def test_no_leak_many_call_lazy():
# Verify no memory leaks when calling a function a lot of times
# This isn't really a unit test, you have to run it and look at top to
# see if there's a leak
def build_graph(x, depth=5):
z = x
for d in range(depth):
z = ifelse(z.mean() > 0.5, -z, z)
return z
def time_linker(name, linker):
steps_a = 10
x = tensor.dvector()
a = build_graph(x, steps_a)
f_a = function([x], a, mode=Mode(optimizer=None, linker=linker()))
inp = np.random.rand(1000000)
for i in range(100):
f_a(inp)
# this doesn't seem to work, prints 0 for everything
# import resource
#
# pre = resource.getrusage(resource.RUSAGE_SELF)
# post = resource.getrusage(resource.RUSAGE_SELF)
# print(pre.ru_ixrss, post.ru_ixrss)
# print(pre.ru_idrss, post.ru_idrss)
# print(pre.ru_maxrss, post.ru_maxrss)
print(1)
time_linker("vmLinker_C", lambda: VMLinker(allow_gc=False, use_cloop=True))
print(2)
time_linker("vmLinker", lambda: VMLinker(allow_gc=False, use_cloop=False))
def test_reallocation():
x = tensor.scalar("x")
y = tensor.scalar("y")
z = tensor.tanh(3 * x + y) + tensor.cosh(x + 5 * y)
# The functinality is currently implement for non lazy and non c VM only.
for linker in [
VMLinker(allow_gc=False, lazy=False, use_cloop=False),
VMLinker(allow_gc=True, lazy=False, use_cloop=False),
]:
m = theano.compile.get_mode(theano.Mode(linker=linker))
m = m.excluding("fusion", "inplace")
f = theano.function([x, y], z, name="test_reduce_memory", mode=m)
output = f(1, 2)
assert output
storage_map = f.fn.storage_map
def check_storage(storage_map):
from theano.tensor.var import TensorConstant
for i in storage_map:
if not isinstance(i, TensorConstant):
keys_copy = list(storage_map.keys())[:]
keys_copy.remove(i)
for o in keys_copy:
if storage_map[i][
0] and storage_map[i][0] is storage_map[o][0]:
return [True, storage_map[o][0]]
return [False, None]
assert check_storage(storage_map)[0]
assert len({id(v) for v in storage_map.values()}) < len(storage_map)
def test_no_leak_many_call_nonlazy():
# Verify no memory leaks when calling a function a lot of times
# This isn't really a unit test, you have to run it and look at top to
# see if there's a leak.
def build_graph(x, depth=5):
z = x
for d in range(depth):
z = tensor.sin(-z + 1)
return z
def time_linker(name, linker):
steps_a = 10
x = tensor.dvector()
a = build_graph(x, steps_a)
f_a = function([x], a, mode=Mode(optimizer=None, linker=linker()))
inp = np.random.rand(1000000)
for i in range(500):
f_a(inp)
print(1)
time_linker("vmLinker_C",
lambda: VMLinker(allow_gc=False, use_cloop=True))
print(2)
time_linker("vmLinker",
lambda: VMLinker(allow_gc=False, use_cloop=False))
def test_no_recycling():
x = theano.tensor.vector()
for lnk in [
VMLinker(use_cloop=True),
VMLinker(use_cloop=False, lazy=True),
VMLinker(use_cloop=False, lazy=False, allow_gc=True),
VMLinker(use_cloop=False, lazy=False, allow_gc=False),
]:
mode = theano.Mode(optimizer="fast_compile", linker=lnk)
f = theano.function([x], x + 1, mode=mode)
f2 = theano.function([x], (x + 1) * 2, mode=mode)
m1 = f.fn.thunks[0].thunk.module
m2 = f2.fn.thunks[0].thunk.module
assert m1 is m2
def test_partial_function_with_updates():
