TemporaryAllocator is a singleton used for allocating memory for Cudnn. Since it is a singleton, we can delete it for better performance in memory.

We replace TemporaryAllocator by CUDADeviceContextAllocator and CUDADeviceContextAllocation, which uses stream callback to delete the memory allocated for the stream to avoid singleton.

Also added data_feed_proto to operator to fix CI in CPU compilation

test=develop

* add scatter_nd op, test=document_preview test=develop

* fixed the document, test=document_preview test=develop

* modify the notes, test=document_preview test=develop

* remove the ShareDataWith, test=develop

* fixed the code for coverage

* fixed the document,test=document_preview test=develop

* test=develop
The scatter op has a calc bug when the indices has same index, the scatter op use overwrite mode to calculate the same index, fix this bug by using the accumulate mode to calculate the same index.At the same time, the gather op has the same bug when the op calc the grad. And we use the lib of open-blas and eigen to optimize the time cost in accumulate mode.

* test=develop
Fix some code format problem, and the same time add the test case in gather and scatter op

* gather_op support int64_t index by adding a template typename

* add UT and rename typename

test=develop

* Add int type support for stack_op

* Improve gather op to support index with shape N x 1

test=develop

* Fix stack_op kernel's registry

test=develop

gather scatter gpu