EvaluationScript/DrawTransferBench.py at main · TrinityEmulator/EvaluationScript · GitHub

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
import numpy as np
import matplotlib
import matplotlib.pyplot as plot
import os

matplotlib.rc('font', family='arial')

BASE_PATH = "DataTransferBench"
DATA_FOLDER = ["async_polling", "async_polling_aggregation", "sync_vm-exit",
               "sync_vm-exit_aggregation", "sync_polling", "sync_polling_aggregation"]
COLOR_LIST = ["blue", "black", "blue", "black"]
MAX_THREAD_NUM = 4
SIZE_STR_L = ["4 KB", "8 KB", "16 KB", "32 KB", "64 KB", "128 KB", "256 KB",
              "512KB", "1MB", "2MB", "4MB", "8MB", "16MB", "32MB", "64MB", "128MB"]
OUTPUT_DPI = 600


def read_all_strategy_data():
    # if there are more than 4 thread
    data = [dict() for i in range(MAX_THREAD_NUM)]
    # folder represent the emulator type
    for folder in DATA_FOLDER:
        for datafolder in os.listdir(folder):
            thread_id = int(datafolder.split('-')[1]) - 1
            data_size = int(datafolder.split('-')[0])
            # separate the num of thread 1~4 -> 0~3
            if data_size not in data[thread_id]:
                data[thread_id][data_size] = [DATA_FOLDER.index(
                    folder), read_throughput(os.path.join(folder, datafolder))]
            elif data_size in data[thread_id] and read_throughput(os.path.join(folder, datafolder)) > \
                    data[thread_id][data_size][1]:
                data[thread_id][data_size] = [DATA_FOLDER.index(
                    folder), read_throughput(os.path.join(folder, datafolder))]
    return data


def read_teleporting_data(folder_name, other_data):
    data = [dict() for i in range(MAX_THREAD_NUM)]
    for datafolder in os.listdir(folder_name):
        thread_id = int(datafolder.split('-')[1]) - 1
        data_size = int(datafolder.split('-')[0])
        if data_size > 524288:
            data[thread_id][data_size] = [len(DATA_FOLDER), read_and_find_nearest(
                os.path.join(folder_name, datafolder), other_data[thread_id][data_size][1])]
        else:
            data[thread_id][data_size] = [len(DATA_FOLDER), read_throughput(
                os.path.join(folder_name, datafolder))]
    return data


def read_goldfish_data(folder_name):
    data = [dict() for i in range(MAX_THREAD_NUM)]
    for datafolder in os.listdir(folder_name):
        thread_id = int(datafolder.split('-')[1]) - 1
        data_size = int(datafolder.split('-')[0])
        data[thread_id][data_size] = [len(DATA_FOLDER), read_throughput(
            os.path.join(folder_name, datafolder))]
    return data


def read_and_find_nearest(path, other_max):
    with open(f"{path}/throughput.txt") as f:
        l = f.readline()
    measurements = [float(m) for m in l.split()]
    diff = np.mean(measurements) / 1048576
    nearest = 0
    for m in measurements:
        if np.abs(m / 1048576 - other_max) < diff:
            diff = np.abs(m - other_max)
            nearest = m / 1048576
    return nearest


def read_throughput(path):
    with open(f"{path}/throughput.txt") as f:
        l = f.readline()
    measurements = [float(m) for m in l.split()]
    return np.mean(measurements) / 1048576


def draw_scatter_plot(fig_name, data, our_data, google_data, num_threads, compare_target, font_size=14,
                      save_type='png'):
    plot.figure(figsize=(8.5, 7))

    plot.rcParams['xtick.direction'] = 'in'
    plot.rcParams['ytick.direction'] = 'in'

    ax = plot.gca()

    ax.spines['bottom'].set_linewidth(2)
    ax.spines['left'].set_linewidth(2)
    ax.spines['right'].set_linewidth(2)
    ax.spines['top'].set_linewidth(2)

