Add a useless opt pass. Need to time.

theano/compile/mode.py

@@ -150,6 +150,16 @@ optdb = gof.SequenceDB()
 optdb.register('merge1', gof.MergeOptimizer(),
                0, 'fast_run', 'fast_compile', 'merge')
 
+# After scan1 opt at 0.5 and before ShapeOpt at 1
+# This should only remove nodes.
+# The opt should not do anything that need shape inference.
+# New nodes that don't have infer_shape need that the original node
+# also don't have infer_shape
+optdb.register('useless', gof.EquilibriumDB(ignore_newtrees=False),
+               0.6, 'fast_run', 'fast_compile')
+optdb.register('merge1.1', gof.MergeOptimizer(),
+               0.65, 'fast_run', 'fast_compile', 'merge')
+
 # rearranges elemwise expressions
 optdb.register('canonicalize', gof.EquilibriumDB(ignore_newtrees=False),
                1, 'fast_run', 'fast_compile', 'canonicalize_db')

theano/tensor/opt.py

@@ -409,6 +409,18 @@ compile.optdb.register('inplace_elemwise_opt', inplace_elemwise_optimizer, 75,
                        'fast_run', 'inplace')
 
 
+def register_useless(lopt, *tags, **kwargs):
+    if type(lopt) == str:
+        def register(inner_lopt):
+            return register_useless(inner_lopt, lopt, *tags, **kwargs)
+        return register
+    else:
+        name = kwargs.pop('name', None) or lopt.__name__
+        compile.optdb['useless'].register(name, lopt, 'fast_run',
+                                          *tags, **kwargs)
+        return lopt
+
+
 def register_canonicalize(lopt, *tags, **kwargs):
     if type(lopt) == str:
         def register(inner_lopt):
@@ -1756,6 +1768,7 @@ compile.optdb.register('local_elemwise_alloc',
 
 
 @register_canonicalize("fast_compile")
+@register_useless("fast_compile")
 @gof.local_optimizer([T.fill])
 def local_useless_fill(node):
     """fill(s,v) -> v
@@ -1776,6 +1789,7 @@ def local_useless_fill(node):
 @register_specialize
 @register_stabilize
 @register_canonicalize
+@register_useless
 @gof.local_optimizer([T.alloc])
 def local_useless_alloc(node):
     """
@@ -1925,6 +1939,7 @@ def local_subtensor_remove_broadcastable_index(node):
 
 @register_specialize
 @register_canonicalize('fast_compile_gpu')
+#@register_useless
 @gof.local_optimizer([Subtensor, AdvancedSubtensor1])
 def local_subtensor_make_vector(node):
     """
@@ -2009,6 +2024,7 @@ def local_subtensor_make_vector(node):
 
 # TODO: the other optimization for and, or, xor, le and ge see ticket #496.
 
+@register_useless('fast_compile')
 @register_canonicalize('fast_compile')
 @register_specialize
 @gof.local_optimizer([T.Elemwise])
@@ -2428,6 +2444,7 @@ def local_upcast_elemwise_constant_inputs(node):
 ##################
 
 
+@register_useless
 @register_canonicalize
 @register_specialize
 @gof.local_optimizer([IncSubtensor])
@@ -2518,6 +2535,7 @@ def local_set_to_inc_subtensor(node):
         return [ret]
 
 
+@register_useless
 @register_canonicalize
 @register_specialize
 @gof.local_optimizer([Subtensor])
@@ -2547,6 +2565,7 @@ def local_useless_slice(node):
             return [out]
 
 
+@register_useless
 @register_canonicalize
 @register_specialize
 @gof.local_optimizer([Subtensor, AdvancedSubtensor1])
@@ -3373,6 +3392,7 @@ def local_adv_sub1_adv_inc_sub1(node):
 @register_specialize
 @register_stabilize
 @register_canonicalize
+@register_useless
 @gof.local_optimizer([IncSubtensor,
                       AdvancedIncSubtensor,
                       AdvancedIncSubtensor1])
@@ -3484,6 +3504,7 @@ def local_useless_inc_subtensor_alloc(node):
 # Rebroadcast opts #
 ####################
 
+@register_useless
 @register_canonicalize
 @register_specialize
 @gof.local_optimizer([T.Rebroadcast])
@@ -3611,6 +3632,7 @@ def apply_rebroadcast_opt(rval):
 #############
 @register_specialize
 @register_canonicalize
+@register_useless
 @gof.local_optimizer([T.Join])
 def local_join_1(node):
     """Join(i, x) => x
@@ -3627,6 +3649,7 @@ def local_join_1(node):
         return [tensors[0]]
 
 
+# TODO: merge in local_useless_join
 @register_specialize
 @register_canonicalize
 @gof.local_optimizer([T.Join])
@@ -3683,6 +3706,7 @@ def local_join_empty(node):
 
 @register_specialize
 @register_canonicalize
+@register_useless
 @gof.local_optimizer([T.Join])
 def local_join_make_vector(node):
     """Join(0, make_vector1, make_vector2, ...) => Join(0, make_vector12, ...)
@@ -3785,6 +3809,7 @@ def local_expm1(node):
 ###############
 # Switch opts #
 ###############
+@register_useless('fast_compile')
 @register_canonicalize('fast_compile', 'local_remove_switch_const_cond')
 @register_specialize
 @gof.local_optimizer([T.Elemwise])
@@ -4053,6 +4078,7 @@ def local_merge_switch_same_cond(node):
 #############
 # Tile Opts #
 #############
+@register_useless
 @register_canonicalize
 @register_stabilize
 @gof.local_optimizer([T.Tile])
@@ -4099,6 +4125,7 @@ def local_useless_tile(node):
 ##############
 # Split Opts #
 ##############
+@register_useless
 @register_canonicalize
 @register_specialize
 @gof.local_optimizer([T.Split])
@@ -4179,6 +4206,7 @@ register_canonicalize(local_reshape_chain(T.Reshape),
                       name='local_reshape_chain')
 
 
+@register_useless
 @register_canonicalize
 @register_stabilize
 @gof.local_optimizer([T.Reshape])
@@ -4987,6 +5015,7 @@ def local_elemwise_sub_zeros(node):
         return [T.zeros_like(node.inputs[0])]
 
 
+@register_useless
 @register_specialize
 @register_stabilize
 @register_canonicalize
@@ -5435,9 +5464,10 @@ def local_reduce_join(node):
         return [ret]
 
 
-@register_canonicalize('fast_compile')
+@register_canonicalize('fast_compile', 'local_cut_useless_reduce')
+@register_useless
 @gof.local_optimizer(ALL_REDUCE)
-def local_cut_useless_reduce(node):
+def local_useless_reduce(node):
     """Sum(a, axis=[]) -> a  """
     if isinstance(node.op, T.CAReduce):
         summed, = node.inputs