Float64

Add an option to desactivate conv2d_fft at the op level

3dfftconv

Don't make the kernels debuggable by default, but add those arguments to the 'nvcc-safe' list.

Conv gemm non-square kernel support

Extend documentation (plus small fix)

Cleanup CUDA convolution tests

Add a blocksparse multiplication implementation

Remove the restriction on indexing a broadcastable dimension.

Update opt to replace Conv3D, ConvGrad3D and ConvTransp3D instead of GpuConv3D, GpuConvGrad3D and GpuConvTransp3D.

Conflicts:
	doc/library/tensor/nnet/conv.txt

@as_op() documentation