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
|
#!/usr/bin/env python3
"""
Copyright (c) 2015, Intel Corporation
Redistribution and use in source and binary forms, with or without
modification, are permitted provided that the following conditions are met:
* Redistributions of source code must retain the above copyright notice,
this list of conditions and the following disclaimer.
* Redistributions in binary form must reproduce the above copyright
notice, this list of conditions and the following disclaimer in the
documentation and/or other materials provided with the distribution.
* Neither the name of Intel Corporation nor the names of its contributors
may be used to endorse or promote products derived from this software
without specific prior written permission.
THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE
FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR
SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER
CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY,
OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
"""
from ez_ut import *
import argparse
import datetime
import pygit2
import time
import sys
import os
# Import ezbench from the utils/ folder
ezbench_dir = os.path.dirname(os.path.dirname(os.path.abspath(__file__)))
sys.path.append(os.path.join(ezbench_dir, 'python-modules'))
from ezbench.smartezbench import *
# parse the options
parser = argparse.ArgumentParser()
parser.add_argument("-r", dest='reuse_data', action="store_true",
help="Do not reset the state")
args = parser.parse_args()
def create_new_branch(repo, base_commit, branch_name, template, state, max_commits):
repo.create_branch(branch_name, base_commit)
ref = 'refs/heads/'+branch_name
last_commit = None
for i in range(0, max_commits):
# Treat every variable independently
changes = False
if random.random() > 0.98:
state['perf'] = int(state['perf'] * (1.0075 + np.random.randn() / 10))
changes = True
if not state['build_broken'] and random.random() > 0.995:
state['build_broken'] = True
changes = True
elif state['build_broken'] and random.random() > 0.75:
state['build_broken'] = False
changes = True
if not state['exec_broken'] and random.random() > 0.995:
state['exec_broken'] = True
changes = True
elif state['exec_broken'] and random.random() > 0.75:
state['exec_broken'] = False
changes = True
if not changes:
state['noise_commit'] += 1
last_commit = create_new_commit(repo, ref, template, state)
return last_commit
# Generate the git history
repo = pygit2.Repository(perf_bisect_repo_dir())
initial_commit = next(repo.walk(repo.revparse_single('HEAD').oid, pygit2.GIT_SORT_REVERSE))
if not args.reuse_data:
random.seed(42)
repo.checkout(repo.lookup_reference('refs/heads/master'))
for branch in repo.listall_branches(pygit2.GIT_BRANCH_LOCAL):
if branch.startswith("tmp_"):
repo.lookup_branch(branch).delete() # Clean all the tmp branches we create
branch_name = datetime.now().strftime('tmp_%Y-%m-%d_%H-%M-%S')
template = repo[initial_commit.tree['perf.py'].id].data
# Initial state
state = dict()
state['perf'] = 100
state['sample_distribution_mode'] = 0
state['variance'] = 1
state['build_broken'] = False
state['exec_broken'] = False
state['noise_commit'] = 0
# implement a state machine
HEAD = create_new_branch(repo, initial_commit, branch_name,
template, copy.deepcopy(state), 1000)
repo.checkout('refs/heads/'+branch_name)
# List the events
max_variance = 0.025
l = list(repo.walk(repo.revparse_single('HEAD').oid, pygit2.GIT_SORT_REVERSE))
initial_commit = l[1]
report = create_report(initial_commit.hex,
(repo.revparse_single('HEAD')).hex, max_variance,
args.reuse_data)
# test the variance of every commit
variance_too_high = 0
sample_error = []
for commit in report.commits:
# FIXME: instead of skipping this commit, replace the template commit by the
# actual first sample!
if commit.full_name.endswith("template"):
continue
for result in commit.results.values():
# Check that the results are for the right test
if result.test.full_name != 'perf_bisect':
continue
e_perf, e_var, e_build, e_exec = commit_info(commit)
# if the commit was expected to fail running, force the performance to 0
if e_build or e_exec:
e_perf = 0
if e_perf > 0:
perf = result.result().mean()
error = abs(e_perf - perf) * 100.0 / e_perf
else:
error = 0
sample_error.append(error)
if error > max_variance * 100:
msg = "Commit {}'s performance differs significantly from the target, {} vs {}"
print(msg.format(commit.sha1, perf, e_perf))
variance_too_high += 1
false_positive = 0
relative_error = []
for e in report.events:
if type(e) is EventPerfChange:
if result.test.full_name != 'perf_bisect':
continue
o_perf, o_variance, o_build, o_exec = commit_info(e.commit_range.old)
n_perf, n_variance, n_build, n_exec = commit_info(e.commit_range.new)
wanted_diff = compute_perf_difference("fps", o_perf, n_perf) - 100
if (not o_exec and n_exec) or (o_exec and not n_exec):
if e.diff() != -1 and e.diff() != float("+inf"):
print("False positive: {} => real perf change was {} to {}".format(e, o_perf, n_perf))
false_positive += 1
else:
rel_error = abs(wanted_diff - e.diff() * 100.0)
relative_error.append(rel_error)
if o_perf == n_perf or rel_error > max_variance * 100.0:
print("False positive: {} => real perf change was {} to {}".format(e, o_perf, n_perf))
false_positive += 1
total_changes = 0
false_negatives = 0
p_perf = 100
p_build = False
p_exec = False
r = 0
HEAD = repo.revparse_single('HEAD')
for commit in repo.walk(HEAD.oid, pygit2.GIT_SORT_REVERSE):
# Skip the first commit
if commit.message == "template\n":
continue
c_perf, c_var, c_build, c_exec = parse_commit_title(commit.message)
if p_perf != c_perf:
e = find_event_from_commit(repo, report, EventPerfChange, commit.hex)
if not e:
print("False negative: commit {} changed perf from {} to {}".format(commit.hex, p_perf, c_perf))
false_negatives += 1
total_changes += 1
#if (not p_exec and c_perf) or (p_exec and not c_perf):
# Copy the new values to the previous ones before iterating
p_perf = c_perf
p_build = c_build
p_exec = c_exec
print("Stats (max error wanted = {:.2f}%):".format(max_variance * 100.0))
print(" Average sampling error: {}".format(do_stats(sample_error, '%')))
print(" Average perf. rel. error: {}".format(do_stats(relative_error, '%')))
print("")
print("Tests:")
print(" Too large variance: {} / {}".format(variance_too_high, len(sample_error)))
print(" False positives: {} / {}".format(false_positive, len(report.events)))
print(" False negatives: {} / {}".format(false_negatives, total_changes))
sys.exit(variance_too_high == 0 and false_positive == 0)
|
