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
|
"""
Generate a torchbench test report from a file containing the PR body.
Currently, only supports running tests on specified model names
Testing environment:
- Intel Xeon 8259CL @ 2.50 GHz, 24 Cores with disabled Turbo and HT
- Nvidia Tesla T4
- Nvidia Driver 470.82.01
- Python 3.8
- CUDA 11.3
"""
# Known issues:
# 1. Does not reuse the build artifact in other CI workflows
# 2. CI jobs are serialized because there is only one worker
import os
import boto3 # type: ignore[import]
import git # type: ignore[import]
import pathlib
import argparse
import subprocess
from pathlib import Path
from typing import List, Tuple
TORCHBENCH_CONFIG_NAME = "config.yaml"
TORCHBENCH_USERBENCHMARK_CONFIG_NAME = "ub-config.yaml"
MAGIC_PREFIX = "RUN_TORCHBENCH:"
MAGIC_TORCHBENCH_PREFIX = "TORCHBENCH_BRANCH:"
ABTEST_CONFIG_TEMPLATE = """# This config is automatically generated by run_torchbench.py
start: {control}
end: {treatment}
threshold: 100
direction: decrease
timeout: 720
tests:"""
S3_BUCKET = "ossci-metrics"
S3_PREFIX = "torchbench-pr-test"
S3_URL_BASE = f"https://{S3_BUCKET}.s3.amazonaws.com/"
class S3Client:
def __init__(self, bucket: str = S3_BUCKET, prefix: str = S3_PREFIX):
self.s3 = boto3.client('s3')
self.resource = boto3.resource('s3')
self.bucket = bucket
self.prefix = prefix
def upload_file(self, file_path: Path, filekey_prefix: str) -> None:
assert file_path.is_file(), f"Specified file path {file_path} does not exist or not file."
file_name = file_path.name
s3_key = f"{self.prefix}/{filekey_prefix}/{file_name}"
print(f"Uploading file {file_name} to S3 with key: {s3_key}")
self.s3.upload_file(str(file_path), self.bucket, s3_key)
# output the result URL
print(f"Uploaded the result file {file_name} to {S3_URL_BASE}{s3_key}")
def gen_abtest_config(control: str, treatment: str, models: List[str]) -> str:
d = {}
d["control"] = control
d["treatment"] = treatment
config = ABTEST_CONFIG_TEMPLATE.format(**d)
if models == ["ALL"]:
return config + "\n"
for model in models:
config = f"{config}\n - {model}"
config = config + "\n"
return config
def setup_gha_env(name: str, val: str) -> None:
fname = os.environ["GITHUB_ENV"]
content = f"{name}={val}\n"
with open(fname, "a") as fo:
fo.write(content)
def find_current_branch(repo_path: str) -> str:
repo = git.Repo(repo_path)
name: str = repo.active_branch.name
return name
def deploy_torchbench_config(output_dir: str, config: str, config_name: str = TORCHBENCH_CONFIG_NAME) -> None:
# Create test dir if needed
pathlib.Path(output_dir).mkdir(exist_ok=True)
# TorchBench config file name
config_path = os.path.join(output_dir, config_name)
with open(config_path, "w") as fp:
fp.write(config)
def get_valid_models(torchbench_path: str) -> List[str]:
benchmark_path = os.path.join(torchbench_path, "torchbenchmark", "models")
valid_models = [model for model in os.listdir(benchmark_path) if os.path.isdir(os.path.join(benchmark_path, model))]
return valid_models
def get_valid_userbenchmarks(torchbench_path: str) -> List[str]:
def is_valid_ub_dir(ub_path: str) -> bool:
return os.path.isdir(ub_path) and os.path.exists(os.path.join(ub_path, "__init__.py"))
ub_path = os.path.join(os.path.abspath(torchbench_path), "userbenchmark")
ubs = list(filter(is_valid_ub_dir, [os.path.join(ub_path, ubdir) for ubdir in os.listdir(ub_path)]))
valid_ubs = list(map(lambda x: os.path.basename(x), ubs))
return valid_ubs
def extract_models_from_pr(torchbench_path: str, prbody_file: str) -> Tuple[List[str], List[str]]:
model_list = []
userbenchmark_list = []
pr_list = []
with open(prbody_file, "r") as pf:
lines = map(lambda x: x.strip(), pf.read().splitlines())
magic_lines = list(filter(lambda x: x.startswith(MAGIC_PREFIX), lines))
if magic_lines:
# Only the first magic line will be recognized.
pr_list = list(map(lambda x: x.strip(), magic_lines[0][len(MAGIC_PREFIX):].split(",")))
valid_models = get_valid_models(torchbench_path)
valid_ubs = get_valid_userbenchmarks(torchbench_path)
for pr_bm in pr_list:
if pr_bm in valid_models or pr_bm == "ALL":
model_list.append(pr_bm)
elif pr_bm in valid_ubs:
userbenchmark_list.append(pr_bm)
else:
print(f"The model or benchmark {pr_bm} you specified does not exist in TorchBench suite. Please double check.")
