benchmark-ips 2.3.0 → 2.11.0
Sign up to get free protection for your applications and to get access to all the features.
- checksums.yaml +5 -5
- data/History.md +228 -0
- data/LICENSE +20 -0
- data/README.md +129 -29
- data/examples/advanced.rb +20 -0
- data/examples/hold.rb +41 -0
- data/examples/save.rb +50 -0
- data/examples/simple.rb +47 -0
- data/lib/benchmark/compare.rb +59 -23
- data/lib/benchmark/ips/job/entry.rb +95 -0
- data/lib/benchmark/ips/job/noop_report.rb +27 -0
- data/lib/benchmark/ips/job/stdout_report.rb +64 -0
- data/lib/benchmark/ips/job.rb +211 -150
- data/lib/benchmark/ips/noop_suite.rb +25 -0
- data/lib/benchmark/ips/report.rb +53 -30
- data/lib/benchmark/ips/share.rb +50 -0
- data/lib/benchmark/ips/stats/bootstrap.rb +58 -0
- data/lib/benchmark/ips/stats/sd.rb +45 -0
- data/lib/benchmark/ips/stats/stats_metric.rb +21 -0
- data/lib/benchmark/ips.rb +91 -24
- data/lib/benchmark/timing.rb +39 -16
- metadata +24 -31
- data/.autotest +0 -23
- data/.gemtest +0 -0
- data/History.txt +0 -87
- data/Manifest.txt +0 -11
- data/Rakefile +0 -26
- data/test/test_benchmark_ips.rb +0 -161
checksums.yaml
CHANGED
@@ -1,7 +1,7 @@
|
|
1
1
|
---
|
2
|
-
|
3
|
-
metadata.gz:
|
4
|
-
data.tar.gz:
|
2
|
+
SHA256:
|
3
|
+
metadata.gz: 194a6da5977a23dc733ade6dadaefe1a7d7215678d6135ff33d5c37304bec993
|
4
|
+
data.tar.gz: e072ffd46009a79e13990e7435e1415ec24fd0407f86d566208fc597b8e33f7d
|
5
5
|
SHA512:
|
6
|
-
metadata.gz:
|
7
|
-
data.tar.gz:
|
6
|
+
metadata.gz: 5b3a8e41f223d3945daf11329e7cd68b1cc41f9b08b9f3f2e9e5a471a37134782f4e1f78caca102b7b3b08eb179ac4d78868bb1f97bfab4ac1ccef22aa38d42d
|
7
|
+
data.tar.gz: e5cd27c03050929b8edf402264c0f5a32560550ee6d79456d2513ae9b4a76f763f669a32ba4cc3bf7a45e7c086426631c2ba26724a749192fc637046bfd6a495
|
data/History.md
ADDED
@@ -0,0 +1,228 @@
|
|
1
|
+
### 2.11.0 / 2023-02-15
|
2
|
+
|
3
|
+
* Feature
|
4
|
+
* Adds .json! method to the ips block argument, allowing you to print the output as JSON to a file or STDOUT.
|
5
|
+
|
6
|
+
### 2.10.0 / 2022-02-17
|
7
|
+
|
8
|
+
* Feature
|
9
|
+
* Adds :order option to compare, with new `:baseline` order which compares all
|
10
|
+
variations against the first option benchmarked.
|
11
|
+
|
12
|
+
### 2.9.3 / 2022-01-25
|
13
|
+
|
14
|
+
* Bug fix
|
15
|
+
* All warmups and benchmarks must run at least once
|
16
|
+
|
17
|
+
### 2.9.2 / 2021-10-10
|
18
|
+
|
19
|
+
* Bug fix
|
20
|
+
* Fix a problem with certain configs of quiet mode
|
21
|
+
|
22
|
+
### 2.9.1 / 2021-05-24
|
23
|
+
|
24
|
+
* Bug fix
|
25
|
+
* Include all files in gem
|
26
|
+
|
27
|
+
### 2.9.0 / 2021-05-21
|
28
|
+
|
29
|
+
* Features
|
30
|
+
* Suite can now be set via an accessor
|
31
|
+
* Default SHARE_URL is now `ips.fastruby.io`, operated by Ombu Labs.
