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
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
|
"""Script for testing the performance of pickling/unpickling.
This will pickle/unpickle several real world-representative objects a few
thousand times. The methodology below was chosen for was chosen to be similar
to real-world scenarios which operate on single objects at a time. Note that if
we did something like
pickle.dumps([dict(some_dict) for _ in range(10000)])
this isn't equivalent to dumping the dict 10000 times: pickle uses a
highly-efficient encoding for the n-1 following copies.
"""
import datetime
import random
import sys
import pyperf
from memray_helper import get_tracker
IS_PYPY = pyperf.python_implementation() == "pypy"
__author__ = "collinwinter@google.com (Collin Winter)"
DICT = {
"ads_flags": 0,
"age": 18,
"birthday": datetime.date(1980, 5, 7),
"bulletin_count": 0,
"comment_count": 0,
"country": "BR",
"encrypted_id": "G9urXXAJwjE",
"favorite_count": 9,
"first_name": "",
"flags": 412317970704,
"friend_count": 0,
"gender": "m",
"gender_for_display": "Male",
"id": 302935349,
"is_custom_profile_icon": 0,
"last_name": "",
"locale_preference": "pt_BR",
"member": 0,
"tags": ["a", "b", "c", "d", "e", "f", "g"],
"profile_foo_id": 827119638,
"secure_encrypted_id": "Z_xxx2dYx3t4YAdnmfgyKw",
"session_number": 2,
"signup_id": "201-19225-223",
"status": "A",
"theme": 1,
"time_created": 1225237014,
"time_updated": 1233134493,
"unread_message_count": 0,
"user_group": "0",
"username": "collinwinter",
"play_count": 9,
"view_count": 7,
"zip": "",
}
TUPLE = (
[
265867233,
265868503,
265252341,
265243910,
265879514,
266219766,
266021701,
265843726,
265592821,
265246784,
265853180,
45526486,
265463699,
265848143,
265863062,
265392591,
265877490,
265823665,
265828884,
265753032,
],
60,
)
def mutate_dict(orig_dict, random_source):
new_dict = dict(orig_dict)
for key, value in new_dict.items():
rand_val = random_source.random() * sys.maxsize
if isinstance(key, (int, bytes, str)):
new_dict[key] = type(key)(rand_val)
return new_dict
random_source = random.Random(5) # Fixed seed.
DICT_GROUP = [mutate_dict(DICT, random_source) for _ in range(3)]
def bench_pickle(loops, pickle, options):
range_it = range(loops)
# micro-optimization: use fast local variables
dumps = pickle.dumps
objs = (DICT, TUPLE, DICT_GROUP)
protocol = options.protocol
with get_tracker():
t0 = pyperf.perf_counter()
for _ in range_it:
for obj in objs:
# 20 dumps
dumps(obj, protocol)
dumps(obj, protocol)
dumps(obj, protocol)
dumps(obj, protocol)
dumps(obj, protocol)
dumps(obj, protocol)
dumps(obj, protocol)
dumps(obj, protocol)
dumps(obj, protocol)
dumps(obj, protocol)
dumps(obj, protocol)
dumps(obj, protocol)
dumps(obj, protocol)
dumps(obj, protocol)
dumps(obj, protocol)
dumps(obj, protocol)
dumps(obj, protocol)
dumps(obj, protocol)
dumps(obj, protocol)
dumps(obj, protocol)
return pyperf.perf_counter() - t0
def bench_unpickle(loops, pickle, options):
pickled_dict = pickle.dumps(DICT, options.protocol)
pickled_tuple = pickle.dumps(TUPLE, options.protocol)
pickled_dict_group = pickle.dumps(DICT_GROUP, options.protocol)
range_it = range(loops)
# micro-optimization: use fast local variables
loads = pickle.loads
objs = (pickled_dict, pickled_tuple, pickled_dict_group)
with get_tracker():
t0 = pyperf.perf_counter()
for _ in range_it:
for obj in objs:
# 20 loads dict
loads(obj)
loads(obj)
loads(obj)
loads(obj)
loads(obj)
loads(obj)
loads(obj)
loads(obj)
loads(obj)
loads(obj)
loads(obj)
loads(obj)