    plot.rcParams['pdf.fonttype'] = 42
    plot.rcParams['ps.fonttype'] = 42
    plot.xlabel("Data Chunk Size", fontsize=30)
    plot.ylabel("Throughput (GB/s)", fontsize=30)

    plot.yticks(fontsize=22)
    plot.tick_params(labelsize=20, width=2)
    markers = ["o", "s"]
    for thread_id in num_threads:
        plot.yticks([0, 4, 8, 12, 16, 20, 24])
        if fig_name == 'Teleporting_goldfish_12_threads':
            plot.ylim(0, 20)
        elif fig_name == 'Teleporting_goldfish_34_threads':
            plot.ylim(0, 22)
        elif fig_name == 'Teleporting_Exhaustion_12_threads':
            plot.ylim(4, 18)
        elif fig_name == 'Teleporting_Exhaustion_34_threads':
            plot.ylim(8, 20)
        chunk_size = sorted(data[thread_id])
        x = np.array([str(int(i / 1024 / 1024)) + " MB" if int(i / 1024)
                                                           >= 1024 else str(int(i / 1024)) + " KB" for i in chunk_size])
        teleporting_x = sorted(our_data[thread_id])
        teleporting_y = np.array([our_data[thread_id][data_size][1]
                                  for data_size in teleporting_x])
        plot.plot(x, teleporting_y, marker=markers[thread_id % 2], color='green',
                 label="Data Teleporting " + str(
                     thread_id + 1) + (" threads" if thread_id > 0 else " thread"), zorder=1, linewidth=2.5,
                 markersize=10)

        if compare_target == "Exhaustion":
            y = np.array([data[thread_id][data_size][1] for data_size in chunk_size])
            plot.plot(x, y, marker=markers[thread_id % 2], color='b', linestyle='--',
                     label="Strategy Exhaustion " + str(
                         thread_id + 1) + (" threads" if thread_id > 0 else " thread"), zorder=1, linewidth=2.5,
                     markersize=10)
        if compare_target == "goldfish":
            google_x = sorted(google_data[thread_id])
            google_y = np.array([google_data[thread_id][data_size][1]
                                 for data_size in google_x])
            plot.plot(x, google_y, marker=markers[thread_id % 2], linestyle='--', color=(128/255, 0/255, 128/255),
                     label="goldfish-pipe " + str(thread_id + 1) + (" threads" if thread_id > 0 else " thread"),
                     zorder=1, linewidth=2.5, markersize=10)
        plot.xticks(x, rotation=45, fontsize=22)

    plot.tight_layout()
    handles,labels = ax.get_legend_handles_labels()
    handles = [handles[0], handles[2], handles[1], handles[3]]
    labels = [labels[0], labels[2], labels[1], labels[3]]
    legend = plot.legend(handles, labels, loc='upper left', fontsize=21, edgecolor=(0.5, 0.5, 0.5), borderpad=0, borderaxespad=1)
    legend.get_frame().set_linewidth(1)
    legend.get_frame().set_boxstyle('Square')
    plot.savefig(f"fig/{fig_name}.{save_type}", dpi=OUTPUT_DPI, bbox_inches='tight')
    plot.show()


if __name__ == "__main__":
    os.chdir("DataTransferBench")
    strategy_data = read_all_strategy_data()
    teleporting_data = read_teleporting_data("teleporting", strategy_data)
    goldfish_data = read_goldfish_data("goldfish-pipe")
    draw_scatter_plot("Teleporting_goldfish_12_threads", strategy_data, teleporting_data, goldfish_data, [0, 1],
                      "goldfish", save_type='pdf')
    draw_scatter_plot("Teleporting_goldfish_34_threads", strategy_data, teleporting_data, goldfish_data, [2, 3],
                      "goldfish", save_type='pdf')
    draw_scatter_plot("Teleporting_Exhaustion_12_threads", strategy_data, teleporting_data, goldfish_data, [0, 1],
                      "Exhaustion", save_type='pdf')
    draw_scatter_plot("Teleporting_Exhaustion_34_threads", strategy_data, teleporting_data, goldfish_data, [2, 3],
                      "Exhaustion", save_type='pdf')