Add optimization removing Rebroadcast when the pattern is not actually changed

theano/tensor/opt.py

@@ -692,6 +692,25 @@ def local_inplace_setsubtensor(node):
 compile.optdb.register('inplace_setsubtensor', TopoOptimizer(local_inplace_setsubtensor,
     failure_callback=TopoOptimizer.warn_inplace), 60, 'fast_run', 'inplace') #DEBUG
 
+
+####################
+# Rebroadcast opts #
+####################
+
+@register_canonicalize
+@register_specialize
+@gof.local_optimizer([T.Rebroadcast])
+def local_useless_rebroadcast(node):
+    """
+    Remove Rebroadcast if id does not actually change the broadcasting pattern
+    """
+    if isinstance(node.op, T.Rebroadcast):
+        x = node.inputs[0]
+        if numpy.all(x.broadcastable == node.outputs[0].broadcastable):
+            return [x]
+
+
+
 ##################
 # Reshape opts   #
 ##################

theano/tensor/tests/test_opt.py

@@ -1126,6 +1126,15 @@ def test_local_mul_specialize():
     assert nodes == [T.mul]
 
 
+def test_local_useless_rebroadcast():
+    v1 = T.vector()
+    v2 = T.vector()
+    j = T.join(0, v1, v2)
+    f = theano.function([v1, v2], j)
+    f([1,2], [3,4,5])
+    e = f.maker.env.toposort()
+    assert len([n for n in e if isinstance(n.op, T.Rebroadcast)]) == 0
+
 if __name__ == '__main__':
 #    unittest.main()
     test_fusion().tes_memory_leak()