|
2052.3.2
by John Arbash Meinel
Change Copyright .. by Canonical to Copyright ... Canonical |
1 |
# Copyright (C) 2006 Canonical Ltd
|
|
1911.2.1
by John Arbash Meinel
Cache encode/decode operations, saves memory and time. Especially when committing a new kernel tree with 7.7M new lines to annotate |
2 |
#
|
3 |
# This program is free software; you can redistribute it and/or modify
|
|
|
2052.3.1
by John Arbash Meinel
Add tests to cleanup the copyright of all source files |
4 |
# it under the terms of the GNU General Public License as published by
|
5 |
# the Free Software Foundation; either version 2 of the License, or
|
|
6 |
# (at your option) any later version.
|
|
|
1911.2.1
by John Arbash Meinel
Cache encode/decode operations, saves memory and time. Especially when committing a new kernel tree with 7.7M new lines to annotate |
7 |
#
|
8 |
# This program is distributed in the hope that it will be useful,
|
|
9 |
# but WITHOUT ANY WARRANTY; without even the implied warranty of
|
|
10 |
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
|
|
11 |
# GNU General Public License for more details.
|
|
12 |
#
|
|
13 |
# You should have received a copy of the GNU General Public License
|
|
14 |
# along with this program; if not, write to the Free Software
|
|
15 |
# Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
|
|
16 |
||
|
2052.3.1
by John Arbash Meinel
Add tests to cleanup the copyright of all source files |
17 |
|
|
1911.2.1
by John Arbash Meinel
Cache encode/decode operations, saves memory and time. Especially when committing a new kernel tree with 7.7M new lines to annotate |
18 |
"""Tests for encoding performance."""
|
19 |
||
20 |
from bzrlib import ( |
|
|
1911.2.3
by John Arbash Meinel
Moving everything into a new location so that we can cache more than just revision ids |
21 |
cache_utf8, |
|
1911.2.1
by John Arbash Meinel
Cache encode/decode operations, saves memory and time. Especially when committing a new kernel tree with 7.7M new lines to annotate |
22 |
osutils, |
23 |
)
|
|
24 |
||
25 |
from bzrlib.benchmarks import Benchmark |
|
26 |
||
27 |
||
28 |
_normal_revision_id = (u'john@arbash-meinel.com-20060801200018' |
|
29 |
u'-cafa6272d9b8cac4') |
|
30 |
_unicode_revision_id = (u'\u062c\u0648\u062c\u0648@\xe5rbash-meinel.com-' |
|
31 |
u'\xb5\xb5\xb5-20060801200018-cafa6272d9b8cac4') |
|
32 |
||
33 |
_normal_revision_id_utf8 = _normal_revision_id.encode('utf-8') |
|
34 |
_unicode_revision_id_utf8 = _unicode_revision_id.encode('utf-8') |
|
35 |
||
36 |
||
37 |
class EncodingBenchmark(Benchmark): |
|
38 |
||
39 |
def setUp(self): |
|
40 |
super(EncodingBenchmark, self).setUp() |
|
41 |
# Make sure we start and end with a clean cache
|
|
|
1911.2.3
by John Arbash Meinel
Moving everything into a new location so that we can cache more than just revision ids |
42 |
cache_utf8.clear_encoding_cache() |
43 |
self.addCleanup(cache_utf8.clear_encoding_cache) |
|
|
1911.2.1
by John Arbash Meinel
Cache encode/decode operations, saves memory and time. Especially when committing a new kernel tree with 7.7M new lines to annotate |
44 |
|
45 |
def encode_1M(self, revision_id): |
|
46 |
"""Encode the given revision id 1 million times"""