exit(-1)
# Shortcut: if pr_list is ["ALL"], run all the model tests
if "ALL" in model_list:
model_list = ["ALL"]
return model_list, userbenchmark_list
def find_torchbench_branch(prbody_file: str) -> str:
branch_name: str = ""
with open(prbody_file, "r") as pf:
lines = map(lambda x: x.strip(), pf.read().splitlines())
magic_lines = list(filter(lambda x: x.startswith(MAGIC_TORCHBENCH_PREFIX), lines))
if magic_lines:
# Only the first magic line will be recognized.
branch_name = magic_lines[0][len(MAGIC_TORCHBENCH_PREFIX):].strip()
# If not specified, use main as the default branch
if not branch_name:
branch_name = "main"
return branch_name
def run_torchbench(pytorch_path: str, torchbench_path: str, output_dir: str) -> None:
# Copy system environment so that we will not override
env = dict(os.environ)
command = ["python", "bisection.py", "--work-dir", output_dir,
"--pytorch-src", pytorch_path, "--torchbench-src", torchbench_path,
"--config", os.path.join(output_dir, TORCHBENCH_CONFIG_NAME),
"--output", os.path.join(output_dir, "result.txt")]
print(f"Running torchbench command: {command}")
subprocess.check_call(command, cwd=torchbench_path, env=env)
def run_userbenchmarks(pytorch_path: str, torchbench_path: str, base_sha: str, head_sha: str,
userbenchmark: str, output_dir: str) -> None:
# Copy system environment so that we will not override
env = dict(os.environ)
command = ["python", "./.github/scripts/abtest.py",
"--pytorch-repo", pytorch_path,
"--base", base_sha,
"--head", head_sha,
"--userbenchmark", userbenchmark,
"--output-dir", output_dir]
print(f"Running torchbench userbenchmark command: {command}")
subprocess.check_call(command, cwd=torchbench_path, env=env)
def process_upload_s3(result_dir: str) -> None:
# validate result directory
result_dir_path = Path(result_dir)
assert result_dir_path.exists(), f"Specified result directory {result_dir} doesn't exist."
# upload all files to S3 bucket oss-ci-metrics
files = [x for x in result_dir_path.iterdir() if x.is_file()]
# upload file to S3 bucket
s3_client: S3Client = S3Client()
filekey_prefix = result_dir_path.name
for f in files:
s3_client.upload_file(f, filekey_prefix)
if __name__ == "__main__":
parser = argparse.ArgumentParser(description='Run TorchBench tests based on PR')
parser.add_argument('--pr-body', help="The file that contains body of a Pull Request")
subparsers = parser.add_subparsers(dest='command')
# parser for setup the torchbench branch name env
branch_parser = subparsers.add_parser("set-torchbench-branch")
# parser to run the torchbench branch
run_parser = subparsers.add_parser("run")
run_parser.add_argument('--pr-num', required=True, type=str, help="The Pull Request number")
run_parser.add_argument('--pr-base-sha', required=True, type=str, help="The Pull Request base hash")
run_parser.add_argument('--pr-head-sha', required=True, type=str, help="The Pull Request head hash")
run_parser.add_argument('--pytorch-path', required=True, type=str, help="Path to pytorch repository")
run_parser.add_argument('--torchbench-path', required=True, type=str, help="Path to TorchBench repository")
# parser to upload results to S3
upload_parser = subparsers.add_parser("upload-s3")
upload_parser.add_argument('--result-dir', required=True, type=str, help="Path to benchmark output")
args = parser.parse_args()
if args.command == 'set-torchbench-branch':
branch_name = find_torchbench_branch(args.pr_body)
# env name: "TORCHBENCH_BRANCH"
setup_gha_env(MAGIC_TORCHBENCH_PREFIX[:-1], branch_name)
elif args.command == 'run':
output_dir: str = os.path.join(os.environ["HOME"], ".torchbench", "bisection", f"pr{args.pr_num}")
# Assert the current branch in args.torchbench_path is the same as the one specified in pr body
branch_name = find_torchbench_branch(args.pr_body)
current_branch = find_current_branch(args.torchbench_path)
assert branch_name == current_branch, f"Torchbench repo {args.torchbench_path} is on branch {current_branch}, \
but user specified to run on branch {branch_name}."
print(f"Ready to run TorchBench with benchmark. Result will be saved in the directory: {output_dir}.")
# Identify the specified models and userbenchmarks
models, userbenchmarks = extract_models_from_pr(args.torchbench_path, args.pr_body)
if models:
torchbench_config = gen_abtest_config(args.pr_base_sha, args.pr_head_sha, models)
deploy_torchbench_config(output_dir, torchbench_config)
run_torchbench(pytorch_path=args.pytorch_path, torchbench_path=args.torchbench_path, output_dir=output_dir)
if userbenchmarks:
assert len(userbenchmarks) == 1, \
"We don't support running multiple userbenchmarks in single workflow yet." \
"If you need, please submit a feature request."
run_userbenchmarks(pytorch_path=args.pytorch_path, torchbench_path=args.torchbench_path,
base_sha=args.pr_base_sha, head_sha=args.pr_head_sha,
userbenchmark=userbenchmarks[0], output_dir=output_dir)
if not models and not userbenchmarks:
print("Can't parse valid models or userbenchmarks from the pr body. Quit.")
exit(-1)
elif args.command == 'upload-s3':
process_upload_s3(args.result_dir)
else:
print(f"The command {args.command} is not supported.")
exit(-1)
|
