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
|
#!/usr/bin/env python
# Plot bars comparing different implementations of mail pipeline.
#
# This script can be improved to account for arbitrary number of data sets, but
# it is not doing it right now.
import json
import numpy as np
import matplotlib.pyplot as plt
# make a prettier graph
from mpltools import style
style.use('ggplot')
OUTPUT_FILENAME = 'legacy-vs-blobs.png'
# each value below will generate one bar for each for each (amount, size) pair.
# The script expects to find files in ./data/SET/ for each set of
# implementations.
#
# The baseline values will be the legacy results in ./data/no-cache/.
graphs = [
'1_10000k',
'10_1000k',
'100_100k',
'1000_10k',
]
# the JSON structure returned by the following function is ugly, but the
# original JSONs are even uglier, so this is here just to make the life of the
# script easier.
#
# We want to have something like:
#
# data[variation][graph][implementation] = <stats>
#
# Where:
#
# - variation is one data set under ./data (i.e. no-cache, cache, persistent,
# etc).
# - graph is one of the values in graphs variable above.
# - implementation is either legacy or blobs (we just need legacy for the
# no-cache variation, as that is the one we are using as baseline.
def get_data():
folders = ['cache', 'no-cache', 'persistent']
data = {}
for folder in folders:
data[folder] = {}
for graph in graphs:
with open('data/%s/%s.json' % (folder, graph)) as f:
d = json.loads(f.read())
benchmarks = d['benchmarks']
data[folder][graph] = {}
for t in ['blobs', 'legacy']:
result = filter(lambda b: t in b['name'], benchmarks)
if result:
result = result.pop()
data[folder][graph][t] = result['stats']
return data
def plot_data(data):
N = 4
# this is our baseline (i.e. legacy / legacy)
absolutes = (1, 1, 1, 1)
ind = np.arange(N) # the x locations for the groups
width = 0.20 # the width of the bars
fig, ax = plt.subplots()
rects1 = ax.bar(ind, absolutes, width)
# for each graph, calculate the ratios
ratios = {'no-cache': [], 'cache': [], 'persistent': []}
for graph in graphs:
legacy = data['no-cache'][graph]['legacy']['mean']
# calculate ratios for no-cache / legacy
ratio = data['no-cache'][graph]['blobs']['mean'] / legacy
ratios['no-cache'].append(ratio)
# calculate ratios for cache / legacy
ratio = data['cache'][graph]['blobs']['mean'] / legacy
ratios['cache'].append(ratio)
# calculate ratios for persistent / legacy
ratio = data['persistent'][graph]['blobs']['mean'] / legacy
ratios['persistent'].append(ratio)
# create the boxes with the ratios
nocache = tuple(ratios['no-cache'])
rects2 = ax.bar(ind + width, nocache, width)
cache = tuple(ratios['cache'])
rects3 = ax.bar(ind + (2 * width), cache, width)
persistent = tuple(ratios['persistent'])
rects4 = ax.bar(ind + (3 * width), persistent, width)
# add some text for labels, title and axes ticks
ax.set_ylabel('Normalized execution time')
ax.set_title('Legacy vs Blobs mail pipeline')
ax.set_xticks(ind + (1.5 * width))
ax.set_xticklabels(tuple(map(lambda name: name.replace('_', ' '), graphs)))
ax.legend(
(rects1[0], rects2[0], rects3[0], rects4[0]),
('legacy', 'blobs', 'blobs + session cache',
'blobs + session cache + persistent http'))
# ax.grid()
plt.savefig(OUTPUT_FILENAME)
# plt.show()
if __name__ == '__main__':
data = get_data()
plot_data(data)
|
