Fix optimization crash and test crash due to new gpu conv default.

theano/sandbox/cuda/opt.py

@@ -1123,11 +1123,19 @@ def local_gpu_conv(node):
     if theano.sandbox.cuda.dnn.dnn_available():
         repl = local_gpu_conv_legacy.transform(node)
         if repl:
+            if isinstance(repl[0].owner.op, GpuConv):
+                n = repl[0].owner
+                add_transfer = False
+            else:
                 n = repl[0].owner.inputs[0].owner
                 assert isinstance(n.op, GpuConv)
+                add_transfer = True
             ret = theano.sandbox.cuda.dnn.local_conv_dnn.transform(n)
             if ret:
+                if add_transfer:
                     return [host_from_gpu(ret[0])]
+                else:
+                    return ret
     # If dnn isn't avail, the local_gpu_conv_legacy wil introduce the
     # legacy opt. Then the local_conv_gemm will convert it to gemm
     # opt.

theano/sandbox/cuda/tests/test_conv_cuda_ndarray.py

@@ -747,6 +747,10 @@ def test_dnn_subsample():
 
 
 class TestConv2DGPU(unittest.TestCase):
+    conv_ops = (cuda.blas.GpuConv,
+                cuda.dnn.GpuDnnConvBase,
+                cuda.blas.BaseGpuCorrMM)
+
     def test_logical_shapes(self):
         seed_rng()
         for stride in range(1, 4):
@@ -773,7 +777,7 @@ class TestConv2DGPU(unittest.TestCase):
 
             func = theano.function([a, A], image_estimate, mode=theano_mode)
             #theano.printing.debugprint(func,)
-            assert any([isinstance(node.op, theano.sandbox.cuda.blas.GpuConv)
+            assert any([isinstance(node.op, self.conv_ops)
                         for node in func.maker.fgraph.toposort()])
 
             a_in = numpy.random.randn(*featshp).astype("float32")

theano/sandbox/cuda/tests/test_mlp.py

@@ -396,7 +396,11 @@ def build_conv_nnet2_classif(use_gpu, isize, ksize, n_batch,
     if use_gpu:
         # Check that GpuConv is used
         topo = train.maker.fgraph.toposort()
-        assert len([n for n in topo if isinstance(n.op, tcn.blas.GpuConv)]) > 0
+        conv_ops = (tcn.blas.GpuConv,
+                    tcn.dnn.GpuDnnConvBase,
+                    tcn.blas.BaseGpuCorrMM)
+
+        assert len([n for n in topo if isinstance(n.op, conv_ops)]) > 0
 
     shape_target = (n_batch, n_out)
     return train, params, shape_img, shape_target, mode