提交 816a83e3 authored 作者: nouiz's avatar nouiz

Merge pull request #1271 from lamblin/arch_in_cache

Always put arch bitwidth in cache
...@@ -364,7 +364,7 @@ import theano and print the config variable, as in: ...@@ -364,7 +364,7 @@ import theano and print the config variable, as in:
.. attribute:: compiledir_format .. attribute:: compiledir_format
Default: "compiledir_%(platform)s-%(processor)s-%(python_version)s" Default: "compiledir_%(platform)s-%(processor)s-%(python_version)s-%(python_bitwidth)s"
This is a Python format string that specifies the subdirectory This is a Python format string that specifies the subdirectory
of ``config.base_compiledir`` in which to store platform-dependent of ``config.base_compiledir`` in which to store platform-dependent
......
...@@ -548,7 +548,7 @@ class Test_pfunc(unittest.TestCase): ...@@ -548,7 +548,7 @@ class Test_pfunc(unittest.TestCase):
def test_default_updates_input(self): def test_default_updates_input(self):
x = shared(0) x = shared(0)
y = shared(1) y = shared(1)
if theano.gof.cmodule.python_int_bitwidth() == 32: if theano.gof.python_int_bitwidth() == 32:
a = iscalar('a') a = iscalar('a')
else: else:
a = lscalar('a') a = lscalar('a')
......
...@@ -18,7 +18,7 @@ class Test_SharedVariable(unittest.TestCase): ...@@ -18,7 +18,7 @@ class Test_SharedVariable(unittest.TestCase):
assert shared(7, dtype='float64').type == Scalar('float64') assert shared(7, dtype='float64').type == Scalar('float64')
else: else:
if theano.gof.cmodule.python_int_bitwidth() == 32: if theano.gof.python_int_bitwidth() == 32:
assert shared(7).type == theano.tensor.iscalar, shared(7).type assert shared(7).type == theano.tensor.iscalar, shared(7).type
else: else:
assert shared(7).type == theano.tensor.lscalar, shared(7).type assert shared(7).type == theano.tensor.lscalar, shared(7).type
......
...@@ -38,7 +38,9 @@ e-mail thread "What is gof?" ...@@ -38,7 +38,9 @@ e-mail thread "What is gof?"
from theano.gof.cc import \ from theano.gof.cc import \
CLinker, OpWiseCLinker, DualLinker CLinker, OpWiseCLinker, DualLinker
import theano.gof.compiledir # adds config vars # Also adds config vars
from theano.gof.compiledir import \
local_bitwidth, python_int_bitwidth
from theano.gof.fg import \ from theano.gof.fg import \
InconsistencyError, MissingInputError, FunctionGraph InconsistencyError, MissingInputError, FunctionGraph
...@@ -77,4 +79,3 @@ from theano.gof.type import \ ...@@ -77,4 +79,3 @@ from theano.gof.type import \
from theano.gof.utils import \ from theano.gof.utils import \
object2, MethodNotDefined object2, MethodNotDefined
...@@ -8,7 +8,6 @@ import os ...@@ -8,7 +8,6 @@ import os
import shutil import shutil
import stat import stat
import StringIO import StringIO
import struct
import subprocess import subprocess
import sys import sys
import tempfile import tempfile
...@@ -27,7 +26,7 @@ from theano.misc.windows import call_subprocess_Popen ...@@ -27,7 +26,7 @@ from theano.misc.windows import call_subprocess_Popen
# we will abuse the lockfile mechanism when reading and writing the registry # we will abuse the lockfile mechanism when reading and writing the registry
from theano.gof import compilelock from theano.gof import compilelock
from theano.gof.compiledir import gcc_version_str from theano.gof.compiledir import gcc_version_str, local_bitwidth
from theano.configparser import AddConfigVar, BoolParam from theano.configparser import AddConfigVar, BoolParam
...@@ -55,29 +54,6 @@ AddConfigVar('cmodule.compilation_warning', ...@@ -55,29 +54,6 @@ AddConfigVar('cmodule.compilation_warning',
BoolParam(False)) BoolParam(False))
def local_bitwidth():
"""
Return 32 for 32bit arch, 64 for 64bit arch
By "architecture", we mean the size of memory pointers (size_t in C),
*not* the size of long int, as it can be different.
"""
# Note that according to Python documentation, `platform.architecture()` is
# not reliable on OS X with universal binaries.
# Also, sys.maxsize does not exist in Python < 2.6.
# 'P' denotes a void*, and the size is expressed in bytes.
return struct.calcsize('P') * 8
def python_int_bitwidth():
"""
Return the bit width of Python int (C long int).
Note that it can be different from the size of a memory pointer.