|
32
|
+
|
33
|
+
### 2.8.4 / 2020-12-03
|
34
|
+
|
35
|
+
* Bug fix
|
36
|
+
* Fixed hold! when results file does not exist.
|
37
|
+
|
38
|
+
### 2.8.3 / 2020-08-28
|
39
|
+
|
40
|
+
* Bug fix
|
41
|
+
* Fixed inaccuracy caused by integer overflows.
|
42
|
+
|
43
|
+
### 2.8.2 / 2020-05-04
|
44
|
+
|
45
|
+
* Bug fix
|
46
|
+
* Fixed problems with Manifest.txt.
|
47
|
+
* Empty interim results files are ignored.
|
48
|
+
|
49
|
+
### 2.8.0 / 2020-05-01
|
50
|
+
|
51
|
+
* Feature
|
52
|
+
* Allow running with empty ips block.
|
53
|
+
* Added save! method for saving interim results.
|
54
|
+
* Run more than just 1 cycle during warmup to reduce overhead.
|
55
|
+
* Optimized Job::Entry hot-path for fairer results on JRuby/TruffleRuby.
|
56
|
+
|
57
|
+
* Bug fix
|
58
|
+
* Removed the warmup section if set to 0.
|
59
|
+
* Added some RDoc docs.
|
60
|
+
* Added some examples in examples/
|
61
|
+
|
62
|
+
### 2.7.2 / 2016-08-18
|
63
|
+
|
64
|
+
* 1 bug fix:
|
65
|
+
* Restore old accessors. Fixes #76
|
66
|
+
|
67
|
+
### 2.7.1 / 2016-08-08
|
68
|
+
|
69
|
+
Add missing files
|
70
|
+
|
71
|
+
### 2.7.0 / 2016-08-05
|
72
|
+
|
73
|
+
* 1 minor features:
|
74
|
+
* Add support for confidence intervals
|
75
|
+
|
76
|
+
* 1 bug fixes:
|
77
|
+
* Cleanup a few coding patterns
|
78
|
+
|
79
|
+
* 2 doc fixes:
|
80
|
+
* Add infos about benchark.fyi to Readme
|
81
|
+
* Remove ancient releases
|
82
|
+
|
83
|
+
* 3 merged PRs:
|
84
|
+
* Merge pull request #65 from kbrock/fixup_inject
|
85
|
+
* Merge pull request #67 from benoittgt/master
|
86
|
+
* Merge pull request #69 from chrisseaton/kalibera-confidence-intervals
|
87
|
+
|
88
|
+
### MISSING 2.6.0 and 2.6.1
|
89
|
+
|
90
|
+
### 2.5.0 / 2016-02-14
|
91
|
+
|
92
|
+
* 1 minor feature:
|
93
|
+
* Add iterations option.
|
94
|
+
|
95
|
+
* 1 bug fixes:
|
96
|
+
* Don't tell people something is slower if it's within the error.
|
97
|
+
|
98
|
+
* 2 merged PRs:
|
99
|
+
* Merge pull request #58 from chrisseaton/iterations
|
100
|
+
* Merge pull request #60 from chrisseaton/significance
|
101
|
+
|
102
|
+
### 2.4.1 / 2016-02-12
|
103
|
+
|
104
|
+
* 1 bug fix:
|
105
|
+
* Add missing files to gem
|
106
|
+
|
107
|
+
### 2.4.0 / 2016-02-12
|
108
|
+
|
109
|
+
* 1 minor features
|
110
|
+
* Add support for hold! and independent invocations.
|
111
|
+
|
112
|
+
* 6 bug fixes
|
113
|
+
* Separate messages for warming up and calculating.
|
114
|
+
* Tighten timing loop.
|
115
|
+
* Pass simple types into Job#create_report
|
116
|
+
* More concise sorting
|
117
|
+
* Fix runtime comparison
|
118
|
+
* Use runtime if ips is not available
|
119
|
+
|
120
|
+
* 5 doc fixes
|
121
|
+
* Fix typo unsed --> used
|
122
|
+
* Better document Report::Entry
|
123
|
+
* Fix some typos in docs
|
124
|
+
* Don't calculate mean 2 times
|
125
|
+
* Add more tolerance to tests
|
126
|
+
|
127
|
+
* 13 merged PRs
|
128
|
+
* Merge pull request #44 from kbrock/job_extract
|
129
|
+
* Merge pull request #45 from kbrock/runtime_only
|
130
|
+
* Merge pull request #47 from kbrock/use_avg
|
131
|
+
* Merge pull request #46 from kbrock/report_stdout
|
132
|
+
* Merge pull request #48 from bquorning/fix-label-for-runtime-comparison
|
133
|
+
* Merge pull request #50 from tjschuck/fix_typo
|
134
|
+
* Merge pull request #51 from bquorning/all-reports-respond-to-ips
|
135
|
+
* Merge pull request #52 from kbrock/document_reports
|
136
|
+
* Merge pull request #53 from kbrock/interface_create_report
|
137
|
+
* Merge pull request #54 from PragTob/patch-2
|
138
|
+
* Merge pull request #55 from chrisseaton/messages
|
139
|
+
* Merge pull request #56 from chrisseaton/independence
|
140
|
+
* Merge pull request #57 from chrisseaton/tighten-loop
|
141
|
+
|
142
|
+
### 2.3.0 / 2015-07-20
|
143
|
+
|
144
|
+
* 2 minor features:
|
145
|
+
* Support keyword arguments
|
146
|
+
* Allow any datatype for labels (use #to_s conversion)
|
147
|
+
|
148
|
+
* 1 doc/test changes:
|
149
|
+
* Newer Travis for 1.8.7, ree, and 2.2.2
|
150
|
+
|
151
|
+
* 3 PRs merged:
|
152
|
+
* Merge pull request #41 from kbrock/kwargs-support
|
153
|
+
* Merge pull request #42 from kbrock/newer_travis
|
154
|
+
* Merge pull request #43 from kbrock/non_to_s_labels
|
155
|
+
|
156
|
+
### 2.2.0 / 2015-05-09
|
157
|
+
|
158
|
+
* 1 minor features:
|
159
|
+
* Fix quiet mode
|
160
|
+
* Allow passing a custom suite via config
|
161
|
+
* Silent a job if a suite was passed and is quiet
|
162
|
+
* Export report to json file.
|
163
|
+
* Accept symbol as report's argument.
|
164
|
+
|
165
|
+
* 2 doc fixes:
|
166
|
+
* Squish duplicate `to` in README
|
167
|
+
* Update copyright to 2015. [ci skip]
|
168
|
+
|
169
|
+
* 9 PRs merged:
|
170
|
+
* Merge pull request #37 from splattael/patch-1
|
171
|
+
* Merge pull request #36 from kirs/quiet-mode
|
172
|
+
* Merge pull request #35 from JuanitoFatas/doc/suite
|
173
|
+
* Merge pull request #34 from splattael/config-suite
|
174
|
+
* Merge pull request #33 from splattael/suite-quiet
|
175
|
+
* Merge pull request #32 from O-I/remove-gemfile-lock
|
176
|
+
* Merge pull request #31 from JuanitoFatas/doc/bump-copyright-year
|
177
|
+
* Merge pull request #29 from JuanitoFatas/feature/json-export
|
178
|
+
* Merge pull request #26 from JuanitoFatas/feature/takes-symbol-as-report-parameter
|
179
|
+
|
180
|
+
### 2.1.1 / 2015-01-12
|
181
|
+
|
182
|
+
* 1 minor fix:
|
183
|
+
* Don't send label through printf so that % work directly
|
184
|
+
|
185
|
+
* 1 documenation changes:
|
186
|
+
* Use HEREDOC and wrap at 80 chars for example result description
|
187
|
+
|
188
|
+
* 1 usage fix:
|
189
|
+
* Add gemspec for use via bundler git
|
190
|
+
|
191
|
+
* 1 PR merged:
|
192
|
+
* Merge pull request #24 from zzak/simple-format-result-description
|
193
|
+
|
194
|
+
### 2.1.0 / 2014-11-10
|
195
|
+
|
196
|
+
* Documentation changes:
|
197
|
+
* Many documentation fixes by Juanito Fatas!