|
|
47 |
# In a real kernel tree there are 7.7M lines of code
|
|
48 |
# so the initial import actually has to encode a revision
|
|
49 |
# id to store annotated lines one time for every line.
|
|
50 |
for i in xrange(1000000): |
|
51 |
revision_id.encode('utf8') |
|
52 |
||
53 |
def encode_cached_1M(self, revision_id): |
|
54 |
"""Encode the given revision id 1 million times using the cache"""
|
|
|
1911.2.3
by John Arbash Meinel
Moving everything into a new location so that we can cache more than just revision ids |
55 |
encode = cache_utf8.encode |
|
1911.2.1
by John Arbash Meinel
Cache encode/decode operations, saves memory and time. Especially when committing a new kernel tree with 7.7M new lines to annotate |
56 |
for i in xrange(1000000): |
|
1911.2.3
by John Arbash Meinel
Moving everything into a new location so that we can cache more than just revision ids |
57 |
encode(revision_id) |
|
1911.2.1
by John Arbash Meinel
Cache encode/decode operations, saves memory and time. Especially when committing a new kernel tree with 7.7M new lines to annotate |
58 |
|
59 |
def encode_multi(self, revision_list, count): |
|
60 |
"""Encode each entry in the list count times"""
|
|
61 |
for i in xrange(count): |
|
62 |
for revision_id in revision_list: |
|
63 |
revision_id.encode('utf-8') |
|
64 |
||
65 |
def encode_cached_multi(self, revision_list, count): |
|
66 |
"""Encode each entry in the list count times"""
|
|
|
1911.2.3
by John Arbash Meinel
Moving everything into a new location so that we can cache more than just revision ids |
67 |
encode = cache_utf8.encode |
|
1911.2.1
by John Arbash Meinel
Cache encode/decode operations, saves memory and time. Especially when committing a new kernel tree with 7.7M new lines to annotate |
68 |
for i in xrange(count): |
69 |
for revision_id in revision_list: |
|
|
1911.2.3
by John Arbash Meinel
Moving everything into a new location so that we can cache more than just revision ids |
70 |
encode(revision_id) |
|
1911.2.1
by John Arbash Meinel
Cache encode/decode operations, saves memory and time. Especially when committing a new kernel tree with 7.7M new lines to annotate |
71 |
|
72 |
def test_encode_1_by_1M_ascii(self): |
|
73 |
"""Test encoding a single revision id 1 million times."""
|
|
74 |
self.time(self.encode_1M, _normal_revision_id) |
|
75 |
||
76 |
def test_encode_1_by_1M_ascii_cached(self): |
|
77 |
"""Test encoding a single revision id 1 million times."""
|
|
78 |
self.time(self.encode_cached_1M, _normal_revision_id) |
|
79 |
||
80 |
def test_encode_1_by_1M_ascii_str(self): |
|
81 |
# We have places that think they have a unicode revision id
|
|
82 |
# but actually, they have a plain string. So .encode(utf8)
|
|
83 |
# actually has to decode from ascii, and then encode into utf8
|
|
84 |
self.time(self.encode_1M, str(_normal_revision_id)) |
|
85 |
||
86 |
def test_encode_1_by_1M_ascii_str_cached(self): |
|
87 |
self.time(self.encode_cached_1M, str(_normal_revision_id)) |
|
88 |
||
89 |
def test_encode_1_by_1M_unicode(self): |
|
90 |
"""Test encoding a single revision id 1 million times."""
|
|
91 |
self.time(self.encode_1M, _unicode_revision_id) |
|
92 |
||
93 |
def test_encode_1_by_1M_unicode_cached(self): |
|
94 |
"""Test encoding a single revision id 1 million times."""
|
|
95 |
self.time(self.encode_cached_1M, _unicode_revision_id) |
|
96 |
||
97 |
def test_encode_1k_by_1k_ascii(self): |
|
98 |
"""Test encoding 5 revisions 100k times"""
|
|
99 |
revisions = [unicode(osutils.rand_chars(60)) for x in xrange(1000)] |
|
100 |
self.time(self.encode_multi, revisions, 1000) |
|
101 |
||
102 |
def test_encode_1k_by_1k_ascii_cached(self): |
|
103 |
"""Test encoding 5 revisions 100k times"""
|
|
104 |
revisions = [unicode(osutils.rand_chars(60)) for x in xrange(1000)] |
|
105 |
self.time(self.encode_cached_multi, revisions, 1000) |
|
106 |
||
107 |
def test_encode_1k_by_1k_unicode(self): |
|
108 |
"""Test encoding 5 revisions 100k times"""
|
|
|
2152.1.1
by John Arbash Meinel
(Dmitry Vasiliev) Update and improve benchmarks for cache_utf8 |
109 |
revisions = [u'\u062c\u0648\u062c\u0648' + |
|
1911.2.1
by John Arbash Meinel
Cache encode/decode operations, saves memory and time. Especially when committing a new kernel tree with 7.7M new lines to annotate |
110 |