"""
# 'l' denotes a C long int, and the size is expressed in bytes.
return struct.calcsize('l') * 8
_logger = logging.getLogger("theano.gof.cmodule") _logger = logging.getLogger("theano.gof.cmodule")
_logger.setLevel(logging.WARNING) _logger.setLevel(logging.WARNING)
...@@ -176,14 +152,14 @@ static struct PyModuleDef moduledef = {{ ...@@ -176,14 +152,14 @@ static struct PyModuleDef moduledef = {{
}}; }};
""".format(name=self.name) """.format(name=self.name)
print >> stream, "PyMODINIT_FUNC PyInit_%s(void) {" % self.name print >> stream, "PyMODINIT_FUNC PyInit_%s(void) {" % self.name
for b in self.init_blocks: for block in self.init_blocks:
print >> stream, ' ', b print >> stream, ' ', block
print >> stream, " PyObject *m = PyModule_Create(&moduledef);" print >> stream, " PyObject *m = PyModule_Create(&moduledef);"
print >> stream, " return m;" print >> stream, " return m;"
else: else:
print >> stream, "PyMODINIT_FUNC init%s(void){" % self.name print >> stream, "PyMODINIT_FUNC init%s(void){" % self.name
for b in self.init_blocks: for block in self.init_blocks:
print >> stream, ' ', b print >> stream, ' ', block
print >> stream, ' ', ('(void) Py_InitModule("%s", MyMethods);' print >> stream, ' ', ('(void) Py_InitModule("%s", MyMethods);'
% self.name) % self.name)
print >> stream, "}" print >> stream, "}"
...@@ -1564,7 +1540,8 @@ class GCC_compiler(object): ...@@ -1564,7 +1540,8 @@ class GCC_compiler(object):
lines = stdout + stderr lines = stdout + stderr
return lines return lines
# The '-' at the end is needed. Otherwise, g++ do not output enough information. # The '-' at the end is needed. Otherwise, g++ do not output
# enough information.
native_lines = get_lines("g++ -march=native -E -v -") native_lines = get_lines("g++ -march=native -E -v -")
_logger.info("g++ -march=native selected lines: %s", native_lines) _logger.info("g++ -march=native selected lines: %s", native_lines)
if len(native_lines) != 1: if len(native_lines) != 1:
...@@ -1619,6 +1596,39 @@ class GCC_compiler(object): ...@@ -1619,6 +1596,39 @@ class GCC_compiler(object):
cxxflags.append("-D NPY_ARRAY_UPDATE_ALL=NPY_UPDATE_ALL") cxxflags.append("-D NPY_ARRAY_UPDATE_ALL=NPY_UPDATE_ALL")
cxxflags.append("-D NPY_ARRAY_C_CONTIGUOUS=NPY_C_CONTIGUOUS") cxxflags.append("-D NPY_ARRAY_C_CONTIGUOUS=NPY_C_CONTIGUOUS")
cxxflags.append("-D NPY_ARRAY_F_CONTIGUOUS=NPY_F_CONTIGUOUS") cxxflags.append("-D NPY_ARRAY_F_CONTIGUOUS=NPY_F_CONTIGUOUS")
# Platform-specific flags.
# We put them here, rather than in compile_str(), so they en up
# in the key of the compiled module, avoiding potential conflicts.
# Figure out whether the current Python executable is 32
# or 64 bit and compile accordingly.
n_bits = local_bitwidth()
cxxflags.append('-m%d' % n_bits)
_logger.debug("Compiling for %s bit architecture", n_bits)
if sys.platform != 'win32':
# Under Windows it looks like fPIC is useless. Compiler warning:
# '-fPIC ignored for target (all code is position independent)'
cxxflags.append('-fPIC')
if sys.platform == 'win32' and local_bitwidth() == 64:
# Under 64-bit Windows installation, sys.platform is 'win32'.
# We need to define MS_WIN64 for the preprocessor to be able to
# link with libpython.
cxxflags.append('-DMS_WIN64')
#DSE Patch 1 for supporting OSX frameworks; add -framework Python
if sys.platform == 'darwin':
cxxflags.extend(['-undefined', 'dynamic_lookup'])
python_inc = distutils.sysconfig.get_python_inc()
# link with the framework library *if specifically requested*
# config.mac_framework_link is by default False, since on some mac
# installs linking with -framework causes a Bus Error
if (python_inc.count('Python.framework') > 0 and
config.cmodule.mac_framework_link):
cxxflags.extend(['-framework', 'Python'])
return cxxflags return cxxflags
@staticmethod @staticmethod
...@@ -1744,40 +1754,10 @@ class GCC_compiler(object): ...@@ -1744,40 +1754,10 @@ class GCC_compiler(object):
else: else:
preargs = list(preargs) preargs = list(preargs)
if sys.platform != 'win32':
# Under Windows it looks like fPIC is useless. Compiler warning:
# '-fPIC ignored for target (all code is position independent)'
preargs.append('-fPIC')
if sys.platform == 'win32' and local_bitwidth() == 64:
# Under 64-bit Windows installation, sys.platform is 'win32'.
# We need to define MS_WIN64 for the preprocessor to be able to
# link with libpython.
preargs.append('-DMS_WIN64')
# We also add "-m64", in case the installed gcc is 32-bit
preargs.append('-m64')
include_dirs = include_dirs + std_include_dirs() include_dirs = include_dirs + std_include_dirs()
libs = std_libs() + libs libs = std_libs() + libs
lib_dirs = std_lib_dirs() + lib_dirs lib_dirs = std_lib_dirs() + lib_dirs
#DSE Patch 1 for supporting OSX frameworks; add -framework Python
if sys.platform == 'darwin':
preargs.extend(['-undefined', 'dynamic_lookup'])
python_inc = distutils.sysconfig.get_python_inc()
# link with the framework library *if specifically requested*
# config.mac_framework_link is by default False, since on some mac
# installs linking with -framework causes a Bus Error
if (python_inc.count('Python.framework') > 0 and
config.cmodule.mac_framework_link):
preargs.extend(['-framework', 'Python'])
# Figure out whether the current Python executable is 32
# or 64 bit and compile accordingly.
n_bits = local_bitwidth()
preargs.extend(['-m%s' % n_bits])
_logger.debug("OS X: compiling for %s bit architecture", n_bits)
# sometimes, the linker cannot find -lpython so we need to tell it # sometimes, the linker cannot find -lpython so we need to tell it
# explicitly where it is located # explicitly where it is located
# this returns somepath/lib/python2.x # this returns somepath/lib/python2.x
......
...@@ -4,6 +4,7 @@ import os ...@@ -4,6 +4,7 @@ import os
import platform import platform
import re import re
import shutil import shutil
import struct
import subprocess import subprocess
import sys import sys
import textwrap import textwrap
...@@ -32,16 +33,44 @@ except OSError: ...@@ -32,16 +33,44 @@ except OSError:
del p del p
del dummy_err del dummy_err
compiledir_format_dict = {"platform": platform.platform(),
"processor": platform.processor(), def local_bitwidth():
"python_version": platform.python_version(), """
"theano_version": theano.__version__, Return 32 for 32bit arch, 64 for 64bit arch
"numpy_version": numpy.__version__,
"gxx_version": gcc_version_str.replace(" ", "_"), By "architecture", we mean the size of memory pointers (size_t in C),
} *not* the size of long int, as it can be different.
"""
# Note that according to Python documentation, `platform.architecture()` is
# not reliable on OS X with universal binaries.
# Also, sys.maxsize does not exist in Python < 2.6.
# 'P' denotes a void*, and the size is expressed in bytes.
return struct.calcsize('P') * 8
def python_int_bitwidth():
"""
Return the bit width of Python int (C long int).
Note that it can be different from the size of a memory pointer.
"""
# 'l' denotes a C long int, and the size is expressed in bytes.
return struct.calcsize('l') * 8
compiledir_format_dict = {
"platform": platform.platform(),
"processor": platform.processor(),
"python_version": platform.python_version(),
"python_bitwidth": local_bitwidth(),
"python_int_bitwidth": python_int_bitwidth(),
"theano_version": theano.__version__,
"numpy_version": numpy.__version__,
"gxx_version": gcc_version_str.replace(" ", "_"),
}
compiledir_format_keys = ", ".join(sorted(compiledir_format_dict.keys())) compiledir_format_keys = ", ".join(sorted(compiledir_format_dict.keys()))
default_compiledir_format =\ default_compiledir_format = ("compiledir_%(platform)s-%(processor)s-"
"compiledir_%(platform)s-%(processor)s-%(python_version)s" "%(python_version)s-%(python_bitwidth)s")
AddConfigVar("compiledir_format", AddConfigVar("compiledir_format",
textwrap.fill(textwrap.dedent("""\ textwrap.fill(textwrap.dedent("""\
......
import re
# import op import traceback
# import variable
from theano import config from theano import config
import re, traceback
def add_tag_trace(thing): def add_tag_trace(thing):
"""Add tag.trace to an node or variable. """Add tag.trace to an node or variable.
...@@ -11,15 +10,18 @@ def add_tag_trace(thing): ...@@ -11,15 +10,18 @@ def add_tag_trace(thing):
The argument is returned after being affected (inplace). The argument is returned after being affected (inplace).
