Corrected check for whether cuda is active or not and compile only one addition kernel instead of many in the perform() method.

Mostly adapted Op and tests to new backend. TODO: Remove faulty python implementation from _dev20 version of op

Copied GpuAdvancedIncSubtensor1() and GpuAdvancedIncSubtensor1_dev() ops and their tests to sandbox/gpuarray

Corrected check for whether cuda is active or not and compile only one addition kernel instead of many in the perform() method.

Cuda ndarray fix

Remove overflow and detect it if it happen again.

Otherwise, it didn't cause other problems, as we reported the error to the user.

close gh-1814

mrg_rng for gpuarray

Make scan work with new backend.

Release GIL while waiting for the GPU

Fix tests in python3.

Opt Reduce(join())

It should also not break the old cuda backend.

Port op GpuImages2Neibs to new GPU backend

Mostly adapted Op and tests to new backend. TODO: Remove faulty python implementation from _dev20 version of op

Minor fixes