def check_updates(linker_name):
x = tensor.lscalar("input")
y = theano.shared(np.asarray(1, "int64"), name="global")
f = theano.function(
[x],
[x, x + 34],
updates=[(y, x + 1)],
mode=Mode(optimizer=None, linker=linker_name),
)
g = theano.function(
[x],
[x - 6],
updates=[(y, y + 3)],
mode=Mode(optimizer=None, linker=linker_name),
)
assert f(3, output_subset=[]) == []
assert y.get_value() == 4
assert g(30, output_subset=[0]) == [24]
assert g(40, output_subset=[]) == []
assert y.get_value() == 10
check_updates(VMLinker(allow_partial_eval=True, use_cloop=False))
check_updates("cvm")
def test_VMLinker_make_vm_cvm():
# We don't want this at module level, since CXX might not be present
from theano.link.c.cvm import CVM
a = tensor.scalar()
linker = VMLinker(allow_gc=False, use_cloop=True)
f = function([a], a, mode=Mode(optimizer=None, linker=linker))
assert isinstance(f.fn, CVM)
def test_callback_with_ifelse(self):
a, b, c = tensor.scalars("abc")
f = function(
[a, b, c],
ifelse(a, 2 * b, 2 * c),
mode=Mode(optimizer=None, linker=VMLinker(callback=self.callback)),
)
f(1, 2, 3)
assert self.n_callbacks["IfElse"] == 2
def test_callback(self):
a, b, c = tensor.scalars("abc")
f = function(
[a, b, c],
(a + b) + c,
mode=Mode(optimizer=None, linker=VMLinker(callback=self.callback)),
)
f(1, 2, 3)
assert sum(self.n_callbacks.values()) == len(f.maker.fgraph.toposort())
f(1, 2, 3)
assert sum(self.n_callbacks.values()) == len(f.maker.fgraph.toposort()) * 2
def test_speed_lazy():
def build_graph(x, depth=5):
z = x
for d in range(depth):
z = ifelse(z[0] > 0, -z, z)
return z
def time_linker(name, linker):
steps_a = 10
steps_b = 100
x = tensor.vector()
a = build_graph(x, steps_a)
b = build_graph(x, steps_b)
f_a = function([x], a, mode=Mode(optimizer=None, linker=linker()))
f_b = function([x], b, mode=Mode(optimizer=None, linker=linker()))
f_a([2.0])
t0 = time.time()
f_a([2.0])
t1 = time.time()
f_b([2.0])
t2 = time.time()
f_b([2.0])
t3 = time.time()
t_a = t1 - t0
t_b = t3 - t2
print(
f"{name} takes {1000 * (t_b - t_a) / (steps_b - steps_a):f} s/Kop")
time_linker("vmLinker", VMLinker)
time_linker("vmLinker_nogc", lambda: VMLinker(allow_gc=False))
if theano.config.cxx:
time_linker("vmLinker_C",
lambda: VMLinker(allow_gc=False, use_cloop=True))
def test_partial_function_with_output_keys():
def check_partial_function_output_keys(linker_name):
x = tensor.scalar("input")
y = 3 * x
f = theano.function(
[x], {"a": y * 5, "b": y - 7}, mode=Mode(optimizer=None, linker=linker_name)
)
assert f(5, output_subset=["a"])["a"] == f(5)["a"]
check_partial_function_output_keys(
VMLinker(allow_partial_eval=True, use_cloop=False)
)
check_partial_function_output_keys("cvm")
def test_VMLinker_no_cxx():
from importlib import reload
from unittest.mock import patch
with config.change_flags(cxx=""):
with pytest.raises(MissingGXX):
import theano.link.c.cvm
reload(theano.link.c.cvm)
with patch.dict("sys.modules", {"theano.link.c.cvm": None}):
linker = VMLinker(allow_gc=False, use_cloop=True)
a = tensor.scalar()
with pytest.raises(ModuleNotFoundError):
_ = function([a], a, mode=Mode(optimizer=None, linker=linker))
def test_vm_gc():