|
198
|
+
* Minor readme fix by Will Leinweber
|
199
|
+
|
200
|
+
* 2 minor features:
|
201
|
+
* Displaying the total runtime for a job is suppressed unless interesting
|
202
|
+
* Formatting of large values improved (human vs raw mode)
|
203
|
+
* Contributed by Charles Oliver Nutter
|
204
|
+
|
205
|
+
### 2.0.0 / 2014-06-18
|
206
|
+
|
207
|
+
* The 'Davy Stevenson' release!
|
208
|
+
* Codename: Springtime Hummingbird Dance
|
209
|
+
|
210
|
+
* Big API refactoring so the internal bits are easier to use
|
211
|
+
* Bump to 2.0 because return types changed to make the API better
|
212
|
+
|
213
|
+
* Contributors added:
|
214
|
+
* Davy Stevenson
|
215
|
+
* Juanito Fatas
|
216
|
+
* Benoit Daloze
|
217
|
+
* Matias
|
218
|
+
* Tony Arcieri
|
219
|
+
* Vipul A M
|
220
|
+
* Zachary Scott
|
221
|
+
* schneems (Richard Schneeman)
|
222
|
+
|
223
|
+
### 1.0.0 / 2012-03-23
|
224
|
+
|
225
|
+
* 1 major enhancement
|
226
|
+
|
227
|
+
* Birthday!
|
228
|
+
|
data/LICENSE
ADDED
@@ -0,0 +1,20 @@
|
|
1
|
+
Copyright (c) 2015 Evan Phoenix
|
2
|
+
|
3
|
+
Permission is hereby granted, free of charge, to any person obtaining
|
4
|
+
a copy of this software and associated documentation files (the
|
5
|
+
'Software'), to deal in the Software without restriction, including
|
6
|
+
without limitation the rights to use, copy, modify, merge, publish,
|
7
|
+
distribute, sublicense, and/or sell copies of the Software, and to
|
8
|
+
permit persons to whom the Software is furnished to do so, subject to
|
9
|
+
the following conditions:
|
10
|
+
|
11
|
+
The above copyright notice and this permission notice shall be
|
12
|
+
included in all copies or substantial portions of the Software.
|
13
|
+
|
14
|
+
THE SOFTWARE IS PROVIDED 'AS IS', WITHOUT WARRANTY OF ANY KIND,
|
15
|
+
EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
|
16
|
+
MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.
|
17
|
+
IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY
|
18
|
+
CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT,
|
19
|
+
TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE
|
20
|
+
SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
|
data/README.md
CHANGED
@@ -1,16 +1,17 @@
|
|
1
|
+
# benchmark-ips
|
2
|
+
|
3
|
+
* rdoc :: http://rubydoc.info/gems/benchmark-ips
|
4
|
+
* home :: https://github.com/evanphx/benchmark-ips
|
5
|
+
|
1
6
|
[![Gem Version](https://badge.fury.io/rb/benchmark-ips.svg)](http://badge.fury.io/rb/benchmark-ips)
|
2
7
|
[![Build Status](https://secure.travis-ci.org/evanphx/benchmark-ips.svg)](http://travis-ci.org/evanphx/benchmark-ips)
|
3
8
|
[![Inline docs](http://inch-ci.org/github/evanphx/benchmark-ips.svg)](http://inch-ci.org/github/evanphx/benchmark-ips)
|
4
9
|
|
5
|
-
# benchmark-ips
|
6
|
-
|
7
10
|
* https://github.com/evanphx/benchmark-ips
|
8
11
|
|
9
|
-
* [documentation](http://rubydoc.info/gems/benchmark-ips)
|
10
|
-
|
11
12
|
## DESCRIPTION:
|
12
13
|
|
13
|
-
|
14
|
+
An iterations per second enhancement to Benchmark.
