def test_assert_ne_with_deferrables(self):
self.assertTrue(sn.assert_ne(1, make_deferrable(2)))
self.assertTrue(sn.assert_ne(make_deferrable(1), False))
self.assertRaisesRegex(SanityError, '1 == 1', evaluate,
sn.assert_ne(make_deferrable(1), 1))
self.assertRaisesRegex(SanityError, '1 == True', evaluate,
sn.assert_ne(make_deferrable(1), True))
def test_assert_ne(self):
self.assertTrue(sn.assert_ne(1, 2))
self.assertTrue(sn.assert_ne(1, False))
self.assertRaisesRegex(SanityError, '1 == 1', evaluate,
sn.assert_ne(1, 1))
self.assertRaisesRegex(SanityError, '1 == True', evaluate,
sn.assert_ne(1, True))
def test_assert_ne_with_deferrables():
assert sn.assert_ne(1, sn.defer(2))
assert sn.assert_ne(sn.defer(1), False)
with pytest.raises(SanityError, match='1 == 1'):
sn.evaluate(sn.assert_ne(sn.defer(1), 1))
with pytest.raises(SanityError, match='1 == True'):
sn.evaluate(sn.assert_ne(sn.defer(1), True))
def test_assert_ne():
assert sn.assert_ne(1, 2)
assert sn.assert_ne(1, False)
with pytest.raises(SanityError, match='1 == 1'):
sn.evaluate(sn.assert_ne(1, 1))
with pytest.raises(SanityError, match='1 == True'):
sn.evaluate(sn.assert_ne(1, True))
def assert_successful_execution(self):
nids = sn.extractall(r'nid(?P<nid>\d+)', self.stdout, 'nid', str)
return sn.all([
sn.assert_ne(nids, []),
sn.assert_ne(nids[0], nids[1]),
sn.assert_found(r'IPCluster is ready\!\s+', self.stdout),
sn.assert_found(r'slope=\S+', self.stdout)
])
def __init__(self):
self.descr = 'Distributed training with TensorFlow using ipyparallel'
self.valid_systems = ['daint:gpu', 'dom:gpu']
self.valid_prog_environs = ['PrgEnv-gnu']
cray_cdt_version = osext.cray_cdt_version()
# FIXME: The following will not be needed after the Daint upgrade
if self.current_system.name == 'dom':
self.modules = [
'ipcmagic',
f'Horovod/0.21.0-CrayGNU-{cray_cdt_version}-tf-2.4.0'
]
else:
self.modules = [
'ipcmagic', 'Horovod/0.19.1-CrayGNU-20.08-tf-2.2.0'
]
self.num_tasks = 2
self.num_tasks_per_node = 1
self.executable = 'ipython'
self.executable_opts = ['tf-hvd-sgd-ipc-tf2.py']
nids = sn.extractall(r'nid(?P<nid>\d+)', self.stdout, 'nid', str)
self.sanity_patterns = sn.all(
[sn.assert_ne(nids, []),
sn.assert_ne(nids[0], nids[1])])
self.reference = {
'daint:gpu': {
'slope': (2.0, -0.1, 0.1, None),
'offset': (0.0, -0.1, 0.1, None),
'retries': (0, None, None, None),
'time': (10, None, None, 's'),
},
'dom:gpu': {
'slope': (2.0, -0.1, 0.1, None),
'offset': (0.0, -0.1, 0.1, None),
'retries': (0, None, None, None),
'time': (10, None, None, 's'),
}
}
self.perf_patterns = {
'slope':
sn.extractsingle(r'slope=(?P<slope>\S+)', self.stdout, 'slope',
float),
'offset':
sn.extractsingle(r'offset=(?P<offset>\S+)', self.stdout, 'offset',
float),
'retries':
4 -
sn.count(sn.findall(r'IPCluster is already running', self.stdout)),
'time':
sn.extractsingle(
r'IPCluster is ready\!\s+'
r'\((?P<time>\d+) seconds\)', self.stdout, 'time', float)
}
self.maintainers = ['RS', 'TR']
self.tags = {'production'}
def __init__(self):
self.descr = 'Distributed training with TensorFlow using ipyparallel'
self.valid_systems = ['daint:gpu', 'dom:gpu']
self.valid_prog_environs = ['PrgEnv-gnu']
self.modules = ['ipcmagic']
self.prerun_cmds = [
'module unload dask',
'module load Horovod/0.16.4-CrayGNU-19.10-tf-1.14.0'
]
self.num_tasks = 2
self.num_tasks_per_node = 1
self.executable = 'ipython'
self.executable_opts = ['tf-hvd-sgd-ipc-tf-1.14.py']
nids = sn.extractall(r'nid(?P<nid>\d+)', self.stdout, 'nid', str)
self.sanity_patterns = sn.all(
[sn.assert_ne(nids, []),
sn.assert_ne(nids[0], nids[1])])
self.reference = {
'daint:gpu': {
'slope': (2.0, -0.1, 0.1, None),
'offset': (0.0, -0.1, 0.1, None),
'retries': (0, None, None, None),
'time': (10, None, None, 's'),
},
'dom:gpu': {
'slope': (2.0, -0.1, 0.1, None),
'offset': (0.0, -0.1, 0.1, None),
'retries': (0, None, None, None),
'time': (10, None, None, 's'),
}
}
self.perf_patterns = {
'slope':
sn.extractsingle(r'slope=(?P<slope>\S+)', self.stdout, 'slope',
float),
'offset':
sn.extractsingle(r'offset=(?P<offset>\S+)', self.stdout, 'offset',
float),
'retries':
4 -
sn.count(sn.findall(r'IPCluster is already running', self.stdout)),
'time':
sn.extractsingle(
r'IPCluster is ready\!\s+'
r'\((?P<time>\d+) seconds\)', self.stdout, 'time', float)
}
self.maintainers = ['RS', 'TR']
self.tags = {'production'}
def assert_successful_execution(self):
'''Checks that the program is running on 2 different nodes (nids
are different), that IPCMagic is configured and returns the correct
end-of-program message (returns the slope parameter in the end).'''
nids = sn.extractall(r'nid(?P<nid>\d+)', self.stdout, 'nid', str)
return sn.all([
sn.assert_eq(sn.len(nids), 2),
sn.assert_ne(nids[0], nids[1]),
sn.assert_found(r'slope=\S+', self.stdout)
])
