vector<std::unique_ptr<vitis::ai::PerformanceTestRunner>>();
for (auto i = 0; i < num_of_threads; ++i) {
LOG(INFO) << "create runner ... " << i << "/" << num_of_threads;
runners.emplace_back(
std::make_unique<MyPerformanceTestRunner>(scale, cls, group));
}
return std::make_unique<vitis::ai::PerformanceTest>()->main(
argc, argv, std::move(runners));
}
|
