forked from ryujaehun/pytorch-gpu-benchmark
-
Notifications
You must be signed in to change notification settings - Fork 1
/
Copy pathplot_benchmark_results.py
83 lines (68 loc) · 3.54 KB
/
plot_benchmark_results.py
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
import os
import re
import pandas as pd
import matplotlib.pyplot as plt
def substring_before(s, delim):
return s.partition(delim)[0]
def substring_after(s, delim):
return s.partition(delim)[2]
def num_sort(test_string):
return list(map(int, re.findall(r'\d+', test_string)))[0]
# "results/AMD/AMD_Radeon_780M/AMD_Radeon_780M__float_model_train_benchmark.csv",
# "results/Nvidia/TITAN_RTX/TITAN_RTX__float_model_train_benchmark.csv",
# "results/Nvidia/GeForce_GTX_1080_TI/GeForce_GTX_1080_TI__float_model_train_benchmark.csv",
# "results/Nvidia/GeForce_RTX_2080_TI/GeForce_RTX_2080_TI__float_model_train_benchmark.csv",
# "results/AMD/AMD_Radeon_780M/AMD_Radeon_780M__float_model_train_benchmark.csv",
# "results/Nvidia/GeForce_RTX_2080_TI/GeForce_RTX_2080_TI__float_model_train_benchmark.csv",
# "results/Nvidia/GeForce_RTX_2060_MaxQ/GeForce_RTX_2060_MaxQ__float_model_train_benchmark.csv",
# "results/AMD/AMD_Radeon_680M/AMD_Radeon_680M__float_model_train_benchmark.csv",
# "results/AMD/AMD_Radeon_780M/AMD_Radeon_780M__float_model_train_benchmark.csv",
# "new_results/0/AMD_Radeon_Graphics/AMD_Radeon_Graphics__float_model_train_benchmark.csv",
_result_filename_arr = [
"results/Nvidia/GeForce_GTX_1080_TI/GeForce_GTX_1080_TI__float_model_train_benchmark.csv",
"results/Nvidia/GeForce_RTX_2080_TI/GeForce_RTX_2080_TI__float_model_train_benchmark.csv",
"results/Nvidia/GeForce_RTX_3090/GeForce_RTX_3090__float_model_train_benchmark.csv",
"results/AMD/AMD_Radeon_RX_7700S/AMD_Radeon_RX_7700S__float_model_train_benchmark.csv",
"results/AMD/AMD_Radeon_RX_7800_XT/AMD_Radeon_RX_7800_XT__float_model_train_benchmark.csv",
"results/AMD/AMD_Radeon_RX_7900_XTX/AMD_Radeon_RX_7900_XTX__float_model_train_benchmark.csv",
]
# "results/Nvidia/A100_SXM4_40GB/A100_SXM4_40GB__float_model_train_benchmark.csv",
"""
_result_filename_arr = [
"test1.csv",
"test2.csv",
]
"""
### Filter file name list for files ending with .csv
_result_filename_arr = [file for file in _result_filename_arr if '.csv' in file]
_benchmark_name_arr = ["resnet", "resnext", "densenet", "squeezenet", "vgg", "mobilenet"]
_benchmark_values_dict = {}
_df_means_appended = []
basename_arr = []
### Loop over all files
for result_filename in _result_filename_arr:
### Read .csv file and append to list
df = pd.read_csv(result_filename)
basename = os.path.basename(result_filename)
basename = substring_before(basename, "__")
basename_arr.append(basename)
tittle_str = substring_after(result_filename, "__")
tittle_str = substring_before(tittle_str, "_benchmark")
# split benchmarks for each device to multiple result arrays
# and then collect resutls for multiple gpus to same arrays
for benchmark_name in _benchmark_name_arr:
_benchmark_values_dict[benchmark_name] = _benchmark_values_dict.get(benchmark_name, [])
_df_means_appended = _benchmark_values_dict[benchmark_name]
cols = [cc for cc in df.columns if cc.startswith(benchmark_name)]
cols.sort(key=num_sort)
df2 = df[cols]
means = df2.mean()
_df_means_appended.append(means)
for benchmark_name in _benchmark_name_arr:
_df_means_appended = _benchmark_values_dict.get(benchmark_name, [])
df_mean = pd.concat(_df_means_appended, axis=1)
df_mean.columns = basename_arr
df_mean.plot(xlabel='benchmark model', ylabel='execution time (msec), smaller value is better', title=tittle_str)
# plt.plot(df_mean)
### Generate the plot
plt.show()