""" """
limit = config.traceback.limit limit = config.traceback.limit
if limit == -1: limit = None if limit == -1:
limit = None
thing.tag.trace = traceback.extract_stack(limit=limit)[:-1] thing.tag.trace = traceback.extract_stack(limit=limit)[:-1]
return thing return thing
def hashgen(): def hashgen():
hashgen.next += 1 hashgen.next += 1
return hashgen.next return hashgen.next
hashgen.next = 0 hashgen.next = 0
class MethodNotDefined(Exception): class MethodNotDefined(Exception):
""" """
To be raised by functions defined as part of an interface. To be raised by functions defined as part of an interface.
...@@ -28,6 +30,7 @@ class MethodNotDefined(Exception): ...@@ -28,6 +30,7 @@ class MethodNotDefined(Exception):
function has been left out of an implementation class. function has been left out of an implementation class.
""" """
class object2(object): class object2(object):
__slots__ = [] __slots__ = []
if 0: if 0:
...@@ -36,23 +39,30 @@ class object2(object): ...@@ -36,23 +39,30 @@ class object2(object):
if hasattr(self, '__eq__') or hasattr(self, '__cmp__'): if hasattr(self, '__eq__') or hasattr(self, '__cmp__'):
raise TypeError("unhashable object: %s" % self) raise TypeError("unhashable object: %s" % self)
return id(self) return id(self)
def __ne__(self, other): def __ne__(self, other):
return not self == other return not self == other
class scratchpad: class scratchpad:
def clear(self): def clear(self):
self.__dict__.clear() self.__dict__.clear()
def __update__(self, other): def __update__(self, other):
self.__dict__.update(other.__dict__) self.__dict__.update(other.__dict__)
return self return self
def __str__(self): def __str__(self):
return "scratchpad" + str(self.__dict__) return "scratchpad" + str(self.__dict__)
def __repr__(self): def __repr__(self):
return "scratchpad" + str(self.__dict__) return "scratchpad" + str(self.__dict__)
def info(self): def info(self):
print "<theano.gof.utils.scratchpad instance at %i>"%id(self) print "<theano.gof.utils.scratchpad instance at %i>" % id(self)
for k,v in self.__dict__.items(): for k, v in self.__dict__.items():
print " %s: %s" % (k,v) print " %s: %s" % (k, v)
class D: class D:
def __init__(self, **d): def __init__(self, **d):
...@@ -63,6 +73,7 @@ def memoize(f): ...@@ -63,6 +73,7 @@ def memoize(f):
"""Cache the return value for each tuple of arguments """Cache the return value for each tuple of arguments
(which must be hashable) """ (which must be hashable) """
cache = {} cache = {}
def rval(*args, **kwargs): def rval(*args, **kwargs):
kwtup = tuple(kwargs.items()) kwtup = tuple(kwargs.items())
key = (args, kwtup) key = (args, kwtup)
...@@ -72,8 +83,8 @@ def memoize(f): ...@@ -72,8 +83,8 @@ def memoize(f):
else: else:
val = cache[key] val = cache[key]
return val return val
return rval
return rval
def deprecated(filename, msg=''): def deprecated(filename, msg=''):
...@@ -92,6 +103,7 @@ def deprecated(filename, msg=''): ...@@ -92,6 +103,7 @@ def deprecated(filename, msg=''):
""" """
def _deprecated(f): def _deprecated(f):
printme = [True] printme = [True]
def g(*args, **kwargs): def g(*args, **kwargs):
if printme[0]: if printme[0]:
print 'WARNING: %s.%s deprecated. %s'\ print 'WARNING: %s.%s deprecated. %s'\
...@@ -99,19 +111,23 @@ def deprecated(filename, msg=''): ...@@ -99,19 +111,23 @@ def deprecated(filename, msg=''):
printme[0] = False printme[0] = False
return f(*args, **kwargs) return f(*args, **kwargs)
return g return g
return _deprecated return _deprecated
def uniq(seq): def uniq(seq):
#TODO: consider building a set out of seq so that the if condition is constant time -JB #TODO: consider building a set out of seq so that the if condition
#is constant time -JB
return [x for i, x in enumerate(seq) if seq.index(x) == i] return [x for i, x in enumerate(seq) if seq.index(x) == i]
def difference(seq1, seq2): def difference(seq1, seq2):
""" """
Returns all elements in seq1 which are not in seq2: i.e seq1\seq2 Returns all elements in seq1 which are not in seq2: i.e seq1\seq2
