Merge pull request #5794 from ReyhaneAskari/faster_topo

doc/library/config.txt

@@ -225,6 +225,20 @@ import theano and print the config variable, as in:
 
     If True, we will print extra scan debug information.
 
+.. attribute:: cycle_detection
+
+    String value, either ``regular`` or ``fast```
+
+    Default: ``regular``
+
+    If :attr:`cycle_detection` is set to ``regular``, most inplaces are allowed,
+    but it is slower. If :attr:`cycle_detection` is set to ``faster``,
+    less inplaces are allowed, but it makes the compilation faster.
+
+    The interaction of which one give the lower peak memory usage is complicated and
+    not predictable, so if you are close to the peak memory usage, triyng both
+    could give you a small gain.
+
 .. attribute:: openmp
 
     Bool value: either ``True`` or ``False``

theano/compile/profiling.py

@@ -830,7 +830,6 @@ class ProfileStats(object):
 
             """
             from theano.gpuarray import GpuArrayType
-
             # Initial Mem info values [CPU, GPU]
             node_memory_size = [0, 0]
             running_memory_size = [0, 0]

theano/configdefaults.py

@@ -1476,6 +1476,17 @@ AddConfigVar('compile.wait',
              IntParam(5, lambda i: i > 0, allow_override=False),
              in_c_key=False)
 
+AddConfigVar('cycle_detection',
+             "If cycle_detection is set to regular, most inplaces are allowed,"
+             "but it is slower. If cycle_detection is set to faster, less inplaces"
+             "are allowed, but it makes the compilation faster."
+
+             "The interaction of which one give the lower peak memory usage is"
+             "complicated and not predictable, so if you are close to the peak"
+             "memory usage, triyng both could give you a small gain. ",
+             EnumStr('regular', 'fast'),
+             in_c_key=False)
+
 
 def _timeout_default():
     return theano.config.compile.wait * 24

theano/gof/fg.py

@@ -654,8 +654,9 @@ class FunctionGraph(utils.object2):
         take care of computing dependencies by itself.
 
         """
-        ords = OrderedDict()
         assert isinstance(self._features, list)
+        all_orderings = []
+
         for feature in self._features:
             if hasattr(feature, 'orderings'):
                 orderings = feature.orderings(self)
@@ -664,16 +665,23 @@ class FunctionGraph(utils.object2):
                                     str(feature.orderings) +
                                     ". Nondeterministic object is " +
                                     str(orderings))
+                if len(orderings) > 0:
+                    all_orderings.append(orderings)
                     for node, prereqs in iteritems(orderings):
                         if not isinstance(prereqs, (list, OrderedSet)):
                             raise TypeError(
                                 "prereqs must be a type with a "
                                 "deterministic iteration order, or toposort "
                                 " will be non-deterministic.")
+        if len(all_orderings) == 1:
+            # If there is only 1 ordering, we reuse it directly.
+            return all_orderings[0].copy()
+        else:
+            # If there is more than 1 ordering, combine them.
+            ords = OrderedDict()
+            for orderings in all_orderings:
+                for node, prereqs in iteritems(orderings):
                     ords.setdefault(node, []).extend(prereqs)
-        # eliminate duplicate prereqs
-        for (node, prereqs) in iteritems(ords):
-            ords[node] = list(OrderedSet(prereqs))
             return ords
 
     def check_integrity(self):