|
14
15
|
|
15
16
|
## FEATURES/PROBLEMS:
|
16
17
|
|
@@ -136,6 +137,129 @@ Benchmark.ips do |x|
|
|
136
137
|
end
|
137
138
|
```
|
138
139
|
|
140
|
+
### Independent benchmarking
|
141
|
+
|
142
|
+
If you are comparing multiple implementations of a piece of code you may want
|
143
|
+
to benchmark them in separate invocations of Ruby so that the measurements
|
144
|
+
are independent of each other. You can do this with the `hold!` command.
|
145
|
+
|
146
|
+
```ruby
|
147
|
+
Benchmark.ips do |x|
|
148
|
+
|
149
|
+
# Hold results between multiple invocations of Ruby
|
150
|
+
x.hold! 'filename'
|
151
|
+
|
152
|
+
end
|
153
|
+
```
|
154
|
+
|
155
|
+
This will run only one benchmarks each time you run the command, storing
|
156
|
+
results in the specified file. The file is deleted when all results have been
|
157
|
+
gathered and the report is shown.
|
158
|
+
|
159
|
+
Alternatively, if you prefer a different approach, the `save!` command is
|
160
|
+
available. Examples for [hold!](examples/hold.rb) and [save!](examples/save.rb) are available in
|
161
|
+
the `examples/` directory.
|
162
|
+
|
163
|
+
|
164
|
+
### Multiple iterations
|
165
|
+
|
166
|
+
In some cases you may want to run multiple iterations of the warmup and
|
167
|
+
calculation stages and take only the last result for comparison. This is useful
|
168
|
+
if you are benchmarking with an implementation of Ruby that optimizes using
|
169
|
+
tracing or on-stack-replacement, because to those implementations the
|
170
|
+
calculation phase may appear as new, unoptimized code.
|
171
|
+
|
172
|
+
You can do this with the `iterations` option, which by default is `1`. The
|
173
|
+
total time spent will then be `iterations * warmup + iterations * time` seconds.
|
174
|
+
|
175
|
+
```ruby
|
176
|
+
Benchmark.ips do |x|
|
177
|
+
|
178
|
+
x.config(:iterations => 3)
|
179
|
+
|
180
|
+
# or
|
181
|
+
|
182
|
+
x.iterations = 3
|
183
|
+
|
184
|
+
end
|
185
|
+
```
|
186
|
+
|
187
|
+
### Online sharing
|
188
|
+
|
189
|
+
If you want to quickly share your benchmark result with others, run you benchmark
|
190
|
+
with `SHARE=1` argument. For example: `SHARE=1 ruby my_benchmark.rb`.
|
191
|
+
|
192
|
+
Result will be sent to [benchmark.fyi](https://ips.fastruby.io/) and benchmark-ips
|
193
|
+
will display the link to share the benchmark's result.
|
194
|
+
|
195
|
+
If you want to run your own instance of [benchmark.fyi](https://github.com/evanphx/benchmark.fyi)
|
196
|
+
and share it to that instance, you can do this: `SHARE_URL=https://ips.example.com ruby my_benchmark.rb`
|
197
|
+
|
198
|
+
### Advanced Statistics
|
199
|
+
|
200
|
+
By default, the margin of error shown is plus-minus one standard deviation. If
|
201
|
+
a more advanced statistical test is wanted, a bootstrap confidence interval
|
202
|
+
can be calculated instead. A bootstrap confidence interval has the advantages of
|
203
|
+
arguably being more mathematically sound for this application than a standard
|
204
|
+
deviation, it additionally produces an error for relative slowdowns, which the
|
205
|
+
standard deviation does not, and it is arguably more intuitive and actionable.
