Convert SetSubtensor to IncSubtensor when possible

This did come up in an implementation of adagrad at some point and will be much faster on GPU.

Do not move the GpuSubtensor to the GPU when this do not make sence

[BUG] Fix bug in {inc,set}_subtensor in advanced1 case

Join opt. Remove empty constant as inputs and merge MakeVector inputs.

do the join change for gh-2328

fix gh-2327

Remove useless warning and make the graph more optimized in old supported cases.

When we do this on an inputs/shared var and that this is the only use
of it in the graph, we should only transfer to the GPU the small
computation.

Reported by Stanislas Lauly

Document nvcc.fastmath and fix cxx description.

Tests

Fix problems with CDataType in DebugMode

Run part of travis on Python 3

- fix install of pydot by using the right version of pyparsing on python 2
- pydot does not run on python 3 for the moment
- run tests on the installed files, not the source

Add DictMixin/OrderedDict to theano.compat in PY3.

Reordered for ease of reading (same order in if/else).