""" """
try: try:
# try to use O(const * len(seq1)) algo # try to use O(const * len(seq1)) algo
if len(seq2) < 4: # I'm guessing this threshold -JB if len(seq2) < 4: # I'm guessing this threshold -JB
raise Exception('not worth it') raise Exception('not worth it')
set2 = set(seq2) set2 = set(seq2)
return [x for x in seq1 if x not in set2] return [x for x in seq1 if x not in set2]
...@@ -132,13 +148,16 @@ def partition(f, seq): ...@@ -132,13 +148,16 @@ def partition(f, seq):
seqf.append(elem) seqf.append(elem)
return seqt, seqf return seqt, seqf
def attr_checker(*attrs): def attr_checker(*attrs):
def f(candidate): def f(candidate):
for attr in attrs: for attr in attrs:
if not hasattr(candidate, attr): if not hasattr(candidate, attr):
return False return False
return True return True
f.__doc__ = "Checks that the candidate has the following attributes: %s" % ", ".join(["'%s'"%attr for attr in attrs])
f.__doc__ = ("Checks that the candidate has the following attributes: %s"
% ", ".join(["'%s'" % attr for attr in attrs]))
return f return f
...@@ -149,11 +168,10 @@ def all_bases(cls, accept): ...@@ -149,11 +168,10 @@ def all_bases(cls, accept):
return [cls for cls in rval if accept(cls)] return [cls for cls in rval if accept(cls)]
def all_bases_collect(cls, raw_name): def all_bases_collect(cls, raw_name):
rval = set() rval = set()
name = "__%s__" % raw_name name = "__%s__" % raw_name
if name in cls.__dict__: # don't use hasattr if name in cls.__dict__: # don't use hasattr
rval.add(getattr(cls, name)) rval.add(getattr(cls, name))
cut = "__%s_override__" % raw_name cut = "__%s_override__" % raw_name
if not cls.__dict__.get(cut, False): if not cls.__dict__.get(cut, False):
...@@ -162,7 +180,7 @@ def all_bases_collect(cls, raw_name): ...@@ -162,7 +180,7 @@ def all_bases_collect(cls, raw_name):
return rval return rval
def camelcase_to_separated(string, sep = "_"): def camelcase_to_separated(string, sep="_"):
return re.sub('(.)([A-Z])', '\\1%s\\2' % sep, string).lower() return re.sub('(.)([A-Z])', '\\1%s\\2' % sep, string).lower()
...@@ -172,6 +190,7 @@ def to_return_values(values): ...@@ -172,6 +190,7 @@ def to_return_values(values):
else: else:
return values return values
def from_return_values(values): def from_return_values(values):
if isinstance(values, (list, tuple)): if isinstance(values, (list, tuple)):
return values return values
...@@ -186,7 +205,8 @@ class ClsInit(type): ...@@ -186,7 +205,8 @@ class ClsInit(type):
Validate and initialize the L{Op} subclass 'cls' Validate and initialize the L{Op} subclass 'cls'
This function: This function:
- changes class attributes input_names and output_names to be lists if they are single strings. - changes class attributes input_names and output_names to be lists
if they are single strings.
""" """
type.__init__(cls, name, bases, dct) type.__init__(cls, name, bases, dct)
...@@ -195,8 +215,10 @@ class ClsInit(type): ...@@ -195,8 +215,10 @@ class ClsInit(type):
def toposort(prereqs_d): def toposort(prereqs_d):
""" """
Sorts prereqs_d.keys() topologically. prereqs_d[x] contains all the elements Sorts prereqs_d.keys() topologically.
that must come before x in the ordering.
prereqs_d[x] contains all the elements that must come before x
in the ordering.
""" """
# all1 = set(prereqs_d.keys()) # all1 = set(prereqs_d.keys())
...@@ -223,19 +245,26 @@ def toposort(prereqs_d): ...@@ -223,19 +245,26 @@ def toposort(prereqs_d):
if not prereqs_d[postreq].difference(done): if not prereqs_d[postreq].difference(done):
next.add(postreq) next.add(postreq)
if len(prereqs_d) != len(seq): if len(prereqs_d) != len(seq):
raise Exception("Cannot sort topologically: there might be cycles, " + \ raise Exception("Cannot sort topologically: there might be cycles, "
"prereqs_d does not have a key for each element or " + \ "prereqs_d does not have a key for each element or "
"some orderings contain invalid elements.") "some orderings contain invalid elements.")
