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
|
import mmap
import os
import tempfile
from memray import AllocatorType
from memray import FileReader
try:
from memray._test import MemoryAllocator
except ImportError:
from memray import MemoryAllocator
from memray import Tracker
from .benchmarking.cases import async_tree_base
from .benchmarking.cases import fannkuch_base
from .benchmarking.cases import mdp_base
from .benchmarking.cases import pprint_format_base
from .benchmarking.cases import raytrace_base
LOOPS = 1000
class TracebackBenchmarks:
def setup(self):
self.tempfile = tempfile.NamedTemporaryFile()
allocator = MemoryAllocator()
def fac(n):
if n == 1:
allocator.valloc(1234)
allocator.free()
return 1
return n * fac(n - 1)
os.unlink(self.tempfile.name)
with Tracker(self.tempfile.name):
fac(300)
(self.record,) = [
record
for record in FileReader(self.tempfile.name).get_allocation_records()
if record.allocator == AllocatorType.VALLOC
]
def time_get_stack_trace(self):
self.record.stack_trace()
class AllocatorBenchmarks:
def setup(self):
self.tempfile = tempfile.NamedTemporaryFile()
self.allocator = MemoryAllocator()
def time_malloc(self):
os.unlink(self.tempfile.name)
with Tracker(self.tempfile.name):
for _ in range(LOOPS):
if self.allocator.malloc(1234):
self.allocator.free()
def time_posix_memalign(self):
os.unlink(self.tempfile.name)
with Tracker(self.tempfile.name):
for _ in range(LOOPS):
if self.allocator.posix_memalign(1234):
self.allocator.free()
def time_posix_realloc(self):
os.unlink(self.tempfile.name)
with Tracker(self.tempfile.name):
for _ in range(LOOPS):
if self.allocator.posix_memalign(1234):
self.allocator.free()
def time_calloc(self):
os.unlink(self.tempfile.name)
with Tracker(self.tempfile.name):
for _ in range(LOOPS):
if self.allocator.calloc(1234):
self.allocator.free()
def time_pvalloc(self):
os.unlink(self.tempfile.name)
with Tracker(self.tempfile.name):
for _ in range(LOOPS):
if self.allocator.pvalloc(1234):
self.allocator.free()
def time_valloc(self):
os.unlink(self.tempfile.name)
with Tracker(self.tempfile.name):
for _ in range(LOOPS):
if self.allocator.valloc(1234):
self.allocator.free()
def time_realloc(self):
os.unlink(self.tempfile.name)
with Tracker(self.tempfile.name):
for _ in range(LOOPS):
if self.allocator.realloc(1234):
self.allocator.free()
def time_mmap(self):
os.unlink(self.tempfile.name)
with Tracker(self.tempfile.name):
for _ in range(LOOPS):
with mmap.mmap(-1, length=2048, access=mmap.ACCESS_WRITE) as mmap_obj:
mmap_obj[0:100] = b"a" * 100
class ParserBenchmarks:
def setup(self):
self.tempfile = tempfile.NamedTemporaryFile()
self.allocator = MemoryAllocator()
os.unlink(self.tempfile.name)
self.tracker = Tracker(self.tempfile.name)
with self.tracker:
for _ in range(LOOPS):
self.allocator.valloc(1234)
self.allocator.free()
def time_end_to_end_parsing(self):
list(FileReader(self.tempfile.name).get_allocation_records())
def recursive(n, chunk_size):
"""Mimics generally-increasing but spiky usage"""
if not n:
return
allocator = MemoryAllocator()
allocator.valloc(n * chunk_size)
# Don't keep allocated memory when recursing, ~50% of the calls.
if n % 2:
allocator.free()
recursive(n - 1, chunk_size)
else:
recursive(n - 1, chunk_size)
allocator.free()
class HighWatermarkBenchmarks:
def setup(self):
self.tempfile = tempfile.NamedTemporaryFile()
os.unlink(self.tempfile.name)
self.tracker = Tracker(self.tempfile.name)
with self.tracker:
recursive(700, 99)
def time_high_watermark(self):
list(
FileReader(self.tempfile.name).get_high_watermark_allocation_records(
merge_threads=False
)
)
class MacroBenchmarksBase:
def __init_subclass__(cls) -> None:
for name in dir(cls):
if name.startswith("bench_"):
bench_name = name[len("bench_") :]
setattr(cls, f"time_{bench_name}", getattr(cls, name))
def bench_async_tree_cpu(self):
with self.tracker:
async_tree_base.run_benchmark("none")
def bench_async_tree_io(self):
with self.tracker:
async_tree_base.run_benchmark("io")
def bench_async_tree_memoization(self):
with self.tracker:
async_tree_base.run_benchmark("memoization")
def bench_async_tree_cpu_io_mixed(self):
with self.tracker:
async_tree_base.run_benchmark("cpu_io_mixed")
def bench_fannkuch(self):
with self.tracker:
fannkuch_base.run_benchmark()
def bench_mdp(self):
with self.tracker:
mdp_base.run_benchmark()
def bench_pprint_format(self):
with self.tracker:
pprint_format_base.run_benchmark()
def bench_raytrace(self):
with self.tracker:
raytrace_base.run_benchmark()
class MacroBenchmarksDefault(MacroBenchmarksBase):
def setup(self):
self.tracker_args = ("/dev/null",)
self.tracker_kwargs = {}
self.tracker = Tracker(*self.tracker_args, **self.tracker_kwargs)
class MacroBenchmarksPythonAllocators(MacroBenchmarksBase):
def setup(self):
self.tracker_args = ("/dev/null",)
self.tracker_kwargs = {"trace_python_allocators": True}
self.tracker = Tracker(*self.tracker_args, **self.tracker_kwargs)
class MacroBenchmarksPythonNative(MacroBenchmarksBase):
def setup(self):
self.tracker_args = ("/dev/null",)
self.tracker_kwargs = {"native_traces": True}
self.tracker = Tracker(*self.tracker_args, **self.tracker_kwargs)
class MacroBenchmarksPythonAll(MacroBenchmarksBase):
def setup(self):
self.tracker_args = ("/dev/null",)
self.tracker_kwargs = {"native_traces": True, "trace_python_allocators": True}
self.tracker = Tracker(*self.tracker_args, **self.tracker_kwargs)
class FileSizeBenchmarks:
def setup(self):
self.tempfile = tempfile.NamedTemporaryFile()
os.unlink(self.tempfile.name)
self.tracker = Tracker(self.tempfile.name)
def track_async_tree_cpu(self):
with self.tracker:
async_tree_base.run_benchmark("none")
return os.stat(self.tempfile.name).st_size
def track_async_tree_io(self):
with self.tracker:
async_tree_base.run_benchmark("io")
return os.stat(self.tempfile.name).st_size
def track_async_tree_memoization(self):
with self.tracker:
async_tree_base.run_benchmark("memoization")
return os.stat(self.tempfile.name).st_size
def track_async_tree_cpu_io_mixed(self):
with self.tracker:
async_tree_base.run_benchmark("cpu_io_mixed")
return os.stat(self.tempfile.name).st_size
def track_fannkuch(self):
with self.tracker:
fannkuch_base.run_benchmark()
return os.stat(self.tempfile.name).st_size
def track_mdp(self):
with self.tracker:
mdp_base.run_benchmark()
return os.stat(self.tempfile.name).st_size
def track_pprint_format(self):
with self.tracker:
pprint_format_base.run_benchmark()
return os.stat(self.tempfile.name).st_size
def track_raytrace(self):
with self.tracker:
raytrace_base.run_benchmark()
return os.stat(self.tempfile.name).st_size
|
