Update following code review

theano/misc/elemwise_openmp_speedup.py

@@ -49,7 +49,7 @@ if __name__ == '__main__':
     else:
         costlySpeed = costlyTimeOpenmp / costlyTime
         costlySpeedstring = "slowdown"
-    print("Timmed with vector of %d elements" % options.N)
+    print("Timed with vector of %d elements" % options.N)
     print("Fast op time without openmp %fs with openmp %fs %s %2.2f" % (
         cheapTime, cheapTimeOpenmp, cheapSpeedstring, cheapSpeed))
 

theano/sandbox/cuda/opt.py

@@ -1116,8 +1116,13 @@ def local_gpu_incsubtensor(node):
             incsubt = host_output.owner.op
             x, y = host_output.owner.inputs[0:2]
             coords = host_output.owner.inputs[2:]
-            if x.dtype != "float32" or y.dtype != "float32":
+            if x.dtype != "float32":
                 return
+            if y.dtype != "float32":
+                # The IncSubtensor upcast to float32 y, so we do it
+                # explicitly to move it to the GPU.
+                y = y.astype('float32')
+
             return [GpuIncSubtensor(
                 incsubt.idx_list,
                 inplace=incsubt.inplace,