unicode(osutils.rand_chars(60)) for x in xrange(1000)] |
111 |
self.time(self.encode_multi, revisions, 1000) |
|
112 |
||
113 |
def test_encode_1k_by_1k_unicode_cached(self): |
|
114 |
"""Test encoding 5 revisions 100k times"""
|
|
|
2152.1.1
by John Arbash Meinel
(Dmitry Vasiliev) Update and improve benchmarks for cache_utf8 |
115 |
revisions = [u'\u062c\u0648\u062c\u0648' + |
|
1911.2.1
by John Arbash Meinel
Cache encode/decode operations, saves memory and time. Especially when committing a new kernel tree with 7.7M new lines to annotate |
116 |
unicode(osutils.rand_chars(60)) for x in xrange(1000)] |
117 |
self.time(self.encode_cached_multi, revisions, 1000) |
|
118 |
||
|
2152.1.1
by John Arbash Meinel
(Dmitry Vasiliev) Update and improve benchmarks for cache_utf8 |
119 |
def test_encode_500K_by_1_ascii(self): |
120 |
revisions = [unicode("test%07d" % x) for x in xrange(500000)] |
|
121 |
self.time(self.encode_multi, revisions, 1) |
|
122 |
||
123 |
def test_encode_500K_by_1_ascii_cached(self): |
|
124 |
revisions = [unicode("test%07d" % x) for x in xrange(500000)] |
|
125 |
self.time(self.encode_cached_multi, revisions, 1) |
|
126 |
||
127 |
def test_encode_500K_by_1_unicode(self): |
|
128 |
revisions = [u'\u062c\u0648\u062c\u0648' + |
|
129 |
unicode("%07d" % x) for x in xrange(500000)] |
|
130 |
self.time(self.encode_multi, revisions, 1) |
|
131 |
||
132 |
def test_encode_500K_by_1_unicode_cached(self): |
|
133 |
revisions = [u'\u062c\u0648\u062c\u0648' + |
|
134 |
unicode("%07d" % x) for x in xrange(500000)] |
|
135 |
self.time(self.encode_cached_multi, revisions, 1) |
|
|
1911.2.1
by John Arbash Meinel
Cache encode/decode operations, saves memory and time. Especially when committing a new kernel tree with 7.7M new lines to annotate |
136 |
|
137 |
class DecodingBenchmarks(Benchmark): |
|
138 |
||
139 |
def setUp(self): |
|
140 |
super(DecodingBenchmarks, self).setUp() |
|
141 |
# Make sure we start and end with a clean cache
|
|
|
1911.2.3
by John Arbash Meinel
Moving everything into a new location so that we can cache more than just revision ids |
142 |
cache_utf8.clear_encoding_cache() |
143 |
self.addCleanup(cache_utf8.clear_encoding_cache) |
|
|
1911.2.1
by John Arbash Meinel
Cache encode/decode operations, saves memory and time. Especially when committing a new kernel tree with 7.7M new lines to annotate |
144 |
|
145 |
def decode_1M(self, revision_id): |
|
146 |
for i in xrange(1000000): |
|
147 |
revision_id.decode('utf8') |
|
148 |
||
149 |
def decode_cached_1M(self, revision_id): |
|
|
1911.2.3
by John Arbash Meinel
Moving everything into a new location so that we can cache more than just revision ids |
150 |
decode = cache_utf8.decode |
|
1911.2.1
by John Arbash Meinel
Cache encode/decode operations, saves memory and time. Especially when committing a new kernel tree with 7.7M new lines to annotate |
151 |
for i in xrange(1000000): |
|
1911.2.3
by John Arbash Meinel
Moving everything into a new location so that we can cache more than just revision ids |
152 |
decode(revision_id) |
|
1911.2.1
by John Arbash Meinel
Cache encode/decode operations, saves memory and time. Especially when committing a new kernel tree with 7.7M new lines to annotate |
153 |
|
154 |
def decode_multi(self, revision_list, count): |
|
155 |
for i in xrange(count): |
|
156 |
for revision_id in revision_list: |
|
157 |
revision_id.decode('utf-8') |
|
158 |
||
159 |
def decode_cached_multi(self, revision_list, count): |
|
|
1911.2.3
by John Arbash Meinel
Moving everything into a new location so that we can cache more than just revision ids |
160 |
decode = cache_utf8.decode |
|
1911.2.1
by John Arbash Meinel
Cache encode/decode operations, saves memory and time. Especially when committing a new kernel tree with 7.7M new lines to annotate |
161 |
for i in xrange(count): |
162 |
for revision_id in revision_list: |
|
|
1911.2.3
by John Arbash Meinel
Moving everything into a new location so that we can cache more than just revision ids |
163 |
decode(revision_id) |
|
1911.2.1
by John Arbash Meinel
Cache encode/decode operations, saves memory and time. Especially when committing a new kernel tree with 7.7M new lines to annotate |
164 |
|
165 |
def test_decode_1_by_1M_ascii(self): |
|
166 |
"""Test decoding a single revision id 1 million times."""