return seq return seq
def print_for_dot(self): def print_for_dot(self):
#TODO: popen2("dot -Tpng | display") and actually make the graph window pop up #TODO: popen2("dot -Tpng | display") and actually make the graph window
print "digraph unix { size = '6,6'; node [color = lightblue2; style = filled];" #pop up
for op in self.order: print ("digraph unix { size = '6,6'; node [color = lightblue2;"
for input in op.inputs: "style = filled];")
if input.owner: for op in self.order:
print input.owner.__class__.__name__ + str(abs(id(input.owner))), " -> ", op.__class__.__name__ + str(abs(id(op))), ";" for input in op.inputs:
if input.owner:
print ' '.join((
input.owner.__class__.__name__ + str(abs(id(input.owner))),
" -> ",
op.__class__.__name__ + str(abs(id(op))),
";"))
class Keyword: class Keyword:
...@@ -263,9 +292,11 @@ simple_types = (int, float, str, bool, None.__class__, Keyword) ...@@ -263,9 +292,11 @@ simple_types = (int, float, str, bool, None.__class__, Keyword)
ANY_TYPE = Keyword("ANY_TYPE") ANY_TYPE = Keyword("ANY_TYPE")
FALL_THROUGH = Keyword("FALL_THROUGH") FALL_THROUGH = Keyword("FALL_THROUGH")
def comm_guard(type1, type2): def comm_guard(type1, type2):
def wrap(f): def wrap(f):
old_f = f.func_globals[f.__name__] old_f = f.func_globals[f.__name__]
def new_f(arg1, arg2, *rest): def new_f(arg1, arg2, *rest):
if (type1 is ANY_TYPE or isinstance(arg1, type1)) \ if (type1 is ANY_TYPE or isinstance(arg1, type1)) \
and (type2 is ANY_TYPE or isinstance(arg2, type2)): and (type2 is ANY_TYPE or isinstance(arg2, type2)):
...@@ -283,6 +314,7 @@ def comm_guard(type1, type2): ...@@ -283,6 +314,7 @@ def comm_guard(type1, type2):
return variable return variable
new_f.__name__ = f.__name__ new_f.__name__ = f.__name__
def typename(type): def typename(type):
if isinstance(type, Keyword): if isinstance(type, Keyword):
return str(type) return str(type)
...@@ -290,14 +322,19 @@ def comm_guard(type1, type2): ...@@ -290,14 +322,19 @@ def comm_guard(type1, type2):
return "(" + ", ".join([x.__name__ for x in type]) + ")" return "(" + ", ".join([x.__name__ for x in type]) + ")"
else: else:
return type.__name__ return type.__name__
new_f.__doc__ = str(old_f.__doc__) + "\n" + ", ".join([typename(type) for type in (type1, type2)]) + "\n" + str(f.__doc__ or "")
new_f.__doc__ = (str(old_f.__doc__) + "\n" +
", ".join([typename(type) for type in (type1, type2)]) +
"\n" + str(f.__doc__ or ""))
return new_f return new_f
return wrap return wrap
def type_guard(type1): def type_guard(type1):
def wrap(f): def wrap(f):
old_f = f.func_globals[f.__name__] old_f = f.func_globals[f.__name__]
def new_f(arg1, *rest): def new_f(arg1, *rest):
if (type1 is ANY_TYPE or isinstance(arg1, type1)): if (type1 is ANY_TYPE or isinstance(arg1, type1)):
variable = f(arg1, *rest) variable = f(arg1, *rest)
...@@ -308,8 +345,8 @@ def type_guard(type1): ...@@ -308,8 +345,8 @@ def type_guard(type1):
else: else:
return old_f(arg1, *rest) return old_f(arg1, *rest)
new_f.__name__ = f.__name__ new_f.__name__ = f.__name__
def typename(type): def typename(type):
if isinstance(type, Keyword): if isinstance(type, Keyword):
return str(type) return str(type)
...@@ -317,8 +354,12 @@ def type_guard(type1): ...@@ -317,8 +354,12 @@ def type_guard(type1):
return "(" + ", ".join([x.__name__ for x in type]) + ")" return "(" + ", ".join([x.__name__ for x in type]) + ")"
else: else:
return type.__name__ return type.__name__
new_f.__doc__ = str(old_f.__doc__) + "\n" + ", ".join([typename(type) for type in (type1,)]) + "\n" + str(f.__doc__ or "")
new_f.__doc__ = (str(old_f.__doc__) + "\n" +
", ".join([typename(type) for type in (type1,)]) +
"\n" + str(f.__doc__ or ""))
return new_f return new_f
return wrap return wrap
...@@ -331,15 +372,18 @@ def flatten(a): ...@@ -331,15 +372,18 @@ def flatten(a):
else: else:
return [a] return [a]
def unique(x): def unique(x):
return len(set(x)) == len(x) return len(set(x)) == len(x)
def hist(coll): def hist(coll):
counts = {} counts = {}
for elem in coll: for elem in coll:
counts[elem] = counts.get(elem, 0) + 1 counts[elem] = counts.get(elem, 0) + 1
return counts return counts
def give_variables_names(variables): def give_variables_names(variables):
""" Gives unique names to an iterable of variables. Modifies input. """ Gives unique names to an iterable of variables. Modifies input.