|
206
|
+
|
207
|
+
When a bootstrap confidence interval is used, a median of the interval is used
|
208
|
+
rather than the mean of the samples, which is what you get with the default
|
209
|
+
standard deviation.
|
210
|
+
|
211
|
+
The bootstrap confidence interval used is the one described by Tomas Kalibera.
|
212
|
+
Note that for this technique to be valid your benchmark should have reached a
|
213
|
+
non-periodic steady state with statistically independent samples (it should
|
214
|
+
have warmed up) by the time measurements start.
|
215
|
+
|
216
|
+
Using a bootstrap confidence internal requires that the 'kalibera' gem is
|
217
|
+
installed separately. This gem is not a formal dependency, as by default it is
|
218
|
+
not needed.
|
219
|
+
|
220
|
+
```
|
221
|
+
gem install kalibera
|
222
|
+
```
|
223
|
+
|
224
|
+
```ruby
|
225
|
+
Benchmark.ips do |x|
|
226
|
+
|
227
|
+
# The default is :stats => :sd, which doesn't have a configurable confidence
|
228
|
+
x.config(:stats => :bootstrap, :confidence => 95)
|
229
|
+
|
230
|
+
# or
|
231
|
+
|
232
|
+
x.stats = :bootstrap
|
233
|
+
x.confidence = 95
|
234
|
+
|
235
|
+
# confidence is 95% by default, so it can be omitted
|
236
|
+
|
237
|
+
end
|
238
|
+
```
|
239
|
+
|
240
|
+
### Output as JSON
|
241
|
+
|
242
|
+
You can generate output in JSON. If you want to write JSON to a file, pass filename to `json!` method:
|
243
|
+
|
244
|
+
```ruby
|
245
|
+
Benchmark.ips do |x|
|
246
|
+
x.report("some report") { }
|
247
|
+
x.json! 'filename.json'
|
248
|
+
end
|
249
|
+
```
|
250
|
+
|
251
|
+
If you want to write JSON to STDOUT, pass `STDOUT` to `json!` method and set `quiet = true` before `json!`:
|
252
|
+
|
253
|
+
```ruby
|
254
|
+
Benchmark.ips do |x|
|
255
|
+
x.report("some report") { }
|
256
|
+
x.quiet = true
|
257
|
+
x.json! STDOUT
|
258
|
+
end
|
259
|
+
```
|
260
|
+
|
261
|
+
This is useful when the output from `benchmark-ips` becomes an input of other tools via stdin.
|
262
|
+
|
139
263
|
## REQUIREMENTS:
|
140
264
|
|
141
265
|
* None!
|
@@ -153,27 +277,3 @@ After checking out the source, run:
|
|
153
277
|
This task will install any missing dependencies, run the tests/specs,
|
154
278
|
and generate the RDoc.
|
155
279
|
|
156
|
-
## LICENSE:
|
157
|
-
|
158
|
-
(The MIT License)
|
159
|
-
|
160
|
-
Copyright (c) 2015 Evan Phoenix
|
161
|
-
|
162
|
-
Permission is hereby granted, free of charge, to any person obtaining
|
163
|
-
a copy of this software and associated documentation files (the
|
164
|
-
'Software'), to deal in the Software without restriction, including
|
165
|
-
without limitation the rights to use, copy, modify, merge, publish,
|
166
|
-
distribute, sublicense, and/or sell copies of the Software, and to
|
167
|
-
permit persons to whom the Software is furnished to do so, subject to
|
168
|
-
the following conditions:
|
169
|
-
|
170
|
-
The above copyright notice and this permission notice shall be
|
171
|
-
included in all copies or substantial portions of the Software.
|
172
|
-
|
173
|
-
THE SOFTWARE IS PROVIDED 'AS IS', WITHOUT WARRANTY OF ANY KIND,
|
174
|
-
EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
|
175
|
-
MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.