|
|
167 |
self.time(self.decode_1M, _normal_revision_id_utf8) |
|
168 |
||
169 |
def test_decode_1_by_1M_ascii_cached(self): |
|
170 |
"""Test decoding a single revision id 1 million times."""
|
|
171 |
self.time(self.decode_cached_1M, _normal_revision_id_utf8) |
|
172 |
||
173 |
def test_decode_1_by_1M_unicode(self): |
|
174 |
"""Test decoding a single revision id 1 million times."""
|
|
175 |
self.time(self.decode_1M, _unicode_revision_id_utf8) |
|
176 |
||
177 |
def test_decode_1_by_1M_unicode_cached(self): |
|
178 |
"""Test decoding a single revision id 1 million times."""
|
|
179 |
self.time(self.decode_cached_1M, _unicode_revision_id_utf8) |
|
180 |
||
181 |
def test_decode_1k_by_1k_ascii(self): |
|
182 |
"""Test decoding 5 revisions 100k times"""
|
|
183 |
revisions = [osutils.rand_chars(60) for x in xrange(1000)] |
|
184 |
self.time(self.decode_multi, revisions, 1000) |
|
185 |
||
186 |
def test_decode_1k_by_1k_ascii_cached(self): |
|
187 |
"""Test decoding 5 revisions 100k times"""
|
|
188 |
revisions = [osutils.rand_chars(60) for x in xrange(1000)] |
|
189 |
self.time(self.decode_cached_multi, revisions, 1000) |
|
190 |
||
191 |
def test_decode_1k_by_1k_unicode(self): |
|
192 |
"""Test decoding 5 revisions 100k times"""
|
|
|
2152.1.1
by John Arbash Meinel
(Dmitry Vasiliev) Update and improve benchmarks for cache_utf8 |
193 |
revisions = [(u'\u062c\u0648\u062c\u0648' + |
|
1911.2.1
by John Arbash Meinel
Cache encode/decode operations, saves memory and time. Especially when committing a new kernel tree with 7.7M new lines to annotate |
194 |
unicode(osutils.rand_chars(60))).encode('utf8') |
195 |
for x in xrange(1000)] |
|
196 |
self.time(self.decode_multi, revisions, 1000) |
|
197 |
||
198 |
def test_decode_1k_by_1k_unicode_cached(self): |
|
199 |
"""Test decoding 5 revisions 100k times"""
|
|
|
2152.1.1
by John Arbash Meinel
(Dmitry Vasiliev) Update and improve benchmarks for cache_utf8 |
200 |
revisions = [(u'\u062c\u0648\u062c\u0648' + |
|
1911.2.1
by John Arbash Meinel
Cache encode/decode operations, saves memory and time. Especially when committing a new kernel tree with 7.7M new lines to annotate |
201 |
unicode(osutils.rand_chars(60))).encode('utf8') |
202 |
for x in xrange(1000)] |
|
203 |
self.time(self.decode_cached_multi, revisions, 1000) |
|
|
2152.1.1
by John Arbash Meinel
(Dmitry Vasiliev) Update and improve benchmarks for cache_utf8 |
204 |
|
205 |
def test_decode_500K_by_1_ascii(self): |
|
206 |
revisions = [("test%07d" % x) for x in xrange(500000)] |
|
207 |
self.time(self.decode_multi, revisions, 1) |
|
208 |
||
209 |
def test_decode_500K_by_1_ascii_cached(self): |
|
210 |
revisions = [("test%07d" % x) for x in xrange(500000)] |
|
211 |
self.time(self.decode_cached_multi, revisions, 1) |
|
212 |
||
213 |
def test_decode_500K_by_1_unicode(self): |
|
214 |
revisions = [(u'\u062c\u0648\u062c\u0648' + |
|
215 |
unicode("%07d" % x)).encode('utf-8') |
|
216 |
for x in xrange(500000)] |
|
217 |
self.time(self.decode_multi, revisions, 1) |
|
218 |
||
219 |
def test_decode_500K_by_1_unicode_cached(self): |
|
220 |
revisions = [(u'\u062c\u0648\u062c\u0648' + |
|
221 |
unicode("%07d" % x)).encode('utf-8') |
|
222 |
for x in xrange(500000)] |
|
223 |
self.time(self.decode_cached_multi, revisions, 1) |