...@@ -349,10 +393,10 @@ def give_variables_names(variables): ...@@ -349,10 +393,10 @@ def give_variables_names(variables):
bad_var = lambda var: not var.name or h[var.name] > 1 bad_var = lambda var: not var.name or h[var.name] > 1
for i, var in enumerate(filter(bad_var, variables)): for i, var in enumerate(filter(bad_var, variables)):
var.name = (var.name or "") + "_%d"%i var.name = (var.name or "") + "_%d" % i
if not unique(map(str, variables)): if not unique(map(str, variables)):
raise ValueError("Not all variables have unique names." raise ValueError("Not all variables have unique names."
"Maybe you've named some of the variables identically") "Maybe you've named some of the variables identically")
return variables return variables
...@@ -9,11 +9,11 @@ import warnings ...@@ -9,11 +9,11 @@ import warnings
import numpy import numpy
import theano from theano.gof import local_bitwidth
from theano.gof.cc import hash_from_file from theano.gof.cc import hash_from_file
from theano.gof.cmodule import (std_libs, std_lib_dirs, from theano.gof.cmodule import (std_libs, std_lib_dirs,
std_include_dirs, dlimport, std_include_dirs, dlimport,
get_lib_extension, local_bitwidth) get_lib_extension)
from theano.gof.python25 import any from theano.gof.python25 import any
from theano.misc.windows import call_subprocess_Popen from theano.misc.windows import call_subprocess_Popen
...@@ -245,8 +245,6 @@ class NVCC_compiler(object): ...@@ -245,8 +245,6 @@ class NVCC_compiler(object):
cppfile = file(cppfilename, 'w') cppfile = file(cppfilename, 'w')
_logger.debug('Writing module C++ code to %s', cppfilename) _logger.debug('Writing module C++ code to %s', cppfilename)
ofiles = []
rval = None
cppfile.write(src_code) cppfile.write(src_code)
cppfile.close() cppfile.close()
......
...@@ -114,12 +114,12 @@ class BinCountOp(theano.Op): ...@@ -114,12 +114,12 @@ class BinCountOp(theano.Op):
# Some dtypes are not supported by numpy's implementation of bincount. # Some dtypes are not supported by numpy's implementation of bincount.
# Until another one is available, we should fail at graph construction # Until another one is available, we should fail at graph construction
# time, not wait for execution. # time, not wait for execution.
int_bitwidth = theano.gof.cmodule.python_int_bitwidth() int_bitwidth = theano.gof.python_int_bitwidth()
if int_bitwidth == 64: if int_bitwidth == 64:
numpy_unsupported_dtypes = ('uint64',) numpy_unsupported_dtypes = ('uint64',)
if int_bitwidth == 32: if int_bitwidth == 32:
numpy_unsupported_dtypes = ('uint32', 'int64', 'uint64') numpy_unsupported_dtypes = ('uint32', 'int64', 'uint64')
intp_bitwidth = theano.gof.cmodule.local_bitwidth() intp_bitwidth = theano.gof.local_bitwidth()
if intp_bitwidth == 32: if intp_bitwidth == 32:
out_type = basic.ivector() out_type = basic.ivector()
elif intp_bitwidth == 64: elif intp_bitwidth == 64:
...@@ -246,7 +246,7 @@ class RepeatOp(theano.Op): ...@@ -246,7 +246,7 @@ class RepeatOp(theano.Op):
# Some dtypes are not supported by numpy's implementation of repeat. # Some dtypes are not supported by numpy's implementation of repeat.
# Until another one is available, we should fail at graph construction # Until another one is available, we should fail at graph construction
# time, not wait for execution. # time, not wait for execution.
int_bitwidth = theano.gof.cmodule.python_int_bitwidth() int_bitwidth = theano.gof.python_int_bitwidth()
if int_bitwidth == 64: if int_bitwidth == 64:
numpy_unsupported_dtypes = ('uint64',) numpy_unsupported_dtypes = ('uint64',)
if int_bitwidth == 32: if int_bitwidth == 32:
...@@ -259,7 +259,7 @@ class RepeatOp(theano.Op): ...@@ -259,7 +259,7 @@ class RepeatOp(theano.Op):
% numpy_unsupported_dtypes), repeats.dtype) % numpy_unsupported_dtypes), repeats.dtype)
if self.axis is None: if self.axis is None:
broadcastable=[False] broadcastable = [False]
else: else:
try: try:
const_reps = basic.get_scalar_constant_value(repeats) const_reps = basic.get_scalar_constant_value(repeats)
......