|
176
|
-
IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY
|
177
|
-
CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT,
|
178
|
-
TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE
|
179
|
-
SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
|
@@ -0,0 +1,20 @@
|
|
1
|
+
#!/usr/bin/env ruby
|
2
|
+
|
3
|
+
require 'benchmark/ips'
|
4
|
+
|
5
|
+
Benchmark.ips do |x|
|
6
|
+
|
7
|
+
# Use bootstrap confidence intervals
|
8
|
+
x.stats = :bootstrap
|
9
|
+
|
10
|
+
# Set confidence to 95%
|
11
|
+
x.confidence = 95
|
12
|
+
|
13
|
+
# Run multiple iterations for better warmup
|
14
|
+
x.iterations = 3
|
15
|
+
|
16
|
+
x.report("mul") { 2 * 2 * 2 * 2 * 2 * 2 * 2 * 2 }
|
17
|
+
x.report("pow") { 2 ** 8 }
|
18
|
+
|
19
|
+
x.compare!
|
20
|
+
end
|
data/examples/hold.rb
ADDED
@@ -0,0 +1,41 @@
|
|
1
|
+
#!/usr/bin/env ruby
|
2
|
+
# example to explain hold! usage https://github.com/evanphx/benchmark-ips/issues/85
|
3
|
+
# The hold! feature expects to be run twice, generally with different Rubys.
|
4
|
+
# hold! can also be used to compare modules changes which impact the run time
|
5
|
+
# RUN_1: ruby examples/hold.rb
|
6
|
+
# Warming up --------------------------------------
|
7
|
+
# without 172.168k i/100ms
|
8
|
+
# Calculating -------------------------------------
|
9
|
+
# without 2.656M (± 3.3%) i/s - 13.429M in 5.062098s
|
10
|
+
#
|
11
|
+
# RUN_2: WITH_MODULE=true ruby examples/hold.rb
|
12
|
+
# Warming up --------------------------------------
|
13
|
+
# with 92.087k i/100ms
|
14
|
+
# Calculating -------------------------------------
|
15
|
+
# with 1.158M (± 1.4%) i/s - 5.801M in 5.010084s
|
16
|
+
#
|
17
|
+
# Comparison:
|
18
|
+
# without: 2464721.3 i/s
|
19
|
+
# with: 1158179.6 i/s - 2.13x slower
|
20
|
+
require 'benchmark/ips'
|
21
|
+
|
22
|
+
Benchmark.ips do |x|
|
23
|
+
x.report('without') do
|
24
|
+
'Bruce'.inspect
|
25
|
+
end
|
26
|
+
|
27
|
+
if ENV['WITH_MODULE'] == 'true'
|
28
|
+
class String
|
29
|
+
def inspect
|
30
|
+
result = %w[Bruce Wayne is Batman]
|
31
|
+
result.join(' ')
|
32
|
+
end
|
33
|
+
end
|
34
|
+
end
|
35
|
+
|
36
|
+
x.report('with') do
|
37
|
+
'Bruce'.inspect
|
38
|
+
end
|
39
|
+
x.hold! 'temp_results'
|
40
|
+
x.compare!
