* add expand_v2/expand_as_v2 for kunlun

* update expand_as_v2

* update expand_as_v2

* support float16/bool

* update xpu.cmake

* suport fetch lodtensor array, test=develop

* refine, test=develop

* refine, test=develop

* refine, test=develop

* [NPU] reorganization for device API abstraction

* [NPU] delete old files

* [NPU] fix npu_collective_helper

* [NPU] fix collective_helper

* [NPU] fix ut

* [NPU] mod memory allocation and hccl_helper

* [NPU] fix place_type

* [NPU] split enfoce.h

* move acl* call into npu_info

* merge conflict

* fix merge

* merge conflict

* merge conflict

* test

* test

* rm test

* update

* update

* update

* add unittest

* update

* update save

* add scalar and scalar_array

* remove DenseTensor include from Scalar and ScalarArray

* remove inner header from scalar_array

* refactor the method of fill_constant and add some comment

* add fill_constant kernel using ScalarArray

* modify some prompt

* remove fill_constant kernel with no shape

* [NPU] add int64 support for argsort op

* [NPU] delete debug codes

* [NPU] add NPU kernel for prior_box op

* [NPU] delete debug codes

* Added GradTensorHolder to Eager Dygraph

* Added accumulation codes to Eager Dygraph

* Fix windows-ci issue

* Fix NPU-CI issue

* Fixed CI-Coverage issue

* clear LoDTensorArray

* fix  bugs

* fix

* fix gpu

* Add second batch of deprecated mkldnn namespace and macro changes

* Unlock CI

* Fix temporary namespace alias placing

* elementwise_mul refactor

* perfect code in test

* delete redundant code

* fix bugs when run test_multiply

* adjust the location of macro

* fix bugs when run ci

* add scalar and scalar_array

* remove DenseTensor include from Scalar and ScalarArray

* remove inner header from scalar_array

* refactor the method of fill_constant and add some comment

* [XPU] Reorganize xpu device codes in platform, test=develop

* fix xpu_header.h, test=develop

Add support for bias is none for fused_attention op.

* enhance scatter err msg check

* fix ci error

* elementwise_div refactor

* fix compile bugs in windows ci

* Added HCCL backend support in dynamic graph mode

* fix segmentation fault

* add ut

* Add transfer_layout/dtype op

* clean useless codes

* fix unused var

* add optest in white.txt

* split into data_transfer.cc

* fix cmake

* modify according reviewer comment

* replace cast_op with transfer_dtype_op

* disable copying of datatype when sharing buffer between two tensors.
* fix for mkldnn operator kernels (elementwise_add, sum, softplus, softmax, scale, activation), mannually set the data type when reusing memory by ShareBufferWith.

* add isclose op, test=develop

* add isclose op, test=develop

* add isclose api, test=develop

* rm useless code

* rm useless code

* update python api of isclose

* add some unittest of isclose op, test=develop

* support zero dim for slice op

* support zero dim Tensor in set_value op

* polish some debug log

* add cast kernel

* add cast cuda kernel

* add cast kernel

* make cast kernel output dtype undefined

* get cast dtype from vardesc

* move cast to manipulation and add test case

* add castinfershape

* avoid reinitilaze variable

* InitializeVariable support datatype

* merge develop branch

* fix merge bug

* revert modify initializeVariable

* revert modify on InitializeVariable

* revert modify on InitializeVariable

* mutable support reset dtype

* enable make pten tensor from variable when def_arg.type is undefined

* fix build pten ctx start_idx error

* copy pten out tensor to variable

* merge develop branch

* fix non pten kernel cast failed

* add reset allocation place for remake tensor

* fix inplace realloc error

* add mutable on pten kernles and remove unused cast files

* rename function names

* fix output type error

* fix conflict with develop branch

* set data type to variable with pten's dtype

* fix test_cast_api type mismatch

* densorTensro mutable_data support 0 bytes value

* fix the inplace bug of reshape kernel

* fix pten.backend != variable.place when moving storage, palce mismatch bug

* fix conflict with develop branch

* Fix bug of paddle::experimental::MovesStorage

* fix ReMakePtenDenseTensor place mismatch bug

* Revert "fix ReMakePtenDenseTensor place mismatch bug"

This reverts commit 86336032f60b8a15eacd2c1ff2fa513f5d8dfd1a.

* fix ReMakePtenDenseTensor place mismatch bug

* reverts the set_lod interface, test=develop

* modify by the review options

* modify error message

* add & for const input arguments

* add reference in params

* elementwise_sub add mutable_data

* fix ResetHolderWithType check size bug

* add dependence pten_tensor to test_cast_api object

* remove unused code to pass ci coverage
Co-authored-by: NChen Weihang <chenweihang@baidu.com>
Co-authored-by: NYuanRisheng <yuanrisheng@baidu.com>
Co-authored-by: Nshixiaowei02 <39303645+Shixiaowei02@users.noreply.github.com>

* optimize cache-key by replace GraphToProgram to Dot string

* fix compile failure bug

* GeneratePass support attr condition and mapping, test=develop

* fix coverage, test=develop

* Add fuse_resnet_unit pass, test=develop

* fix CI errors, test=develop

* fix CI errors, test=develop

* fix unittest error when compiling without CUDA, test=develop

* fix static ci error, test=develop

* limit kernel size must equal 1, test=develop

* add cpu version, using set: sum, min, max

* add cpu version: mean

* improve cpu code and fix dynamic memory allcation problem

* fix arg error, add index judge, delete fp16

* fix bug in CudaAtomicMax and CudaAtomicMin

* add CUDA version

* fix grad_op bug for index

* add op test, add correct cpu grad op

* Add correct CUDA Mean grad

* [Add] Successful MEAN and SUM

* [Add] Successful MIN and MAX in CPU

* [Add] Successful MIN and MAX in CUDA

* fix windows dtype ci

* fix ROCM ci by adding HIP flag

* rename fused_gather_scatter to send_recv

* unify name as send and recv

* change zero index return time

* add send_recv incubate api

* fix index data type, add unittest case for API

* delete redundant input tensor

* fix en example and docs, add default value in pool_type

* add shape judge and max grid judge

* fix comment

* fix index type bug

* add const &

* fix en docs

* delete numpy in examples

* add unittest for int input

* fix send_recv comment

* change send_recv to graph_send_recv

* fix bug to support dropout eval grad computing.

* Remove useless code.

* elementwise_add kernel refactor

* fix compile bugs in elementwise_add refactor

* fix compile bugs when run in npu/xpu

* fix bugs when run unit test

* fix bugs when run ci-windows

* modify code as recommended

* code format adjust

* fix bugs when run ci

* fix compile bug when run in ci-windwos

* elementwise_sub refactor

* add PD_DLL_DECL for elementwise_sub

* fix bugs when compilei

* Add the `GetFetchNames` method in CinnGraphSymbolization.

* Use unordered_set instead vector as the type of fetch_var_names.

* Reuse the definition of kCompilationKey.

* Use CompileOptions to set fetch_var_ids.

* Update the argument passing of GraphCompiler.Build.

* Fix some bugs in CinnGraphSymbolization::GetFetchIds.

topk中有cub和手写kernel两种实现，而cub是通过排序来获取topk，通过多组数据发现只有当input_width>=128且k超过input_width 75%的时候性能会比手写的更好。