forked from apache/systemds
-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathsystemml-spark-submit.py
executable file
·162 lines (137 loc) · 6.62 KB
/
systemml-spark-submit.py
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
#!/usr/bin/env python
# -------------------------------------------------------------
#
# Licensed to the Apache Software Foundation (ASF) under one
# or more contributor license agreements. See the NOTICE file
# distributed with this work for additional information
# regarding copyright ownership. The ASF licenses this file
# to you under the Apache License, Version 2.0 (the
# "License"); you may not use this file except in compliance
# with the License. You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing,
# software distributed under the License is distributed on an
# "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
# KIND, either express or implied. See the License for the
# specific language governing permissions and limitations
# under the License.
#
# -------------------------------------------------------------
import os
import glob
from os.path import join
import platform
import argparse
from utils import get_env_systemml_home, get_env_spark_home, find_dml_file, log4j_path, config_path
def default_jars(systemml_home):
"""
return: String
Location of systemml and jcuda jars
"""
build_dir = join(systemml_home, 'target')
lib_dir = join(build_dir, 'lib')
systemml_jar = build_dir + os.sep + "SystemML.jar"
jcuda_jars = glob.glob(lib_dir + os.sep + "jcu*.jar")
target_jars = ','.join(jcuda_jars)
return target_jars, systemml_jar
def spark_submit_entry(master, deploy_mode, driver_memory, num_executors,
executor_memory, executor_cores, conf,
nvargs, args, config, explain, debug, stats, gpu, f):
"""
This function is responsible for the execution of arguments via
subprocess call in hybrid_spark mode
"""
spark_home = get_env_spark_home()
systemml_home = get_env_systemml_home()
spark_path = join(spark_home, 'bin', 'spark-submit')
script_file = find_dml_file(systemml_home, f)
# Jars
cuda_jars, systemml_jars = default_jars(systemml_home)
# Log4j
log4j = log4j_path(systemml_home)
log4j_properties_path = 'spark.driver.extraJavaOptions=-Dlog4j.configuration=file:{}'.format(log4j)
if conf is None:
default_conf = log4j_properties_path
else:
default_conf = ' --conf '.join(conf + [log4j_properties_path])
# Config XML
if config is None:
default_config = config_path(systemml_home)
else:
default_config = ' -config '.join([config] + [config_path(systemml_home)])
if platform.system() == 'Windows':
default_conf = default_conf.replace('\\', '//')
# optional arguments
ml_options = []
if nvargs is not None:
ml_options.append('-nvargs')
ml_options.append(' '.join(nvargs))
if args is not None:
ml_options.append('-args')
ml_options.append(' '.join(args))
if explain is not None:
ml_options.append('-explain')
ml_options.append(explain)
if debug is not False:
ml_options.append('-debug')
if stats is not None:
ml_options.append('-stats')
ml_options.append(stats)
if gpu is not None:
ml_options.append('-gpu')
if gpu is not 'no_option':
ml_options.append(gpu)
if len(ml_options) < 1:
ml_options = ''
# stats, explain, target_jars
cmd_spark = [spark_path, '--class', 'org.apache.sysml.api.DMLScript',
'--master', master, '--deploy-mode', deploy_mode,
'--driver-memory', driver_memory,
'--conf', default_conf,
'--jars', cuda_jars, systemml_jars]
if num_executors is not None:
cmd_spark = cmd_spark + ['--num-executors', num_executors]
if executor_memory is not None:
cmd_spark = cmd_spark + ['--executor-memory', executor_memory]
if executor_cores is not None:
cmd_spark = cmd_spark + ['--executor-cores', executor_cores]
cmd_system_ml = ['-config', default_config,
'-exec', 'hybrid_spark', '-f', script_file, ' '.join(ml_options)]
cmd = cmd_spark + cmd_system_ml
# Debug
print(' '.join(cmd))
return_code = os.system(' '.join(cmd))
return return_code
if __name__ == '__main__':
cparser = argparse.ArgumentParser(formatter_class=argparse.ArgumentDefaultsHelpFormatter,
description='System-ML Spark Submit Script')
# SPARK-SUBMIT Options
cparser.add_argument('--master', default='local[*]', help='local, yarn', metavar='')
cparser.add_argument('--deploy-mode', help='client, cluster', default='client', metavar='')
cparser.add_argument('--driver-memory', default='8G', help='Memory for driver (e.g. 512M, 1G)', metavar='')
cparser.add_argument('--num-executors', nargs=1, help='Number of executors to launch', metavar='')
cparser.add_argument('--executor-memory', nargs=1, help='Memory per executor', metavar='')
cparser.add_argument('--executor-cores', nargs=1, help='Number of executor cores', metavar='')
cparser.add_argument('--conf', help='Spark configuration file', nargs='+', metavar='')
# SYSTEM-ML Options
cparser.add_argument('-nvargs', help='List of attributeName-attributeValue pairs', nargs='+', metavar='')
cparser.add_argument('-args', help='List of positional argument values', metavar='', nargs='+')
cparser.add_argument('-config', help='System-ML configuration file (e.g SystemML-config.xml)', metavar='')
cparser.add_argument('-explain', help='explains plan levels can be hops, runtime, '
'recompile_hops, recompile_runtime', nargs='?', const='runtime', metavar='')
cparser.add_argument('-debug', help='runs in debug mode', action='store_true')
cparser.add_argument('-stats', help='Monitor and report caching/recompilation statistics, '
'heavy hitter <count> is 10 unless overridden', nargs='?', const='10',
metavar='')
cparser.add_argument('-gpu', help='uses CUDA instructions when reasonable, '
'set <force> option to skip conservative memory estimates '
'and use GPU wherever possible', nargs='?', const='no_option')
cparser.add_argument('-f', required=True, help='specifies dml/pydml file to execute; '
'path can be local/hdfs/gpfs', metavar='')
args = cparser.parse_args()
arg_dict = vars(args)
return_code = spark_submit_entry(**arg_dict)
if return_code != 0:
print('Failed to run SystemML. Exit code :' + str(return_code))