Try to fix error related to multiples instances of overloaded function "fabs" with float16.

theano/gpuarray/elemwise.py

@@ -948,8 +948,7 @@ class GpuCAReduceCuda(GpuKernelBase, HideC, CAReduceDtype):
             assert isinstance(self.scalar_op, (scalar.Maximum,
                                                scalar.Minimum))
             if self.pre_scalar_op:  # TODO: multiple dtypes
-                # dtype = node.inputs[0].dtype
-                dtype = 'float32'
+                dtype = self._acc_dtype(self.acc_dtype)
 
                 dummy_var = scalar.Scalar(dtype=dtype)()
 
@@ -1719,6 +1718,7 @@ class GpuCAReduceCuda(GpuKernelBase, HideC, CAReduceDtype):
         for i in node.inputs + node.outputs:
             version.extend(Scalar(dtype=i.type.dtype).c_code_cache_version())
         version.extend(self.kernel_version(node))
+        version.extend(self._acc_dtype(self.acc_dtype))
         if all(version):
             return tuple(version)
         else: