The type of tensor z should be bool. And there's no need to call mutable_data because ElementwiseComputeEx will do it.

There are mainly following fixes:

- take `DeviceContext` as the template parameter of math functors and OpKernel instead of `Place`
- remove `eigen_device` interface in base class  `DeviceContext`
- remove `GetEigenDevice` interface in `ExecutionContext` and base class `DeviceContext`
- remove unused `platform::EigenDeviceConverter`
- rename `REGISTER_OP_GPU_KERNEL` to `REGISTER_OP_CUDA_KERNEL`
- rename `USE_GPU_ONLY_OP` to `USE_CUDA_ONLY_OP`

* Adding greater than and less than equal ops to compare op
* Changing the name of the less_than_equal and greater_than_equal op
* Also changing the name of the functors

* Compare Operator

* Follow comments