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
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
|
.. _cherrypyspeed:
**********************
How fast is CherryPy ?
**********************
Introduction
============
When people ask this question, they usually mean "how fast will my CherryPy-based application be ?".
In 99% of the cases, the answer is "this depends on your actual application code, not on !CherryPy itself".
The reason is that, for 99% of the real-world dynamic applications, most of the time spent to return a page will be spent in your actual application code, and the time actually spent in the CherryPy code will be negligible.
For instance, a typical page that requires a few database calls to be built might take in total 200ms to be served. Out of these 200ms, about 2ms will be spent by CherryPy itself, and 198ms will be spend in your actual database calls and page rendering...
So you can see that, if you want to optimize anything, you should really optimize your actual application code before you try to optimize !CherryPy
Raw speed of the CherryPy HTTP server
=====================================
Despite the real-life most common scenario explained in the introduction, some people still want to know the raw speed of the CherryPy HTTP server.
So I sat down and did some benchmarking...
About the benchmark
-------------------
This benchmarking only makes sense on very small documents, otherwise we're no longer measuring the raw speed of the HTTP server, but also the speed of the application ...
This benchmarking was performed on a laptop in the following environment:
* Processor: Pentium M 1.6 Ghz
* RAM: 1GB
* Windows XP 2
* Load testing tool: ab from Apache2
* CherryPy version: SVN snapshot on 2005/01/13
Note that "ab" was running on the same machine as the CherryPy server, so CherryPy is probably a bit faster than what we're getting.
Test 1: Dynamic content / single threaded server
------------------------------------------------
I used the following basic CherryPy app::
#!python
from cherrypy import cpg
class Root:
def index(self):
return "OK"
index.exposed = True
cpg.root = Root()
cpg.server.start(configMap = {'socketPort': 10000})
Here are the "ab" results::
$ ./ab.exe -n 1000 http://localhost:10000/
This is ApacheBench, Version 2.0.41-dev <$Revision: 1.121.2.12 $> apache-2.0
Copyright (c) 1996 Adam Twiss, Zeus Technology Ltd, http://www.zeustech.net/
Copyright (c) 1998-2002 The Apache Software Foundation, http://www.apache.org/
Benchmarking localhost (be patient)
Completed 100 requests
Completed 200 requests
Completed 300 requests
Completed 400 requests
Completed 500 requests
Completed 600 requests
Completed 700 requests
Completed 800 requests
Completed 900 requests
Finished 1000 requests
Server Software: CherryPy/2.0.0b
Server Hostname: localhost
Server Port: 10000
Document Path: /
Document Length: 2 bytes
Concurrency Level: 1
Time taken for tests: 1.789044 seconds
Complete requests: 1000
Failed requests: 0
Write errors: 0
Total transferred: 127000 bytes
HTML transferred: 2000 bytes
Requests per second: 558.96 [#/sec] (mean)
Time per request: 1.789 [ms] (mean)
Time per request: 1.789 [ms] (mean, across all concurrent requests)
Transfer rate: 69.31 [Kbytes/sec] received
Connection Times (ms)
min mean[+/-sd] median max
Connect: 0 0 1.9 0 15
Processing: 0 1 4.2 0 15
Waiting: 0 0 0.8 0 15
Total: 0 1 4.5 0 15
Percentage of the requests served within a certain time (ms)
50% 0
66% 0
75% 0
80% 0
90% 15
95% 15
98% 15
99% 15
100% 15 (longest request)
As you can see, CherryPy averaged 558 requests/second, which is pretty good ...
Test 2: Dynamic content / multi threaded server
-----------------------------------------------
I used the same code as test 1, but started CherryPy in thread-pool mode, with 10 threads.
I also told "ab" to simulate 10 concurrent users ...
Here are the "ab" results::
$ ./ab.exe -c 10 -n 1000 http://localhost:10000/
This is ApacheBench, Version 2.0.41-dev <$Revision: 1.121.2.12 $> apache-2.0
Copyright (c) 1996 Adam Twiss, Zeus Technology Ltd, http://www.zeustech.net/
Copyright (c) 1998-2002 The Apache Software Foundation, http://www.apache.org/
Benchmarking localhost (be patient)
Completed 100 requests
Completed 200 requests
Completed 300 requests
Completed 400 requests
Completed 500 requests
Completed 600 requests
Completed 700 requests
Completed 800 requests
Completed 900 requests
Finished 1000 requests
Server Software: CherryPy/2.0.0b
Server Hostname: localhost
Server Port: 10000
Document Path: /
Document Length: 2 bytes
Concurrency Level: 10
Time taken for tests: 2.327670 seconds
Complete requests: 1000
Failed requests: 0
Write errors: 0
Total transferred: 127000 bytes
HTML transferred: 2000 bytes
Requests per second: 429.61 [#/sec] (mean)
Time per request: 23.277 [ms] (mean)
Time per request: 2.328 [ms] (mean, across all concurrent requests)
Transfer rate: 53.27 [Kbytes/sec] received
Connection Times (ms)
min mean[+/-sd] median max
Connect: 0 0 2.3 0 15
Processing: 15 21 8.9 15 47
Waiting: 0 16 6.2 15 47
Total: 15 22 9.0 15 47
Percentage of the requests served within a certain time (ms)
50% 15
66% 31
75% 31
80% 31
90% 31
95% 31
98% 47
99% 47
100% 47 (longest request)
As you can see, CherryPy averaged 429 requests/second, which is a bit less than test 1 (there is a small thread-switching overhead), but is still pretty good ...
Test 3: Static content / single threaded server
-----------------------------------------------
This time, I used CherryPy to serve a static file from disc.
The file was a simple text containing "OK".
Here was the config file for CherryPy::
[server]
socketPort = 10000
[staticContent]
static.html = static.html
Here are the "ab" results::
$ ./ab.exe -n 1000 http://localhost:10000/static.html
This is ApacheBench, Version 2.0.41-dev <$Revision: 1.121.2.12 $> apache-2.0
Copyright (c) 1996 Adam Twiss, Zeus Technology Ltd, http://www.zeustech.net/
Copyright (c) 1998-2002 The Apache Software Foundation, http://www.apache.org/
Benchmarking localhost (be patient)
Completed 100 requests
Completed 200 requests
Completed 300 requests
Completed 400 requests
Completed 500 requests
Completed 600 requests
Completed 700 requests
Completed 800 requests
Completed 900 requests
Finished 1000 requests
Server Software: CherryPy/2.0.0b
Server Hostname: localhost
Server Port: 10000
Document Path: /static.html
Document Length: 4 bytes
Concurrency Level: 1
Time taken for tests: 1.979130 seconds
Complete requests: 1000
Failed requests: 0
Write errors: 0
Total transferred: 175000 bytes
HTML transferred: 4000 bytes
Requests per second: 505.27 [#/sec] (mean)
Time per request: 1.979 [ms] (mean)
Time per request: 1.979 [ms] (mean, across all concurrent requests)
Transfer rate: 85.90 [Kbytes/sec] received
Connection Times (ms)
min mean[+/-sd] median max
Connect: 0 0 2.2 0 15
Processing: 0 1 4.3 0 15
Waiting: 0 0 0.5 0 15
Total: 0 1 4.8 0 15
Percentage of the requests served within a certain time (ms)
50% 0
66% 0
75% 0
80% 0
90% 15
95% 15
98% 15
99% 15
100% 15 (longest request)
As you can see, CherryPy averaged 505 requests/second. Again it is a little bit less than a dynamic page, but it is still pretty good ...
|