|
41
|
+
end
|
data/examples/save.rb
ADDED
@@ -0,0 +1,50 @@
|
|
1
|
+
#!/usr/bin/env ruby
|
2
|
+
|
3
|
+
# example to explain save!
|
4
|
+
# The save! feature expects to be run twice, generally with different Rubys.
|
5
|
+
# save! can also be used to compare modules changes which impact the run time
|
6
|
+
#
|
7
|
+
# If you're comparing ruby versions, Just use the version in the label
|
8
|
+
#
|
9
|
+
# x.report("ruby #{RUBY_VERSION}") { 'Bruce'.inspect }
|
10
|
+
#
|
11
|
+
# Or use a hash
|
12
|
+
#
|
13
|
+
# x.report("version" => RUBY_VERSION, "method" => 'bruce') { 'Bruce'.inspect }
|
14
|
+
#
|
15
|
+
# RUN_1: SAVE_FILE='run1.out' ruby examples/hold.rb
|
16
|
+
# Warming up --------------------------------------
|
17
|
+
# without 172.168k i/100ms
|
18
|
+
# Calculating -------------------------------------
|
19
|
+
# without 2.656M (± 3.3%) i/s - 13.429M in 5.062098s
|
20
|
+
#
|
21
|
+
# RUN_2: SAVE_FILE='run1.out' WITH_MODULE=true ruby examples/hold.rb
|
22
|
+
# Warming up --------------------------------------
|
23
|
+
# with 92.087k i/100ms
|
24
|
+
# Calculating -------------------------------------
|
25
|
+
# with 1.158M (± 1.4%) i/s - 5.801M in 5.010084s
|
26
|
+
#
|
27
|
+
# Comparison:
|
28
|
+
# without: 2464721.3 i/s
|
29
|
+
# with: 1158179.6 i/s - 2.13x slower
|
30
|
+
# CLEANUP: rm run1.out
|
31
|
+
|
32
|
+
require 'benchmark/ips'
|
33
|
+
|
34
|
+
Benchmark.ips do |x|
|
35
|
+
x.report(ENV['WITH_MODULE'] == 'true' ? 'with' : 'without') do
|
36
|
+
'Bruce'.inspect
|
37
|
+
end
|
38
|
+
|
39
|
+
if ENV['WITH_MODULE'] == 'true'
|
40
|
+
class String
|
41
|
+
def inspect
|
42
|
+
result = %w[Bruce Wayne is Batman]
|
43
|
+
result.join(' ')
|
44
|
+
end
|
45
|
+
end
|
46
|
+
end
|
47
|
+
|
48
|
+
x.save! ENV['SAVE_FILE'] if ENV['SAVE_FILE']
|
49
|
+
x.compare!
|
50
|
+
end
|
data/examples/simple.rb
ADDED
@@ -0,0 +1,47 @@
|
|
1
|
+
#!/usr/bin/env ruby
|
2
|
+
|
3
|
+
require 'benchmark/ips'
|
4
|
+
|
5
|
+
Benchmark.ips do |x|
|
6
|
+
|
7
|
+
# Configure the number of seconds used during
|
8
|
+
# the warmup phase and calculation phase
|
9
|
+
x.config(:time => 5, :warmup => 2)
|
10
|
+
|
11
|
+
# These parameters can also be configured this way
|
12
|
+
x.time = 5
|
13
|
+
x.warmup = 2
|
14
|
+
|
15
|
+
# Typical mode, runs the block as many times as it can
|
16
|
+
x.report("addition") { 1 + 2 }
|
17
|
+
|
18
|
+
# To reduce overhead, the number of iterations is passed in
|
19
|
+
# and the block must run the code the specific number of times.
|
20
|
+
# Used for when the workload is very small and any overhead
|
21
|
+
# introduces incorrectable errors.
|
22
|
+
x.report(:addition2) do |times|
|
23
|
+
i = 0
|
24
|
+
while i < times
|
25
|
+
1 + 2
|
26
|
+
i += 1
|
27
|
+
end
|
28
|
+
end
|
29
|
+
|
30
|
+
# To reduce overhead even more, grafts the code given into
|
31
|
+
# the loop that performs the iterations internally to reduce
|
32
|
+
# overhead. Typically not needed, use the |times| form instead.
|
33
|
+
x.report("addition3", "1 + 2")
|
34
|
+
|
35
|
+
# Really long labels should be formatted correctly
|
36
|
+
x.report("addition-test-long-label") { 1 + 2 }
|
37
|
+
|
38
|
+
x.compare!
|
39
|
+
end
|
40
|
+
|
41
|
+
puts <<-EOD
|
42
|
+
Typical results will show addition2 & addition3 to be the most performant, and
|
43
|
+
they should perform reasonably similarly. You should see addition and
|
44
|
+
addition-test-long-label to perform very similarly to each other (as they are
|
45
|
+
running the same test, just with different labels), and they should both run in
|
46
|
+
the neighborhood of 3.5 times slower than addition2 and addition3."
|
47
|
+
EOD
|