loads(obj)
loads(obj)
loads(obj)
loads(obj)
loads(obj)
loads(obj)
loads(obj)
loads(obj)
return pyperf.perf_counter() - t0
LIST = [[list(range(10)), list(range(10))] for _ in range(10)]
def bench_pickle_list(loops, pickle, options):
range_it = range(loops)
# micro-optimization: use fast local variables
dumps = pickle.dumps
obj = LIST
protocol = options.protocol
with get_tracker():
t0 = pyperf.perf_counter()
for _ in range_it:
# 10 dumps list
dumps(obj, protocol)
dumps(obj, protocol)
dumps(obj, protocol)
dumps(obj, protocol)
dumps(obj, protocol)
dumps(obj, protocol)
dumps(obj, protocol)
dumps(obj, protocol)
dumps(obj, protocol)
dumps(obj, protocol)
return pyperf.perf_counter() - t0
def bench_unpickle_list(loops, pickle, options):
pickled_list = pickle.dumps(LIST, options.protocol)
range_it = range(loops)
# micro-optimization: use fast local variables
loads = pickle.loads
with get_tracker():
t0 = pyperf.perf_counter()
for _ in range_it:
# 10 loads list
loads(pickled_list)
loads(pickled_list)
loads(pickled_list)
loads(pickled_list)
loads(pickled_list)
loads(pickled_list)
loads(pickled_list)
loads(pickled_list)
loads(pickled_list)
loads(pickled_list)
return pyperf.perf_counter() - t0
MICRO_DICT = dict((key, dict.fromkeys(range(10))) for key in range(100))
def bench_pickle_dict(loops, pickle, options):
range_it = range(loops)
# micro-optimization: use fast local variables
protocol = options.protocol
obj = MICRO_DICT
with get_tracker():
t0 = pyperf.perf_counter()
for _ in range_it:
# 5 dumps dict
pickle.dumps(obj, protocol)
pickle.dumps(obj, protocol)
pickle.dumps(obj, protocol)
pickle.dumps(obj, protocol)
pickle.dumps(obj, protocol)
return pyperf.perf_counter() - t0
BENCHMARKS = {
# 20 inner-loops: don't count the 3 pickled objects
"pickle": (bench_pickle, 20),
# 20 inner-loops: don't count the 3 unpickled objects
"unpickle": (bench_unpickle, 20),
"pickle_list": (bench_pickle_list, 10),
"unpickle_list": (bench_unpickle_list, 10),
"pickle_dict": (bench_pickle_dict, 5),
}
def is_accelerated_module(module):
return getattr(module.Pickler, "__module__", "<jython>") != "pickle"
def add_cmdline_args(cmd, args):
if args.pure_python:
cmd.append("--pure-python")
cmd.extend(("--protocol", str(args.protocol)))
cmd.append(args.benchmark)
if __name__ == "__main__":
runner = pyperf.Runner(add_cmdline_args=add_cmdline_args)
runner.metadata["description"] = "Test the performance of pickling."
parser = runner.argparser
parser.add_argument(
"--pure-python", action="store_true", help="Use the C version of pickle."
)
parser.add_argument(
"--protocol",
action="store",
default=None,
type=int,
help="Which protocol to use (default: highest protocol).",
)
benchmarks = sorted(BENCHMARKS)
parser.add_argument("benchmark", choices=benchmarks)
options = runner.parse_args()
benchmark, inner_loops = BENCHMARKS[options.benchmark]
name = options.benchmark
if options.pure_python:
name += "_pure_python"
if not (options.pure_python or IS_PYPY):
# C accelerators are enabled by default on 3.x
import pickle
if not is_accelerated_module(pickle):
raise RuntimeError("Missing C accelerators for pickle")
else:
sys.modules["_pickle"] = None
import pickle
if is_accelerated_module(pickle):
raise RuntimeError("Unexpected C accelerators for pickle")
if options.protocol is None:
options.protocol = pickle.HIGHEST_PROTOCOL
runner.metadata["pickle_protocol"] = str(options.protocol)
runner.metadata["pickle_module"] = pickle.__name__
runner.bench_time_func(name, benchmark, pickle, options, inner_loops=inner_loops)
|
