[ENH] for rc3, Graph clean up, faster compilation, conv.assert_shape Theano flags, add unsafe in useless opt

Remove warp-synchronous programming.

Prepare release 0.9.0rc2

Fixes #5623

Remove a warning. It happen frequently enough that people need to wor…

Remove warning by warn_float64 when we want the upcast to float64.

Explicitly specify the output ndim in reshape

Remove a warning. It happen frequently enough that people need to work around g++ bugs or strange installation environment that it is more annoing then useful. Also people don't set it if they don't have problem, so it don't really help.

Add in the doc the optional dependencies of pycuda and skcuda.

Add test for the previous fix and add more check to prevent other crash with bad neib_shape or neib_step input value.

Fix dnn conv grad issues

The whole code expects the shape to be of length 4, and the output to be
4D already

Updates to solve potential problems with compiler-specific casting

Print scan profile again

Remove the old benchmark directory.

Remove the old benchmark directory. They use the module system that was removed years ago. Now the benchmark are outside Theano.

Fix various GPU unit tests on Windows