# This already caused a bug in the trunk of Theano.
#
# The bug was introduced in the trunk on July 5th, 2012 and fixed on
# July 30th.
x = theano.tensor.vector()
p = RunOnce()(x)
mode = theano.Mode(linker=VMLinker(lazy=True))
f = theano.function([theano.In(x, mutable=True)], [p + 1, p + 2], mode=mode)
f([1, 2, 3])
p = RunOnce()(x)
pp = p + p
f = theano.function([x], [pp + pp], mode=mode)
f([1, 2, 3])
def test_c_thunks():
a = tensor.scalars("a")
b, c = tensor.vectors("bc")
cases = [False]
if theano.config.cxx:
cases.append(True)
for c_thunks in cases:
f = function(
[a, b, c],
ifelse(a, a * b, b * c),
mode=Mode(optimizer=None,
linker=VMLinker(c_thunks=c_thunks, use_cloop=False)),
)
f(1, [2], [3, 2])
with pytest.raises(ValueError):
f(0, [2], [3, 4])
assert any([hasattr(t, "cthunk") for t in f.fn.thunks]) == c_thunks
def test_partial_function():
from tests import unittest_tools as utt
def check_partial_function(linker_name):
x = tensor.scalar("input")
y = x**2
f = theano.function([x], [y + 7, y - 9, y / 14.0],
mode=Mode(optimizer=None, linker=linker_name))
assert f(3, output_subset=[0, 1, 2]) == f(3)
assert f(4, output_subset=[0, 2]) == [f(4)[0], f(4)[2]]
utt.assert_allclose(f(5), np.array([32.0, 16.0, 1.7857142857142858]))
check_partial_function(VMLinker(allow_partial_eval=True, use_cloop=False))
if not theano.config.cxx:
pytest.skip("Need cxx for this test")
check_partial_function("cvm")
def test_VMLinker_make_vm_no_cvm():
from importlib import reload
from unittest.mock import patch
with config.change_flags(cxx=""):
# Make sure that GXX isn't present
with pytest.raises(MissingGXX):
import theano.link.c.cvm
reload(theano.link.c.cvm)
# Make sure that `cvm` module is missing
with patch.dict("sys.modules", {"theano.link.c.cvm": None}):
a = tensor.scalar()
linker = VMLinker(allow_gc=False, use_cloop=True)
with pytest.raises(ModuleNotFoundError):
import theano.link.c.cvm
f = function([a], a, mode=Mode(optimizer=None, linker=linker))
assert isinstance(f.fn, Loop)
def test_speed():
def build_graph(x, depth=5):
z = x
for d in range(depth):
z = z + z
return z
def numpy_version(x, depth):
z = x
for d in range(depth):
z = z + z
return z
def time_numpy():
steps_a = 5
steps_b = 100
x = np.asarray([2.0, 3.0], dtype=theano.config.floatX)
numpy_version(x, steps_a)
t0 = time.time()
# print numpy_version(x, steps_a)
t1 = time.time()
t2 = time.time()
# print numpy_version(x, steps_b)
t3 = time.time()
t_a = t1 - t0
t_b = t3 - t2
print(
f"numpy takes {1000 * (t_b - t_a) / (steps_b - steps_a):f} s/Kop")
def time_linker(name, linker):
steps_a = 5
steps_b = 100
x = tensor.vector()
a = build_graph(x, steps_a)
b = build_graph(x, steps_b)
f_a = function([x], a, mode=Mode(optimizer=None, linker=linker()))
f_b = function([x], b, mode=Mode(optimizer=None, linker=linker()))
f_a([2.0, 3.0])
t0 = time.time()
f_a([2.0, 3.0])
t1 = time.time()
f_b([2.0, 3.0])
t2 = time.time()
f_b([2.0, 3.0])
t3 = time.time()
t_a = t1 - t0
t_b = t3 - t2
print(
f"{name} takes {1000 * (t_b - t_a) / (steps_b - steps_a):f} s/Kop")
from theano.link.c.basic import OpWiseCLinker
time_linker("c|py", OpWiseCLinker)
time_linker("vmLinker", VMLinker)
time_linker("vmLinker_nogc", lambda: VMLinker(allow_gc=False))
if theano.config.cxx:
time_linker("vmLinker_CLOOP",
lambda: VMLinker(allow_gc=False, use_cloop=True))
time_numpy()
from theano.link.basic import PerformLinker
from theano.link.c.basic import CLinker, OpWiseCLinker
from theano.link.jax import JAXLinker
from theano.link.vm import VMLinker
_logger = logging.getLogger("theano.compile.mode")
# If a string is passed as the linker argument in the constructor for
# Mode, it will be used as the key to retrieve the real linker in this
# dictionary
predefined_linkers = {
"py": PerformLinker(), # Use allow_gc Theano flag
"c": CLinker(), # Don't support gc. so don't check allow_gc
"c|py": OpWiseCLinker(), # Use allow_gc Theano flag
"c|py_nogc": OpWiseCLinker(allow_gc=False),
"vm": VMLinker(use_cloop=False), # Use allow_gc Theano flag
"cvm": VMLinker(use_cloop=True), # Use allow_gc Theano flag
"vm_nogc": VMLinker(allow_gc=False, use_cloop=False),
"cvm_nogc": VMLinker(allow_gc=False, use_cloop=True),
"jax": JAXLinker(),
}
def register_linker(name, linker):
"""Add a `Linker` which can be referred to by `name` in `Mode`."""
if name in predefined_linkers:
raise ValueError(f"Linker name already taken: {name}")
predefined_linkers[name] = linker
# If a string is passed as the optimizer argument in the constructor
def __init__(self, record=None, **kwargs):
"""
Takes either a Record object or the keyword arguments to make one.
Parameters
----------
record : Record
The existing Record object to use.
kwargs : pointer?
Keyword arguments to construct new object.
"""
if record is None:
record = Record(**kwargs)
else:
assert len(kwargs.keys()) == 0
self.set_record(record)
def handle_line(fgraph, line, i, node, fn):
"""
Records new node computation.
Parameters
----------
line : string
Line to record. For example, the function name or node name.
i : integer
Node number in the toposort order.
node : Apply,
The Apply node which created the entry.
fn : Function,
Function related to Apply node.
"""
try:
self.record.handle_line(line)
except MismatchError as e:
print("Got this MismatchError:")
print(e)
print(f"while processing node i={i}:")
print(f"str(node):{node}")
print("Symbolic inputs: ")
for elem in node.inputs:
print(min_informative_str(elem))
print("str(output) of outputs: ")
for elem in fn.outputs:
assert isinstance(elem, list)
(elem, ) = elem
print(str(elem))
print(f"function name: {fgraph.name}")
raise MismatchError("Non-determinism detected by WrapLinker")
def callback(fgraph, i, node, fn):
"""
Function called by Apply nodes at the end of each computation?
"""
if fgraph.name is None:
raise ValueError(
"Un-named functions are not allowed with RecordMode, "
"because they make it impossible to tell if the same function is "
"running during the playback.")
if fgraph not in self.known_fgraphs:
assert not any(
[elem.name == fgraph.name for elem in self.known_fgraphs])
self.known_fgraphs.add(fgraph)
num_app = len(fgraph.apply_nodes)
line = f"Function {fgraph.name} has {num_app} apply nodes.\n"
handle_line(fgraph, line, i, node, fn)
line = f"Function name: {fgraph.name}\n"
handle_line(fgraph, line, i, node, fn)
line = f"Node {i}:{node}\n"
handle_line(fgraph, line, i, node, fn)
assert all(
[isinstance(x, list) and len(x) == 1 for x in fn.inputs])
def digest(x):
x = x[0]
return hex_digest(x)
inputs_digest = " ".join([digest(x) for x in fn.inputs])
line = f"Inputs: {inputs_digest}\n"
handle_line(fgraph, line, i, node, fn)
fn()
outputs_digest = " ".join([digest(x) for x in fn.outputs])
line = f"Outputs: {outputs_digest}\n"
handle_line(fgraph, line, i, node, fn)
# linker = theano.link.c.basic.OpWiseCLinker()
linker = VMLinker(use_cloop=bool(config.cxx))
wrap_linker = WrapLinkerMany([linker], [callback])
super().__init__(wrap_linker, optimizer="fast_run")
