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
|
import csv
# makes math far easier
import numpy as np
def filter_results_by_name(results, substr):
perfs = []
for name, perf in results.items():
if substr in name:
perfs.append(perf)
return np.array(perfs)
def main():
# using dicts seems easier at first but we'd have to eventually convert
# to a list for easy elementwise division or use an ugly loop
funcs = []
main = []
new = []
with open("perf.csv", "r") as f:
csv_reader = csv.reader(f)
# skip header row
next(csv_reader)
# csv is sorted by name so we can assume even rows are main and odd are new
for indx, (func, median) in enumerate(csv_reader):
func = func.split("::")[-1]
# indx is even
if not (indx % 2):
funcs.append(func)
main.append(float(median))
else:
new.append(float(median))
main = np.array(main)
new = np.array(new)
relative_perfs = new / main
# more stats will be included when needed
print(
f"The new branch is {round(relative_perfs.mean(), 4)}x faster than the main branch aross all tests!"
)
results = {}
for func_name, relative_perf in zip(funcs, relative_perfs):
# this allows us to categorize classes of results, such as encode vs decode
results[func_name] = relative_perf
encode_perfs = filter_results_by_name(results, "encode")
decode_perfs = filter_results_by_name(results, "decode")
print(
f"The new branch is {round(encode_perfs.mean(), 4)}x faster than the main branch aross encode tests!"
)
print(
f"The new branch is {round(decode_perfs.mean(), 4)}x faster than the main branch aross decode tests!"
)
if __name__ == "__main__":
main()
|