...@@ -13,6 +13,7 @@ from theano import config, tensor, function ...@@ -13,6 +13,7 @@ from theano import config, tensor, function
numpy_ver = [int(n) for n in numpy.__version__.split('.')[:2]] numpy_ver = [int(n) for n in numpy.__version__.split('.')[:2]]
numpy_16 = bool(numpy_ver >= [1, 6]) numpy_16 = bool(numpy_ver >= [1, 6])
class TestBinCountOp(utt.InferShapeTester): class TestBinCountOp(utt.InferShapeTester):
def setUp(self): def setUp(self):
super(TestBinCountOp, self).setUp() super(TestBinCountOp, self).setUp()
...@@ -25,7 +26,7 @@ class TestBinCountOp(utt.InferShapeTester): ...@@ -25,7 +26,7 @@ class TestBinCountOp(utt.InferShapeTester):
'uint8', 'uint16', 'uint32', 'uint64'): 'uint8', 'uint16', 'uint32', 'uint64'):
# uint64 always fails # uint64 always fails
# int64 and uint32 also fail if python int are 32-bit # int64 and uint32 also fail if python int are 32-bit
int_bitwidth = theano.gof.cmodule.python_int_bitwidth() int_bitwidth = theano.gof.python_int_bitwidth()
if int_bitwidth == 64: if int_bitwidth == 64:
numpy_unsupported_dtypes = ('uint64',) numpy_unsupported_dtypes = ('uint64',)
if int_bitwidth == 32: if int_bitwidth == 32:
...@@ -57,7 +58,7 @@ class TestBinCountOp(utt.InferShapeTester): ...@@ -57,7 +58,7 @@ class TestBinCountOp(utt.InferShapeTester):
for dtype in tensor.discrete_dtypes: for dtype in tensor.discrete_dtypes:
# uint64 always fails # uint64 always fails
# int64 and uint32 also fail if python int are 32-bit # int64 and uint32 also fail if python int are 32-bit
int_bitwidth = theano.gof.cmodule.python_int_bitwidth() int_bitwidth = theano.gof.python_int_bitwidth()
if int_bitwidth == 64: if int_bitwidth == 64:
numpy_unsupported_dtypes = ('uint64',) numpy_unsupported_dtypes = ('uint64',)
if int_bitwidth == 32: if int_bitwidth == 32:
...@@ -188,7 +189,6 @@ class SqueezeTester(utt.InferShapeTester): ...@@ -188,7 +189,6 @@ class SqueezeTester(utt.InferShapeTester):
def test_grad(self): def test_grad(self):
for shape, broadcast in zip(self.shape_list, self.broadcast_list): for shape, broadcast in zip(self.shape_list, self.broadcast_list):
data = numpy.random.random(size=shape).astype(theano.config.floatX) data = numpy.random.random(size=shape).astype(theano.config.floatX)
variable = tensor.TensorType(theano.config.floatX, broadcast)()
utt.verify_grad(self.op, [data]) utt.verify_grad(self.op, [data])
...@@ -203,7 +203,7 @@ class TestRepeatOp(utt.InferShapeTester): ...@@ -203,7 +203,7 @@ class TestRepeatOp(utt.InferShapeTester):
self.op = RepeatOp() self.op = RepeatOp()
# uint64 always fails # uint64 always fails
# int64 and uint32 also fail if python int are 32-bit # int64 and uint32 also fail if python int are 32-bit
int_bitwidth = theano.gof.cmodule.python_int_bitwidth() int_bitwidth = theano.gof.python_int_bitwidth()
if int_bitwidth == 64: if int_bitwidth == 64:
self.numpy_unsupported_dtypes = ('uint64',) self.numpy_unsupported_dtypes = ('uint64',)
if int_bitwidth == 32: if int_bitwidth == 32:
...@@ -287,11 +287,12 @@ class TestRepeatOp(utt.InferShapeTester): ...@@ -287,11 +287,12 @@ class TestRepeatOp(utt.InferShapeTester):
x = T.TensorType(config.floatX, [False, True, False])() x = T.TensorType(config.floatX, [False, True, False])()
r = RepeatOp(axis=1)(x, 2) r = RepeatOp(axis=1)(x, 2)
self.assertEqual(r.broadcastable, (False, False, False)) self.assertEqual(r.broadcastable, (False, False, False))
r = RepeatOp(axis=1)(x, 1) r = RepeatOp(axis=1)(x, 1)
self.assertEqual(r.broadcastable, (False, True, False)) self.assertEqual(r.broadcastable, (False, True, False))
r = RepeatOp(axis=0)(x, 2) r = RepeatOp(axis=0)(x, 2)
self.assertEqual(r.broadcastable, (False, True, False)) self.assertEqual(r.broadcastable, (False, True, False))
class TestBartlett(utt.InferShapeTester): class TestBartlett(utt.InferShapeTester):
def setUp(self): def setUp(self):
......
Markdown 格式
0%
您添加了 0 到此讨论。请谨慎行事。
请先完成此评论的编辑!
注册 或者 后发表评论