Fix crash for MRG_RandomStreams.seed on GPU

theano/sandbox/rng_mrg.py

@@ -1133,8 +1133,7 @@ class MRG_RandomStreams(object):
             if nstreams is None:
                 nstreams = self.n_streams(size)
             rstates = self.get_substream_rstates(nstreams)
-
-            if self.use_cuda and dtype == 'float32':
+            if self.use_cuda and new_r.owner.outputs[1].dtype == 'float32':
                 rstates = rstates.flatten()
                 # HACK - we use fact that int32 and float32 have same size to
                 # sneak ints into the CudaNdarray type.
@@ -1142,12 +1141,12 @@ class MRG_RandomStreams(object):
                 tmp_float_buf = numpy.frombuffer(rstates.data, dtype='float32')
                 assert tmp_float_buf.shape == rstates.shape
                 assert (tmp_float_buf.view('int32') == rstates).all()
-                rstate = tmp_float_buf
+                rstates = tmp_float_buf
             assert (old_r.get_value(borrow=True,
                                     return_internal_type=True).shape ==
                     rstates.shape)
-            old_r.set_value(rstates,
-                            borrow=True)
+            assert rstates.dtype == old_r.dtype
+            old_r.set_value(rstates, borrow=True)
 
     def inc_rstate(self):
         """Update self.rstate to be skipped 2^134 steps forward to the next stream start"""

theano/sandbox/test_rng_mrg.py

@@ -938,12 +938,22 @@ def test_seed_fn():
         test_use_cuda.append(True)
     idx = tensor.ivector()
     for use_cuda in test_use_cuda:
+        if config.mode == 'FAST_COMPILE' and use_cuda:
+            mode = 'FAST_RUN'
+        else:
+            mode = config.mode
+
         for new_seed, same in [(234, True), (None, True), (23, False)]:
-            random = MRG_RandomStreams(234)
-            fn1 = theano.function([], random.uniform((2, 2)))
-            fn2 = theano.function([], random.uniform((3, 3), nstreams=2))
+            random = MRG_RandomStreams(234, use_cuda=use_cuda)
+            fn1 = theano.function([], random.uniform((2, 2), dtype='float32'),
+                                  mode=mode)
+            fn2 = theano.function([], random.uniform((3, 3), nstreams=2,
+                                                     dtype='float32'),
+                                  mode=mode)
             fn3 = theano.function([idx],
-                                  random.uniform(idx, nstreams=3, ndim=1))
+                                  random.uniform(idx, nstreams=3, ndim=1,
+                                                 dtype='float32'),
+                                  mode=mode)
 
             fn1_val0 = fn1()
             fn1_val1 = fn1()